diff options
Diffstat (limited to 'drivers/gpu/drm')
91 files changed, 6787 insertions, 1712 deletions
diff --git a/drivers/gpu/drm/drm_bufs.c b/drivers/gpu/drm/drm_bufs.c index 4fcbc445a8e5..a5c9ce93bbcb 100644 --- a/drivers/gpu/drm/drm_bufs.c +++ b/drivers/gpu/drm/drm_bufs.c @@ -948,7 +948,7 @@ int drm_addbufs_pci(struct drm_device * dev, struct drm_buf_desc * request) dma->buflist[i + dma->buf_count] = &entry->buflist[i]; } - /* No allocations failed, so now we can replace the orginal pagelist + /* No allocations failed, so now we can replace the original pagelist * with the new one. */ if (dma->page_count) { diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 994d23beeb1d..57cea01c4ffb 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -1840,8 +1840,10 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev, ret = copy_from_user(clips, clips_ptr, num_clips * sizeof(*clips)); - if (ret) + if (ret) { + ret = -EFAULT; goto out_err2; + } } if (fb->funcs->dirty) { diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index 764401951041..9b2a54117c91 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -860,19 +860,24 @@ static void output_poll_execute(struct slow_work *work) } } -void drm_kms_helper_poll_init(struct drm_device *dev) +void drm_kms_helper_poll_disable(struct drm_device *dev) +{ + if (!dev->mode_config.poll_enabled) + return; + delayed_slow_work_cancel(&dev->mode_config.output_poll_slow_work); +} +EXPORT_SYMBOL(drm_kms_helper_poll_disable); + +void drm_kms_helper_poll_enable(struct drm_device *dev) { - struct drm_connector *connector; bool poll = false; + struct drm_connector *connector; int ret; list_for_each_entry(connector, &dev->mode_config.connector_list, head) { if (connector->polled) poll = true; } - slow_work_register_user(THIS_MODULE); - delayed_slow_work_init(&dev->mode_config.output_poll_slow_work, - &output_poll_ops); if (poll) { ret = delayed_slow_work_enqueue(&dev->mode_config.output_poll_slow_work, DRM_OUTPUT_POLL_PERIOD); @@ -880,11 +885,22 @@ void drm_kms_helper_poll_init(struct drm_device *dev) DRM_ERROR("delayed enqueue failed %d\n", ret); } } +EXPORT_SYMBOL(drm_kms_helper_poll_enable); + +void drm_kms_helper_poll_init(struct drm_device *dev) +{ + slow_work_register_user(THIS_MODULE); + delayed_slow_work_init(&dev->mode_config.output_poll_slow_work, + &output_poll_ops); + dev->mode_config.poll_enabled = true; + + drm_kms_helper_poll_enable(dev); +} EXPORT_SYMBOL(drm_kms_helper_poll_init); void drm_kms_helper_poll_fini(struct drm_device *dev) { - delayed_slow_work_cancel(&dev->mode_config.output_poll_slow_work); + drm_kms_helper_poll_disable(dev); slow_work_unregister_user(THIS_MODULE); } EXPORT_SYMBOL(drm_kms_helper_poll_fini); diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index b3779d243aef..719662034bbf 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -146,7 +146,7 @@ static bool drm_fb_helper_connector_parse_command_line(struct drm_fb_helper_conn cvt = 1; break; case 'R': - if (!cvt) + if (cvt) rb = 1; break; case 'm': @@ -264,7 +264,7 @@ bool drm_fb_helper_force_kernel_mode(void) int drm_fb_helper_panic(struct notifier_block *n, unsigned long ununsed, void *panic_str) { - DRM_ERROR("panic occurred, switching back to text console\n"); + printk(KERN_ERR "panic occurred, switching back to text console\n"); return drm_fb_helper_force_kernel_mode(); return 0; } @@ -315,8 +315,9 @@ static void drm_fb_helper_on(struct fb_info *info) struct drm_device *dev = fb_helper->dev; struct drm_crtc *crtc; struct drm_crtc_helper_funcs *crtc_funcs; + struct drm_connector *connector; struct drm_encoder *encoder; - int i; + int i, j; /* * For each CRTC in this fb, turn the crtc on then, @@ -332,7 +333,14 @@ static void drm_fb_helper_on(struct fb_info *info) crtc_funcs->dpms(crtc, DRM_MODE_DPMS_ON); - + /* Walk the connectors & encoders on this fb turning them on */ + for (j = 0; j < fb_helper->connector_count; j++) { + connector = fb_helper->connector_info[j]->connector; + connector->dpms = DRM_MODE_DPMS_ON; + drm_connector_property_set_value(connector, + dev->mode_config.dpms_property, + DRM_MODE_DPMS_ON); + } /* Found a CRTC on this fb, now find encoders */ list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { if (encoder->crtc == crtc) { @@ -352,8 +360,9 @@ static void drm_fb_helper_off(struct fb_info *info, int dpms_mode) struct drm_device *dev = fb_helper->dev; struct drm_crtc *crtc; struct drm_crtc_helper_funcs *crtc_funcs; + struct drm_connector *connector; struct drm_encoder *encoder; - int i; + int i, j; /* * For each CRTC in this fb, find all associated encoders @@ -367,6 +376,14 @@ static void drm_fb_helper_off(struct fb_info *info, int dpms_mode) if (!crtc->enabled) continue; + /* Walk the connectors on this fb and mark them off */ + for (j = 0; j < fb_helper->connector_count; j++) { + connector = fb_helper->connector_info[j]->connector; + connector->dpms = dpms_mode; + drm_connector_property_set_value(connector, + dev->mode_config.dpms_property, + dpms_mode); + } /* Found a CRTC on this fb, now find encoders */ list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { if (encoder->crtc == crtc) { @@ -1024,11 +1041,18 @@ static struct drm_display_mode *drm_pick_cmdline_mode(struct drm_fb_helper_conne } create_mode: - mode = drm_cvt_mode(fb_helper_conn->connector->dev, cmdline_mode->xres, - cmdline_mode->yres, - cmdline_mode->refresh_specified ? cmdline_mode->refresh : 60, - cmdline_mode->rb, cmdline_mode->interlace, - cmdline_mode->margins); + if (cmdline_mode->cvt) + mode = drm_cvt_mode(fb_helper_conn->connector->dev, + cmdline_mode->xres, cmdline_mode->yres, + cmdline_mode->refresh_specified ? cmdline_mode->refresh : 60, + cmdline_mode->rb, cmdline_mode->interlace, + cmdline_mode->margins); + else + mode = drm_gtf_mode(fb_helper_conn->connector->dev, + cmdline_mode->xres, cmdline_mode->yres, + cmdline_mode->refresh_specified ? cmdline_mode->refresh : 60, + cmdline_mode->interlace, + cmdline_mode->margins); drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V); list_add(&mode->head, &fb_helper_conn->connector->modes); return mode; diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c index 14d9d829ef27..86118a742231 100644 --- a/drivers/gpu/drm/drm_sysfs.c +++ b/drivers/gpu/drm/drm_sysfs.c @@ -193,8 +193,9 @@ static ssize_t enabled_show(struct device *device, "disabled"); } -static ssize_t edid_show(struct kobject *kobj, struct bin_attribute *attr, - char *buf, loff_t off, size_t count) +static ssize_t edid_show(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, char *buf, loff_t off, + size_t count) { struct device *connector_dev = container_of(kobj, struct device, kobj); struct drm_connector *connector = to_drm_connector(connector_dev); diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index 95639017bdbe..da78f2c0d909 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -22,6 +22,7 @@ i915-y := i915_drv.o i915_dma.o i915_irq.o i915_mem.o \ intel_fb.o \ intel_tv.o \ intel_dvo.o \ + intel_ringbuffer.o \ intel_overlay.o \ dvo_ch7xxx.o \ dvo_ch7017.o \ diff --git a/drivers/gpu/drm/i915/dvo_tfp410.c b/drivers/gpu/drm/i915/dvo_tfp410.c index 66c697bc9b22..56f66426207f 100644 --- a/drivers/gpu/drm/i915/dvo_tfp410.c +++ b/drivers/gpu/drm/i915/dvo_tfp410.c @@ -208,7 +208,7 @@ static enum drm_connector_status tfp410_detect(struct intel_dvo_device *dvo) uint8_t ctl2; if (tfp410_readb(dvo, TFP410_CTL_2, &ctl2)) { - if (ctl2 & TFP410_CTL_2_HTPLG) + if (ctl2 & TFP410_CTL_2_RSEN) ret = connector_status_connected; else ret = connector_status_disconnected; diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 322070c0c631..aee83fa178f6 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -77,7 +77,7 @@ static int i915_gem_object_list_info(struct seq_file *m, void *data) case ACTIVE_LIST: seq_printf(m, "Active:\n"); lock = &dev_priv->mm.active_list_lock; - head = &dev_priv->mm.active_list; + head = &dev_priv->render_ring.active_list; break; case INACTIVE_LIST: seq_printf(m, "Inactive:\n"); @@ -129,7 +129,8 @@ static int i915_gem_request_info(struct seq_file *m, void *data) struct drm_i915_gem_request *gem_request; seq_printf(m, "Request:\n"); - list_for_each_entry(gem_request, &dev_priv->mm.request_list, list) { + list_for_each_entry(gem_request, &dev_priv->render_ring.request_list, + list) { seq_printf(m, " %d @ %d\n", gem_request->seqno, (int) (jiffies - gem_request->emitted_jiffies)); @@ -143,9 +144,9 @@ static int i915_gem_seqno_info(struct seq_file *m, void *data) struct drm_device *dev = node->minor->dev; drm_i915_private_t *dev_priv = dev->dev_private; - if (dev_priv->hw_status_page != NULL) { + if (dev_priv->render_ring.status_page.page_addr != NULL) { seq_printf(m, "Current sequence: %d\n", - i915_get_gem_seqno(dev)); + i915_get_gem_seqno(dev, &dev_priv->render_ring)); } else { seq_printf(m, "Current sequence: hws uninitialized\n"); } @@ -195,9 +196,9 @@ static int i915_interrupt_info(struct seq_file *m, void *data) } seq_printf(m, "Interrupts received: %d\n", atomic_read(&dev_priv->irq_received)); - if (dev_priv->hw_status_page != NULL) { + if (dev_priv->render_ring.status_page.page_addr != NULL) { seq_printf(m, "Current sequence: %d\n", - i915_get_gem_seqno(dev)); + i915_get_gem_seqno(dev, &dev_priv->render_ring)); } else { seq_printf(m, "Current sequence: hws uninitialized\n"); } @@ -251,7 +252,7 @@ static int i915_hws_info(struct seq_file *m, void *data) int i; volatile u32 *hws; - hws = (volatile u32 *)dev_priv->hw_status_page; + hws = (volatile u32 *)dev_priv->render_ring.status_page.page_addr; if (hws == NULL) return 0; @@ -287,7 +288,8 @@ static int i915_batchbuffer_info(struct seq_file *m, void *data) spin_lock(&dev_priv->mm.active_list_lock); - list_for_each_entry(obj_priv, &dev_priv->mm.active_list, list) { + list_for_each_entry(obj_priv, &dev_priv->render_ring.active_list, + list) { obj = &obj_priv->base; if (obj->read_domains & I915_GEM_DOMAIN_COMMAND) { ret = i915_gem_object_get_pages(obj, 0); @@ -317,14 +319,14 @@ static int i915_ringbuffer_data(struct seq_file *m, void *data) u8 *virt; uint32_t *ptr, off; - if (!dev_priv->ring.ring_obj) { + if (!dev_priv->render_ring.gem_object) { seq_printf(m, "No ringbuffer setup\n"); return 0; } - virt = dev_priv->ring.virtual_start; + virt = dev_priv->render_ring.virtual_start; - for (off = 0; off < dev_priv->ring.Size; off += 4) { + for (off = 0; off < dev_priv->render_ring.size; off += 4) { ptr = (uint32_t *)(virt + off); seq_printf(m, "%08x : %08x\n", off, *ptr); } @@ -344,7 +346,7 @@ static int i915_ringbuffer_info(struct seq_file *m, void *data) seq_printf(m, "RingHead : %08x\n", head); seq_printf(m, "RingTail : %08x\n", tail); - seq_printf(m, "RingSize : %08lx\n", dev_priv->ring.Size); + seq_printf(m, "RingSize : %08lx\n", dev_priv->render_ring.size); seq_printf(m, "Acthd : %08x\n", I915_READ(IS_I965G(dev) ? ACTHD_I965 : ACTHD)); return 0; @@ -489,11 +491,14 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused) struct drm_device *dev = node->minor->dev; drm_i915_private_t *dev_priv = dev->dev_private; u16 rgvswctl = I915_READ16(MEMSWCTL); + u16 rgvstat = I915_READ16(MEMSTAT_ILK); - seq_printf(m, "Last command: 0x%01x\n", (rgvswctl >> 13) & 0x3); - seq_printf(m, "Command status: %d\n", (rgvswctl >> 12) & 1); - seq_printf(m, "P%d DELAY 0x%02x\n", (rgvswctl >> 8) & 0xf, - rgvswctl & 0x3f); + seq_printf(m, "Requested P-state: %d\n", (rgvswctl >> 8) & 0xf); + seq_printf(m, "Requested VID: %d\n", rgvswctl & 0x3f); + seq_printf(m, "Current VID: %d\n", (rgvstat & MEMSTAT_VID_MASK) >> + MEMSTAT_VID_SHIFT); + seq_printf(m, "Current P-state: %d\n", + (rgvstat & MEMSTAT_PSTATE_MASK) >> MEMSTAT_PSTATE_SHIFT); return 0; } @@ -508,7 +513,8 @@ static int i915_delayfreq_table(struct seq_file *m, void *unused) for (i = 0; i < 16; i++) { delayfreq = I915_READ(PXVFREQ_BASE + i * 4); - seq_printf(m, "P%02dVIDFREQ: 0x%08x\n", i, delayfreq); + seq_printf(m, "P%02dVIDFREQ: 0x%08x (VID: %d)\n", i, delayfreq, + (delayfreq & PXVFREQ_PX_MASK) >> PXVFREQ_PX_SHIFT); } return 0; @@ -541,6 +547,8 @@ static int i915_drpc_info(struct seq_file *m, void *unused) struct drm_device *dev = node->minor->dev; drm_i915_private_t *dev_priv = dev->dev_private; u32 rgvmodectl = I915_READ(MEMMODECTL); + u32 rstdbyctl = I915_READ(MCHBAR_RENDER_STANDBY); + u16 crstandvid = I915_READ16(CRSTANDVID); seq_printf(m, "HD boost: %s\n", (rgvmodectl & MEMMODE_BOOST_EN) ? "yes" : "no"); @@ -555,9 +563,13 @@ static int i915_drpc_info(struct seq_file *m, void *unused) rgvmodectl & MEMMODE_RCLK_GATE ? "yes" : "no"); seq_printf(m, "Starting frequency: P%d\n", (rgvmodectl & MEMMODE_FSTART_MASK) >> MEMMODE_FSTART_SHIFT); - seq_printf(m, "Max frequency: P%d\n", + seq_printf(m, "Max P-state: P%d\n", (rgvmodectl & MEMMODE_FMAX_MASK) >> MEMMODE_FMAX_SHIFT); - seq_printf(m, "Min frequency: P%d\n", (rgvmodectl & MEMMODE_FMIN_MASK)); + seq_printf(m, "Min P-state: P%d\n", (rgvmodectl & MEMMODE_FMIN_MASK)); + seq_printf(m, "RS1 VID: %d\n", (crstandvid & 0x3f)); + seq_printf(m, "RS2 VID: %d\n", ((crstandvid >> 8) & 0x3f)); + seq_printf(m, "Render standby enabled: %s\n", + (rstdbyctl & RCX_SW_EXIT) ? "no" : "yes"); return 0; } @@ -608,7 +620,7 @@ static int i915_sr_status(struct seq_file *m, void *unused) drm_i915_private_t *dev_priv = dev->dev_private; bool sr_enabled = false; - if (IS_I965G(dev) || IS_I945G(dev) || IS_I945GM(dev)) + if (IS_I965GM(dev) || IS_I945G(dev) || IS_I945GM(dev)) sr_enabled = I915_READ(FW_BLC_SELF) & FW_BLC_SELF_EN; else if (IS_I915GM(dev)) sr_enabled = I915_READ(INSTPM) & INSTPM_SELF_EN; @@ -621,6 +633,36 @@ static int i915_sr_status(struct seq_file *m, void *unused) return 0; } +static int i915_emon_status(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + drm_i915_private_t *dev_priv = dev->dev_private; + unsigned long temp, chipset, gfx; + + temp = i915_mch_val(dev_priv); + chipset = i915_chipset_val(dev_priv); + gfx = i915_gfx_val(dev_priv); + + seq_printf(m, "GMCH temp: %ld\n", temp); + seq_printf(m, "Chipset power: %ld\n", chipset); + seq_printf(m, "GFX power: %ld\n", gfx); + seq_printf(m, "Total power: %ld\n", chipset + gfx); + + return 0; +} + +static int i915_gfxec(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + drm_i915_private_t *dev_priv = dev->dev_private; + + seq_printf(m, "GFXEC: %ld\n", (unsigned long)I915_READ(0x112f4)); + + return 0; +} + static int i915_wedged_open(struct inode *inode, struct file *filp) @@ -743,6 +785,8 @@ static struct drm_info_list i915_debugfs_list[] = { {"i915_delayfreq_table", i915_delayfreq_table, 0}, {"i915_inttoext_table", i915_inttoext_table, 0}, {"i915_drpc_info", i915_drpc_info, 0}, + {"i915_emon_status", i915_emon_status, 0}, + {"i915_gfxec", i915_gfxec, 0}, {"i915_fbc_status", i915_fbc_status, 0}, {"i915_sr_status", i915_sr_status, 0}, }; diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index 9bed5617e0ea..92898035845d 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -41,84 +41,6 @@ #include <linux/vga_switcheroo.h> #include <linux/slab.h> -/* Really want an OS-independent resettable timer. Would like to have - * this loop run for (eg) 3 sec, but have the timer reset every time - * the head pointer changes, so that EBUSY only happens if the ring - * actually stalls for (eg) 3 seconds. - */ -int i915_wait_ring(struct drm_device * dev, int n, const char *caller) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - drm_i915_ring_buffer_t *ring = &(dev_priv->ring); - u32 acthd_reg = IS_I965G(dev) ? ACTHD_I965 : ACTHD; - u32 last_acthd = I915_READ(acthd_reg); - u32 acthd; - u32 last_head = I915_READ(PRB0_HEAD) & HEAD_ADDR; - int i; - - trace_i915_ring_wait_begin (dev); - - for (i = 0; i < 100000; i++) { - ring->head = I915_READ(PRB0_HEAD) & HEAD_ADDR; - acthd = I915_READ(acthd_reg); - ring->space = ring->head - (ring->tail + 8); - if (ring->space < 0) - ring->space += ring->Size; - if (ring->space >= n) { - trace_i915_ring_wait_end (dev); - return 0; - } - - if (dev->primary->master) { - struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv; - if (master_priv->sarea_priv) - master_priv->sarea_priv->perf_boxes |= I915_BOX_WAIT; - } - - - if (ring->head != last_head) - i = 0; - if (acthd != last_acthd) - i = 0; - - last_head = ring->head; - last_acthd = acthd; - msleep_interruptible(10); - - } - - trace_i915_ring_wait_end (dev); - return -EBUSY; -} - -/* As a ringbuffer is only allowed to wrap between instructions, fill - * the tail with NOOPs. - */ -int i915_wrap_ring(struct drm_device *dev) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - volatile unsigned int *virt; - int rem; - - rem = dev_priv->ring.Size - dev_priv->ring.tail; - if (dev_priv->ring.space < rem) { - int ret = i915_wait_ring(dev, rem, __func__); - if (ret) - return ret; - } - dev_priv->ring.space -= rem; - - virt = (unsigned int *) - (dev_priv->ring.virtual_start + dev_priv->ring.tail); - rem /= 4; - while (rem--) - *virt++ = MI_NOOP; - - dev_priv->ring.tail = 0; - - return 0; -} - /** * Sets up the hardware status page for devices that need a physical address * in the register. @@ -134,10 +56,11 @@ static int i915_init_phys_hws(struct drm_device *dev) DRM_ERROR("Can not allocate hardware status page\n"); return -ENOMEM; } - dev_priv->hw_status_page = dev_priv->status_page_dmah->vaddr; + dev_priv->render_ring.status_page.page_addr + = dev_priv->status_page_dmah->vaddr; dev_priv->dma_status_page = dev_priv->status_page_dmah->busaddr; - memset(dev_priv->hw_status_page, 0, PAGE_SIZE); + memset(dev_priv->render_ring.status_page.page_addr, 0, PAGE_SIZE); if (IS_I965G(dev)) dev_priv->dma_status_page |= (dev_priv->dma_status_page >> 28) & @@ -160,8 +83,8 @@ static void i915_free_hws(struct drm_device *dev) dev_priv->status_page_dmah = NULL; } - if (dev_priv->status_gfx_addr) { - dev_priv->status_gfx_addr = 0; + if (dev_priv->render_ring.status_page.gfx_addr) { + dev_priv->render_ring.status_page.gfx_addr = 0; drm_core_ioremapfree(&dev_priv->hws_map, dev); } @@ -173,7 +96,7 @@ void i915_kernel_lost_context(struct drm_device * dev) { drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_master_private *master_priv; - drm_i915_ring_buffer_t *ring = &(dev_priv->ring); + struct intel_ring_buffer *ring = &dev_priv->render_ring; /* * We should never lose context on the ring with modesetting @@ -186,7 +109,7 @@ void i915_kernel_lost_context(struct drm_device * dev) ring->tail = I915_READ(PRB0_TAIL) & TAIL_ADDR; ring->space = ring->head - (ring->tail + 8); if (ring->space < 0) - ring->space += ring->Size; + ring->space += ring->size; if (!dev->primary->master) return; @@ -206,12 +129,11 @@ static int i915_dma_cleanup(struct drm_device * dev) if (dev->irq_enabled) drm_irq_uninstall(dev); - if (dev_priv->ring.virtual_start) { - drm_core_ioremapfree(&dev_priv->ring.map, dev); - dev_priv->ring.virtual_start = NULL; - dev_priv->ring.map.handle = NULL; - dev_priv->ring.map.size = 0; - } + mutex_lock(&dev->struct_mutex); + intel_cleanup_ring_buffer(dev, &dev_priv->render_ring); + if (HAS_BSD(dev)) + intel_cleanup_ring_buffer(dev, &dev_priv->bsd_ring); + mutex_unlock(&dev->struct_mutex); /* Clear the HWS virtual address at teardown */ if (I915_NEED_GFX_HWS(dev)) @@ -234,24 +156,24 @@ static int i915_initialize(struct drm_device * dev, drm_i915_init_t * init) } if (init->ring_size != 0) { - if (dev_priv->ring.ring_obj != NULL) { + if (dev_priv->render_ring.gem_object != NULL) { i915_dma_cleanup(dev); DRM_ERROR("Client tried to initialize ringbuffer in " "GEM mode\n"); return -EINVAL; } - dev_priv->ring.Size = init->ring_size; + dev_priv->render_ring.size = init->ring_size; - dev_priv->ring.map.offset = init->ring_start; - dev_priv->ring.map.size = init->ring_size; - dev_priv->ring.map.type = 0; - dev_priv->ring.map.flags = 0; - dev_priv->ring.map.mtrr = 0; + dev_priv->render_ring.map.offset = init->ring_start; + dev_priv->render_ring.map.size = init->ring_size; + dev_priv->render_ring.map.type = 0; + dev_priv->render_ring.map.flags = 0; + dev_priv->render_ring.map.mtrr = 0; - drm_core_ioremap_wc(&dev_priv->ring.map, dev); + drm_core_ioremap_wc(&dev_priv->render_ring.map, dev); - if (dev_priv->ring.map.handle == NULL) { + if (dev_priv->render_ring.map.handle == NULL) { i915_dma_cleanup(dev); DRM_ERROR("can not ioremap virtual address for" " ring buffer\n"); @@ -259,7 +181,7 @@ static int i915_initialize(struct drm_device * dev, drm_i915_init_t * init) } } - dev_priv->ring.virtual_start = dev_priv->ring.map.handle; + dev_priv->render_ring.virtual_start = dev_priv->render_ring.map.handle; dev_priv->cpp = init->cpp; dev_priv->back_offset = init->back_offset; @@ -279,26 +201,29 @@ static int i915_dma_resume(struct drm_device * dev) { drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; + struct intel_ring_buffer *ring; DRM_DEBUG_DRIVER("%s\n", __func__); - if (dev_priv->ring.map.handle == NULL) { + ring = &dev_priv->render_ring; + + if (ring->map.handle == NULL) { DRM_ERROR("can not ioremap virtual address for" " ring buffer\n"); return -ENOMEM; } /* Program Hardware Status Page */ - if (!dev_priv->hw_status_page) { + if (!ring->status_page.page_addr) { DRM_ERROR("Can not find hardware status page\n"); return -EINVAL; } DRM_DEBUG_DRIVER("hw status page @ %p\n", - dev_priv->hw_status_page); - - if (dev_priv->status_gfx_addr != 0) - I915_WRITE(HWS_PGA, dev_priv->status_gfx_addr); + ring->status_page.page_addr); + if (ring->status_page.gfx_addr != 0) + ring->setup_status_page(dev, ring); else I915_WRITE(HWS_PGA, dev_priv->dma_status_page); + DRM_DEBUG_DRIVER("Enabled hardware status page\n"); return 0; @@ -408,9 +333,8 @@ static int i915_emit_cmds(struct drm_device * dev, int *buffer, int dwords) { drm_i915_private_t *dev_priv = dev->dev_private; int i; - RING_LOCALS; - if ((dwords+1) * sizeof(int) >= dev_priv->ring.Size - 8) + if ((dwords+1) * sizeof(int) >= dev_priv->render_ring.size - 8) return -EINVAL; BEGIN_LP_RING((dwords+1)&~1); @@ -443,9 +367,7 @@ i915_emit_box(struct drm_device *dev, struct drm_clip_rect *boxes, int i, int DR1, int DR4) { - drm_i915_private_t *dev_priv = dev->dev_private; struct drm_clip_rect box = boxes[i]; - RING_LOCALS; if (box.y2 <= box.y1 || box.x2 <= box.x1 || box.y2 <= 0 || box.x2 <= 0) { DRM_ERROR("Bad box %d,%d..%d,%d\n", @@ -482,7 +404,6 @@ static void i915_emit_breadcrumb(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv; - RING_LOCALS; dev_priv->counter++; if (dev_priv->counter > 0x7FFFFFFFUL) @@ -536,10 +457,8 @@ static int i915_dispatch_batchbuffer(struct drm_device * dev, drm_i915_batchbuffer_t * batch, struct drm_clip_rect *cliprects) { - drm_i915_private_t *dev_priv = dev->dev_private; int nbox = batch->num_cliprects; int i = 0, count; - RING_LOCALS; if ((batch->start | batch->used) & 0x7) { DRM_ERROR("alignment"); @@ -588,7 +507,6 @@ static int i915_dispatch_flip(struct drm_device * dev) drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv; - RING_LOCALS; if (!master_priv->sarea_priv) return -EINVAL; @@ -641,7 +559,8 @@ static int i915_quiescent(struct drm_device * dev) drm_i915_private_t *dev_priv = dev->dev_private; i915_kernel_lost_context(dev); - return i915_wait_ring(dev, dev_priv->ring.Size - 8, __func__); + return intel_wait_ring_buffer(dev, &dev_priv->render_ring, + dev_priv->render_ring.size - 8); } static int i915_flush_ioctl(struct drm_device *dev, void *data, @@ -828,6 +747,9 @@ static int i915_getparam(struct drm_device *dev, void *data, /* depends on GEM */ value = dev_priv->has_gem; break; + case I915_PARAM_HAS_BSD: + value = HAS_BSD(dev); + break; default: DRM_DEBUG_DRIVER("Unknown parameter %d\n", param->param); @@ -883,6 +805,7 @@ static int i915_set_status_page(struct drm_device *dev, void *data, { drm_i915_private_t *dev_priv = dev->dev_private; drm_i915_hws_addr_t *hws = data; + struct intel_ring_buffer *ring = &dev_priv->render_ring; if (!I915_NEED_GFX_HWS(dev)) return -EINVAL; @@ -899,7 +822,7 @@ static int i915_set_status_page(struct drm_device *dev, void *data, DRM_DEBUG_DRIVER("set status page addr 0x%08x\n", (u32)hws->addr); - dev_priv->status_gfx_addr = hws->addr & (0x1ffff<<12); + ring->status_page.gfx_addr = hws->addr & (0x1ffff<<12); dev_priv->hws_map.offset = dev->agp->base + hws->addr; dev_priv->hws_map.size = 4*1024; @@ -910,19 +833,19 @@ static int i915_set_status_page(struct drm_device *dev, void *data, drm_core_ioremap_wc(&dev_priv->hws_map, dev); if (dev_priv->hws_map.handle == NULL) { i915_dma_cleanup(dev); - dev_priv->status_gfx_addr = 0; + ring->status_page.gfx_addr = 0; DRM_ERROR("can not ioremap virtual address for" " G33 hw status page\n"); return -ENOMEM; } - dev_priv->hw_status_page = dev_priv->hws_map.handle; + ring->status_page.page_addr = dev_priv->hws_map.handle; + memset(ring->status_page.page_addr, 0, PAGE_SIZE); + I915_WRITE(HWS_PGA, ring->status_page.gfx_addr); - memset(dev_priv->hw_status_page, 0, PAGE_SIZE); - I915_WRITE(HWS_PGA, dev_priv->status_gfx_addr); DRM_DEBUG_DRIVER("load hws HWS_PGA with gfx mem 0x%x\n", - dev_priv->status_gfx_addr); + ring->status_page.gfx_addr); DRM_DEBUG_DRIVER("load hws at %p\n", - dev_priv->hw_status_page); + ring->status_page.page_addr); return 0; } @@ -1309,7 +1232,7 @@ static void i915_warn_stolen(struct drm_device *dev) static void i915_setup_compression(struct drm_device *dev, int size) { struct drm_i915_private *dev_priv = dev->dev_private; - struct drm_mm_node *compressed_fb, *compressed_llb; + struct drm_mm_node *compressed_fb, *uninitialized_var(compressed_llb); unsigned long cfb_base; unsigned long ll_base = 0; @@ -1400,12 +1323,14 @@ static void i915_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_ struct drm_device *dev = pci_get_drvdata(pdev); pm_message_t pmm = { .event = PM_EVENT_SUSPEND }; if (state == VGA_SWITCHEROO_ON) { - printk(KERN_INFO "i915: switched off\n"); + printk(KERN_INFO "i915: switched on\n"); /* i915 resume handler doesn't set to D0 */ pci_set_power_state(dev->pdev, PCI_D0); i915_resume(dev); + drm_kms_helper_poll_enable(dev); } else { printk(KERN_ERR "i915: switched off\n"); + drm_kms_helper_poll_disable(dev); i915_suspend(dev, pmm); } } @@ -1480,19 +1405,23 @@ static int i915_load_modeset_init(struct drm_device *dev, /* if we have > 1 VGA cards, then disable the radeon VGA resources */ ret = vga_client_register(dev->pdev, dev, NULL, i915_vga_set_decode); if (ret) - goto destroy_ringbuffer; + goto cleanup_ringbuffer; ret = vga_switcheroo_register_client(dev->pdev, i915_switcheroo_set_state, i915_switcheroo_can_switch); if (ret) - goto destroy_ringbuffer; + goto cleanup_vga_client; + + /* IIR "flip pending" bit means done if this bit is set */ + if (IS_GEN3(dev) && (I915_READ(ECOSKPD) & ECO_FLIP_DONE)) + dev_priv->flip_pending_is_done = true; intel_modeset_init(dev); ret = drm_irq_install(dev); if (ret) - goto destroy_ringbuffer; + goto cleanup_vga_switcheroo; /* Always safe in the mode setting case. */ /* FIXME: do pre/post-mode set stuff in core KMS code */ @@ -1504,11 +1433,20 @@ static int i915_load_modeset_init(struct drm_device *dev, I915_WRITE(INSTPM, (1 << 5) | (1 << 21)); - intel_fbdev_init(dev); + ret = intel_fbdev_init(dev); + if (ret) + goto cleanup_irq; + drm_kms_helper_poll_init(dev); return 0; -destroy_ringbuffer: +cleanup_irq: + drm_irq_uninstall(dev); +cleanup_vga_switcheroo: + vga_switcheroo_unregister_client(dev->pdev); +cleanup_vga_client: + vga_client_register(dev->pdev, NULL, NULL, NULL); +cleanup_ringbuffer: mutex_lock(&dev->struct_mutex); i915_gem_cleanup_ringbuffer(dev); mutex_unlock(&dev->struct_mutex); @@ -1540,14 +1478,11 @@ void i915_master_destroy(struct drm_device *dev, struct drm_master *master) master->driver_priv = NULL; } -static void i915_get_mem_freq(struct drm_device *dev) +static void i915_pineview_get_mem_freq(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; u32 tmp; - if (!IS_PINEVIEW(dev)) - return; - tmp = I915_READ(CLKCFG); switch (tmp & CLKCFG_FSB_MASK) { @@ -1576,8 +1511,525 @@ static void i915_get_mem_freq(struct drm_device *dev) dev_priv->mem_freq = 800; break; } + + /* detect pineview DDR3 setting */ + tmp = I915_READ(CSHRDDR3CTL); + dev_priv->is_ddr3 = (tmp & CSHRDDR3CTL_DDR3) ? 1 : 0; +} + +static void i915_ironlake_get_mem_freq(struct drm_device *dev) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + u16 ddrpll, csipll; + + ddrpll = I915_READ16(DDRMPLL1); + csipll = I915_READ16(CSIPLL0); + + switch (ddrpll & 0xff) { + case 0xc: + dev_priv->mem_freq = 800; + break; + case 0x10: + dev_priv->mem_freq = 1066; + break; + case 0x14: + dev_priv->mem_freq = 1333; + break; + case 0x18: + dev_priv->mem_freq = 1600; + break; + default: + DRM_DEBUG_DRIVER("unknown memory frequency 0x%02x\n", + ddrpll & 0xff); + dev_priv->mem_freq = 0; + break; + } + + dev_priv->r_t = dev_priv->mem_freq; + + switch (csipll & 0x3ff) { + case 0x00c: + dev_priv->fsb_freq = 3200; + break; + case 0x00e: + dev_priv->fsb_freq = 3733; + break; + case 0x010: + dev_priv->fsb_freq = 4266; + break; + case 0x012: + dev_priv->fsb_freq = 4800; + break; + case 0x014: + dev_priv->fsb_freq = 5333; + break; + case 0x016: + dev_priv->fsb_freq = 5866; + break; + case 0x018: + dev_priv->fsb_freq = 6400; + break; + default: + DRM_DEBUG_DRIVER("unknown fsb frequency 0x%04x\n", + csipll & 0x3ff); + dev_priv->fsb_freq = 0; + break; + } + + if (dev_priv->fsb_freq == 3200) { + dev_priv->c_m = 0; + } else if (dev_priv->fsb_freq > 3200 && dev_priv->fsb_freq <= 4800) { + dev_priv->c_m = 1; + } else { + dev_priv->c_m = 2; + } +} + +struct v_table { + u8 vid; + unsigned long vd; /* in .1 mil */ + unsigned long vm; /* in .1 mil */ + u8 pvid; +}; + +static struct v_table v_table[] = { + { 0, 16125, 15000, 0x7f, }, + { 1, 16000, 14875, 0x7e, }, + { 2, 15875, 14750, 0x7d, }, + { 3, 15750, 14625, 0x7c, }, + { 4, 15625, 14500, 0x7b, }, + { 5, 15500, 14375, 0x7a, }, + { 6, 15375, 14250, 0x79, }, + { 7, 15250, 14125, 0x78, }, + { 8, 15125, 14000, 0x77, }, + { 9, 15000, 13875, 0x76, }, + { 10, 14875, 13750, 0x75, }, + { 11, 14750, 13625, 0x74, }, + { 12, 14625, 13500, 0x73, }, + { 13, 14500, 13375, 0x72, }, + { 14, 14375, 13250, 0x71, }, + { 15, 14250, 13125, 0x70, }, + { 16, 14125, 13000, 0x6f, }, + { 17, 14000, 12875, 0x6e, }, + { 18, 13875, 12750, 0x6d, }, + { 19, 13750, 12625, 0x6c, }, + { 20, 13625, 12500, 0x6b, }, + { 21, 13500, 12375, 0x6a, }, + { 22, 13375, 12250, 0x69, }, + { 23, 13250, 12125, 0x68, }, + { 24, 13125, 12000, 0x67, }, + { 25, 13000, 11875, 0x66, }, + { 26, 12875, 11750, 0x65, }, + { 27, 12750, 11625, 0x64, }, + { 28, 12625, 11500, 0x63, }, + { 29, 12500, 11375, 0x62, }, + { 30, 12375, 11250, 0x61, }, + { 31, 12250, 11125, 0x60, }, + { 32, 12125, 11000, 0x5f, }, + { 33, 12000, 10875, 0x5e, }, + { 34, 11875, 10750, 0x5d, }, + { 35, 11750, 10625, 0x5c, }, + { 36, 11625, 10500, 0x5b, }, + { 37, 11500, 10375, 0x5a, }, + { 38, 11375, 10250, 0x59, }, + { 39, 11250, 10125, 0x58, }, + { 40, 11125, 10000, 0x57, }, + { 41, 11000, 9875, 0x56, }, + { 42, 10875, 9750, 0x55, }, + { 43, 10750, 9625, 0x54, }, + { 44, 10625, 9500, 0x53, }, + { 45, 10500, 9375, 0x52, }, + { 46, 10375, 9250, 0x51, }, + { 47, 10250, 9125, 0x50, }, + { 48, 10125, 9000, 0x4f, }, + { 49, 10000, 8875, 0x4e, }, + { 50, 9875, 8750, 0x4d, }, + { 51, 9750, 8625, 0x4c, }, + { 52, 9625, 8500, 0x4b, }, + { 53, 9500, 8375, 0x4a, }, + { 54, 9375, 8250, 0x49, }, + { 55, 9250, 8125, 0x48, }, + { 56, 9125, 8000, 0x47, }, + { 57, 9000, 7875, 0x46, }, + { 58, 8875, 7750, 0x45, }, + { 59, 8750, 7625, 0x44, }, + { 60, 8625, 7500, 0x43, }, + { 61, 8500, 7375, 0x42, }, + { 62, 8375, 7250, 0x41, }, + { 63, 8250, 7125, 0x40, }, + { 64, 8125, 7000, 0x3f, }, + { 65, 8000, 6875, 0x3e, }, + { 66, 7875, 6750, 0x3d, }, + { 67, 7750, 6625, 0x3c, }, + { 68, 7625, 6500, 0x3b, }, + { 69, 7500, 6375, 0x3a, }, + { 70, 7375, 6250, 0x39, }, + { 71, 7250, 6125, 0x38, }, + { 72, 7125, 6000, 0x37, }, + { 73, 7000, 5875, 0x36, }, + { 74, 6875, 5750, 0x35, }, + { 75, 6750, 5625, 0x34, }, + { 76, 6625, 5500, 0x33, }, + { 77, 6500, 5375, 0x32, }, + { 78, 6375, 5250, 0x31, }, + { 79, 6250, 5125, 0x30, }, + { 80, 6125, 5000, 0x2f, }, + { 81, 6000, 4875, 0x2e, }, + { 82, 5875, 4750, 0x2d, }, + { 83, 5750, 4625, 0x2c, }, + { 84, 5625, 4500, 0x2b, }, + { 85, 5500, 4375, 0x2a, }, + { 86, 5375, 4250, 0x29, }, + { 87, 5250, 4125, 0x28, }, + { 88, 5125, 4000, 0x27, }, + { 89, 5000, 3875, 0x26, }, + { 90, 4875, 3750, 0x25, }, + { 91, 4750, 3625, 0x24, }, + { 92, 4625, 3500, 0x23, }, + { 93, 4500, 3375, 0x22, }, + { 94, 4375, 3250, 0x21, }, + { 95, 4250, 3125, 0x20, }, + { 96, 4125, 3000, 0x1f, }, + { 97, 4125, 3000, 0x1e, }, + { 98, 4125, 3000, 0x1d, }, + { 99, 4125, 3000, 0x1c, }, + { 100, 4125, 3000, 0x1b, }, + { 101, 4125, 3000, 0x1a, }, + { 102, 4125, 3000, 0x19, }, + { 103, 4125, 3000, 0x18, }, + { 104, 4125, 3000, 0x17, }, + { 105, 4125, 3000, 0x16, }, + { 106, 4125, 3000, 0x15, }, + { 107, 4125, 3000, 0x14, }, + { 108, 4125, 3000, 0x13, }, + { 109, 4125, 3000, 0x12, }, + { 110, 4125, 3000, 0x11, }, + { 111, 4125, 3000, 0x10, }, + { 112, 4125, 3000, 0x0f, }, + { 113, 4125, 3000, 0x0e, }, + { 114, 4125, 3000, 0x0d, }, + { 115, 4125, 3000, 0x0c, }, + { 116, 4125, 3000, 0x0b, }, + { 117, 4125, 3000, 0x0a, }, + { 118, 4125, 3000, 0x09, }, + { 119, 4125, 3000, 0x08, }, + { 120, 1125, 0, 0x07, }, + { 121, 1000, 0, 0x06, }, + { 122, 875, 0, 0x05, }, + { 123, 750, 0, 0x04, }, + { 124, 625, 0, 0x03, }, + { 125, 500, 0, 0x02, }, + { 126, 375, 0, 0x01, }, + { 127, 0, 0, 0x00, }, +}; + +struct cparams { + int i; + int t; + int m; + int c; +}; + +static struct cparams cparams[] = { + { 1, 1333, 301, 28664 }, + { 1, 1066, 294, 24460 }, + { 1, 800, 294, 25192 }, + { 0, 1333, 276, 27605 }, + { 0, 1066, 276, 27605 }, + { 0, 800, 231, 23784 }, +}; + +unsigned long i915_chipset_val(struct drm_i915_private *dev_priv) +{ + u64 total_count, diff, ret; + u32 count1, count2, count3, m = 0, c = 0; + unsigned long now = jiffies_to_msecs(jiffies), diff1; + int i; + + diff1 = now - dev_priv->last_time1; + + count1 = I915_READ(DMIEC); + count2 = I915_READ(DDREC); + count3 = I915_READ(CSIEC); + + total_count = count1 + count2 + count3; + + /* FIXME: handle per-counter overflow */ + if (total_count < dev_priv->last_count1) { + diff = ~0UL - dev_priv->last_count1; + diff += total_count; + } else { + diff = total_count - dev_priv->last_count1; + } + + for (i = 0; i < ARRAY_SIZE(cparams); i++) { + if (cparams[i].i == dev_priv->c_m && + cparams[i].t == dev_priv->r_t) { + m = cparams[i].m; + c = cparams[i].c; + break; + } + } + + div_u64(diff, diff1); + ret = ((m * diff) + c); + div_u64(ret, 10); + + dev_priv->last_count1 = total_count; + dev_priv->last_time1 = now; + + return ret; } +unsigned long i915_mch_val(struct drm_i915_private *dev_priv) +{ + unsigned long m, x, b; + u32 tsfs; + + tsfs = I915_READ(TSFS); + + m = ((tsfs & TSFS_SLOPE_MASK) >> TSFS_SLOPE_SHIFT); + x = I915_READ8(TR1); + + b = tsfs & TSFS_INTR_MASK; + + return ((m * x) / 127) - b; +} + +static unsigned long pvid_to_extvid(struct drm_i915_private *dev_priv, u8 pxvid) +{ + unsigned long val = 0; + int i; + + for (i = 0; i < ARRAY_SIZE(v_table); i++) { + if (v_table[i].pvid == pxvid) { + if (IS_MOBILE(dev_priv->dev)) + val = v_table[i].vm; + else + val = v_table[i].vd; + } + } + + return val; +} + +void i915_update_gfx_val(struct drm_i915_private *dev_priv) +{ + struct timespec now, diff1; + u64 diff; + unsigned long diffms; + u32 count; + + getrawmonotonic(&now); + diff1 = timespec_sub(now, dev_priv->last_time2); + + /* Don't divide by 0 */ + diffms = diff1.tv_sec * 1000 + diff1.tv_nsec / 1000000; + if (!diffms) + return; + + count = I915_READ(GFXEC); + + if (count < dev_priv->last_count2) { + diff = ~0UL - dev_priv->last_count2; + diff += count; + } else { + diff = count - dev_priv->last_count2; + } + + dev_priv->last_count2 = count; + dev_priv->last_time2 = now; + + /* More magic constants... */ + diff = diff * 1181; + div_u64(diff, diffms * 10); + dev_priv->gfx_power = diff; +} + +unsigned long i915_gfx_val(struct drm_i915_private *dev_priv) +{ + unsigned long t, corr, state1, corr2, state2; + u32 pxvid, ext_v; + + pxvid = I915_READ(PXVFREQ_BASE + (dev_priv->cur_delay * 4)); + pxvid = (pxvid >> 24) & 0x7f; + ext_v = pvid_to_extvid(dev_priv, pxvid); + + state1 = ext_v; + + t = i915_mch_val(dev_priv); + + /* Revel in the empirically derived constants */ + + /* Correction factor in 1/100000 units */ + if (t > 80) + corr = ((t * 2349) + 135940); + else if (t >= 50) + corr = ((t * 964) + 29317); + else /* < 50 */ + corr = ((t * 301) + 1004); + + corr = corr * ((150142 * state1) / 10000 - 78642); + corr /= 100000; + corr2 = (corr * dev_priv->corr); + + state2 = (corr2 * state1) / 10000; + state2 /= 100; /* convert to mW */ + + i915_update_gfx_val(dev_priv); + + return dev_priv->gfx_power + state2; +} + +/* Global for IPS driver to get at the current i915 device */ +static struct drm_i915_private *i915_mch_dev; +/* + * Lock protecting IPS related data structures + * - i915_mch_dev + * - dev_priv->max_delay + * - dev_priv->min_delay + * - dev_priv->fmax + * - dev_priv->gpu_busy + */ +DEFINE_SPINLOCK(mchdev_lock); + +/** + * i915_read_mch_val - return value for IPS use + * + * Calculate and return a value for the IPS driver to use when deciding whether + * we have thermal and power headroom to increase CPU or GPU power budget. + */ +unsigned long i915_read_mch_val(void) +{ + struct drm_i915_private *dev_priv; + unsigned long chipset_val, graphics_val, ret = 0; + + spin_lock(&mchdev_lock); + if (!i915_mch_dev) + goto out_unlock; + dev_priv = i915_mch_dev; + + chipset_val = i915_chipset_val(dev_priv); + graphics_val = i915_gfx_val(dev_priv); + + ret = chipset_val + graphics_val; + +out_unlock: + spin_unlock(&mchdev_lock); + + return ret; +} +EXPORT_SYMBOL_GPL(i915_read_mch_val); + +/** + * i915_gpu_raise - raise GPU frequency limit + * + * Raise the limit; IPS indicates we have thermal headroom. + */ +bool i915_gpu_raise(void) +{ + struct drm_i915_private *dev_priv; + bool ret = true; + + spin_lock(&mchdev_lock); + if (!i915_mch_dev) { + ret = false; + goto out_unlock; + } + dev_priv = i915_mch_dev; + + if (dev_priv->max_delay > dev_priv->fmax) + dev_priv->max_delay--; + +out_unlock: + spin_unlock(&mchdev_lock); + + return ret; +} +EXPORT_SYMBOL_GPL(i915_gpu_raise); + +/** + * i915_gpu_lower - lower GPU frequency limit + * + * IPS indicates we're close to a thermal limit, so throttle back the GPU + * frequency maximum. + */ +bool i915_gpu_lower(void) +{ + struct drm_i915_private *dev_priv; + bool ret = true; + + spin_lock(&mchdev_lock); + if (!i915_mch_dev) { + ret = false; + goto out_unlock; + } + dev_priv = i915_mch_dev; + + if (dev_priv->max_delay < dev_priv->min_delay) + dev_priv->max_delay++; + +out_unlock: + spin_unlock(&mchdev_lock); + + return ret; +} +EXPORT_SYMBOL_GPL(i915_gpu_lower); + +/** + * i915_gpu_busy - indicate GPU business to IPS + * + * Tell the IPS driver whether or not the GPU is busy. + */ +bool i915_gpu_busy(void) +{ + struct drm_i915_private *dev_priv; + bool ret = false; + + spin_lock(&mchdev_lock); + if (!i915_mch_dev) + goto out_unlock; + dev_priv = i915_mch_dev; + + ret = dev_priv->busy; + +out_unlock: + spin_unlock(&mchdev_lock); + + return ret; +} +EXPORT_SYMBOL_GPL(i915_gpu_busy); + +/** + * i915_gpu_turbo_disable - disable graphics turbo + * + * Disable graphics turbo by resetting the max frequency and setting the + * current frequency to the default. + */ +bool i915_gpu_turbo_disable(void) +{ + struct drm_i915_private *dev_priv; + bool ret = true; + + spin_lock(&mchdev_lock); + if (!i915_mch_dev) { + ret = false; + goto out_unlock; + } + dev_priv = i915_mch_dev; + + dev_priv->max_delay = dev_priv->fstart; + + if (!ironlake_set_drps(dev_priv->dev, dev_priv->fstart)) + ret = false; + +out_unlock: + spin_unlock(&mchdev_lock); + + return ret; +} +EXPORT_SYMBOL_GPL(i915_gpu_turbo_disable); + /** * i915_driver_load - setup chip and create an initial config * @dev: DRM device @@ -1595,7 +2047,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) resource_size_t base, size; int ret = 0, mmio_bar; uint32_t agp_size, prealloc_size, prealloc_start; - /* i915 has 4 more counters */ dev->counters += 4; dev->types[6] = _DRM_STAT_IRQ; @@ -1673,6 +2124,13 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) dev_priv->has_gem = 0; } + if (dev_priv->has_gem == 0 && + drm_core_check_feature(dev, DRIVER_MODESET)) { + DRM_ERROR("kernel modesetting requires GEM, disabling driver.\n"); + ret = -ENODEV; + goto out_iomapfree; + } + dev->driver->get_vblank_counter = i915_get_vblank_counter; dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */ if (IS_G4X(dev) || IS_IRONLAKE(dev) || IS_GEN6(dev)) { @@ -1692,7 +2150,10 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) goto out_workqueue_free; } - i915_get_mem_freq(dev); + if (IS_PINEVIEW(dev)) + i915_pineview_get_mem_freq(dev); + else if (IS_IRONLAKE(dev)) + i915_ironlake_get_mem_freq(dev); /* On the 945G/GM, the chipset reports the MSI capability on the * integrated graphics even though the support isn't actually there @@ -1710,7 +2171,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) spin_lock_init(&dev_priv->user_irq_lock); spin_lock_init(&dev_priv->error_lock); - dev_priv->user_irq_refcount = 0; dev_priv->trace_irq_seqno = 0; ret = drm_vblank_init(dev, I915_NUM_PIPE); @@ -1739,6 +2199,12 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) setup_timer(&dev_priv->hangcheck_timer, i915_hangcheck_elapsed, (unsigned long) dev); + + spin_lock(&mchdev_lock); + i915_mch_dev = dev_priv; + dev_priv->mchdev_lock = &mchdev_lock; + spin_unlock(&mchdev_lock); + return 0; out_workqueue_free: @@ -1760,6 +2226,10 @@ int i915_driver_unload(struct drm_device *dev) i915_destroy_error_state(dev); + spin_lock(&mchdev_lock); + i915_mch_dev = NULL; + spin_unlock(&mchdev_lock); + destroy_workqueue(dev_priv->wq); del_timer_sync(&dev_priv->hangcheck_timer); diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index b7aecf5ea1fc..65d3f3e8475b 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -60,95 +60,95 @@ extern int intel_agp_enabled; .subdevice = PCI_ANY_ID, \ .driver_data = (unsigned long) info } -const static struct intel_device_info intel_i830_info = { +static const struct intel_device_info intel_i830_info = { .is_i8xx = 1, .is_mobile = 1, .cursor_needs_physical = 1, }; -const static struct intel_device_info intel_845g_info = { +static const struct intel_device_info intel_845g_info = { .is_i8xx = 1, }; -const static struct intel_device_info intel_i85x_info = { +static const struct intel_device_info intel_i85x_info = { .is_i8xx = 1, .is_i85x = 1, .is_mobile = 1, .cursor_needs_physical = 1, }; -const static struct intel_device_info intel_i865g_info = { +static const struct intel_device_info intel_i865g_info = { .is_i8xx = 1, }; -const static struct intel_device_info intel_i915g_info = { +static const struct intel_device_info intel_i915g_info = { .is_i915g = 1, .is_i9xx = 1, .cursor_needs_physical = 1, }; -const static struct intel_device_info intel_i915gm_info = { +static const struct intel_device_info intel_i915gm_info = { .is_i9xx = 1, .is_mobile = 1, .cursor_needs_physical = 1, }; -const static struct intel_device_info intel_i945g_info = { +static const struct intel_device_info intel_i945g_info = { .is_i9xx = 1, .has_hotplug = 1, .cursor_needs_physical = 1, }; -const static struct intel_device_info intel_i945gm_info = { +static const struct intel_device_info intel_i945gm_info = { .is_i945gm = 1, .is_i9xx = 1, .is_mobile = 1, .has_hotplug = 1, .cursor_needs_physical = 1, }; -const static struct intel_device_info intel_i965g_info = { +static const struct intel_device_info intel_i965g_info = { .is_i965g = 1, .is_i9xx = 1, .has_hotplug = 1, }; -const static struct intel_device_info intel_i965gm_info = { +static const struct intel_device_info intel_i965gm_info = { .is_i965g = 1, .is_mobile = 1, .is_i965gm = 1, .is_i9xx = 1, .is_mobile = 1, .has_fbc = 1, .has_rc6 = 1, .has_hotplug = 1, }; -const static struct intel_device_info intel_g33_info = { +static const struct intel_device_info intel_g33_info = { .is_g33 = 1, .is_i9xx = 1, .need_gfx_hws = 1, .has_hotplug = 1, }; -const static struct intel_device_info intel_g45_info = { +static const struct intel_device_info intel_g45_info = { .is_i965g = 1, .is_g4x = 1, .is_i9xx = 1, .need_gfx_hws = 1, .has_pipe_cxsr = 1, .has_hotplug = 1, }; -const static struct intel_device_info intel_gm45_info = { +static const struct intel_device_info intel_gm45_info = { .is_i965g = 1, .is_mobile = 1, .is_g4x = 1, .is_i9xx = 1, .is_mobile = 1, .need_gfx_hws = 1, .has_fbc = 1, .has_rc6 = 1, .has_pipe_cxsr = 1, .has_hotplug = 1, }; -const static struct intel_device_info intel_pineview_info = { +static const struct intel_device_info intel_pineview_info = { .is_g33 = 1, .is_pineview = 1, .is_mobile = 1, .is_i9xx = 1, .need_gfx_hws = 1, .has_hotplug = 1, }; -const static struct intel_device_info intel_ironlake_d_info = { +static const struct intel_device_info intel_ironlake_d_info = { .is_ironlake = 1, .is_i965g = 1, .is_i9xx = 1, .need_gfx_hws = 1, .has_pipe_cxsr = 1, .has_hotplug = 1, }; -const static struct intel_device_info intel_ironlake_m_info = { +static const struct intel_device_info intel_ironlake_m_info = { .is_ironlake = 1, .is_mobile = 1, .is_i965g = 1, .is_i9xx = 1, .need_gfx_hws = 1, .has_rc6 = 1, .has_hotplug = 1, }; -const static struct intel_device_info intel_sandybridge_d_info = { +static const struct intel_device_info intel_sandybridge_d_info = { .is_i965g = 1, .is_i9xx = 1, .need_gfx_hws = 1, .has_hotplug = 1, .is_gen6 = 1, }; -const static struct intel_device_info intel_sandybridge_m_info = { +static const struct intel_device_info intel_sandybridge_m_info = { .is_i965g = 1, .is_mobile = 1, .is_i9xx = 1, .need_gfx_hws = 1, .has_hotplug = 1, .is_gen6 = 1, }; -const static struct pci_device_id pciidlist[] = { +static const struct pci_device_id pciidlist[] = { INTEL_VGA_DEVICE(0x3577, &intel_i830_info), INTEL_VGA_DEVICE(0x2562, &intel_845g_info), INTEL_VGA_DEVICE(0x3582, &intel_i85x_info), @@ -340,7 +340,7 @@ int i965_reset(struct drm_device *dev, u8 flags) /* * Clear request list */ - i915_gem_retire_requests(dev); + i915_gem_retire_requests(dev, &dev_priv->render_ring); if (need_display) i915_save_display(dev); @@ -370,6 +370,7 @@ int i965_reset(struct drm_device *dev, u8 flags) } } else { DRM_ERROR("Error occurred. Don't know how to reset this chip.\n"); + mutex_unlock(&dev->struct_mutex); return -ENODEV; } @@ -388,33 +389,10 @@ int i965_reset(struct drm_device *dev, u8 flags) * switched away). */ if (drm_core_check_feature(dev, DRIVER_MODESET) || - !dev_priv->mm.suspended) { - drm_i915_ring_buffer_t *ring = &dev_priv->ring; - struct drm_gem_object *obj = ring->ring_obj; - struct drm_i915_gem_object *obj_priv = to_intel_bo(obj); + !dev_priv->mm.suspended) { + struct intel_ring_buffer *ring = &dev_priv->render_ring; dev_priv->mm.suspended = 0; - - /* Stop the ring if it's running. */ - I915_WRITE(PRB0_CTL, 0); - I915_WRITE(PRB0_TAIL, 0); - I915_WRITE(PRB0_HEAD, 0); - - /* Initialize the ring. */ - I915_WRITE(PRB0_START, obj_priv->gtt_offset); - I915_WRITE(PRB0_CTL, - ((obj->size - 4096) & RING_NR_PAGES) | - RING_NO_REPORT | - RING_VALID); - if (!drm_core_check_feature(dev, DRIVER_MODESET)) - i915_kernel_lost_context(dev); - else { - ring->head = I915_READ(PRB0_HEAD) & HEAD_ADDR; - ring->tail = I915_READ(PRB0_TAIL) & TAIL_ADDR; - ring->space = ring->head - (ring->tail + 8); - if (ring->space < 0) - ring->space += ring->Size; - } - + ring->init(dev, ring); mutex_unlock(&dev->struct_mutex); drm_irq_uninstall(dev); drm_irq_install(dev); diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 7f797ef1ab39..d147ab2f5bfc 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -32,6 +32,7 @@ #include "i915_reg.h" #include "intel_bios.h" +#include "intel_ringbuffer.h" #include <linux/io-mapping.h> /* General customization: @@ -55,6 +56,8 @@ enum plane { #define I915_NUM_PIPE 2 +#define I915_GEM_GPU_DOMAINS (~(I915_GEM_DOMAIN_CPU | I915_GEM_DOMAIN_GTT)) + /* Interface history: * * 1.1: Original. @@ -89,16 +92,6 @@ struct drm_i915_gem_phys_object { struct drm_gem_object *cur_obj; }; -typedef struct _drm_i915_ring_buffer { - unsigned long Size; - u8 *virtual_start; - int head; - int tail; - int space; - drm_local_map_t map; - struct drm_gem_object *ring_obj; -} drm_i915_ring_buffer_t; - struct mem_block { struct mem_block *next; struct mem_block *prev; @@ -241,17 +234,15 @@ typedef struct drm_i915_private { void __iomem *regs; struct pci_dev *bridge_dev; - drm_i915_ring_buffer_t ring; + struct intel_ring_buffer render_ring; + struct intel_ring_buffer bsd_ring; drm_dma_handle_t *status_page_dmah; - void *hw_status_page; void *seqno_page; dma_addr_t dma_status_page; uint32_t counter; - unsigned int status_gfx_addr; unsigned int seqno_gfx_addr; drm_local_map_t hws_map; - struct drm_gem_object *hws_obj; struct drm_gem_object *seqno_obj; struct drm_gem_object *pwrctx; @@ -267,8 +258,6 @@ typedef struct drm_i915_private { atomic_t irq_received; /** Protects user_irq_refcount and irq_mask_reg */ spinlock_t user_irq_lock; - /** Refcount for i915_user_irq_get() versus i915_user_irq_put(). */ - int user_irq_refcount; u32 trace_irq_seqno; /** Cached value of IMR to avoid reads in updating the bitfield */ u32 irq_mask_reg; @@ -289,6 +278,7 @@ typedef struct drm_i915_private { struct mem_block *agp_heap; unsigned int sr01, adpa, ppcr, dvob, dvoc, lvds; int vblank_pipe; + int num_pipe; /* For hangcheck timer */ #define DRM_I915_HANGCHECK_PERIOD 75 /* in jiffies */ @@ -334,7 +324,7 @@ typedef struct drm_i915_private { int fence_reg_start; /* 4 if userland hasn't ioctl'd us yet */ int num_fence_regs; /* 8 on pre-965, 16 otherwise */ - unsigned int fsb_freq, mem_freq; + unsigned int fsb_freq, mem_freq, is_ddr3; spinlock_t error_lock; struct drm_i915_error_state *first_error; @@ -514,18 +504,7 @@ typedef struct drm_i915_private { */ struct list_head shrink_list; - /** - * List of objects currently involved in rendering from the - * ringbuffer. - * - * Includes buffers having the contents of their GPU caches - * flushed, not necessarily primitives. last_rendering_seqno - * represents when the rendering involved will be completed. - * - * A reference is held on the buffer while on this list. - */ spinlock_t active_list_lock; - struct list_head active_list; /** * List of objects which are not in the ringbuffer but which @@ -563,12 +542,6 @@ typedef struct drm_i915_private { struct list_head fence_list; /** - * List of breadcrumbs associated with GPU requests currently - * outstanding. - */ - struct list_head request_list; - - /** * We leave the user IRQ off as much as possible, * but this means that requests will finish and never * be retired once the system goes idle. Set a timer to @@ -623,6 +596,7 @@ typedef struct drm_i915_private { struct drm_crtc *plane_to_crtc_mapping[2]; struct drm_crtc *pipe_to_crtc_mapping[2]; wait_queue_head_t pending_flip_queue; + bool flip_pending_is_done; /* Reclocking support */ bool render_reclock_avail; @@ -644,6 +618,18 @@ typedef struct drm_i915_private { u8 cur_delay; u8 min_delay; u8 max_delay; + u8 fmax; + u8 fstart; + + u64 last_count1; + unsigned long last_time1; + u64 last_count2; + struct timespec last_time2; + unsigned long gfx_power; + int c_m; + int r_t; + u8 corr; + spinlock_t *mchdev_lock; enum no_fbc_reason no_fbc_reason; @@ -671,19 +657,64 @@ struct drm_i915_gem_object { * (has pending rendering), and is not set if it's on inactive (ready * to be unbound). */ - int active; + unsigned int active : 1; /** * This is set if the object has been written to since last bound * to the GTT */ - int dirty; + unsigned int dirty : 1; + + /** + * Fence register bits (if any) for this object. Will be set + * as needed when mapped into the GTT. + * Protected by dev->struct_mutex. + * + * Size: 4 bits for 16 fences + sign (for FENCE_REG_NONE) + */ + int fence_reg : 5; + + /** + * Used for checking the object doesn't appear more than once + * in an execbuffer object list. + */ + unsigned int in_execbuffer : 1; + + /** + * Advice: are the backing pages purgeable? + */ + unsigned int madv : 2; + + /** + * Refcount for the pages array. With the current locking scheme, there + * are at most two concurrent users: Binding a bo to the gtt and + * pwrite/pread using physical addresses. So two bits for a maximum + * of two users are enough. + */ + unsigned int pages_refcount : 2; +#define DRM_I915_GEM_OBJECT_MAX_PAGES_REFCOUNT 0x3 + + /** + * Current tiling mode for the object. + */ + unsigned int tiling_mode : 2; + + /** How many users have pinned this object in GTT space. The following + * users can each hold at most one reference: pwrite/pread, pin_ioctl + * (via user_pin_count), execbuffer (objects are not allowed multiple + * times for the same batchbuffer), and the framebuffer code. When + * switching/pageflipping, the framebuffer code has at most two buffers + * pinned per crtc. + * + * In the worst case this is 1 + 1 + 1 + 2*2 = 7. That would fit into 3 + * bits with absolutely no headroom. So use 4 bits. */ + int pin_count : 4; +#define DRM_I915_GEM_OBJECT_MAX_PIN_COUNT 0xf /** AGP memory structure for our GTT binding. */ DRM_AGP_MEM *agp_mem; struct page **pages; - int pages_refcount; /** * Current offset of the object in GTT space. @@ -692,26 +723,18 @@ struct drm_i915_gem_object { */ uint32_t gtt_offset; + /* Which ring is refering to is this object */ + struct intel_ring_buffer *ring; + /** * Fake offset for use by mmap(2) */ uint64_t mmap_offset; - /** - * Fence register bits (if any) for this object. Will be set - * as needed when mapped into the GTT. - * Protected by dev->struct_mutex. - */ - int fence_reg; - - /** How many users have pinned this object in GTT space */ - int pin_count; - /** Breadcrumb of last rendering to the buffer. */ uint32_t last_rendering_seqno; - /** Current tiling mode for the object. */ - uint32_t tiling_mode; + /** Current tiling stride for the object, if it's tiled. */ uint32_t stride; /** Record of address bit 17 of each page at last unbind. */ @@ -734,17 +757,6 @@ struct drm_i915_gem_object { struct drm_i915_gem_phys_object *phys_obj; /** - * Used for checking the object doesn't appear more than once - * in an execbuffer object list. - */ - int in_execbuffer; - - /** - * Advice: are the backing pages purgeable? - */ - int madv; - - /** * Number of crtcs where this object is currently the fb, but * will be page flipped away on the next vblank. When it * reaches 0, dev_priv->pending_flip_queue will be woken up. @@ -765,6 +777,9 @@ struct drm_i915_gem_object { * an emission time with seqnos for tracking how far ahead of the GPU we are. */ struct drm_i915_gem_request { + /** On Which ring this request was generated */ + struct intel_ring_buffer *ring; + /** GEM sequence number associated with this request. */ uint32_t seqno; @@ -821,6 +836,11 @@ extern int i915_emit_box(struct drm_device *dev, struct drm_clip_rect *boxes, int i, int DR1, int DR4); extern int i965_reset(struct drm_device *dev, u8 flags); +extern unsigned long i915_chipset_val(struct drm_i915_private *dev_priv); +extern unsigned long i915_mch_val(struct drm_i915_private *dev_priv); +extern unsigned long i915_gfx_val(struct drm_i915_private *dev_priv); +extern void i915_update_gfx_val(struct drm_i915_private *dev_priv); + /* i915_irq.c */ void i915_hangcheck_elapsed(unsigned long data); @@ -829,9 +849,7 @@ extern int i915_irq_emit(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int i915_irq_wait(struct drm_device *dev, void *data, struct drm_file *file_priv); -void i915_user_irq_get(struct drm_device *dev); void i915_trace_irq_get(struct drm_device *dev, u32 seqno); -void i915_user_irq_put(struct drm_device *dev); extern void i915_enable_interrupt (struct drm_device *dev); extern irqreturn_t i915_driver_irq_handler(DRM_IRQ_ARGS); @@ -849,6 +867,11 @@ extern u32 gm45_get_vblank_counter(struct drm_device *dev, int crtc); extern int i915_vblank_swap(struct drm_device *dev, void *data, struct drm_file *file_priv); extern void i915_enable_irq(drm_i915_private_t *dev_priv, u32 mask); +extern void i915_disable_irq(drm_i915_private_t *dev_priv, u32 mask); +extern void ironlake_enable_graphics_irq(drm_i915_private_t *dev_priv, + u32 mask); +extern void ironlake_disable_graphics_irq(drm_i915_private_t *dev_priv, + u32 mask); void i915_enable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask); @@ -922,11 +945,13 @@ void i915_gem_object_unpin(struct drm_gem_object *obj); int i915_gem_object_unbind(struct drm_gem_object *obj); void i915_gem_release_mmap(struct drm_gem_object *obj); void i915_gem_lastclose(struct drm_device *dev); -uint32_t i915_get_gem_seqno(struct drm_device *dev); +uint32_t i915_get_gem_seqno(struct drm_device *dev, + struct intel_ring_buffer *ring); bool i915_seqno_passed(uint32_t seq1, uint32_t seq2); int i915_gem_object_get_fence_reg(struct drm_gem_object *obj); int i915_gem_object_put_fence_reg(struct drm_gem_object *obj); -void i915_gem_retire_requests(struct drm_device *dev); +void i915_gem_retire_requests(struct drm_device *dev, + struct intel_ring_buffer *ring); void i915_gem_retire_work_handler(struct work_struct *work); void i915_gem_clflush_object(struct drm_gem_object *obj); int i915_gem_object_set_domain(struct drm_gem_object *obj, @@ -937,9 +962,13 @@ void i915_gem_cleanup_ringbuffer(struct drm_device *dev); int i915_gem_do_init(struct drm_device *dev, unsigned long start, unsigned long end); int i915_gem_idle(struct drm_device *dev); -uint32_t i915_add_request(struct drm_device *dev, struct drm_file *file_priv, - uint32_t flush_domains); -int i915_do_wait_request(struct drm_device *dev, uint32_t seqno, int interruptible); +uint32_t i915_add_request(struct drm_device *dev, + struct drm_file *file_priv, + uint32_t flush_domains, + struct intel_ring_buffer *ring); +int i915_do_wait_request(struct drm_device *dev, + uint32_t seqno, int interruptible, + struct intel_ring_buffer *ring); int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf); int i915_gem_object_set_to_gtt_domain(struct drm_gem_object *obj, int write); @@ -1015,7 +1044,7 @@ extern void g4x_disable_fbc(struct drm_device *dev); extern void intel_disable_fbc(struct drm_device *dev); extern void intel_enable_fbc(struct drm_crtc *crtc, unsigned long interval); extern bool intel_fbc_enabled(struct drm_device *dev); - +extern bool ironlake_set_drps(struct drm_device *dev, u8 val); extern void intel_detect_pch (struct drm_device *dev); extern int intel_trans_dp_port_sel (struct drm_crtc *crtc); @@ -1026,7 +1055,8 @@ extern int intel_trans_dp_port_sel (struct drm_crtc *crtc); * has access to the ring. */ #define RING_LOCK_TEST_WITH_RETURN(dev, file_priv) do { \ - if (((drm_i915_private_t *)dev->dev_private)->ring.ring_obj == NULL) \ + if (((drm_i915_private_t *)dev->dev_private)->render_ring.gem_object \ + == NULL) \ LOCK_TEST_WITH_RETURN(dev, file_priv); \ } while (0) @@ -1039,35 +1069,31 @@ extern int intel_trans_dp_port_sel (struct drm_crtc *crtc); #define I915_WRITE64(reg, val) writeq(val, dev_priv->regs + (reg)) #define I915_READ64(reg) readq(dev_priv->regs + (reg)) #define POSTING_READ(reg) (void)I915_READ(reg) +#define POSTING_READ16(reg) (void)I915_READ16(reg) #define I915_VERBOSE 0 -#define RING_LOCALS volatile unsigned int *ring_virt__; - -#define BEGIN_LP_RING(n) do { \ - int bytes__ = 4*(n); \ - if (I915_VERBOSE) DRM_DEBUG("BEGIN_LP_RING(%d)\n", (n)); \ - /* a wrap must occur between instructions so pad beforehand */ \ - if (unlikely (dev_priv->ring.tail + bytes__ > dev_priv->ring.Size)) \ - i915_wrap_ring(dev); \ - if (unlikely (dev_priv->ring.space < bytes__)) \ - i915_wait_ring(dev, bytes__, __func__); \ - ring_virt__ = (unsigned int *) \ - (dev_priv->ring.virtual_start + dev_priv->ring.tail); \ - dev_priv->ring.tail += bytes__; \ - dev_priv->ring.tail &= dev_priv->ring.Size - 1; \ - dev_priv->ring.space -= bytes__; \ +#define BEGIN_LP_RING(n) do { \ + drm_i915_private_t *dev_priv = dev->dev_private; \ + if (I915_VERBOSE) \ + DRM_DEBUG(" BEGIN_LP_RING %x\n", (int)(n)); \ + intel_ring_begin(dev, &dev_priv->render_ring, (n)); \ } while (0) -#define OUT_RING(n) do { \ - if (I915_VERBOSE) DRM_DEBUG(" OUT_RING %x\n", (int)(n)); \ - *ring_virt__++ = (n); \ + +#define OUT_RING(x) do { \ + drm_i915_private_t *dev_priv = dev->dev_private; \ + if (I915_VERBOSE) \ + DRM_DEBUG(" OUT_RING %x\n", (int)(x)); \ + intel_ring_emit(dev, &dev_priv->render_ring, x); \ } while (0) #define ADVANCE_LP_RING() do { \ + drm_i915_private_t *dev_priv = dev->dev_private; \ if (I915_VERBOSE) \ - DRM_DEBUG("ADVANCE_LP_RING %x\n", dev_priv->ring.tail); \ - I915_WRITE(PRB0_TAIL, dev_priv->ring.tail); \ + DRM_DEBUG("ADVANCE_LP_RING %x\n", \ + dev_priv->render_ring.tail); \ + intel_ring_advance(dev, &dev_priv->render_ring); \ } while(0) /** @@ -1085,14 +1111,12 @@ extern int intel_trans_dp_port_sel (struct drm_crtc *crtc); * * The area from dword 0x20 to 0x3ff is available for driver usage. */ -#define READ_HWSP(dev_priv, reg) (((volatile u32*)(dev_priv->hw_status_page))[reg]) +#define READ_HWSP(dev_priv, reg) (((volatile u32 *)\ + (dev_priv->render_ring.status_page.page_addr))[reg]) #define READ_BREADCRUMB(dev_priv) READ_HWSP(dev_priv, I915_BREADCRUMB_INDEX) #define I915_GEM_HWS_INDEX 0x20 #define I915_BREADCRUMB_INDEX 0x21 -extern int i915_wrap_ring(struct drm_device * dev); -extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); - #define INTEL_INFO(dev) (((struct drm_i915_private *) (dev)->dev_private)->info) #define IS_I830(dev) ((dev)->pci_device == 0x3577) @@ -1138,6 +1162,7 @@ extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); (dev)->pci_device == 0x2A42 || \ (dev)->pci_device == 0x2E42) +#define HAS_BSD(dev) (IS_IRONLAKE(dev) || IS_G4X(dev)) #define I915_NEED_GFX_HWS(dev) (INTEL_INFO(dev)->need_gfx_hws) /* With the 945 and later, Y tiling got adjusted so that it was 32 128-byte diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 112699f71fa4..074385882ccf 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -35,8 +35,6 @@ #include <linux/swap.h> #include <linux/pci.h> -#define I915_GEM_GPU_DOMAINS (~(I915_GEM_DOMAIN_CPU | I915_GEM_DOMAIN_GTT)) - static void i915_gem_object_flush_gpu_write_domain(struct drm_gem_object *obj); static void i915_gem_object_flush_gtt_write_domain(struct drm_gem_object *obj); static void i915_gem_object_flush_cpu_write_domain(struct drm_gem_object *obj); @@ -169,7 +167,7 @@ static int i915_gem_object_needs_bit17_swizzle(struct drm_gem_object *obj) obj_priv->tiling_mode != I915_TILING_NONE; } -static inline int +static inline void slow_shmem_copy(struct page *dst_page, int dst_offset, struct page *src_page, @@ -178,25 +176,16 @@ slow_shmem_copy(struct page *dst_page, { char *dst_vaddr, *src_vaddr; - dst_vaddr = kmap_atomic(dst_page, KM_USER0); - if (dst_vaddr == NULL) - return -ENOMEM; - - src_vaddr = kmap_atomic(src_page, KM_USER1); - if (src_vaddr == NULL) { - kunmap_atomic(dst_vaddr, KM_USER0); - return -ENOMEM; - } + dst_vaddr = kmap(dst_page); + src_vaddr = kmap(src_page); memcpy(dst_vaddr + dst_offset, src_vaddr + src_offset, length); - kunmap_atomic(src_vaddr, KM_USER1); - kunmap_atomic(dst_vaddr, KM_USER0); - - return 0; + kunmap(src_page); + kunmap(dst_page); } -static inline int +static inline void slow_shmem_bit17_copy(struct page *gpu_page, int gpu_offset, struct page *cpu_page, @@ -216,15 +205,8 @@ slow_shmem_bit17_copy(struct page *gpu_page, cpu_page, cpu_offset, length); } - gpu_vaddr = kmap_atomic(gpu_page, KM_USER0); - if (gpu_vaddr == NULL) - return -ENOMEM; - - cpu_vaddr = kmap_atomic(cpu_page, KM_USER1); - if (cpu_vaddr == NULL) { - kunmap_atomic(gpu_vaddr, KM_USER0); - return -ENOMEM; - } + gpu_vaddr = kmap(gpu_page); + cpu_vaddr = kmap(cpu_page); /* Copy the data, XORing A6 with A17 (1). The user already knows he's * XORing with the other bits (A9 for Y, A9 and A10 for X) @@ -248,10 +230,8 @@ slow_shmem_bit17_copy(struct page *gpu_page, length -= this_length; } - kunmap_atomic(cpu_vaddr, KM_USER1); - kunmap_atomic(gpu_vaddr, KM_USER0); - - return 0; + kunmap(cpu_page); + kunmap(gpu_page); } /** @@ -427,21 +407,19 @@ i915_gem_shmem_pread_slow(struct drm_device *dev, struct drm_gem_object *obj, page_length = PAGE_SIZE - data_page_offset; if (do_bit17_swizzling) { - ret = slow_shmem_bit17_copy(obj_priv->pages[shmem_page_index], - shmem_page_offset, - user_pages[data_page_index], - data_page_offset, - page_length, - 1); - } else { - ret = slow_shmem_copy(user_pages[data_page_index], - data_page_offset, - obj_priv->pages[shmem_page_index], + slow_shmem_bit17_copy(obj_priv->pages[shmem_page_index], shmem_page_offset, - page_length); + user_pages[data_page_index], + data_page_offset, + page_length, + 1); + } else { + slow_shmem_copy(user_pages[data_page_index], + data_page_offset, + obj_priv->pages[shmem_page_index], + shmem_page_offset, + page_length); } - if (ret) - goto fail_put_pages; remain -= page_length; data_ptr += page_length; @@ -531,25 +509,24 @@ fast_user_write(struct io_mapping *mapping, * page faults */ -static inline int +static inline void slow_kernel_write(struct io_mapping *mapping, loff_t gtt_base, int gtt_offset, struct page *user_page, int user_offset, int length) { - char *src_vaddr, *dst_vaddr; - unsigned long unwritten; + char __iomem *dst_vaddr; + char *src_vaddr; - dst_vaddr = io_mapping_map_atomic_wc(mapping, gtt_base); - src_vaddr = kmap_atomic(user_page, KM_USER1); - unwritten = __copy_from_user_inatomic_nocache(dst_vaddr + gtt_offset, - src_vaddr + user_offset, - length); - kunmap_atomic(src_vaddr, KM_USER1); - io_mapping_unmap_atomic(dst_vaddr); - if (unwritten) - return -EFAULT; - return 0; + dst_vaddr = io_mapping_map_wc(mapping, gtt_base); + src_vaddr = kmap(user_page); + + memcpy_toio(dst_vaddr + gtt_offset, + src_vaddr + user_offset, + length); + + kunmap(user_page); + io_mapping_unmap(dst_vaddr); } static inline int @@ -722,18 +699,11 @@ i915_gem_gtt_pwrite_slow(struct drm_device *dev, struct drm_gem_object *obj, if ((data_page_offset + page_length) > PAGE_SIZE) page_length = PAGE_SIZE - data_page_offset; - ret = slow_kernel_write(dev_priv->mm.gtt_mapping, - gtt_page_base, gtt_page_offset, - user_pages[data_page_index], - data_page_offset, - page_length); - - /* If we get a fault while copying data, then (presumably) our - * source page isn't available. Return the error and we'll - * retry in the slow path. - */ - if (ret) - goto out_unpin_object; + slow_kernel_write(dev_priv->mm.gtt_mapping, + gtt_page_base, gtt_page_offset, + user_pages[data_page_index], + data_page_offset, + page_length); remain -= page_length; offset += page_length; @@ -902,21 +872,19 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev, struct drm_gem_object *obj, page_length = PAGE_SIZE - data_page_offset; if (do_bit17_swizzling) { - ret = slow_shmem_bit17_copy(obj_priv->pages[shmem_page_index], - shmem_page_offset, - user_pages[data_page_index], - data_page_offset, - page_length, - 0); - } else { - ret = slow_shmem_copy(obj_priv->pages[shmem_page_index], + slow_shmem_bit17_copy(obj_priv->pages[shmem_page_index], shmem_page_offset, user_pages[data_page_index], data_page_offset, - page_length); + page_length, + 0); + } else { + slow_shmem_copy(obj_priv->pages[shmem_page_index], + shmem_page_offset, + user_pages[data_page_index], + data_page_offset, + page_length); } - if (ret) - goto fail_put_pages; remain -= page_length; data_ptr += page_length; @@ -973,7 +941,8 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data, if (obj_priv->phys_obj) ret = i915_gem_phys_pwrite(dev, obj, args, file_priv); else if (obj_priv->tiling_mode == I915_TILING_NONE && - dev->gtt_total != 0) { + dev->gtt_total != 0 && + obj->write_domain != I915_GEM_DOMAIN_CPU) { ret = i915_gem_gtt_pwrite_fast(dev, obj, args, file_priv); if (ret == -EFAULT) { ret = i915_gem_gtt_pwrite_slow(dev, obj, args, @@ -1484,11 +1453,14 @@ i915_gem_object_put_pages(struct drm_gem_object *obj) } static void -i915_gem_object_move_to_active(struct drm_gem_object *obj, uint32_t seqno) +i915_gem_object_move_to_active(struct drm_gem_object *obj, uint32_t seqno, + struct intel_ring_buffer *ring) { struct drm_device *dev = obj->dev; drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_gem_object *obj_priv = to_intel_bo(obj); + BUG_ON(ring == NULL); + obj_priv->ring = ring; /* Add a reference if we're newly entering the active list. */ if (!obj_priv->active) { @@ -1497,8 +1469,7 @@ i915_gem_object_move_to_active(struct drm_gem_object *obj, uint32_t seqno) } /* Move from whatever list we were on to the tail of execution. */ spin_lock(&dev_priv->mm.active_list_lock); - list_move_tail(&obj_priv->list, - &dev_priv->mm.active_list); + list_move_tail(&obj_priv->list, &ring->active_list); spin_unlock(&dev_priv->mm.active_list_lock); obj_priv->last_rendering_seqno = seqno; } @@ -1551,6 +1522,7 @@ i915_gem_object_move_to_inactive(struct drm_gem_object *obj) BUG_ON(!list_empty(&obj_priv->gpu_write_list)); obj_priv->last_rendering_seqno = 0; + obj_priv->ring = NULL; if (obj_priv->active) { obj_priv->active = 0; drm_gem_object_unreference(obj); @@ -1560,7 +1532,8 @@ i915_gem_object_move_to_inactive(struct drm_gem_object *obj) static void i915_gem_process_flushing_list(struct drm_device *dev, - uint32_t flush_domains, uint32_t seqno) + uint32_t flush_domains, uint32_t seqno, + struct intel_ring_buffer *ring) { drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_gem_object *obj_priv, *next; @@ -1571,12 +1544,13 @@ i915_gem_process_flushing_list(struct drm_device *dev, struct drm_gem_object *obj = &obj_priv->base; if ((obj->write_domain & flush_domains) == - obj->write_domain) { + obj->write_domain && + obj_priv->ring->ring_flag == ring->ring_flag) { uint32_t old_write_domain = obj->write_domain; obj->write_domain = 0; list_del_init(&obj_priv->gpu_write_list); - i915_gem_object_move_to_active(obj, seqno); + i915_gem_object_move_to_active(obj, seqno, ring); /* update the fence lru list */ if (obj_priv->fence_reg != I915_FENCE_REG_NONE) { @@ -1593,31 +1567,15 @@ i915_gem_process_flushing_list(struct drm_device *dev, } } -#define PIPE_CONTROL_FLUSH(addr) \ - OUT_RING(GFX_OP_PIPE_CONTROL | PIPE_CONTROL_QW_WRITE | \ - PIPE_CONTROL_DEPTH_STALL); \ - OUT_RING(addr | PIPE_CONTROL_GLOBAL_GTT); \ - OUT_RING(0); \ - OUT_RING(0); \ - -/** - * Creates a new sequence number, emitting a write of it to the status page - * plus an interrupt, which will trigger i915_user_interrupt_handler. - * - * Must be called with struct_lock held. - * - * Returned sequence numbers are nonzero on success. - */ uint32_t i915_add_request(struct drm_device *dev, struct drm_file *file_priv, - uint32_t flush_domains) + uint32_t flush_domains, struct intel_ring_buffer *ring) { drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_file_private *i915_file_priv = NULL; struct drm_i915_gem_request *request; uint32_t seqno; int was_empty; - RING_LOCALS; if (file_priv != NULL) i915_file_priv = file_priv->driver_priv; @@ -1626,62 +1584,14 @@ i915_add_request(struct drm_device *dev, struct drm_file *file_priv, if (request == NULL) return 0; - /* Grab the seqno we're going to make this request be, and bump the - * next (skipping 0 so it can be the reserved no-seqno value). - */ - seqno = dev_priv->mm.next_gem_seqno; - dev_priv->mm.next_gem_seqno++; - if (dev_priv->mm.next_gem_seqno == 0) - dev_priv->mm.next_gem_seqno++; - - if (HAS_PIPE_CONTROL(dev)) { - u32 scratch_addr = dev_priv->seqno_gfx_addr + 128; - - /* - * Workaround qword write incoherence by flushing the - * PIPE_NOTIFY buffers out to memory before requesting - * an interrupt. - */ - BEGIN_LP_RING(32); - OUT_RING(GFX_OP_PIPE_CONTROL | PIPE_CONTROL_QW_WRITE | - PIPE_CONTROL_WC_FLUSH | PIPE_CONTROL_TC_FLUSH); - OUT_RING(dev_priv->seqno_gfx_addr | PIPE_CONTROL_GLOBAL_GTT); - OUT_RING(seqno); - OUT_RING(0); - PIPE_CONTROL_FLUSH(scratch_addr); - scratch_addr += 128; /* write to separate cachelines */ - PIPE_CONTROL_FLUSH(scratch_addr); - scratch_addr += 128; - PIPE_CONTROL_FLUSH(scratch_addr); - scratch_addr += 128; - PIPE_CONTROL_FLUSH(scratch_addr); - scratch_addr += 128; - PIPE_CONTROL_FLUSH(scratch_addr); - scratch_addr += 128; - PIPE_CONTROL_FLUSH(scratch_addr); - OUT_RING(GFX_OP_PIPE_CONTROL | PIPE_CONTROL_QW_WRITE | - PIPE_CONTROL_WC_FLUSH | PIPE_CONTROL_TC_FLUSH | - PIPE_CONTROL_NOTIFY); - OUT_RING(dev_priv->seqno_gfx_addr | PIPE_CONTROL_GLOBAL_GTT); - OUT_RING(seqno); - OUT_RING(0); - ADVANCE_LP_RING(); - } else { - BEGIN_LP_RING(4); - OUT_RING(MI_STORE_DWORD_INDEX); - OUT_RING(I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT); - OUT_RING(seqno); - - OUT_RING(MI_USER_INTERRUPT); - ADVANCE_LP_RING(); - } - - DRM_DEBUG_DRIVER("%d\n", seqno); + seqno = ring->add_request(dev, ring, file_priv, flush_domains); request->seqno = seqno; + request->ring = ring; request->emitted_jiffies = jiffies; - was_empty = list_empty(&dev_priv->mm.request_list); - list_add_tail(&request->list, &dev_priv->mm.request_list); + was_empty = list_empty(&ring->request_list); + list_add_tail(&request->list, &ring->request_list); + if (i915_file_priv) { list_add_tail(&request->client_list, &i915_file_priv->mm.request_list); @@ -1693,7 +1603,7 @@ i915_add_request(struct drm_device *dev, struct drm_file *file_priv, * domain we're flushing with our flush. */ if (flush_domains != 0) - i915_gem_process_flushing_list(dev, flush_domains, seqno); + i915_gem_process_flushing_list(dev, flush_domains, seqno, ring); if (!dev_priv->mm.suspended) { mod_timer(&dev_priv->hangcheck_timer, jiffies + DRM_I915_HANGCHECK_PERIOD); @@ -1710,20 +1620,16 @@ i915_add_request(struct drm_device *dev, struct drm_file *file_priv, * before signalling the CPU */ static uint32_t -i915_retire_commands(struct drm_device *dev) +i915_retire_commands(struct drm_device *dev, struct intel_ring_buffer *ring) { - drm_i915_private_t *dev_priv = dev->dev_private; - uint32_t cmd = MI_FLUSH | MI_NO_WRITE_FLUSH; uint32_t flush_domains = 0; - RING_LOCALS; /* The sampler always gets flushed on i965 (sigh) */ if (IS_I965G(dev)) flush_domains |= I915_GEM_DOMAIN_SAMPLER; - BEGIN_LP_RING(2); - OUT_RING(cmd); - OUT_RING(0); /* noop */ - ADVANCE_LP_RING(); + + ring->flush(dev, ring, + I915_GEM_DOMAIN_COMMAND, flush_domains); return flush_domains; } @@ -1743,11 +1649,11 @@ i915_gem_retire_request(struct drm_device *dev, * by the ringbuffer to the flushing/inactive lists as appropriate. */ spin_lock(&dev_priv->mm.active_list_lock); - while (!list_empty(&dev_priv->mm.active_list)) { + while (!list_empty(&request->ring->active_list)) { struct drm_gem_object *obj; struct drm_i915_gem_object *obj_priv; - obj_priv = list_first_entry(&dev_priv->mm.active_list, + obj_priv = list_first_entry(&request->ring->active_list, struct drm_i915_gem_object, list); obj = &obj_priv->base; @@ -1794,35 +1700,33 @@ i915_seqno_passed(uint32_t seq1, uint32_t seq2) } uint32_t -i915_get_gem_seqno(struct drm_device *dev) +i915_get_gem_seqno(struct drm_device *dev, + struct intel_ring_buffer *ring) { - drm_i915_private_t *dev_priv = dev->dev_private; - - if (HAS_PIPE_CONTROL(dev)) - return ((volatile u32 *)(dev_priv->seqno_page))[0]; - else - return READ_HWSP(dev_priv, I915_GEM_HWS_INDEX); + return ring->get_gem_seqno(dev, ring); } /** * This function clears the request list as sequence numbers are passed. */ void -i915_gem_retire_requests(struct drm_device *dev) +i915_gem_retire_requests(struct drm_device *dev, + struct intel_ring_buffer *ring) { drm_i915_private_t *dev_priv = dev->dev_private; uint32_t seqno; - if (!dev_priv->hw_status_page || list_empty(&dev_priv->mm.request_list)) + if (!ring->status_page.page_addr + || list_empty(&ring->request_list)) return; - seqno = i915_get_gem_seqno(dev); + seqno = i915_get_gem_seqno(dev, ring); - while (!list_empty(&dev_priv->mm.request_list)) { + while (!list_empty(&ring->request_list)) { struct drm_i915_gem_request *request; uint32_t retiring_seqno; - request = list_first_entry(&dev_priv->mm.request_list, + request = list_first_entry(&ring->request_list, struct drm_i915_gem_request, list); retiring_seqno = request->seqno; @@ -1840,7 +1744,8 @@ i915_gem_retire_requests(struct drm_device *dev) if (unlikely (dev_priv->trace_irq_seqno && i915_seqno_passed(dev_priv->trace_irq_seqno, seqno))) { - i915_user_irq_put(dev); + + ring->user_irq_put(dev, ring); dev_priv->trace_irq_seqno = 0; } } @@ -1856,15 +1761,22 @@ i915_gem_retire_work_handler(struct work_struct *work) dev = dev_priv->dev; mutex_lock(&dev->struct_mutex); - i915_gem_retire_requests(dev); + i915_gem_retire_requests(dev, &dev_priv->render_ring); + + if (HAS_BSD(dev)) + i915_gem_retire_requests(dev, &dev_priv->bsd_ring); + if (!dev_priv->mm.suspended && - !list_empty(&dev_priv->mm.request_list)) + (!list_empty(&dev_priv->render_ring.request_list) || + (HAS_BSD(dev) && + !list_empty(&dev_priv->bsd_ring.request_list)))) queue_delayed_work(dev_priv->wq, &dev_priv->mm.retire_work, HZ); mutex_unlock(&dev->struct_mutex); } int -i915_do_wait_request(struct drm_device *dev, uint32_t seqno, int interruptible) +i915_do_wait_request(struct drm_device *dev, uint32_t seqno, + int interruptible, struct intel_ring_buffer *ring) { drm_i915_private_t *dev_priv = dev->dev_private; u32 ier; @@ -1875,7 +1787,7 @@ i915_do_wait_request(struct drm_device *dev, uint32_t seqno, int interruptible) if (atomic_read(&dev_priv->mm.wedged)) return -EIO; - if (!i915_seqno_passed(i915_get_gem_seqno(dev), seqno)) { + if (!i915_seqno_passed(ring->get_gem_seqno(dev, ring), seqno)) { if (HAS_PCH_SPLIT(dev)) ier = I915_READ(DEIER) | I915_READ(GTIER); else @@ -1889,19 +1801,21 @@ i915_do_wait_request(struct drm_device *dev, uint32_t seqno, int interruptible) trace_i915_gem_request_wait_begin(dev, seqno); - dev_priv->mm.waiting_gem_seqno = seqno; - i915_user_irq_get(dev); + ring->waiting_gem_seqno = seqno; + ring->user_irq_get(dev, ring); if (interruptible) - ret = wait_event_interruptible(dev_priv->irq_queue, - i915_seqno_passed(i915_get_gem_seqno(dev), seqno) || - atomic_read(&dev_priv->mm.wedged)); + ret = wait_event_interruptible(ring->irq_queue, + i915_seqno_passed( + ring->get_gem_seqno(dev, ring), seqno) + || atomic_read(&dev_priv->mm.wedged)); else - wait_event(dev_priv->irq_queue, - i915_seqno_passed(i915_get_gem_seqno(dev), seqno) || - atomic_read(&dev_priv->mm.wedged)); + wait_event(ring->irq_queue, + i915_seqno_passed( + ring->get_gem_seqno(dev, ring), seqno) + || atomic_read(&dev_priv->mm.wedged)); - i915_user_irq_put(dev); - dev_priv->mm.waiting_gem_seqno = 0; + ring->user_irq_put(dev, ring); + ring->waiting_gem_seqno = 0; trace_i915_gem_request_wait_end(dev, seqno); } @@ -1910,7 +1824,7 @@ i915_do_wait_request(struct drm_device *dev, uint32_t seqno, int interruptible) if (ret && ret != -ERESTARTSYS) DRM_ERROR("%s returns %d (awaiting %d at %d)\n", - __func__, ret, seqno, i915_get_gem_seqno(dev)); + __func__, ret, seqno, ring->get_gem_seqno(dev, ring)); /* Directly dispatch request retiring. While we have the work queue * to handle this, the waiter on a request often wants an associated @@ -1918,7 +1832,7 @@ i915_do_wait_request(struct drm_device *dev, uint32_t seqno, int interruptible) * a separate wait queue to handle that. */ if (ret == 0) - i915_gem_retire_requests(dev); + i915_gem_retire_requests(dev, ring); return ret; } @@ -1928,9 +1842,10 @@ i915_do_wait_request(struct drm_device *dev, uint32_t seqno, int interruptible) * request and object lists appropriately for that event. */ static int -i915_wait_request(struct drm_device *dev, uint32_t seqno) +i915_wait_request(struct drm_device *dev, uint32_t seqno, + struct intel_ring_buffer *ring) { - return i915_do_wait_request(dev, seqno, 1); + return i915_do_wait_request(dev, seqno, 1, ring); } static void @@ -1939,71 +1854,29 @@ i915_gem_flush(struct drm_device *dev, uint32_t flush_domains) { drm_i915_private_t *dev_priv = dev->dev_private; - uint32_t cmd; - RING_LOCALS; - -#if WATCH_EXEC - DRM_INFO("%s: invalidate %08x flush %08x\n", __func__, - invalidate_domains, flush_domains); -#endif - trace_i915_gem_request_flush(dev, dev_priv->mm.next_gem_seqno, - invalidate_domains, flush_domains); - if (flush_domains & I915_GEM_DOMAIN_CPU) drm_agp_chipset_flush(dev); + dev_priv->render_ring.flush(dev, &dev_priv->render_ring, + invalidate_domains, + flush_domains); + + if (HAS_BSD(dev)) + dev_priv->bsd_ring.flush(dev, &dev_priv->bsd_ring, + invalidate_domains, + flush_domains); +} - if ((invalidate_domains | flush_domains) & I915_GEM_GPU_DOMAINS) { - /* - * read/write caches: - * - * I915_GEM_DOMAIN_RENDER is always invalidated, but is - * only flushed if MI_NO_WRITE_FLUSH is unset. On 965, it is - * also flushed at 2d versus 3d pipeline switches. - * - * read-only caches: - * - * I915_GEM_DOMAIN_SAMPLER is flushed on pre-965 if - * MI_READ_FLUSH is set, and is always flushed on 965. - * - * I915_GEM_DOMAIN_COMMAND may not exist? - * - * I915_GEM_DOMAIN_INSTRUCTION, which exists on 965, is - * invalidated when MI_EXE_FLUSH is set. - * - * I915_GEM_DOMAIN_VERTEX, which exists on 965, is - * invalidated with every MI_FLUSH. - * - * TLBs: - * - * On 965, TLBs associated with I915_GEM_DOMAIN_COMMAND - * and I915_GEM_DOMAIN_CPU in are invalidated at PTE write and - * I915_GEM_DOMAIN_RENDER and I915_GEM_DOMAIN_SAMPLER - * are flushed at any MI_FLUSH. - */ - - cmd = MI_FLUSH | MI_NO_WRITE_FLUSH; - if ((invalidate_domains|flush_domains) & - I915_GEM_DOMAIN_RENDER) - cmd &= ~MI_NO_WRITE_FLUSH; - if (!IS_I965G(dev)) { - /* - * On the 965, the sampler cache always gets flushed - * and this bit is reserved. - */ - if (invalidate_domains & I915_GEM_DOMAIN_SAMPLER) - cmd |= MI_READ_FLUSH; - } - if (invalidate_domains & I915_GEM_DOMAIN_INSTRUCTION) - cmd |= MI_EXE_FLUSH; - -#if WATCH_EXEC - DRM_INFO("%s: queue flush %08x to ring\n", __func__, cmd); -#endif - BEGIN_LP_RING(2); - OUT_RING(cmd); - OUT_RING(MI_NOOP); - ADVANCE_LP_RING(); - } +static void +i915_gem_flush_ring(struct drm_device *dev, + uint32_t invalidate_domains, + uint32_t flush_domains, + struct intel_ring_buffer *ring) +{ + if (flush_domains & I915_GEM_DOMAIN_CPU) + drm_agp_chipset_flush(dev); + ring->flush(dev, ring, + invalidate_domains, + flush_domains); } /** @@ -2030,7 +1903,8 @@ i915_gem_object_wait_rendering(struct drm_gem_object *obj) DRM_INFO("%s: object %p wait for seqno %08x\n", __func__, obj, obj_priv->last_rendering_seqno); #endif - ret = i915_wait_request(dev, obj_priv->last_rendering_seqno); + ret = i915_wait_request(dev, + obj_priv->last_rendering_seqno, obj_priv->ring); if (ret != 0) return ret; } @@ -2146,11 +2020,14 @@ i915_gpu_idle(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; bool lists_empty; - uint32_t seqno; + uint32_t seqno1, seqno2; + int ret; spin_lock(&dev_priv->mm.active_list_lock); - lists_empty = list_empty(&dev_priv->mm.flushing_list) && - list_empty(&dev_priv->mm.active_list); + lists_empty = (list_empty(&dev_priv->mm.flushing_list) && + list_empty(&dev_priv->render_ring.active_list) && + (!HAS_BSD(dev) || + list_empty(&dev_priv->bsd_ring.active_list))); spin_unlock(&dev_priv->mm.active_list_lock); if (lists_empty) @@ -2158,11 +2035,25 @@ i915_gpu_idle(struct drm_device *dev) /* Flush everything onto the inactive list. */ i915_gem_flush(dev, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS); - seqno = i915_add_request(dev, NULL, I915_GEM_GPU_DOMAINS); - if (seqno == 0) + seqno1 = i915_add_request(dev, NULL, I915_GEM_GPU_DOMAINS, + &dev_priv->render_ring); + if (seqno1 == 0) return -ENOMEM; + ret = i915_wait_request(dev, seqno1, &dev_priv->render_ring); + + if (HAS_BSD(dev)) { + seqno2 = i915_add_request(dev, NULL, I915_GEM_GPU_DOMAINS, + &dev_priv->bsd_ring); + if (seqno2 == 0) + return -ENOMEM; + + ret = i915_wait_request(dev, seqno2, &dev_priv->bsd_ring); + if (ret) + return ret; + } + - return i915_wait_request(dev, seqno); + return ret; } static int @@ -2175,7 +2066,9 @@ i915_gem_evict_everything(struct drm_device *dev) spin_lock(&dev_priv->mm.active_list_lock); lists_empty = (list_empty(&dev_priv->mm.inactive_list) && list_empty(&dev_priv->mm.flushing_list) && - list_empty(&dev_priv->mm.active_list)); + list_empty(&dev_priv->render_ring.active_list) && + (!HAS_BSD(dev) + || list_empty(&dev_priv->bsd_ring.active_list))); spin_unlock(&dev_priv->mm.active_list_lock); if (lists_empty) @@ -2195,7 +2088,9 @@ i915_gem_evict_everything(struct drm_device *dev) spin_lock(&dev_priv->mm.active_list_lock); lists_empty = (list_empty(&dev_priv->mm.inactive_list) && list_empty(&dev_priv->mm.flushing_list) && - list_empty(&dev_priv->mm.active_list)); + list_empty(&dev_priv->render_ring.active_list) && + (!HAS_BSD(dev) + || list_empty(&dev_priv->bsd_ring.active_list))); spin_unlock(&dev_priv->mm.active_list_lock); BUG_ON(!lists_empty); @@ -2209,8 +2104,13 @@ i915_gem_evict_something(struct drm_device *dev, int min_size) struct drm_gem_object *obj; int ret; + struct intel_ring_buffer *render_ring = &dev_priv->render_ring; + struct intel_ring_buffer *bsd_ring = &dev_priv->bsd_ring; for (;;) { - i915_gem_retire_requests(dev); + i915_gem_retire_requests(dev, render_ring); + + if (HAS_BSD(dev)) + i915_gem_retire_requests(dev, bsd_ring); /* If there's an inactive buffer available now, grab it * and be done. @@ -2234,14 +2134,30 @@ i915_gem_evict_something(struct drm_device *dev, int min_size) * things, wait for the next to finish and hopefully leave us * a buffer to evict. */ - if (!list_empty(&dev_priv->mm.request_list)) { + if (!list_empty(&render_ring->request_list)) { + struct drm_i915_gem_request *request; + + request = list_first_entry(&render_ring->request_list, + struct drm_i915_gem_request, + list); + + ret = i915_wait_request(dev, + request->seqno, request->ring); + if (ret) + return ret; + + continue; + } + + if (HAS_BSD(dev) && !list_empty(&bsd_ring->request_list)) { struct drm_i915_gem_request *request; - request = list_first_entry(&dev_priv->mm.request_list, + request = list_first_entry(&bsd_ring->request_list, struct drm_i915_gem_request, list); - ret = i915_wait_request(dev, request->seqno); + ret = i915_wait_request(dev, + request->seqno, request->ring); if (ret) return ret; @@ -2268,10 +2184,13 @@ i915_gem_evict_something(struct drm_device *dev, int min_size) if (obj != NULL) { uint32_t seqno; - i915_gem_flush(dev, + i915_gem_flush_ring(dev, + obj->write_domain, obj->write_domain, - obj->write_domain); - seqno = i915_add_request(dev, NULL, obj->write_domain); + obj_priv->ring); + seqno = i915_add_request(dev, NULL, + obj->write_domain, + obj_priv->ring); if (seqno == 0) return -ENOMEM; continue; @@ -2299,6 +2218,9 @@ i915_gem_object_get_pages(struct drm_gem_object *obj, struct inode *inode; struct page *page; + BUG_ON(obj_priv->pages_refcount + == DRM_I915_GEM_OBJECT_MAX_PAGES_REFCOUNT); + if (obj_priv->pages_refcount++ != 0) return 0; @@ -2317,7 +2239,7 @@ i915_gem_object_get_pages(struct drm_gem_object *obj, mapping = inode->i_mapping; for (i = 0; i < page_count; i++) { page = read_cache_page_gfp(mapping, i, - mapping_gfp_mask (mapping) | + GFP_HIGHUSER | __GFP_COLD | gfpmask); if (IS_ERR(page)) @@ -2697,6 +2619,14 @@ i915_gem_object_bind_to_gtt(struct drm_gem_object *obj, unsigned alignment) return -EINVAL; } + /* If the object is bigger than the entire aperture, reject it early + * before evicting everything in a vain attempt to find space. + */ + if (obj->size > dev->gtt_total) { + DRM_ERROR("Attempting to bind an object larger than the aperture\n"); + return -E2BIG; + } + search_free: free_space = drm_mm_search_free(&dev_priv->mm.gtt_space, obj->size, alignment, 0); @@ -2807,6 +2737,7 @@ i915_gem_object_flush_gpu_write_domain(struct drm_gem_object *obj) { struct drm_device *dev = obj->dev; uint32_t old_write_domain; + struct drm_i915_gem_object *obj_priv = to_intel_bo(obj); if ((obj->write_domain & I915_GEM_GPU_DOMAINS) == 0) return; @@ -2814,7 +2745,7 @@ i915_gem_object_flush_gpu_write_domain(struct drm_gem_object *obj) /* Queue the GPU write cache flushing we need. */ old_write_domain = obj->write_domain; i915_gem_flush(dev, 0, obj->write_domain); - (void) i915_add_request(dev, NULL, obj->write_domain); + (void) i915_add_request(dev, NULL, obj->write_domain, obj_priv->ring); BUG_ON(obj->write_domain); trace_i915_gem_object_change_domain(obj, @@ -2954,23 +2885,24 @@ i915_gem_object_set_to_display_plane(struct drm_gem_object *obj) DRM_INFO("%s: object %p wait for seqno %08x\n", __func__, obj, obj_priv->last_rendering_seqno); #endif - ret = i915_do_wait_request(dev, obj_priv->last_rendering_seqno, 0); + ret = i915_do_wait_request(dev, + obj_priv->last_rendering_seqno, + 0, + obj_priv->ring); if (ret != 0) return ret; } + i915_gem_object_flush_cpu_write_domain(obj); + old_write_domain = obj->write_domain; old_read_domains = obj->read_domains; - obj->read_domains &= I915_GEM_DOMAIN_GTT; - - i915_gem_object_flush_cpu_write_domain(obj); - /* It should now be out of any other write domains, and we can update * the domain values for our changes. */ BUG_ON((obj->write_domain & ~I915_GEM_DOMAIN_GTT) != 0); - obj->read_domains |= I915_GEM_DOMAIN_GTT; + obj->read_domains = I915_GEM_DOMAIN_GTT; obj->write_domain = I915_GEM_DOMAIN_GTT; obj_priv->dirty = 1; @@ -3354,9 +3286,13 @@ i915_gem_object_pin_and_relocate(struct drm_gem_object *obj, obj_priv->tiling_mode != I915_TILING_NONE; /* Check fence reg constraints and rebind if necessary */ - if (need_fence && !i915_gem_object_fence_offset_ok(obj, - obj_priv->tiling_mode)) - i915_gem_object_unbind(obj); + if (need_fence && + !i915_gem_object_fence_offset_ok(obj, + obj_priv->tiling_mode)) { + ret = i915_gem_object_unbind(obj); + if (ret) + return ret; + } /* Choose the GTT offset for our buffer and put it there. */ ret = i915_gem_object_pin(obj, (uint32_t) entry->alignment); @@ -3370,9 +3306,6 @@ i915_gem_object_pin_and_relocate(struct drm_gem_object *obj, if (need_fence) { ret = i915_gem_object_get_fence_reg(obj); if (ret != 0) { - if (ret != -EBUSY && ret != -ERESTARTSYS) - DRM_ERROR("Failure to install fence: %d\n", - ret); i915_gem_object_unpin(obj); return ret; } @@ -3545,62 +3478,6 @@ i915_gem_object_pin_and_relocate(struct drm_gem_object *obj, return 0; } -/** Dispatch a batchbuffer to the ring - */ -static int -i915_dispatch_gem_execbuffer(struct drm_device *dev, - struct drm_i915_gem_execbuffer2 *exec, - struct drm_clip_rect *cliprects, - uint64_t exec_offset) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - int nbox = exec->num_cliprects; - int i = 0, count; - uint32_t exec_start, exec_len; - RING_LOCALS; - - exec_start = (uint32_t) exec_offset + exec->batch_start_offset; - exec_len = (uint32_t) exec->batch_len; - - trace_i915_gem_request_submit(dev, dev_priv->mm.next_gem_seqno + 1); - - count = nbox ? nbox : 1; - - for (i = 0; i < count; i++) { - if (i < nbox) { - int ret = i915_emit_box(dev, cliprects, i, - exec->DR1, exec->DR4); - if (ret) - return ret; - } - - if (IS_I830(dev) || IS_845G(dev)) { - BEGIN_LP_RING(4); - OUT_RING(MI_BATCH_BUFFER); - OUT_RING(exec_start | MI_BATCH_NON_SECURE); - OUT_RING(exec_start + exec_len - 4); - OUT_RING(0); - ADVANCE_LP_RING(); - } else { - BEGIN_LP_RING(2); - if (IS_I965G(dev)) { - OUT_RING(MI_BATCH_BUFFER_START | - (2 << 6) | - MI_BATCH_NON_SECURE_I965); - OUT_RING(exec_start); - } else { - OUT_RING(MI_BATCH_BUFFER_START | - (2 << 6)); - OUT_RING(exec_start | MI_BATCH_NON_SECURE); - } - ADVANCE_LP_RING(); - } - } - - /* XXX breadcrumb */ - return 0; -} - /* Throttle our rendering by waiting until the ring has completed our requests * emitted over 20 msec ago. * @@ -3629,7 +3506,7 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file_priv) if (time_after_eq(request->emitted_jiffies, recent_enough)) break; - ret = i915_wait_request(dev, request->seqno); + ret = i915_wait_request(dev, request->seqno, request->ring); if (ret != 0) break; } @@ -3786,10 +3663,22 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, uint32_t seqno, flush_domains, reloc_index; int pin_tries, flips; + struct intel_ring_buffer *ring = NULL; + #if WATCH_EXEC DRM_INFO("buffers_ptr %d buffer_count %d len %08x\n", (int) args->buffers_ptr, args->buffer_count, args->batch_len); #endif + if (args->flags & I915_EXEC_BSD) { + if (!HAS_BSD(dev)) { + DRM_ERROR("execbuf with wrong flag\n"); + return -EINVAL; + } + ring = &dev_priv->bsd_ring; + } else { + ring = &dev_priv->render_ring; + } + if (args->buffer_count < 1) { DRM_ERROR("execbuf with %d buffers\n", args->buffer_count); @@ -3902,11 +3791,19 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, if (ret != -ENOSPC || pin_tries >= 1) { if (ret != -ERESTARTSYS) { unsigned long long total_size = 0; - for (i = 0; i < args->buffer_count; i++) + int num_fences = 0; + for (i = 0; i < args->buffer_count; i++) { + obj_priv = object_list[i]->driver_private; + total_size += object_list[i]->size; - DRM_ERROR("Failed to pin buffer %d of %d, total %llu bytes: %d\n", + num_fences += + exec_list[i].flags & EXEC_OBJECT_NEEDS_FENCE && + obj_priv->tiling_mode != I915_TILING_NONE; + } + DRM_ERROR("Failed to pin buffer %d of %d, total %llu bytes, %d fences: %d\n", pinned+1, args->buffer_count, - total_size, ret); + total_size, num_fences, + ret); DRM_ERROR("%d objects [%d pinned], " "%d object bytes [%d pinned], " "%d/%d gtt bytes\n", @@ -3976,9 +3873,16 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, i915_gem_flush(dev, dev->invalidate_domains, dev->flush_domains); - if (dev->flush_domains & I915_GEM_GPU_DOMAINS) + if (dev->flush_domains & I915_GEM_GPU_DOMAINS) { (void)i915_add_request(dev, file_priv, - dev->flush_domains); + dev->flush_domains, + &dev_priv->render_ring); + + if (HAS_BSD(dev)) + (void)i915_add_request(dev, file_priv, + dev->flush_domains, + &dev_priv->bsd_ring); + } } for (i = 0; i < args->buffer_count; i++) { @@ -4015,7 +3919,8 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, #endif /* Exec the batchbuffer */ - ret = i915_dispatch_gem_execbuffer(dev, args, cliprects, exec_offset); + ret = ring->dispatch_gem_execbuffer(dev, ring, args, + cliprects, exec_offset); if (ret) { DRM_ERROR("dispatch failed %d\n", ret); goto err; @@ -4025,7 +3930,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, * Ensure that the commands in the batch buffer are * finished before the interrupt fires */ - flush_domains = i915_retire_commands(dev); + flush_domains = i915_retire_commands(dev, ring); i915_verify_inactive(dev, __FILE__, __LINE__); @@ -4036,12 +3941,13 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, * *some* interrupts representing completion of buffers that we can * wait on when trying to clear up gtt space). */ - seqno = i915_add_request(dev, file_priv, flush_domains); + seqno = i915_add_request(dev, file_priv, flush_domains, ring); BUG_ON(seqno == 0); for (i = 0; i < args->buffer_count; i++) { struct drm_gem_object *obj = object_list[i]; + obj_priv = to_intel_bo(obj); - i915_gem_object_move_to_active(obj, seqno); + i915_gem_object_move_to_active(obj, seqno, ring); #if WATCH_LRU DRM_INFO("%s: move to exec list %p\n", __func__, obj); #endif @@ -4153,7 +4059,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data, exec2.DR4 = args->DR4; exec2.num_cliprects = args->num_cliprects; exec2.cliprects_ptr = args->cliprects_ptr; - exec2.flags = 0; + exec2.flags = I915_EXEC_RENDER; ret = i915_gem_do_execbuffer(dev, data, file_priv, &exec2, exec2_list); if (!ret) { @@ -4239,7 +4145,20 @@ i915_gem_object_pin(struct drm_gem_object *obj, uint32_t alignment) struct drm_i915_gem_object *obj_priv = to_intel_bo(obj); int ret; + BUG_ON(obj_priv->pin_count == DRM_I915_GEM_OBJECT_MAX_PIN_COUNT); + i915_verify_inactive(dev, __FILE__, __LINE__); + + if (obj_priv->gtt_space != NULL) { + if (alignment == 0) + alignment = i915_gem_get_gtt_alignment(obj); + if (obj_priv->gtt_offset & (alignment - 1)) { + ret = i915_gem_object_unbind(obj); + if (ret) + return ret; + } + } + if (obj_priv->gtt_space == NULL) { ret = i915_gem_object_bind_to_gtt(obj, alignment); if (ret) @@ -4392,6 +4311,7 @@ i915_gem_busy_ioctl(struct drm_device *dev, void *data, struct drm_i915_gem_busy *args = data; struct drm_gem_object *obj; struct drm_i915_gem_object *obj_priv; + drm_i915_private_t *dev_priv = dev->dev_private; obj = drm_gem_object_lookup(dev, file_priv, args->handle); if (obj == NULL) { @@ -4406,7 +4326,10 @@ i915_gem_busy_ioctl(struct drm_device *dev, void *data, * actually unmasked, and our working set ends up being larger than * required. */ - i915_gem_retire_requests(dev); + i915_gem_retire_requests(dev, &dev_priv->render_ring); + + if (HAS_BSD(dev)) + i915_gem_retire_requests(dev, &dev_priv->bsd_ring); obj_priv = to_intel_bo(obj); /* Don't count being on the flushing list against the object being @@ -4573,7 +4496,10 @@ i915_gem_idle(struct drm_device *dev) mutex_lock(&dev->struct_mutex); - if (dev_priv->mm.suspended || dev_priv->ring.ring_obj == NULL) { + if (dev_priv->mm.suspended || + (dev_priv->render_ring.gem_object == NULL) || + (HAS_BSD(dev) && + dev_priv->bsd_ring.gem_object == NULL)) { mutex_unlock(&dev->struct_mutex); return 0; } @@ -4654,71 +4580,6 @@ err: return ret; } -static int -i915_gem_init_hws(struct drm_device *dev) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - struct drm_gem_object *obj; - struct drm_i915_gem_object *obj_priv; - int ret; - - /* If we need a physical address for the status page, it's already - * initialized at driver load time. - */ - if (!I915_NEED_GFX_HWS(dev)) - return 0; - - obj = i915_gem_alloc_object(dev, 4096); - if (obj == NULL) { - DRM_ERROR("Failed to allocate status page\n"); - ret = -ENOMEM; - goto err; - } - obj_priv = to_intel_bo(obj); - obj_priv->agp_type = AGP_USER_CACHED_MEMORY; - - ret = i915_gem_object_pin(obj, 4096); - if (ret != 0) { - drm_gem_object_unreference(obj); - goto err_unref; - } - - dev_priv->status_gfx_addr = obj_priv->gtt_offset; - - dev_priv->hw_status_page = kmap(obj_priv->pages[0]); - if (dev_priv->hw_status_page == NULL) { - DRM_ERROR("Failed to map status page.\n"); - memset(&dev_priv->hws_map, 0, sizeof(dev_priv->hws_map)); - ret = -EINVAL; - goto err_unpin; - } - - if (HAS_PIPE_CONTROL(dev)) { - ret = i915_gem_init_pipe_control(dev); - if (ret) - goto err_unpin; - } - - dev_priv->hws_obj = obj; - memset(dev_priv->hw_status_page, 0, PAGE_SIZE); - if (IS_GEN6(dev)) { - I915_WRITE(HWS_PGA_GEN6, dev_priv->status_gfx_addr); - I915_READ(HWS_PGA_GEN6); /* posting read */ - } else { - I915_WRITE(HWS_PGA, dev_priv->status_gfx_addr); - I915_READ(HWS_PGA); /* posting read */ - } - DRM_DEBUG_DRIVER("hws offset: 0x%08x\n", dev_priv->status_gfx_addr); - - return 0; - -err_unpin: - i915_gem_object_unpin(obj); -err_unref: - drm_gem_object_unreference(obj); -err: - return 0; -} static void i915_gem_cleanup_pipe_control(struct drm_device *dev) @@ -4737,146 +4598,46 @@ i915_gem_cleanup_pipe_control(struct drm_device *dev) dev_priv->seqno_page = NULL; } -static void -i915_gem_cleanup_hws(struct drm_device *dev) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - struct drm_gem_object *obj; - struct drm_i915_gem_object *obj_priv; - - if (dev_priv->hws_obj == NULL) - return; - - obj = dev_priv->hws_obj; - obj_priv = to_intel_bo(obj); - - kunmap(obj_priv->pages[0]); - i915_gem_object_unpin(obj); - drm_gem_object_unreference(obj); - dev_priv->hws_obj = NULL; - - memset(&dev_priv->hws_map, 0, sizeof(dev_priv->hws_map)); - dev_priv->hw_status_page = NULL; - - if (HAS_PIPE_CONTROL(dev)) - i915_gem_cleanup_pipe_control(dev); - - /* Write high address into HWS_PGA when disabling. */ - I915_WRITE(HWS_PGA, 0x1ffff000); -} - int i915_gem_init_ringbuffer(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; - struct drm_gem_object *obj; - struct drm_i915_gem_object *obj_priv; - drm_i915_ring_buffer_t *ring = &dev_priv->ring; int ret; - u32 head; - - ret = i915_gem_init_hws(dev); - if (ret != 0) - return ret; - obj = i915_gem_alloc_object(dev, 128 * 1024); - if (obj == NULL) { - DRM_ERROR("Failed to allocate ringbuffer\n"); - i915_gem_cleanup_hws(dev); - return -ENOMEM; - } - obj_priv = to_intel_bo(obj); + dev_priv->render_ring = render_ring; - ret = i915_gem_object_pin(obj, 4096); - if (ret != 0) { - drm_gem_object_unreference(obj); - i915_gem_cleanup_hws(dev); - return ret; + if (!I915_NEED_GFX_HWS(dev)) { + dev_priv->render_ring.status_page.page_addr + = dev_priv->status_page_dmah->vaddr; + memset(dev_priv->render_ring.status_page.page_addr, + 0, PAGE_SIZE); } - /* Set up the kernel mapping for the ring. */ - ring->Size = obj->size; - - ring->map.offset = dev->agp->base + obj_priv->gtt_offset; - ring->map.size = obj->size; - ring->map.type = 0; - ring->map.flags = 0; - ring->map.mtrr = 0; - - drm_core_ioremap_wc(&ring->map, dev); - if (ring->map.handle == NULL) { - DRM_ERROR("Failed to map ringbuffer.\n"); - memset(&dev_priv->ring, 0, sizeof(dev_priv->ring)); - i915_gem_object_unpin(obj); - drm_gem_object_unreference(obj); - i915_gem_cleanup_hws(dev); - return -EINVAL; - } - ring->ring_obj = obj; - ring->virtual_start = ring->map.handle; - - /* Stop the ring if it's running. */ - I915_WRITE(PRB0_CTL, 0); - I915_WRITE(PRB0_TAIL, 0); - I915_WRITE(PRB0_HEAD, 0); - - /* Initialize the ring. */ - I915_WRITE(PRB0_START, obj_priv->gtt_offset); - head = I915_READ(PRB0_HEAD) & HEAD_ADDR; - - /* G45 ring initialization fails to reset head to zero */ - if (head != 0) { - DRM_ERROR("Ring head not reset to zero " - "ctl %08x head %08x tail %08x start %08x\n", - I915_READ(PRB0_CTL), - I915_READ(PRB0_HEAD), - I915_READ(PRB0_TAIL), - I915_READ(PRB0_START)); - I915_WRITE(PRB0_HEAD, 0); - - DRM_ERROR("Ring head forced to zero " - "ctl %08x head %08x tail %08x start %08x\n", - I915_READ(PRB0_CTL), - I915_READ(PRB0_HEAD), - I915_READ(PRB0_TAIL), - I915_READ(PRB0_START)); - } - - I915_WRITE(PRB0_CTL, - ((obj->size - 4096) & RING_NR_PAGES) | - RING_NO_REPORT | - RING_VALID); - - head = I915_READ(PRB0_HEAD) & HEAD_ADDR; - - /* If the head is still not zero, the ring is dead */ - if (head != 0) { - DRM_ERROR("Ring initialization failed " - "ctl %08x head %08x tail %08x start %08x\n", - I915_READ(PRB0_CTL), - I915_READ(PRB0_HEAD), - I915_READ(PRB0_TAIL), - I915_READ(PRB0_START)); - return -EIO; + if (HAS_PIPE_CONTROL(dev)) { + ret = i915_gem_init_pipe_control(dev); + if (ret) + return ret; } - /* Update our cache of the ring state */ - if (!drm_core_check_feature(dev, DRIVER_MODESET)) - i915_kernel_lost_context(dev); - else { - ring->head = I915_READ(PRB0_HEAD) & HEAD_ADDR; - ring->tail = I915_READ(PRB0_TAIL) & TAIL_ADDR; - ring->space = ring->head - (ring->tail + 8); - if (ring->space < 0) - ring->space += ring->Size; - } + ret = intel_init_ring_buffer(dev, &dev_priv->render_ring); + if (ret) + goto cleanup_pipe_control; - if (IS_I9XX(dev) && !IS_GEN3(dev)) { - I915_WRITE(MI_MODE, - (VS_TIMER_DISPATCH) << 16 | VS_TIMER_DISPATCH); + if (HAS_BSD(dev)) { + dev_priv->bsd_ring = bsd_ring; + ret = intel_init_ring_buffer(dev, &dev_priv->bsd_ring); + if (ret) + goto cleanup_render_ring; } return 0; + +cleanup_render_ring: + intel_cleanup_ring_buffer(dev, &dev_priv->render_ring); +cleanup_pipe_control: + if (HAS_PIPE_CONTROL(dev)) + i915_gem_cleanup_pipe_control(dev); + return ret; } void @@ -4884,17 +4645,11 @@ i915_gem_cleanup_ringbuffer(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; - if (dev_priv->ring.ring_obj == NULL) - return; - - drm_core_ioremapfree(&dev_priv->ring.map, dev); - - i915_gem_object_unpin(dev_priv->ring.ring_obj); - drm_gem_object_unreference(dev_priv->ring.ring_obj); - dev_priv->ring.ring_obj = NULL; - memset(&dev_priv->ring, 0, sizeof(dev_priv->ring)); - - i915_gem_cleanup_hws(dev); + intel_cleanup_ring_buffer(dev, &dev_priv->render_ring); + if (HAS_BSD(dev)) + intel_cleanup_ring_buffer(dev, &dev_priv->bsd_ring); + if (HAS_PIPE_CONTROL(dev)) + i915_gem_cleanup_pipe_control(dev); } int @@ -4922,12 +4677,14 @@ i915_gem_entervt_ioctl(struct drm_device *dev, void *data, } spin_lock(&dev_priv->mm.active_list_lock); - BUG_ON(!list_empty(&dev_priv->mm.active_list)); + BUG_ON(!list_empty(&dev_priv->render_ring.active_list)); + BUG_ON(HAS_BSD(dev) && !list_empty(&dev_priv->bsd_ring.active_list)); spin_unlock(&dev_priv->mm.active_list_lock); BUG_ON(!list_empty(&dev_priv->mm.flushing_list)); BUG_ON(!list_empty(&dev_priv->mm.inactive_list)); - BUG_ON(!list_empty(&dev_priv->mm.request_list)); + BUG_ON(!list_empty(&dev_priv->render_ring.request_list)); + BUG_ON(HAS_BSD(dev) && !list_empty(&dev_priv->bsd_ring.request_list)); mutex_unlock(&dev->struct_mutex); drm_irq_install(dev); @@ -4966,18 +4723,20 @@ i915_gem_load(struct drm_device *dev) drm_i915_private_t *dev_priv = dev->dev_private; spin_lock_init(&dev_priv->mm.active_list_lock); - INIT_LIST_HEAD(&dev_priv->mm.active_list); INIT_LIST_HEAD(&dev_priv->mm.flushing_list); INIT_LIST_HEAD(&dev_priv->mm.gpu_write_list); INIT_LIST_HEAD(&dev_priv->mm.inactive_list); - INIT_LIST_HEAD(&dev_priv->mm.request_list); INIT_LIST_HEAD(&dev_priv->mm.fence_list); + INIT_LIST_HEAD(&dev_priv->render_ring.active_list); + INIT_LIST_HEAD(&dev_priv->render_ring.request_list); + if (HAS_BSD(dev)) { + INIT_LIST_HEAD(&dev_priv->bsd_ring.active_list); + INIT_LIST_HEAD(&dev_priv->bsd_ring.request_list); + } for (i = 0; i < 16; i++) INIT_LIST_HEAD(&dev_priv->fence_regs[i].lru_list); INIT_DELAYED_WORK(&dev_priv->mm.retire_work, i915_gem_retire_work_handler); - dev_priv->mm.next_gem_seqno = 1; - spin_lock(&shrink_list_lock); list_add(&dev_priv->mm.shrink_list, &shrink_list); spin_unlock(&shrink_list_lock); @@ -5209,7 +4968,9 @@ i915_gpu_is_active(struct drm_device *dev) spin_lock(&dev_priv->mm.active_list_lock); lists_empty = list_empty(&dev_priv->mm.flushing_list) && - list_empty(&dev_priv->mm.active_list); + list_empty(&dev_priv->render_ring.active_list); + if (HAS_BSD(dev)) + lists_empty &= list_empty(&dev_priv->bsd_ring.active_list); spin_unlock(&dev_priv->mm.active_list_lock); return !lists_empty; @@ -5254,8 +5015,10 @@ rescan: continue; spin_unlock(&shrink_list_lock); + i915_gem_retire_requests(dev, &dev_priv->render_ring); - i915_gem_retire_requests(dev); + if (HAS_BSD(dev)) + i915_gem_retire_requests(dev, &dev_priv->bsd_ring); list_for_each_entry_safe(obj_priv, next_obj, &dev_priv->mm.inactive_list, diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 26792af7e1a1..dba53d4b9fb3 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -53,7 +53,7 @@ I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT) /** Interrupts that we mask and unmask at runtime. */ -#define I915_INTERRUPT_ENABLE_VAR (I915_USER_INTERRUPT) +#define I915_INTERRUPT_ENABLE_VAR (I915_USER_INTERRUPT | I915_BSD_USER_INTERRUPT) #define I915_PIPE_VBLANK_STATUS (PIPE_START_VBLANK_INTERRUPT_STATUS |\ PIPE_VBLANK_INTERRUPT_STATUS) @@ -74,7 +74,7 @@ ironlake_enable_graphics_irq(drm_i915_private_t *dev_priv, u32 mask) } } -static inline void +void ironlake_disable_graphics_irq(drm_i915_private_t *dev_priv, u32 mask) { if ((dev_priv->gt_irq_mask_reg & mask) != mask) { @@ -115,7 +115,7 @@ i915_enable_irq(drm_i915_private_t *dev_priv, u32 mask) } } -static inline void +void i915_disable_irq(drm_i915_private_t *dev_priv, u32 mask) { if ((dev_priv->irq_mask_reg & mask) != mask) { @@ -278,10 +278,9 @@ static void i915_handle_rps_change(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; u32 busy_up, busy_down, max_avg, min_avg; - u16 rgvswctl; u8 new_delay = dev_priv->cur_delay; - I915_WRITE(MEMINTRSTS, I915_READ(MEMINTRSTS) & ~MEMINT_EVAL_CHG); + I915_WRITE16(MEMINTRSTS, MEMINT_EVAL_CHG); busy_up = I915_READ(RCPREVBSYTUPAVG); busy_down = I915_READ(RCPREVBSYTDNAVG); max_avg = I915_READ(RCBMAXAVG); @@ -300,27 +299,8 @@ static void i915_handle_rps_change(struct drm_device *dev) new_delay = dev_priv->min_delay; } - DRM_DEBUG("rps change requested: %d -> %d\n", - dev_priv->cur_delay, new_delay); - - rgvswctl = I915_READ(MEMSWCTL); - if (rgvswctl & MEMCTL_CMD_STS) { - DRM_ERROR("gpu busy, RCS change rejected\n"); - return; /* still busy with another command */ - } - - /* Program the new state */ - rgvswctl = (MEMCTL_CMD_CHFREQ << MEMCTL_CMD_SHIFT) | - (new_delay << MEMCTL_FREQ_SHIFT) | MEMCTL_SFCAVM; - I915_WRITE(MEMSWCTL, rgvswctl); - POSTING_READ(MEMSWCTL); - - rgvswctl |= MEMCTL_CMD_STS; - I915_WRITE(MEMSWCTL, rgvswctl); - - dev_priv->cur_delay = new_delay; - - DRM_DEBUG("rps changed\n"); + if (ironlake_set_drps(dev, new_delay)) + dev_priv->cur_delay = new_delay; return; } @@ -331,6 +311,7 @@ irqreturn_t ironlake_irq_handler(struct drm_device *dev) int ret = IRQ_NONE; u32 de_iir, gt_iir, de_ier, pch_iir; struct drm_i915_master_private *master_priv; + struct intel_ring_buffer *render_ring = &dev_priv->render_ring; /* disable master interrupt before clearing iir */ de_ier = I915_READ(DEIER); @@ -354,13 +335,16 @@ irqreturn_t ironlake_irq_handler(struct drm_device *dev) } if (gt_iir & GT_PIPE_NOTIFY) { - u32 seqno = i915_get_gem_seqno(dev); - dev_priv->mm.irq_gem_seqno = seqno; + u32 seqno = render_ring->get_gem_seqno(dev, render_ring); + render_ring->irq_gem_seqno = seqno; trace_i915_gem_request_complete(dev, seqno); - DRM_WAKEUP(&dev_priv->irq_queue); + DRM_WAKEUP(&dev_priv->render_ring.irq_queue); dev_priv->hangcheck_count = 0; mod_timer(&dev_priv->hangcheck_timer, jiffies + DRM_I915_HANGCHECK_PERIOD); } + if (gt_iir & GT_BSD_USER_INTERRUPT) + DRM_WAKEUP(&dev_priv->bsd_ring.irq_queue); + if (de_iir & DE_GSE) ironlake_opregion_gse_intr(dev); @@ -388,7 +372,7 @@ irqreturn_t ironlake_irq_handler(struct drm_device *dev) } if (de_iir & DE_PCU_EVENT) { - I915_WRITE(MEMINTRSTS, I915_READ(MEMINTRSTS)); + I915_WRITE16(MEMINTRSTS, I915_READ(MEMINTRSTS)); i915_handle_rps_change(dev); } @@ -460,11 +444,15 @@ i915_error_object_create(struct drm_device *dev, for (page = 0; page < page_count; page++) { void *s, *d = kmalloc(PAGE_SIZE, GFP_ATOMIC); + unsigned long flags; + if (d == NULL) goto unwind; - s = kmap_atomic(src_priv->pages[page], KM_USER0); + local_irq_save(flags); + s = kmap_atomic(src_priv->pages[page], KM_IRQ0); memcpy(d, s, PAGE_SIZE); - kunmap_atomic(s, KM_USER0); + kunmap_atomic(s, KM_IRQ0); + local_irq_restore(flags); dst->pages[page] = d; } dst->page_count = page_count; @@ -532,17 +520,18 @@ i915_ringbuffer_last_batch(struct drm_device *dev) */ bbaddr = 0; head = I915_READ(PRB0_HEAD) & HEAD_ADDR; - ring = (u32 *)(dev_priv->ring.virtual_start + head); + ring = (u32 *)(dev_priv->render_ring.virtual_start + head); - while (--ring >= (u32 *)dev_priv->ring.virtual_start) { + while (--ring >= (u32 *)dev_priv->render_ring.virtual_start) { bbaddr = i915_get_bbaddr(dev, ring); if (bbaddr) break; } if (bbaddr == 0) { - ring = (u32 *)(dev_priv->ring.virtual_start + dev_priv->ring.Size); - while (--ring >= (u32 *)dev_priv->ring.virtual_start) { + ring = (u32 *)(dev_priv->render_ring.virtual_start + + dev_priv->render_ring.size); + while (--ring >= (u32 *)dev_priv->render_ring.virtual_start) { bbaddr = i915_get_bbaddr(dev, ring); if (bbaddr) break; @@ -583,7 +572,7 @@ static void i915_capture_error_state(struct drm_device *dev) return; } - error->seqno = i915_get_gem_seqno(dev); + error->seqno = i915_get_gem_seqno(dev, &dev_priv->render_ring); error->eir = I915_READ(EIR); error->pgtbl_er = I915_READ(PGTBL_ER); error->pipeastat = I915_READ(PIPEASTAT); @@ -611,7 +600,9 @@ static void i915_capture_error_state(struct drm_device *dev) batchbuffer[0] = NULL; batchbuffer[1] = NULL; count = 0; - list_for_each_entry(obj_priv, &dev_priv->mm.active_list, list) { + list_for_each_entry(obj_priv, + &dev_priv->render_ring.active_list, list) { + struct drm_gem_object *obj = &obj_priv->base; if (batchbuffer[0] == NULL && @@ -635,7 +626,8 @@ static void i915_capture_error_state(struct drm_device *dev) error->batchbuffer[1] = i915_error_object_create(dev, batchbuffer[1]); /* Record the ringbuffer */ - error->ringbuffer = i915_error_object_create(dev, dev_priv->ring.ring_obj); + error->ringbuffer = i915_error_object_create(dev, + dev_priv->render_ring.gem_object); /* Record buffers on the active list. */ error->active_bo = NULL; @@ -647,7 +639,8 @@ static void i915_capture_error_state(struct drm_device *dev) if (error->active_bo) { int i = 0; - list_for_each_entry(obj_priv, &dev_priv->mm.active_list, list) { + list_for_each_entry(obj_priv, + &dev_priv->render_ring.active_list, list) { struct drm_gem_object *obj = &obj_priv->base; error->active_bo[i].size = obj->size; @@ -699,24 +692,13 @@ void i915_destroy_error_state(struct drm_device *dev) i915_error_state_free(dev, error); } -/** - * i915_handle_error - handle an error interrupt - * @dev: drm device - * - * Do some basic checking of regsiter state at error interrupt time and - * dump it to the syslog. Also call i915_capture_error_state() to make - * sure we get a record and make it available in debugfs. Fire a uevent - * so userspace knows something bad happened (should trigger collection - * of a ring dump etc.). - */ -static void i915_handle_error(struct drm_device *dev, bool wedged) +static void i915_report_and_clear_eir(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; u32 eir = I915_READ(EIR); - u32 pipea_stats = I915_READ(PIPEASTAT); - u32 pipeb_stats = I915_READ(PIPEBSTAT); - i915_capture_error_state(dev); + if (!eir) + return; printk(KERN_ERR "render error detected, EIR: 0x%08x\n", eir); @@ -762,6 +744,9 @@ static void i915_handle_error(struct drm_device *dev, bool wedged) } if (eir & I915_ERROR_MEMORY_REFRESH) { + u32 pipea_stats = I915_READ(PIPEASTAT); + u32 pipeb_stats = I915_READ(PIPEBSTAT); + printk(KERN_ERR "memory refresh error\n"); printk(KERN_ERR "PIPEASTAT: 0x%08x\n", pipea_stats); @@ -818,6 +803,24 @@ static void i915_handle_error(struct drm_device *dev, bool wedged) I915_WRITE(EMR, I915_READ(EMR) | eir); I915_WRITE(IIR, I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT); } +} + +/** + * i915_handle_error - handle an error interrupt + * @dev: drm device + * + * Do some basic checking of regsiter state at error interrupt time and + * dump it to the syslog. Also call i915_capture_error_state() to make + * sure we get a record and make it available in debugfs. Fire a uevent + * so userspace knows something bad happened (should trigger collection + * of a ring dump etc.). + */ +static void i915_handle_error(struct drm_device *dev, bool wedged) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + + i915_capture_error_state(dev); + i915_report_and_clear_eir(dev); if (wedged) { atomic_set(&dev_priv->mm.wedged, 1); @@ -825,7 +828,7 @@ static void i915_handle_error(struct drm_device *dev, bool wedged) /* * Wakeup waiting processes so they don't hang */ - DRM_WAKEUP(&dev_priv->irq_queue); + DRM_WAKEUP(&dev_priv->render_ring.irq_queue); } queue_work(dev_priv->wq, &dev_priv->error_work); @@ -844,6 +847,7 @@ irqreturn_t i915_driver_irq_handler(DRM_IRQ_ARGS) unsigned long irqflags; int irq_received; int ret = IRQ_NONE; + struct intel_ring_buffer *render_ring = &dev_priv->render_ring; atomic_inc(&dev_priv->irq_received); @@ -924,30 +928,42 @@ irqreturn_t i915_driver_irq_handler(DRM_IRQ_ARGS) } if (iir & I915_USER_INTERRUPT) { - u32 seqno = i915_get_gem_seqno(dev); - dev_priv->mm.irq_gem_seqno = seqno; + u32 seqno = + render_ring->get_gem_seqno(dev, render_ring); + render_ring->irq_gem_seqno = seqno; trace_i915_gem_request_complete(dev, seqno); - DRM_WAKEUP(&dev_priv->irq_queue); + DRM_WAKEUP(&dev_priv->render_ring.irq_queue); dev_priv->hangcheck_count = 0; mod_timer(&dev_priv->hangcheck_timer, jiffies + DRM_I915_HANGCHECK_PERIOD); } - if (iir & I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT) + if (HAS_BSD(dev) && (iir & I915_BSD_USER_INTERRUPT)) + DRM_WAKEUP(&dev_priv->bsd_ring.irq_queue); + + if (iir & I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT) { intel_prepare_page_flip(dev, 0); + if (dev_priv->flip_pending_is_done) + intel_finish_page_flip_plane(dev, 0); + } - if (iir & I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT) + if (iir & I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT) { intel_prepare_page_flip(dev, 1); + if (dev_priv->flip_pending_is_done) + intel_finish_page_flip_plane(dev, 1); + } if (pipea_stats & vblank_status) { vblank++; drm_handle_vblank(dev, 0); - intel_finish_page_flip(dev, 0); + if (!dev_priv->flip_pending_is_done) + intel_finish_page_flip(dev, 0); } if (pipeb_stats & vblank_status) { vblank++; drm_handle_vblank(dev, 1); - intel_finish_page_flip(dev, 1); + if (!dev_priv->flip_pending_is_done) + intel_finish_page_flip(dev, 1); } if ((pipea_stats & I915_LEGACY_BLC_EVENT_STATUS) || @@ -980,7 +996,6 @@ static int i915_emit_irq(struct drm_device * dev) { drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv; - RING_LOCALS; i915_kernel_lost_context(dev); @@ -1002,43 +1017,13 @@ static int i915_emit_irq(struct drm_device * dev) return dev_priv->counter; } -void i915_user_irq_get(struct drm_device *dev) -{ - drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; - unsigned long irqflags; - - spin_lock_irqsave(&dev_priv->user_irq_lock, irqflags); - if (dev->irq_enabled && (++dev_priv->user_irq_refcount == 1)) { - if (HAS_PCH_SPLIT(dev)) - ironlake_enable_graphics_irq(dev_priv, GT_PIPE_NOTIFY); - else - i915_enable_irq(dev_priv, I915_USER_INTERRUPT); - } - spin_unlock_irqrestore(&dev_priv->user_irq_lock, irqflags); -} - -void i915_user_irq_put(struct drm_device *dev) -{ - drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; - unsigned long irqflags; - - spin_lock_irqsave(&dev_priv->user_irq_lock, irqflags); - BUG_ON(dev->irq_enabled && dev_priv->user_irq_refcount <= 0); - if (dev->irq_enabled && (--dev_priv->user_irq_refcount == 0)) { - if (HAS_PCH_SPLIT(dev)) - ironlake_disable_graphics_irq(dev_priv, GT_PIPE_NOTIFY); - else - i915_disable_irq(dev_priv, I915_USER_INTERRUPT); - } - spin_unlock_irqrestore(&dev_priv->user_irq_lock, irqflags); -} - void i915_trace_irq_get(struct drm_device *dev, u32 seqno) { drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; + struct intel_ring_buffer *render_ring = &dev_priv->render_ring; if (dev_priv->trace_irq_seqno == 0) - i915_user_irq_get(dev); + render_ring->user_irq_get(dev, render_ring); dev_priv->trace_irq_seqno = seqno; } @@ -1048,6 +1033,7 @@ static int i915_wait_irq(struct drm_device * dev, int irq_nr) drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv; int ret = 0; + struct intel_ring_buffer *render_ring = &dev_priv->render_ring; DRM_DEBUG_DRIVER("irq_nr=%d breadcrumb=%d\n", irq_nr, READ_BREADCRUMB(dev_priv)); @@ -1061,10 +1047,10 @@ static int i915_wait_irq(struct drm_device * dev, int irq_nr) if (master_priv->sarea_priv) master_priv->sarea_priv->perf_boxes |= I915_BOX_WAIT; - i915_user_irq_get(dev); - DRM_WAIT_ON(ret, dev_priv->irq_queue, 3 * DRM_HZ, + render_ring->user_irq_get(dev, render_ring); + DRM_WAIT_ON(ret, dev_priv->render_ring.irq_queue, 3 * DRM_HZ, READ_BREADCRUMB(dev_priv) >= irq_nr); - i915_user_irq_put(dev); + render_ring->user_irq_put(dev, render_ring); if (ret == -EBUSY) { DRM_ERROR("EBUSY -- rec: %d emitted: %d\n", @@ -1083,7 +1069,7 @@ int i915_irq_emit(struct drm_device *dev, void *data, drm_i915_irq_emit_t *emit = data; int result; - if (!dev_priv || !dev_priv->ring.virtual_start) { + if (!dev_priv || !dev_priv->render_ring.virtual_start) { DRM_ERROR("called with no initialization\n"); return -EINVAL; } @@ -1229,9 +1215,12 @@ int i915_vblank_swap(struct drm_device *dev, void *data, return -EINVAL; } -struct drm_i915_gem_request *i915_get_tail_request(struct drm_device *dev) { +struct drm_i915_gem_request * +i915_get_tail_request(struct drm_device *dev) +{ drm_i915_private_t *dev_priv = dev->dev_private; - return list_entry(dev_priv->mm.request_list.prev, struct drm_i915_gem_request, list); + return list_entry(dev_priv->render_ring.request_list.prev, + struct drm_i915_gem_request, list); } /** @@ -1256,8 +1245,10 @@ void i915_hangcheck_elapsed(unsigned long data) acthd = I915_READ(ACTHD_I965); /* If all work is done then ACTHD clearly hasn't advanced. */ - if (list_empty(&dev_priv->mm.request_list) || - i915_seqno_passed(i915_get_gem_seqno(dev), i915_get_tail_request(dev)->seqno)) { + if (list_empty(&dev_priv->render_ring.request_list) || + i915_seqno_passed(i915_get_gem_seqno(dev, + &dev_priv->render_ring), + i915_get_tail_request(dev)->seqno)) { dev_priv->hangcheck_count = 0; return; } @@ -1310,7 +1301,7 @@ static int ironlake_irq_postinstall(struct drm_device *dev) /* enable kind of interrupts always enabled */ u32 display_mask = DE_MASTER_IRQ_CONTROL | DE_GSE | DE_PCH_EVENT | DE_PLANEA_FLIP_DONE | DE_PLANEB_FLIP_DONE; - u32 render_mask = GT_PIPE_NOTIFY; + u32 render_mask = GT_PIPE_NOTIFY | GT_BSD_USER_INTERRUPT; u32 hotplug_mask = SDE_CRT_HOTPLUG | SDE_PORTB_HOTPLUG | SDE_PORTC_HOTPLUG | SDE_PORTD_HOTPLUG; @@ -1324,7 +1315,7 @@ static int ironlake_irq_postinstall(struct drm_device *dev) (void) I915_READ(DEIER); /* user interrupt should be enabled, but masked initial */ - dev_priv->gt_irq_mask_reg = 0xffffffff; + dev_priv->gt_irq_mask_reg = ~render_mask; dev_priv->gt_irq_enable_reg = render_mask; I915_WRITE(GTIIR, I915_READ(GTIIR)); @@ -1387,7 +1378,10 @@ int i915_driver_irq_postinstall(struct drm_device *dev) u32 enable_mask = I915_INTERRUPT_ENABLE_FIX | I915_INTERRUPT_ENABLE_VAR; u32 error_mask; - DRM_INIT_WAITQUEUE(&dev_priv->irq_queue); + DRM_INIT_WAITQUEUE(&dev_priv->render_ring.irq_queue); + + if (HAS_BSD(dev)) + DRM_INIT_WAITQUEUE(&dev_priv->bsd_ring.irq_queue); dev_priv->vblank_pipe = DRM_I915_VBLANK_PIPE_A | DRM_I915_VBLANK_PIPE_B; @@ -1401,29 +1395,10 @@ int i915_driver_irq_postinstall(struct drm_device *dev) dev_priv->pipestat[1] = 0; if (I915_HAS_HOTPLUG(dev)) { - u32 hotplug_en = I915_READ(PORT_HOTPLUG_EN); - - /* Note HDMI and DP share bits */ - if (dev_priv->hotplug_supported_mask & HDMIB_HOTPLUG_INT_STATUS) - hotplug_en |= HDMIB_HOTPLUG_INT_EN; - if (dev_priv->hotplug_supported_mask & HDMIC_HOTPLUG_INT_STATUS) - hotplug_en |= HDMIC_HOTPLUG_INT_EN; - if (dev_priv->hotplug_supported_mask & HDMID_HOTPLUG_INT_STATUS) - hotplug_en |= HDMID_HOTPLUG_INT_EN; - if (dev_priv->hotplug_supported_mask & SDVOC_HOTPLUG_INT_STATUS) - hotplug_en |= SDVOC_HOTPLUG_INT_EN; - if (dev_priv->hotplug_supported_mask & SDVOB_HOTPLUG_INT_STATUS) - hotplug_en |= SDVOB_HOTPLUG_INT_EN; - if (dev_priv->hotplug_supported_mask & CRT_HOTPLUG_INT_STATUS) - hotplug_en |= CRT_HOTPLUG_INT_EN; - /* Ignore TV since it's buggy */ - - I915_WRITE(PORT_HOTPLUG_EN, hotplug_en); - /* Enable in IER... */ enable_mask |= I915_DISPLAY_PORT_INTERRUPT; /* and unmask in IMR */ - i915_enable_irq(dev_priv, I915_DISPLAY_PORT_INTERRUPT); + dev_priv->irq_mask_reg &= ~I915_DISPLAY_PORT_INTERRUPT; } /* @@ -1441,16 +1416,41 @@ int i915_driver_irq_postinstall(struct drm_device *dev) } I915_WRITE(EMR, error_mask); - /* Disable pipe interrupt enables, clear pending pipe status */ - I915_WRITE(PIPEASTAT, I915_READ(PIPEASTAT) & 0x8000ffff); - I915_WRITE(PIPEBSTAT, I915_READ(PIPEBSTAT) & 0x8000ffff); - /* Clear pending interrupt status */ - I915_WRITE(IIR, I915_READ(IIR)); - - I915_WRITE(IER, enable_mask); I915_WRITE(IMR, dev_priv->irq_mask_reg); + I915_WRITE(IER, enable_mask); (void) I915_READ(IER); + if (I915_HAS_HOTPLUG(dev)) { + u32 hotplug_en = I915_READ(PORT_HOTPLUG_EN); + + /* Note HDMI and DP share bits */ + if (dev_priv->hotplug_supported_mask & HDMIB_HOTPLUG_INT_STATUS) + hotplug_en |= HDMIB_HOTPLUG_INT_EN; + if (dev_priv->hotplug_supported_mask & HDMIC_HOTPLUG_INT_STATUS) + hotplug_en |= HDMIC_HOTPLUG_INT_EN; + if (dev_priv->hotplug_supported_mask & HDMID_HOTPLUG_INT_STATUS) + hotplug_en |= HDMID_HOTPLUG_INT_EN; + if (dev_priv->hotplug_supported_mask & SDVOC_HOTPLUG_INT_STATUS) + hotplug_en |= SDVOC_HOTPLUG_INT_EN; + if (dev_priv->hotplug_supported_mask & SDVOB_HOTPLUG_INT_STATUS) + hotplug_en |= SDVOB_HOTPLUG_INT_EN; + if (dev_priv->hotplug_supported_mask & CRT_HOTPLUG_INT_STATUS) { + hotplug_en |= CRT_HOTPLUG_INT_EN; + + /* Programming the CRT detection parameters tends + to generate a spurious hotplug event about three + seconds later. So just do it once. + */ + if (IS_G4X(dev)) + hotplug_en |= CRT_HOTPLUG_ACTIVATION_PERIOD_64; + hotplug_en |= CRT_HOTPLUG_VOLTAGE_COMPARE_50; + } + + /* Ignore TV since it's buggy */ + + I915_WRITE(PORT_HOTPLUG_EN, hotplug_en); + } + opregion_enable_asle(dev); return 0; diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index f3e39cc46f0d..150400f40534 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -178,6 +178,7 @@ #define MI_OVERLAY_OFF (0x2<<21) #define MI_LOAD_SCAN_LINES_INCL MI_INSTR(0x12, 0) #define MI_DISPLAY_FLIP MI_INSTR(0x14, 2) +#define MI_DISPLAY_FLIP_I915 MI_INSTR(0x14, 1) #define MI_DISPLAY_FLIP_PLANE(n) ((n) << 20) #define MI_STORE_DWORD_IMM MI_INSTR(0x20, 1) #define MI_MEM_VIRTUAL (1 << 22) /* 965+ only */ @@ -334,6 +335,7 @@ #define I915_DEBUG_INTERRUPT (1<<2) #define I915_USER_INTERRUPT (1<<1) #define I915_ASLE_INTERRUPT (1<<0) +#define I915_BSD_USER_INTERRUPT (1<<25) #define EIR 0x020b0 #define EMR 0x020b4 #define ESR 0x020b8 @@ -367,7 +369,40 @@ #define CM0_RC_OP_FLUSH_DISABLE (1<<0) #define BB_ADDR 0x02140 /* 8 bytes */ #define GFX_FLSH_CNTL 0x02170 /* 915+ only */ +#define ECOSKPD 0x021d0 +#define ECO_GATING_CX_ONLY (1<<3) +#define ECO_FLIP_DONE (1<<0) + +/* GEN6 interrupt control */ +#define GEN6_RENDER_HWSTAM 0x2098 +#define GEN6_RENDER_IMR 0x20a8 +#define GEN6_RENDER_CONTEXT_SWITCH_INTERRUPT (1 << 8) +#define GEN6_RENDER_PPGTT_PAGE_FAULT (1 << 7) +#define GEN6_RENDER TIMEOUT_COUNTER_EXPIRED (1 << 6) +#define GEN6_RENDER_L3_PARITY_ERROR (1 << 5) +#define GEN6_RENDER_PIPE_CONTROL_NOTIFY_INTERRUPT (1 << 4) +#define GEN6_RENDER_COMMAND_PARSER_MASTER_ERROR (1 << 3) +#define GEN6_RENDER_SYNC_STATUS (1 << 2) +#define GEN6_RENDER_DEBUG_INTERRUPT (1 << 1) +#define GEN6_RENDER_USER_INTERRUPT (1 << 0) + +#define GEN6_BLITTER_HWSTAM 0x22098 +#define GEN6_BLITTER_IMR 0x220a8 +#define GEN6_BLITTER_MI_FLUSH_DW_NOTIFY_INTERRUPT (1 << 26) +#define GEN6_BLITTER_COMMAND_PARSER_MASTER_ERROR (1 << 25) +#define GEN6_BLITTER_SYNC_STATUS (1 << 24) +#define GEN6_BLITTER_USER_INTERRUPT (1 << 22) +/* + * BSD (bit stream decoder instruction and interrupt control register defines + * (G4X and Ironlake only) + */ +#define BSD_RING_TAIL 0x04030 +#define BSD_RING_HEAD 0x04034 +#define BSD_RING_START 0x04038 +#define BSD_RING_CTL 0x0403c +#define BSD_RING_ACTHD 0x04074 +#define BSD_HWS_PGA 0x04080 /* * Framebuffer compression (915+ only) @@ -805,6 +840,10 @@ #define DCC_CHANNEL_XOR_DISABLE (1 << 10) #define DCC_CHANNEL_XOR_BIT_17 (1 << 9) +/** Pineview MCH register contains DDR3 setting */ +#define CSHRDDR3CTL 0x101a8 +#define CSHRDDR3CTL_DDR3 (1 << 2) + /** 965 MCH register controlling DRAM channel configuration */ #define C0DRB3 0x10206 #define C1DRB3 0x10606 @@ -826,6 +865,12 @@ #define CLKCFG_MEM_800 (3 << 4) #define CLKCFG_MEM_MASK (7 << 4) +#define TR1 0x11006 +#define TSFS 0x11020 +#define TSFS_SLOPE_MASK 0x0000ff00 +#define TSFS_SLOPE_SHIFT 8 +#define TSFS_INTR_MASK 0x000000ff + #define CRSTANDVID 0x11100 #define PXVFREQ_BASE 0x11110 /* P[0-15]VIDFREQ (0x1114c) (Ironlake) */ #define PXVFREQ_PX_MASK 0x7f000000 @@ -964,6 +1009,41 @@ #define MEMSTAT_SRC_CTL_STDBY 3 #define RCPREVBSYTUPAVG 0x113b8 #define RCPREVBSYTDNAVG 0x113bc +#define SDEW 0x1124c +#define CSIEW0 0x11250 +#define CSIEW1 0x11254 +#define CSIEW2 0x11258 +#define PEW 0x1125c +#define DEW 0x11270 +#define MCHAFE 0x112c0 +#define CSIEC 0x112e0 +#define DMIEC 0x112e4 +#define DDREC 0x112e8 +#define PEG0EC 0x112ec +#define PEG1EC 0x112f0 +#define GFXEC 0x112f4 +#define RPPREVBSYTUPAVG 0x113b8 +#define RPPREVBSYTDNAVG 0x113bc +#define ECR 0x11600 +#define ECR_GPFE (1<<31) +#define ECR_IMONE (1<<30) +#define ECR_CAP_MASK 0x0000001f /* Event range, 0-31 */ +#define OGW0 0x11608 +#define OGW1 0x1160c +#define EG0 0x11610 +#define EG1 0x11614 +#define EG2 0x11618 +#define EG3 0x1161c +#define EG4 0x11620 +#define EG5 0x11624 +#define EG6 0x11628 +#define EG7 0x1162c +#define PXW 0x11664 +#define PXWL 0x11680 +#define LCFUSE02 0x116c0 +#define LCFUSE_HIV_MASK 0x000000ff +#define CSIPLL0 0x12c10 +#define DDRMPLL1 0X12c20 #define PEG_BAND_GAP_DATA 0x14d68 /* @@ -1054,8 +1134,6 @@ #define CRT_HOTPLUG_DETECT_DELAY_2G (1 << 4) #define CRT_HOTPLUG_DETECT_VOLTAGE_325MV (0 << 2) #define CRT_HOTPLUG_DETECT_VOLTAGE_475MV (1 << 2) -#define CRT_HOTPLUG_MASK (0x3fc) /* Bits 9-2 */ -#define CRT_FORCE_HOTPLUG_MASK 0xfffffe1f #define PORT_HOTPLUG_STAT 0x61114 #define HDMIB_HOTPLUG_INT_STATUS (1 << 29) @@ -2355,6 +2433,8 @@ #define GT_PIPE_NOTIFY (1 << 4) #define GT_SYNC_STATUS (1 << 2) #define GT_USER_INTERRUPT (1 << 0) +#define GT_BSD_USER_INTERRUPT (1 << 5) + #define GTISR 0x44010 #define GTIMR 0x44014 @@ -2690,6 +2770,9 @@ #define SDVO_ENCODING (0) #define TMDS_ENCODING (2 << 10) #define NULL_PACKET_VSYNC_ENABLE (1 << 9) +/* CPT */ +#define HDMI_MODE_SELECT (1 << 9) +#define DVI_MODE_SELECT (0) #define SDVOB_BORDER_ENABLE (1 << 7) #define AUDIO_ENABLE (1 << 6) #define VSYNC_ACTIVE_HIGH (1 << 4) diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h index 9e4c45f68d6e..fab21760dd57 100644 --- a/drivers/gpu/drm/i915/i915_trace.h +++ b/drivers/gpu/drm/i915/i915_trace.h @@ -53,23 +53,6 @@ TRACE_EVENT(i915_gem_object_bind, __entry->obj, __entry->gtt_offset) ); -TRACE_EVENT(i915_gem_object_clflush, - - TP_PROTO(struct drm_gem_object *obj), - - TP_ARGS(obj), - - TP_STRUCT__entry( - __field(struct drm_gem_object *, obj) - ), - - TP_fast_assign( - __entry->obj = obj; - ), - - TP_printk("obj=%p", __entry->obj) -); - TRACE_EVENT(i915_gem_object_change_domain, TP_PROTO(struct drm_gem_object *obj, uint32_t old_read_domains, uint32_t old_write_domain), @@ -132,6 +115,13 @@ DECLARE_EVENT_CLASS(i915_gem_object, TP_printk("obj=%p", __entry->obj) ); +DEFINE_EVENT(i915_gem_object, i915_gem_object_clflush, + + TP_PROTO(struct drm_gem_object *obj), + + TP_ARGS(obj) +); + DEFINE_EVENT(i915_gem_object, i915_gem_object_unbind, TP_PROTO(struct drm_gem_object *obj), diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c index 4c748d8f73d6..96f75d7f6633 100644 --- a/drivers/gpu/drm/i915/intel_bios.c +++ b/drivers/gpu/drm/i915/intel_bios.c @@ -95,6 +95,16 @@ fill_detail_timing_data(struct drm_display_mode *panel_fixed_mode, panel_fixed_mode->clock = dvo_timing->clock * 10; panel_fixed_mode->type = DRM_MODE_TYPE_PREFERRED; + if (dvo_timing->hsync_positive) + panel_fixed_mode->flags |= DRM_MODE_FLAG_PHSYNC; + else + panel_fixed_mode->flags |= DRM_MODE_FLAG_NHSYNC; + + if (dvo_timing->vsync_positive) + panel_fixed_mode->flags |= DRM_MODE_FLAG_PVSYNC; + else + panel_fixed_mode->flags |= DRM_MODE_FLAG_NVSYNC; + /* Some VBTs have bogus h/vtotal values */ if (panel_fixed_mode->hsync_end > panel_fixed_mode->htotal) panel_fixed_mode->htotal = panel_fixed_mode->hsync_end + 1; diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c index e16ac5a28c3c..ee0732b222a1 100644 --- a/drivers/gpu/drm/i915/intel_crt.c +++ b/drivers/gpu/drm/i915/intel_crt.c @@ -217,7 +217,8 @@ static bool intel_crt_detect_hotplug(struct drm_connector *connector) { struct drm_device *dev = connector->dev; struct drm_i915_private *dev_priv = dev->dev_private; - u32 hotplug_en; + u32 hotplug_en, orig, stat; + bool ret = false; int i, tries = 0; if (HAS_PCH_SPLIT(dev)) @@ -232,15 +233,9 @@ static bool intel_crt_detect_hotplug(struct drm_connector *connector) tries = 2; else tries = 1; - hotplug_en = I915_READ(PORT_HOTPLUG_EN); - hotplug_en &= CRT_FORCE_HOTPLUG_MASK; + hotplug_en = orig = I915_READ(PORT_HOTPLUG_EN); hotplug_en |= CRT_HOTPLUG_FORCE_DETECT; - if (IS_G4X(dev)) - hotplug_en |= CRT_HOTPLUG_ACTIVATION_PERIOD_64; - - hotplug_en |= CRT_HOTPLUG_VOLTAGE_COMPARE_50; - for (i = 0; i < tries ; i++) { unsigned long timeout; /* turn on the FORCE_DETECT */ @@ -255,11 +250,17 @@ static bool intel_crt_detect_hotplug(struct drm_connector *connector) } while (time_after(timeout, jiffies)); } - if ((I915_READ(PORT_HOTPLUG_STAT) & CRT_HOTPLUG_MONITOR_MASK) != - CRT_HOTPLUG_MONITOR_NONE) - return true; + stat = I915_READ(PORT_HOTPLUG_STAT); + if ((stat & CRT_HOTPLUG_MONITOR_MASK) != CRT_HOTPLUG_MONITOR_NONE) + ret = true; + + /* clear the interrupt we just generated, if any */ + I915_WRITE(PORT_HOTPLUG_STAT, CRT_HOTPLUG_INT_STATUS); + + /* and put the bits back */ + I915_WRITE(PORT_HOTPLUG_EN, orig); - return false; + return ret; } static bool intel_crt_detect_ddc(struct drm_encoder *encoder) @@ -569,7 +570,7 @@ void intel_crt_init(struct drm_device *dev) (1 << INTEL_ANALOG_CLONE_BIT) | (1 << INTEL_SDVO_LVDS_CLONE_BIT); intel_encoder->crtc_mask = (1 << 0) | (1 << 1); - connector->interlace_allowed = 0; + connector->interlace_allowed = 1; connector->doublescan_allowed = 0; drm_encoder_helper_add(&intel_encoder->enc, &intel_crt_helper_funcs); diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 8c668e3122a5..68dcf36e2793 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -887,9 +887,9 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, memset(best_clock, 0, sizeof(*best_clock)); max_n = limit->n.max; - /* based on hardware requriment prefer smaller n to precision */ + /* based on hardware requirement, prefer smaller n to precision */ for (clock.n = limit->n.min; clock.n <= max_n; clock.n++) { - /* based on hardware requirment prefere larger m1,m2 */ + /* based on hardware requirement, prefere larger m1,m2 */ for (clock.m1 = limit->m1.max; clock.m1 >= limit->m1.min; clock.m1--) { for (clock.m2 = limit->m2.max; @@ -1029,19 +1029,28 @@ static void i8xx_enable_fbc(struct drm_crtc *crtc, unsigned long interval) void i8xx_disable_fbc(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; + unsigned long timeout = jiffies + msecs_to_jiffies(1); u32 fbc_ctl; if (!I915_HAS_FBC(dev)) return; + if (!(I915_READ(FBC_CONTROL) & FBC_CTL_EN)) + return; /* Already off, just return */ + /* Disable compression */ fbc_ctl = I915_READ(FBC_CONTROL); fbc_ctl &= ~FBC_CTL_EN; I915_WRITE(FBC_CONTROL, fbc_ctl); /* Wait for compressing bit to clear */ - while (I915_READ(FBC_STATUS) & FBC_STAT_COMPRESSING) - ; /* nothing */ + while (I915_READ(FBC_STATUS) & FBC_STAT_COMPRESSING) { + if (time_after(jiffies, timeout)) { + DRM_DEBUG_DRIVER("FBC idle timed out\n"); + break; + } + ; /* do nothing */ + } intel_wait_for_vblank(dev); @@ -1239,10 +1248,11 @@ static void intel_update_fbc(struct drm_crtc *crtc, return; out_disable: - DRM_DEBUG_KMS("unsupported config, disabling FBC\n"); /* Multiple disables should be harmless */ - if (intel_fbc_enabled(dev)) + if (intel_fbc_enabled(dev)) { + DRM_DEBUG_KMS("unsupported config, disabling FBC\n"); intel_disable_fbc(dev); + } } static int @@ -1386,7 +1396,8 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y, Start = obj_priv->gtt_offset; Offset = y * crtc->fb->pitch + x * (crtc->fb->bits_per_pixel / 8); - DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d\n", Start, Offset, x, y); + DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d %d\n", + Start, Offset, x, y, crtc->fb->pitch); I915_WRITE(dspstride, crtc->fb->pitch); if (IS_I965G(dev)) { I915_WRITE(dspbase, Offset); @@ -2345,6 +2356,8 @@ static bool intel_crtc_mode_fixup(struct drm_crtc *crtc, if (mode->clock * 3 > 27000 * 4) return MODE_CLOCK_HIGH; } + + drm_mode_set_crtcinfo(adjusted_mode, 0); return true; } @@ -2629,6 +2642,7 @@ static unsigned long intel_calculate_wm(unsigned long clock_in_khz, struct cxsr_latency { int is_desktop; + int is_ddr3; unsigned long fsb_freq; unsigned long mem_freq; unsigned long display_sr; @@ -2638,33 +2652,45 @@ struct cxsr_latency { }; static struct cxsr_latency cxsr_latency_table[] = { - {1, 800, 400, 3382, 33382, 3983, 33983}, /* DDR2-400 SC */ - {1, 800, 667, 3354, 33354, 3807, 33807}, /* DDR2-667 SC */ - {1, 800, 800, 3347, 33347, 3763, 33763}, /* DDR2-800 SC */ - - {1, 667, 400, 3400, 33400, 4021, 34021}, /* DDR2-400 SC */ - {1, 667, 667, 3372, 33372, 3845, 33845}, /* DDR2-667 SC */ - {1, 667, 800, 3386, 33386, 3822, 33822}, /* DDR2-800 SC */ - - {1, 400, 400, 3472, 33472, 4173, 34173}, /* DDR2-400 SC */ - {1, 400, 667, 3443, 33443, 3996, 33996}, /* DDR2-667 SC */ - {1, 400, 800, 3430, 33430, 3946, 33946}, /* DDR2-800 SC */ - - {0, 800, 400, 3438, 33438, 4065, 34065}, /* DDR2-400 SC */ - {0, 800, 667, 3410, 33410, 3889, 33889}, /* DDR2-667 SC */ - {0, 800, 800, 3403, 33403, 3845, 33845}, /* DDR2-800 SC */ - - {0, 667, 400, 3456, 33456, 4103, 34106}, /* DDR2-400 SC */ - {0, 667, 667, 3428, 33428, 3927, 33927}, /* DDR2-667 SC */ - {0, 667, 800, 3443, 33443, 3905, 33905}, /* DDR2-800 SC */ - - {0, 400, 400, 3528, 33528, 4255, 34255}, /* DDR2-400 SC */ - {0, 400, 667, 3500, 33500, 4079, 34079}, /* DDR2-667 SC */ - {0, 400, 800, 3487, 33487, 4029, 34029}, /* DDR2-800 SC */ + {1, 0, 800, 400, 3382, 33382, 3983, 33983}, /* DDR2-400 SC */ + {1, 0, 800, 667, 3354, 33354, 3807, 33807}, /* DDR2-667 SC */ + {1, 0, 800, 800, 3347, 33347, 3763, 33763}, /* DDR2-800 SC */ + {1, 1, 800, 667, 6420, 36420, 6873, 36873}, /* DDR3-667 SC */ + {1, 1, 800, 800, 5902, 35902, 6318, 36318}, /* DDR3-800 SC */ + + {1, 0, 667, 400, 3400, 33400, 4021, 34021}, /* DDR2-400 SC */ + {1, 0, 667, 667, 3372, 33372, 3845, 33845}, /* DDR2-667 SC */ + {1, 0, 667, 800, 3386, 33386, 3822, 33822}, /* DDR2-800 SC */ + {1, 1, 667, 667, 6438, 36438, 6911, 36911}, /* DDR3-667 SC */ + {1, 1, 667, 800, 5941, 35941, 6377, 36377}, /* DDR3-800 SC */ + + {1, 0, 400, 400, 3472, 33472, 4173, 34173}, /* DDR2-400 SC */ + {1, 0, 400, 667, 3443, 33443, 3996, 33996}, /* DDR2-667 SC */ + {1, 0, 400, 800, 3430, 33430, 3946, 33946}, /* DDR2-800 SC */ + {1, 1, 400, 667, 6509, 36509, 7062, 37062}, /* DDR3-667 SC */ + {1, 1, 400, 800, 5985, 35985, 6501, 36501}, /* DDR3-800 SC */ + + {0, 0, 800, 400, 3438, 33438, 4065, 34065}, /* DDR2-400 SC */ + {0, 0, 800, 667, 3410, 33410, 3889, 33889}, /* DDR2-667 SC */ + {0, 0, 800, 800, 3403, 33403, 3845, 33845}, /* DDR2-800 SC */ + {0, 1, 800, 667, 6476, 36476, 6955, 36955}, /* DDR3-667 SC */ + {0, 1, 800, 800, 5958, 35958, 6400, 36400}, /* DDR3-800 SC */ + + {0, 0, 667, 400, 3456, 33456, 4103, 34106}, /* DDR2-400 SC */ + {0, 0, 667, 667, 3428, 33428, 3927, 33927}, /* DDR2-667 SC */ + {0, 0, 667, 800, 3443, 33443, 3905, 33905}, /* DDR2-800 SC */ + {0, 1, 667, 667, 6494, 36494, 6993, 36993}, /* DDR3-667 SC */ + {0, 1, 667, 800, 5998, 35998, 6460, 36460}, /* DDR3-800 SC */ + + {0, 0, 400, 400, 3528, 33528, 4255, 34255}, /* DDR2-400 SC */ + {0, 0, 400, 667, 3500, 33500, 4079, 34079}, /* DDR2-667 SC */ + {0, 0, 400, 800, 3487, 33487, 4029, 34029}, /* DDR2-800 SC */ + {0, 1, 400, 667, 6566, 36566, 7145, 37145}, /* DDR3-667 SC */ + {0, 1, 400, 800, 6042, 36042, 6584, 36584}, /* DDR3-800 SC */ }; -static struct cxsr_latency *intel_get_cxsr_latency(int is_desktop, int fsb, - int mem) +static struct cxsr_latency *intel_get_cxsr_latency(int is_desktop, int is_ddr3, + int fsb, int mem) { int i; struct cxsr_latency *latency; @@ -2675,6 +2701,7 @@ static struct cxsr_latency *intel_get_cxsr_latency(int is_desktop, int fsb, for (i = 0; i < ARRAY_SIZE(cxsr_latency_table); i++) { latency = &cxsr_latency_table[i]; if (is_desktop == latency->is_desktop && + is_ddr3 == latency->is_ddr3 && fsb == latency->fsb_freq && mem == latency->mem_freq) return latency; } @@ -2789,8 +2816,8 @@ static void pineview_update_wm(struct drm_device *dev, int planea_clock, struct cxsr_latency *latency; int sr_clock; - latency = intel_get_cxsr_latency(IS_PINEVIEW_G(dev), dev_priv->fsb_freq, - dev_priv->mem_freq); + latency = intel_get_cxsr_latency(IS_PINEVIEW_G(dev), dev_priv->is_ddr3, + dev_priv->fsb_freq, dev_priv->mem_freq); if (!latency) { DRM_DEBUG_KMS("Unknown FSB/MEM found, disable CxSR\n"); pineview_disable_cxsr(dev); @@ -2943,11 +2970,13 @@ static void i965_update_wm(struct drm_device *dev, int planea_clock, if (srwm < 0) srwm = 1; srwm &= 0x3f; - I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN); + if (IS_I965GM(dev)) + I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN); } else { /* Turn off self refresh if both pipes are enabled */ - I915_WRITE(FW_BLC_SELF, I915_READ(FW_BLC_SELF) - & ~FW_BLC_SELF_EN); + if (IS_I965GM(dev)) + I915_WRITE(FW_BLC_SELF, I915_READ(FW_BLC_SELF) + & ~FW_BLC_SELF_EN); } DRM_DEBUG_KMS("Setting FIFO watermarks - A: 8, B: 8, C: 8, SR %d\n", @@ -3626,6 +3655,11 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, pipeconf &= ~PIPEACONF_DOUBLE_WIDE; } + dspcntr |= DISPLAY_PLANE_ENABLE; + pipeconf |= PIPEACONF_ENABLE; + dpll |= DPLL_VCO_ENABLE; + + /* Disable the panel fitter if it was on our pipe */ if (!HAS_PCH_SPLIT(dev) && intel_panel_fitter_pipe(dev) == pipe) I915_WRITE(PFIT_CONTROL, 0); @@ -3772,6 +3806,18 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, } } + if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) { + pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION; + /* the chip adds 2 halflines automatically */ + adjusted_mode->crtc_vdisplay -= 1; + adjusted_mode->crtc_vtotal -= 1; + adjusted_mode->crtc_vblank_start -= 1; + adjusted_mode->crtc_vblank_end -= 1; + adjusted_mode->crtc_vsync_end -= 1; + adjusted_mode->crtc_vsync_start -= 1; + } else + pipeconf &= ~PIPECONF_INTERLACE_W_FIELD_INDICATION; /* progressive */ + I915_WRITE(htot_reg, (adjusted_mode->crtc_hdisplay - 1) | ((adjusted_mode->crtc_htotal - 1) << 16)); I915_WRITE(hblank_reg, (adjusted_mode->crtc_hblank_start - 1) | @@ -3934,6 +3980,13 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc, DRM_ERROR("failed to pin cursor bo\n"); goto fail_locked; } + + ret = i915_gem_object_set_to_gtt_domain(bo, 0); + if (ret) { + DRM_ERROR("failed to move cursor bo into the GTT\n"); + goto fail_unpin; + } + addr = obj_priv->gtt_offset; } else { ret = i915_gem_attach_phys_object(dev, bo, (pipe == 0) ? I915_GEM_PHYS_CURSOR_0 : I915_GEM_PHYS_CURSOR_1); @@ -3977,6 +4030,8 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc, intel_crtc->cursor_bo = bo; return 0; +fail_unpin: + i915_gem_object_unpin(bo); fail_locked: mutex_unlock(&dev->struct_mutex); fail: @@ -4430,27 +4485,31 @@ static void intel_idle_update(struct work_struct *work) struct drm_device *dev = dev_priv->dev; struct drm_crtc *crtc; struct intel_crtc *intel_crtc; + int enabled = 0; if (!i915_powersave) return; mutex_lock(&dev->struct_mutex); - if (IS_I945G(dev) || IS_I945GM(dev)) { - DRM_DEBUG_DRIVER("enable memory self refresh on 945\n"); - I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN_MASK | FW_BLC_SELF_EN); - } + i915_update_gfx_val(dev_priv); list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { /* Skip inactive CRTCs */ if (!crtc->fb) continue; + enabled++; intel_crtc = to_intel_crtc(crtc); if (!intel_crtc->busy) intel_decrease_pllclock(crtc); } + if ((enabled == 1) && (IS_I945G(dev) || IS_I945GM(dev))) { + DRM_DEBUG_DRIVER("enable memory self refresh on 945\n"); + I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN_MASK | FW_BLC_SELF_EN); + } + mutex_unlock(&dev->struct_mutex); } @@ -4546,10 +4605,10 @@ static void intel_unpin_work_fn(struct work_struct *__work) kfree(work); } -void intel_finish_page_flip(struct drm_device *dev, int pipe) +static void do_intel_finish_page_flip(struct drm_device *dev, + struct drm_crtc *crtc) { drm_i915_private_t *dev_priv = dev->dev_private; - struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe]; struct intel_crtc *intel_crtc = to_intel_crtc(crtc); struct intel_unpin_work *work; struct drm_i915_gem_object *obj_priv; @@ -4564,12 +4623,6 @@ void intel_finish_page_flip(struct drm_device *dev, int pipe) spin_lock_irqsave(&dev->event_lock, flags); work = intel_crtc->unpin_work; if (work == NULL || !work->pending) { - if (work && !work->pending) { - obj_priv = to_intel_bo(work->pending_flip_obj); - DRM_DEBUG_DRIVER("flip finish: %p (%d) not pending?\n", - obj_priv, - atomic_read(&obj_priv->pending_flip)); - } spin_unlock_irqrestore(&dev->event_lock, flags); return; } @@ -4599,6 +4652,22 @@ void intel_finish_page_flip(struct drm_device *dev, int pipe) schedule_work(&work->work); } +void intel_finish_page_flip(struct drm_device *dev, int pipe) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe]; + + do_intel_finish_page_flip(dev, crtc); +} + +void intel_finish_page_flip_plane(struct drm_device *dev, int plane) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + struct drm_crtc *crtc = dev_priv->plane_to_crtc_mapping[plane]; + + do_intel_finish_page_flip(dev, crtc); +} + void intel_prepare_page_flip(struct drm_device *dev, int plane) { drm_i915_private_t *dev_priv = dev->dev_private; @@ -4629,14 +4698,12 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, unsigned long flags; int pipesrc_reg = (intel_crtc->pipe == 0) ? PIPEASRC : PIPEBSRC; int ret, pipesrc; - RING_LOCALS; + u32 flip_mask; work = kzalloc(sizeof *work, GFP_KERNEL); if (work == NULL) return -ENOMEM; - mutex_lock(&dev->struct_mutex); - work->event = event; work->dev = crtc->dev; intel_fb = to_intel_framebuffer(crtc->fb); @@ -4646,10 +4713,10 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, /* We borrow the event spin lock for protecting unpin_work */ spin_lock_irqsave(&dev->event_lock, flags); if (intel_crtc->unpin_work) { - DRM_DEBUG_DRIVER("flip queue: crtc already busy\n"); spin_unlock_irqrestore(&dev->event_lock, flags); kfree(work); - mutex_unlock(&dev->struct_mutex); + + DRM_DEBUG_DRIVER("flip queue: crtc already busy\n"); return -EBUSY; } intel_crtc->unpin_work = work; @@ -4658,13 +4725,19 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, intel_fb = to_intel_framebuffer(fb); obj = intel_fb->obj; + mutex_lock(&dev->struct_mutex); ret = intel_pin_and_fence_fb_obj(dev, obj); if (ret != 0) { - DRM_DEBUG_DRIVER("flip queue: %p pin & fence failed\n", - to_intel_bo(obj)); - kfree(work); - intel_crtc->unpin_work = NULL; mutex_unlock(&dev->struct_mutex); + + spin_lock_irqsave(&dev->event_lock, flags); + intel_crtc->unpin_work = NULL; + spin_unlock_irqrestore(&dev->event_lock, flags); + + kfree(work); + + DRM_DEBUG_DRIVER("flip queue: %p pin & fence failed\n", + to_intel_bo(obj)); return ret; } @@ -4679,15 +4752,28 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, atomic_inc(&obj_priv->pending_flip); work->pending_flip_obj = obj; + if (intel_crtc->plane) + flip_mask = I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT; + else + flip_mask = I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT; + + /* Wait for any previous flip to finish */ + if (IS_GEN3(dev)) + while (I915_READ(ISR) & flip_mask) + ; + BEGIN_LP_RING(4); - OUT_RING(MI_DISPLAY_FLIP | - MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); - OUT_RING(fb->pitch); if (IS_I965G(dev)) { + OUT_RING(MI_DISPLAY_FLIP | + MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); + OUT_RING(fb->pitch); OUT_RING(obj_priv->gtt_offset | obj_priv->tiling_mode); pipesrc = I915_READ(pipesrc_reg); OUT_RING(pipesrc & 0x0fff0fff); } else { + OUT_RING(MI_DISPLAY_FLIP_I915 | + MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); + OUT_RING(fb->pitch); OUT_RING(obj_priv->gtt_offset); OUT_RING(MI_NOOP); } @@ -5023,10 +5109,32 @@ err_unref: return NULL; } +bool ironlake_set_drps(struct drm_device *dev, u8 val) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + u16 rgvswctl; + + rgvswctl = I915_READ16(MEMSWCTL); + if (rgvswctl & MEMCTL_CMD_STS) { + DRM_DEBUG("gpu busy, RCS change rejected\n"); + return false; /* still busy with another command */ + } + + rgvswctl = (MEMCTL_CMD_CHFREQ << MEMCTL_CMD_SHIFT) | + (val << MEMCTL_FREQ_SHIFT) | MEMCTL_SFCAVM; + I915_WRITE16(MEMSWCTL, rgvswctl); + POSTING_READ16(MEMSWCTL); + + rgvswctl |= MEMCTL_CMD_STS; + I915_WRITE16(MEMSWCTL, rgvswctl); + + return true; +} + void ironlake_enable_drps(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - u32 rgvmodectl = I915_READ(MEMMODECTL), rgvswctl; + u32 rgvmodectl = I915_READ(MEMMODECTL); u8 fmax, fmin, fstart, vstart; int i = 0; @@ -5045,13 +5153,21 @@ void ironlake_enable_drps(struct drm_device *dev) fmin = (rgvmodectl & MEMMODE_FMIN_MASK); fstart = (rgvmodectl & MEMMODE_FSTART_MASK) >> MEMMODE_FSTART_SHIFT; + fstart = fmax; + vstart = (I915_READ(PXVFREQ_BASE + (fstart * 4)) & PXVFREQ_PX_MASK) >> PXVFREQ_PX_SHIFT; - dev_priv->max_delay = fstart; /* can't go to fmax w/o IPS */ + dev_priv->fmax = fstart; /* IPS callback will increase this */ + dev_priv->fstart = fstart; + + dev_priv->max_delay = fmax; dev_priv->min_delay = fmin; dev_priv->cur_delay = fstart; + DRM_DEBUG_DRIVER("fmax: %d, fmin: %d, fstart: %d\n", fmax, fmin, + fstart); + I915_WRITE(MEMINTREN, MEMINT_CX_SUPR_EN | MEMINT_EVAL_CHG_EN); /* @@ -5073,20 +5189,19 @@ void ironlake_enable_drps(struct drm_device *dev) } msleep(1); - rgvswctl = (MEMCTL_CMD_CHFREQ << MEMCTL_CMD_SHIFT) | - (fstart << MEMCTL_FREQ_SHIFT) | MEMCTL_SFCAVM; - I915_WRITE(MEMSWCTL, rgvswctl); - POSTING_READ(MEMSWCTL); + ironlake_set_drps(dev, fstart); - rgvswctl |= MEMCTL_CMD_STS; - I915_WRITE(MEMSWCTL, rgvswctl); + dev_priv->last_count1 = I915_READ(0x112e4) + I915_READ(0x112e8) + + I915_READ(0x112e0); + dev_priv->last_time1 = jiffies_to_msecs(jiffies); + dev_priv->last_count2 = I915_READ(0x112f4); + getrawmonotonic(&dev_priv->last_time2); } void ironlake_disable_drps(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - u32 rgvswctl; - u8 fstart; + u16 rgvswctl = I915_READ16(MEMSWCTL); /* Ack interrupts, disable EFC interrupt */ I915_WRITE(MEMINTREN, I915_READ(MEMINTREN) & ~MEMINT_EVAL_CHG_EN); @@ -5096,11 +5211,7 @@ void ironlake_disable_drps(struct drm_device *dev) I915_WRITE(DEIMR, I915_READ(DEIMR) | DE_PCU_EVENT); /* Go back to the starting frequency */ - fstart = (I915_READ(MEMMODECTL) & MEMMODE_FSTART_MASK) >> - MEMMODE_FSTART_SHIFT; - rgvswctl = (MEMCTL_CMD_CHFREQ << MEMCTL_CMD_SHIFT) | - (fstart << MEMCTL_FREQ_SHIFT) | MEMCTL_SFCAVM; - I915_WRITE(MEMSWCTL, rgvswctl); + ironlake_set_drps(dev, dev_priv->fstart); msleep(1); rgvswctl |= MEMCTL_CMD_STS; I915_WRITE(MEMSWCTL, rgvswctl); @@ -5108,6 +5219,92 @@ void ironlake_disable_drps(struct drm_device *dev) } +static unsigned long intel_pxfreq(u32 vidfreq) +{ + unsigned long freq; + int div = (vidfreq & 0x3f0000) >> 16; + int post = (vidfreq & 0x3000) >> 12; + int pre = (vidfreq & 0x7); + + if (!pre) + return 0; + + freq = ((div * 133333) / ((1<<post) * pre)); + + return freq; +} + +void intel_init_emon(struct drm_device *dev) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + u32 lcfuse; + u8 pxw[16]; + int i; + + /* Disable to program */ + I915_WRITE(ECR, 0); + POSTING_READ(ECR); + + /* Program energy weights for various events */ + I915_WRITE(SDEW, 0x15040d00); + I915_WRITE(CSIEW0, 0x007f0000); + I915_WRITE(CSIEW1, 0x1e220004); + I915_WRITE(CSIEW2, 0x04000004); + + for (i = 0; i < 5; i++) + I915_WRITE(PEW + (i * 4), 0); + for (i = 0; i < 3; i++) + I915_WRITE(DEW + (i * 4), 0); + + /* Program P-state weights to account for frequency power adjustment */ + for (i = 0; i < 16; i++) { + u32 pxvidfreq = I915_READ(PXVFREQ_BASE + (i * 4)); + unsigned long freq = intel_pxfreq(pxvidfreq); + unsigned long vid = (pxvidfreq & PXVFREQ_PX_MASK) >> + PXVFREQ_PX_SHIFT; + unsigned long val; + + val = vid * vid; + val *= (freq / 1000); + val *= 255; + val /= (127*127*900); + if (val > 0xff) + DRM_ERROR("bad pxval: %ld\n", val); + pxw[i] = val; + } + /* Render standby states get 0 weight */ + pxw[14] = 0; + pxw[15] = 0; + + for (i = 0; i < 4; i++) { + u32 val = (pxw[i*4] << 24) | (pxw[(i*4)+1] << 16) | + (pxw[(i*4)+2] << 8) | (pxw[(i*4)+3]); + I915_WRITE(PXW + (i * 4), val); + } + + /* Adjust magic regs to magic values (more experimental results) */ + I915_WRITE(OGW0, 0); + I915_WRITE(OGW1, 0); + I915_WRITE(EG0, 0x00007f00); + I915_WRITE(EG1, 0x0000000e); + I915_WRITE(EG2, 0x000e0000); + I915_WRITE(EG3, 0x68000300); + I915_WRITE(EG4, 0x42000000); + I915_WRITE(EG5, 0x00140031); + I915_WRITE(EG6, 0); + I915_WRITE(EG7, 0); + + for (i = 0; i < 8; i++) + I915_WRITE(PXWL + (i * 4), 0); + + /* Enable PMON + select events */ + I915_WRITE(ECR, 0x80000019); + + lcfuse = I915_READ(LCFUSE02); + + dev_priv->corr = (lcfuse & LCFUSE_HIV_MASK); +} + void intel_init_clock_gating(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; @@ -5277,11 +5474,13 @@ static void intel_init_display(struct drm_device *dev) dev_priv->display.update_wm = NULL; } else if (IS_PINEVIEW(dev)) { if (!intel_get_cxsr_latency(IS_PINEVIEW_G(dev), + dev_priv->is_ddr3, dev_priv->fsb_freq, dev_priv->mem_freq)) { DRM_INFO("failed to find known CxSR latency " - "(found fsb freq %d, mem freq %d), " + "(found ddr%s fsb freq %d, mem freq %d), " "disabling CxSR\n", + (dev_priv->is_ddr3 == 1) ? "3": "2", dev_priv->fsb_freq, dev_priv->mem_freq); /* Disable CxSR and never update its watermark again */ pineview_disable_cxsr(dev); @@ -5310,7 +5509,6 @@ static void intel_init_display(struct drm_device *dev) void intel_modeset_init(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - int num_pipe; int i; drm_mode_config_init(dev); @@ -5340,13 +5538,13 @@ void intel_modeset_init(struct drm_device *dev) dev->mode_config.fb_base = pci_resource_start(dev->pdev, 0); if (IS_MOBILE(dev) || IS_I9XX(dev)) - num_pipe = 2; + dev_priv->num_pipe = 2; else - num_pipe = 1; + dev_priv->num_pipe = 1; DRM_DEBUG_KMS("%d display pipe%s available.\n", - num_pipe, num_pipe > 1 ? "s" : ""); + dev_priv->num_pipe, dev_priv->num_pipe > 1 ? "s" : ""); - for (i = 0; i < num_pipe; i++) { + for (i = 0; i < dev_priv->num_pipe; i++) { intel_crtc_init(dev, i); } @@ -5354,8 +5552,10 @@ void intel_modeset_init(struct drm_device *dev) intel_init_clock_gating(dev); - if (IS_IRONLAKE_M(dev)) + if (IS_IRONLAKE_M(dev)) { ironlake_enable_drps(dev); + intel_init_emon(dev); + } INIT_WORK(&dev_priv->idle_work, intel_idle_update); setup_timer(&dev_priv->idle_timer, intel_gpu_idle_timer, diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 6b1c9a27c27a..1aac59e83bff 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -136,6 +136,12 @@ intel_dp_link_required(struct drm_device *dev, } static int +intel_dp_max_data_rate(int max_link_clock, int max_lanes) +{ + return (max_link_clock * max_lanes * 8) / 10; +} + +static int intel_dp_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) { @@ -144,8 +150,11 @@ intel_dp_mode_valid(struct drm_connector *connector, int max_link_clock = intel_dp_link_clock(intel_dp_max_link_bw(intel_encoder)); int max_lanes = intel_dp_max_lane_count(intel_encoder); - if (intel_dp_link_required(connector->dev, intel_encoder, mode->clock) - > max_link_clock * max_lanes) + /* only refuse the mode on non eDP since we have seen some wierd eDP panels + which are outside spec tolerances but somehow work by magic */ + if (!IS_eDP(intel_encoder) && + (intel_dp_link_required(connector->dev, intel_encoder, mode->clock) + > intel_dp_max_data_rate(max_link_clock, max_lanes))) return MODE_CLOCK_HIGH; if (mode->clock < 10000) @@ -506,7 +515,7 @@ intel_dp_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, for (lane_count = 1; lane_count <= max_lane_count; lane_count <<= 1) { for (clock = 0; clock <= max_clock; clock++) { - int link_avail = intel_dp_link_clock(bws[clock]) * lane_count; + int link_avail = intel_dp_max_data_rate(intel_dp_link_clock(bws[clock]), lane_count); if (intel_dp_link_required(encoder->dev, intel_encoder, mode->clock) <= link_avail) { @@ -521,6 +530,18 @@ intel_dp_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, } } } + + if (IS_eDP(intel_encoder)) { + /* okay we failed just pick the highest */ + dp_priv->lane_count = max_lane_count; + dp_priv->link_bw = bws[max_clock]; + adjusted_mode->clock = intel_dp_link_clock(dp_priv->link_bw); + DRM_DEBUG_KMS("Force picking display port link bw %02x lane " + "count %d clock %d\n", + dp_priv->link_bw, dp_priv->lane_count, + adjusted_mode->clock); + return true; + } return false; } @@ -576,7 +597,7 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode, struct intel_encoder *intel_encoder; struct intel_dp_priv *dp_priv; - if (!encoder || encoder->crtc != crtc) + if (encoder->crtc != crtc) continue; intel_encoder = enc_to_intel_encoder(encoder); @@ -675,10 +696,9 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, dp_priv->link_configuration[1] = dp_priv->lane_count; /* - * Check for DPCD version > 1.1, - * enable enahanced frame stuff in that case + * Check for DPCD version > 1.1 and enhanced framing support */ - if (dp_priv->dpcd[0] >= 0x11) { + if (dp_priv->dpcd[0] >= 0x11 && (dp_priv->dpcd[2] & DP_ENHANCED_FRAME_CAP)) { dp_priv->link_configuration[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN; dp_priv->DP |= DP_ENHANCED_FRAMING; } @@ -1208,6 +1228,8 @@ ironlake_dp_detect(struct drm_connector *connector) if (dp_priv->dpcd[0] != 0) status = connector_status_connected; } + DRM_DEBUG_KMS("DPCD: %hx%hx%hx%hx\n", dp_priv->dpcd[0], + dp_priv->dpcd[1], dp_priv->dpcd[2], dp_priv->dpcd[3]); return status; } @@ -1352,7 +1374,7 @@ intel_trans_dp_port_sel (struct drm_crtc *crtc) struct intel_encoder *intel_encoder = NULL; list_for_each_entry(encoder, &mode_config->encoder_list, head) { - if (!encoder || encoder->crtc != crtc) + if (encoder->crtc != crtc) continue; intel_encoder = enc_to_intel_encoder(encoder); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index df931f787665..72206f37c4fb 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -224,6 +224,7 @@ extern void intel_fbdev_fini(struct drm_device *dev); extern void intel_prepare_page_flip(struct drm_device *dev, int plane); extern void intel_finish_page_flip(struct drm_device *dev, int pipe); +extern void intel_finish_page_flip_plane(struct drm_device *dev, int plane); extern void intel_setup_overlay(struct drm_device *dev); extern void intel_cleanup_overlay(struct drm_device *dev); diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c index 6f53cf7fbc50..c3c505244e07 100644 --- a/drivers/gpu/drm/i915/intel_fb.c +++ b/drivers/gpu/drm/i915/intel_fb.c @@ -105,7 +105,11 @@ static int intelfb_create(struct intel_fbdev *ifbdev, } /* Flush everything out, we'll be doing GTT only from now on */ - i915_gem_object_set_to_gtt_domain(fbo, 1); + ret = i915_gem_object_set_to_gtt_domain(fbo, 1); + if (ret) { + DRM_ERROR("failed to bind fb: %d.\n", ret); + goto out_unpin; + } info = framebuffer_alloc(0, device); if (!info) { @@ -241,6 +245,7 @@ int intel_fbdev_init(struct drm_device *dev) { struct intel_fbdev *ifbdev; drm_i915_private_t *dev_priv = dev->dev_private; + int ret; ifbdev = kzalloc(sizeof(struct intel_fbdev), GFP_KERNEL); if (!ifbdev) @@ -249,8 +254,13 @@ int intel_fbdev_init(struct drm_device *dev) dev_priv->fbdev = ifbdev; ifbdev->helper.funcs = &intel_fb_helper_funcs; - drm_fb_helper_init(dev, &ifbdev->helper, 2, - INTELFB_CONN_LIMIT); + ret = drm_fb_helper_init(dev, &ifbdev->helper, + dev_priv->num_pipe, + INTELFB_CONN_LIMIT); + if (ret) { + kfree(ifbdev); + return ret; + } drm_fb_helper_single_add_all_connectors(&ifbdev->helper); drm_fb_helper_initial_config(&ifbdev->helper, 32); diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index 65727f0a79a3..83bd764b000e 100644 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -59,8 +59,11 @@ static void intel_hdmi_mode_set(struct drm_encoder *encoder, SDVO_VSYNC_ACTIVE_HIGH | SDVO_HSYNC_ACTIVE_HIGH; - if (hdmi_priv->has_hdmi_sink) + if (hdmi_priv->has_hdmi_sink) { sdvox |= SDVO_AUDIO_ENABLE; + if (HAS_PCH_CPT(dev)) + sdvox |= HDMI_MODE_SELECT; + } if (intel_crtc->pipe == 1) { if (HAS_PCH_CPT(dev)) diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c index 6a1accd83aec..31df55f0a0a7 100644 --- a/drivers/gpu/drm/i915/intel_lvds.c +++ b/drivers/gpu/drm/i915/intel_lvds.c @@ -983,8 +983,8 @@ void intel_lvds_init(struct drm_device *dev) drm_connector_attach_property(&intel_connector->base, dev->mode_config.scaling_mode_property, - DRM_MODE_SCALE_FULLSCREEN); - lvds_priv->fitting_mode = DRM_MODE_SCALE_FULLSCREEN; + DRM_MODE_SCALE_ASPECT); + lvds_priv->fitting_mode = DRM_MODE_SCALE_ASPECT; /* * LVDS discovery: * 1) check for EDID on DDC diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index b0e17b06eb6e..d7ad5139d17c 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -211,9 +211,8 @@ static void intel_overlay_unmap_regs_atomic(struct intel_overlay *overlay) static int intel_overlay_on(struct intel_overlay *overlay) { struct drm_device *dev = overlay->dev; - drm_i915_private_t *dev_priv = dev->dev_private; int ret; - RING_LOCALS; + drm_i915_private_t *dev_priv = dev->dev_private; BUG_ON(overlay->active); @@ -227,11 +226,13 @@ static int intel_overlay_on(struct intel_overlay *overlay) OUT_RING(MI_NOOP); ADVANCE_LP_RING(); - overlay->last_flip_req = i915_add_request(dev, NULL, 0); + overlay->last_flip_req = + i915_add_request(dev, NULL, 0, &dev_priv->render_ring); if (overlay->last_flip_req == 0) return -ENOMEM; - ret = i915_do_wait_request(dev, overlay->last_flip_req, 1); + ret = i915_do_wait_request(dev, + overlay->last_flip_req, 1, &dev_priv->render_ring); if (ret != 0) return ret; @@ -248,7 +249,6 @@ static void intel_overlay_continue(struct intel_overlay *overlay, drm_i915_private_t *dev_priv = dev->dev_private; u32 flip_addr = overlay->flip_addr; u32 tmp; - RING_LOCALS; BUG_ON(!overlay->active); @@ -265,7 +265,8 @@ static void intel_overlay_continue(struct intel_overlay *overlay, OUT_RING(flip_addr); ADVANCE_LP_RING(); - overlay->last_flip_req = i915_add_request(dev, NULL, 0); + overlay->last_flip_req = + i915_add_request(dev, NULL, 0, &dev_priv->render_ring); } static int intel_overlay_wait_flip(struct intel_overlay *overlay) @@ -274,10 +275,10 @@ static int intel_overlay_wait_flip(struct intel_overlay *overlay) drm_i915_private_t *dev_priv = dev->dev_private; int ret; u32 tmp; - RING_LOCALS; if (overlay->last_flip_req != 0) { - ret = i915_do_wait_request(dev, overlay->last_flip_req, 1); + ret = i915_do_wait_request(dev, overlay->last_flip_req, + 1, &dev_priv->render_ring); if (ret == 0) { overlay->last_flip_req = 0; @@ -296,11 +297,13 @@ static int intel_overlay_wait_flip(struct intel_overlay *overlay) OUT_RING(MI_NOOP); ADVANCE_LP_RING(); - overlay->last_flip_req = i915_add_request(dev, NULL, 0); + overlay->last_flip_req = + i915_add_request(dev, NULL, 0, &dev_priv->render_ring); if (overlay->last_flip_req == 0) return -ENOMEM; - ret = i915_do_wait_request(dev, overlay->last_flip_req, 1); + ret = i915_do_wait_request(dev, overlay->last_flip_req, + 1, &dev_priv->render_ring); if (ret != 0) return ret; @@ -314,9 +317,8 @@ static int intel_overlay_off(struct intel_overlay *overlay) { u32 flip_addr = overlay->flip_addr; struct drm_device *dev = overlay->dev; - drm_i915_private_t *dev_priv = dev->dev_private; + drm_i915_private_t *dev_priv = dev->dev_private; int ret; - RING_LOCALS; BUG_ON(!overlay->active); @@ -336,11 +338,13 @@ static int intel_overlay_off(struct intel_overlay *overlay) OUT_RING(MI_NOOP); ADVANCE_LP_RING(); - overlay->last_flip_req = i915_add_request(dev, NULL, 0); + overlay->last_flip_req = + i915_add_request(dev, NULL, 0, &dev_priv->render_ring); if (overlay->last_flip_req == 0) return -ENOMEM; - ret = i915_do_wait_request(dev, overlay->last_flip_req, 1); + ret = i915_do_wait_request(dev, overlay->last_flip_req, + 1, &dev_priv->render_ring); if (ret != 0) return ret; @@ -354,11 +358,13 @@ static int intel_overlay_off(struct intel_overlay *overlay) OUT_RING(MI_NOOP); ADVANCE_LP_RING(); - overlay->last_flip_req = i915_add_request(dev, NULL, 0); + overlay->last_flip_req = + i915_add_request(dev, NULL, 0, &dev_priv->render_ring); if (overlay->last_flip_req == 0) return -ENOMEM; - ret = i915_do_wait_request(dev, overlay->last_flip_req, 1); + ret = i915_do_wait_request(dev, overlay->last_flip_req, + 1, &dev_priv->render_ring); if (ret != 0) return ret; @@ -390,22 +396,23 @@ int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay, int interruptible) { struct drm_device *dev = overlay->dev; - drm_i915_private_t *dev_priv = dev->dev_private; struct drm_gem_object *obj; + drm_i915_private_t *dev_priv = dev->dev_private; u32 flip_addr; int ret; - RING_LOCALS; if (overlay->hw_wedged == HW_WEDGED) return -EIO; if (overlay->last_flip_req == 0) { - overlay->last_flip_req = i915_add_request(dev, NULL, 0); + overlay->last_flip_req = + i915_add_request(dev, NULL, 0, &dev_priv->render_ring); if (overlay->last_flip_req == 0) return -ENOMEM; } - ret = i915_do_wait_request(dev, overlay->last_flip_req, interruptible); + ret = i915_do_wait_request(dev, overlay->last_flip_req, + interruptible, &dev_priv->render_ring); if (ret != 0) return ret; @@ -429,12 +436,13 @@ int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay, OUT_RING(MI_NOOP); ADVANCE_LP_RING(); - overlay->last_flip_req = i915_add_request(dev, NULL, 0); + overlay->last_flip_req = i915_add_request(dev, NULL, + 0, &dev_priv->render_ring); if (overlay->last_flip_req == 0) return -ENOMEM; ret = i915_do_wait_request(dev, overlay->last_flip_req, - interruptible); + interruptible, &dev_priv->render_ring); if (ret != 0) return ret; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c new file mode 100644 index 000000000000..26362f8495a8 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -0,0 +1,851 @@ +/* + * Copyright © 2008-2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + * Authors: + * Eric Anholt <eric@anholt.net> + * Zou Nan hai <nanhai.zou@intel.com> + * Xiang Hai hao<haihao.xiang@intel.com> + * + */ + +#include "drmP.h" +#include "drm.h" +#include "i915_drv.h" +#include "i915_drm.h" +#include "i915_trace.h" + +static void +render_ring_flush(struct drm_device *dev, + struct intel_ring_buffer *ring, + u32 invalidate_domains, + u32 flush_domains) +{ +#if WATCH_EXEC + DRM_INFO("%s: invalidate %08x flush %08x\n", __func__, + invalidate_domains, flush_domains); +#endif + u32 cmd; + trace_i915_gem_request_flush(dev, ring->next_seqno, + invalidate_domains, flush_domains); + + if ((invalidate_domains | flush_domains) & I915_GEM_GPU_DOMAINS) { + /* + * read/write caches: + * + * I915_GEM_DOMAIN_RENDER is always invalidated, but is + * only flushed if MI_NO_WRITE_FLUSH is unset. On 965, it is + * also flushed at 2d versus 3d pipeline switches. + * + * read-only caches: + * + * I915_GEM_DOMAIN_SAMPLER is flushed on pre-965 if + * MI_READ_FLUSH is set, and is always flushed on 965. + * + * I915_GEM_DOMAIN_COMMAND may not exist? + * + * I915_GEM_DOMAIN_INSTRUCTION, which exists on 965, is + * invalidated when MI_EXE_FLUSH is set. + * + * I915_GEM_DOMAIN_VERTEX, which exists on 965, is + * invalidated with every MI_FLUSH. + * + * TLBs: + * + * On 965, TLBs associated with I915_GEM_DOMAIN_COMMAND + * and I915_GEM_DOMAIN_CPU in are invalidated at PTE write and + * I915_GEM_DOMAIN_RENDER and I915_GEM_DOMAIN_SAMPLER + * are flushed at any MI_FLUSH. + */ + + cmd = MI_FLUSH | MI_NO_WRITE_FLUSH; + if ((invalidate_domains|flush_domains) & + I915_GEM_DOMAIN_RENDER) + cmd &= ~MI_NO_WRITE_FLUSH; + if (!IS_I965G(dev)) { + /* + * On the 965, the sampler cache always gets flushed + * and this bit is reserved. + */ + if (invalidate_domains & I915_GEM_DOMAIN_SAMPLER) + cmd |= MI_READ_FLUSH; + } + if (invalidate_domains & I915_GEM_DOMAIN_INSTRUCTION) + cmd |= MI_EXE_FLUSH; + +#if WATCH_EXEC + DRM_INFO("%s: queue flush %08x to ring\n", __func__, cmd); +#endif + intel_ring_begin(dev, ring, 2); + intel_ring_emit(dev, ring, cmd); + intel_ring_emit(dev, ring, MI_NOOP); + intel_ring_advance(dev, ring); + } +} + +static unsigned int render_ring_get_head(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + return I915_READ(PRB0_HEAD) & HEAD_ADDR; +} + +static unsigned int render_ring_get_tail(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + return I915_READ(PRB0_TAIL) & TAIL_ADDR; +} + +static unsigned int render_ring_get_active_head(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + u32 acthd_reg = IS_I965G(dev) ? ACTHD_I965 : ACTHD; + + return I915_READ(acthd_reg); +} + +static void render_ring_advance_ring(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + I915_WRITE(PRB0_TAIL, ring->tail); +} + +static int init_ring_common(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + u32 head; + drm_i915_private_t *dev_priv = dev->dev_private; + struct drm_i915_gem_object *obj_priv; + obj_priv = to_intel_bo(ring->gem_object); + + /* Stop the ring if it's running. */ + I915_WRITE(ring->regs.ctl, 0); + I915_WRITE(ring->regs.head, 0); + I915_WRITE(ring->regs.tail, 0); + + /* Initialize the ring. */ + I915_WRITE(ring->regs.start, obj_priv->gtt_offset); + head = ring->get_head(dev, ring); + + /* G45 ring initialization fails to reset head to zero */ + if (head != 0) { + DRM_ERROR("%s head not reset to zero " + "ctl %08x head %08x tail %08x start %08x\n", + ring->name, + I915_READ(ring->regs.ctl), + I915_READ(ring->regs.head), + I915_READ(ring->regs.tail), + I915_READ(ring->regs.start)); + + I915_WRITE(ring->regs.head, 0); + + DRM_ERROR("%s head forced to zero " + "ctl %08x head %08x tail %08x start %08x\n", + ring->name, + I915_READ(ring->regs.ctl), + I915_READ(ring->regs.head), + I915_READ(ring->regs.tail), + I915_READ(ring->regs.start)); + } + + I915_WRITE(ring->regs.ctl, + ((ring->gem_object->size - PAGE_SIZE) & RING_NR_PAGES) + | RING_NO_REPORT | RING_VALID); + + head = I915_READ(ring->regs.head) & HEAD_ADDR; + /* If the head is still not zero, the ring is dead */ + if (head != 0) { + DRM_ERROR("%s initialization failed " + "ctl %08x head %08x tail %08x start %08x\n", + ring->name, + I915_READ(ring->regs.ctl), + I915_READ(ring->regs.head), + I915_READ(ring->regs.tail), + I915_READ(ring->regs.start)); + return -EIO; + } + + if (!drm_core_check_feature(dev, DRIVER_MODESET)) + i915_kernel_lost_context(dev); + else { + ring->head = ring->get_head(dev, ring); + ring->tail = ring->get_tail(dev, ring); + ring->space = ring->head - (ring->tail + 8); + if (ring->space < 0) + ring->space += ring->size; + } + return 0; +} + +static int init_render_ring(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + int ret = init_ring_common(dev, ring); + if (IS_I9XX(dev) && !IS_GEN3(dev)) { + I915_WRITE(MI_MODE, + (VS_TIMER_DISPATCH) << 16 | VS_TIMER_DISPATCH); + } + return ret; +} + +#define PIPE_CONTROL_FLUSH(addr) \ +do { \ + OUT_RING(GFX_OP_PIPE_CONTROL | PIPE_CONTROL_QW_WRITE | \ + PIPE_CONTROL_DEPTH_STALL | 2); \ + OUT_RING(addr | PIPE_CONTROL_GLOBAL_GTT); \ + OUT_RING(0); \ + OUT_RING(0); \ +} while (0) + +/** + * Creates a new sequence number, emitting a write of it to the status page + * plus an interrupt, which will trigger i915_user_interrupt_handler. + * + * Must be called with struct_lock held. + * + * Returned sequence numbers are nonzero on success. + */ +static u32 +render_ring_add_request(struct drm_device *dev, + struct intel_ring_buffer *ring, + struct drm_file *file_priv, + u32 flush_domains) +{ + u32 seqno; + drm_i915_private_t *dev_priv = dev->dev_private; + seqno = intel_ring_get_seqno(dev, ring); + + if (IS_GEN6(dev)) { + BEGIN_LP_RING(6); + OUT_RING(GFX_OP_PIPE_CONTROL | 3); + OUT_RING(PIPE_CONTROL_QW_WRITE | + PIPE_CONTROL_WC_FLUSH | PIPE_CONTROL_IS_FLUSH | + PIPE_CONTROL_NOTIFY); + OUT_RING(dev_priv->seqno_gfx_addr | PIPE_CONTROL_GLOBAL_GTT); + OUT_RING(seqno); + OUT_RING(0); + OUT_RING(0); + ADVANCE_LP_RING(); + } else if (HAS_PIPE_CONTROL(dev)) { + u32 scratch_addr = dev_priv->seqno_gfx_addr + 128; + + /* + * Workaround qword write incoherence by flushing the + * PIPE_NOTIFY buffers out to memory before requesting + * an interrupt. + */ + BEGIN_LP_RING(32); + OUT_RING(GFX_OP_PIPE_CONTROL | PIPE_CONTROL_QW_WRITE | + PIPE_CONTROL_WC_FLUSH | PIPE_CONTROL_TC_FLUSH); + OUT_RING(dev_priv->seqno_gfx_addr | PIPE_CONTROL_GLOBAL_GTT); + OUT_RING(seqno); + OUT_RING(0); + PIPE_CONTROL_FLUSH(scratch_addr); + scratch_addr += 128; /* write to separate cachelines */ + PIPE_CONTROL_FLUSH(scratch_addr); + scratch_addr += 128; + PIPE_CONTROL_FLUSH(scratch_addr); + scratch_addr += 128; + PIPE_CONTROL_FLUSH(scratch_addr); + scratch_addr += 128; + PIPE_CONTROL_FLUSH(scratch_addr); + scratch_addr += 128; + PIPE_CONTROL_FLUSH(scratch_addr); + OUT_RING(GFX_OP_PIPE_CONTROL | PIPE_CONTROL_QW_WRITE | + PIPE_CONTROL_WC_FLUSH | PIPE_CONTROL_TC_FLUSH | + PIPE_CONTROL_NOTIFY); + OUT_RING(dev_priv->seqno_gfx_addr | PIPE_CONTROL_GLOBAL_GTT); + OUT_RING(seqno); + OUT_RING(0); + ADVANCE_LP_RING(); + } else { + BEGIN_LP_RING(4); + OUT_RING(MI_STORE_DWORD_INDEX); + OUT_RING(I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT); + OUT_RING(seqno); + + OUT_RING(MI_USER_INTERRUPT); + ADVANCE_LP_RING(); + } + return seqno; +} + +static u32 +render_ring_get_gem_seqno(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; + if (HAS_PIPE_CONTROL(dev)) + return ((volatile u32 *)(dev_priv->seqno_page))[0]; + else + return intel_read_status_page(ring, I915_GEM_HWS_INDEX); +} + +static void +render_ring_get_user_irq(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; + unsigned long irqflags; + + spin_lock_irqsave(&dev_priv->user_irq_lock, irqflags); + if (dev->irq_enabled && (++ring->user_irq_refcount == 1)) { + if (HAS_PCH_SPLIT(dev)) + ironlake_enable_graphics_irq(dev_priv, GT_PIPE_NOTIFY); + else + i915_enable_irq(dev_priv, I915_USER_INTERRUPT); + } + spin_unlock_irqrestore(&dev_priv->user_irq_lock, irqflags); +} + +static void +render_ring_put_user_irq(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; + unsigned long irqflags; + + spin_lock_irqsave(&dev_priv->user_irq_lock, irqflags); + BUG_ON(dev->irq_enabled && ring->user_irq_refcount <= 0); + if (dev->irq_enabled && (--ring->user_irq_refcount == 0)) { + if (HAS_PCH_SPLIT(dev)) + ironlake_disable_graphics_irq(dev_priv, GT_PIPE_NOTIFY); + else + i915_disable_irq(dev_priv, I915_USER_INTERRUPT); + } + spin_unlock_irqrestore(&dev_priv->user_irq_lock, irqflags); +} + +static void render_setup_status_page(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + if (IS_GEN6(dev)) { + I915_WRITE(HWS_PGA_GEN6, ring->status_page.gfx_addr); + I915_READ(HWS_PGA_GEN6); /* posting read */ + } else { + I915_WRITE(HWS_PGA, ring->status_page.gfx_addr); + I915_READ(HWS_PGA); /* posting read */ + } + +} + +void +bsd_ring_flush(struct drm_device *dev, + struct intel_ring_buffer *ring, + u32 invalidate_domains, + u32 flush_domains) +{ + intel_ring_begin(dev, ring, 2); + intel_ring_emit(dev, ring, MI_FLUSH); + intel_ring_emit(dev, ring, MI_NOOP); + intel_ring_advance(dev, ring); +} + +static inline unsigned int bsd_ring_get_head(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + return I915_READ(BSD_RING_HEAD) & HEAD_ADDR; +} + +static inline unsigned int bsd_ring_get_tail(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + return I915_READ(BSD_RING_TAIL) & TAIL_ADDR; +} + +static inline unsigned int bsd_ring_get_active_head(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + return I915_READ(BSD_RING_ACTHD); +} + +static inline void bsd_ring_advance_ring(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + I915_WRITE(BSD_RING_TAIL, ring->tail); +} + +static int init_bsd_ring(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + return init_ring_common(dev, ring); +} + +static u32 +bsd_ring_add_request(struct drm_device *dev, + struct intel_ring_buffer *ring, + struct drm_file *file_priv, + u32 flush_domains) +{ + u32 seqno; + seqno = intel_ring_get_seqno(dev, ring); + intel_ring_begin(dev, ring, 4); + intel_ring_emit(dev, ring, MI_STORE_DWORD_INDEX); + intel_ring_emit(dev, ring, + I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT); + intel_ring_emit(dev, ring, seqno); + intel_ring_emit(dev, ring, MI_USER_INTERRUPT); + intel_ring_advance(dev, ring); + + DRM_DEBUG_DRIVER("%s %d\n", ring->name, seqno); + + return seqno; +} + +static void bsd_setup_status_page(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + I915_WRITE(BSD_HWS_PGA, ring->status_page.gfx_addr); + I915_READ(BSD_HWS_PGA); +} + +static void +bsd_ring_get_user_irq(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + /* do nothing */ +} +static void +bsd_ring_put_user_irq(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + /* do nothing */ +} + +static u32 +bsd_ring_get_gem_seqno(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + return intel_read_status_page(ring, I915_GEM_HWS_INDEX); +} + +static int +bsd_ring_dispatch_gem_execbuffer(struct drm_device *dev, + struct intel_ring_buffer *ring, + struct drm_i915_gem_execbuffer2 *exec, + struct drm_clip_rect *cliprects, + uint64_t exec_offset) +{ + uint32_t exec_start; + exec_start = (uint32_t) exec_offset + exec->batch_start_offset; + intel_ring_begin(dev, ring, 2); + intel_ring_emit(dev, ring, MI_BATCH_BUFFER_START | + (2 << 6) | MI_BATCH_NON_SECURE_I965); + intel_ring_emit(dev, ring, exec_start); + intel_ring_advance(dev, ring); + return 0; +} + + +static int +render_ring_dispatch_gem_execbuffer(struct drm_device *dev, + struct intel_ring_buffer *ring, + struct drm_i915_gem_execbuffer2 *exec, + struct drm_clip_rect *cliprects, + uint64_t exec_offset) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + int nbox = exec->num_cliprects; + int i = 0, count; + uint32_t exec_start, exec_len; + exec_start = (uint32_t) exec_offset + exec->batch_start_offset; + exec_len = (uint32_t) exec->batch_len; + + trace_i915_gem_request_submit(dev, dev_priv->mm.next_gem_seqno + 1); + + count = nbox ? nbox : 1; + + for (i = 0; i < count; i++) { + if (i < nbox) { + int ret = i915_emit_box(dev, cliprects, i, + exec->DR1, exec->DR4); + if (ret) + return ret; + } + + if (IS_I830(dev) || IS_845G(dev)) { + intel_ring_begin(dev, ring, 4); + intel_ring_emit(dev, ring, MI_BATCH_BUFFER); + intel_ring_emit(dev, ring, + exec_start | MI_BATCH_NON_SECURE); + intel_ring_emit(dev, ring, exec_start + exec_len - 4); + intel_ring_emit(dev, ring, 0); + } else { + intel_ring_begin(dev, ring, 4); + if (IS_I965G(dev)) { + intel_ring_emit(dev, ring, + MI_BATCH_BUFFER_START | (2 << 6) + | MI_BATCH_NON_SECURE_I965); + intel_ring_emit(dev, ring, exec_start); + } else { + intel_ring_emit(dev, ring, MI_BATCH_BUFFER_START + | (2 << 6)); + intel_ring_emit(dev, ring, exec_start | + MI_BATCH_NON_SECURE); + } + } + intel_ring_advance(dev, ring); + } + + /* XXX breadcrumb */ + return 0; +} + +static void cleanup_status_page(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + struct drm_gem_object *obj; + struct drm_i915_gem_object *obj_priv; + + obj = ring->status_page.obj; + if (obj == NULL) + return; + obj_priv = to_intel_bo(obj); + + kunmap(obj_priv->pages[0]); + i915_gem_object_unpin(obj); + drm_gem_object_unreference(obj); + ring->status_page.obj = NULL; + + memset(&dev_priv->hws_map, 0, sizeof(dev_priv->hws_map)); +} + +static int init_status_page(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + struct drm_gem_object *obj; + struct drm_i915_gem_object *obj_priv; + int ret; + + obj = i915_gem_alloc_object(dev, 4096); + if (obj == NULL) { + DRM_ERROR("Failed to allocate status page\n"); + ret = -ENOMEM; + goto err; + } + obj_priv = to_intel_bo(obj); + obj_priv->agp_type = AGP_USER_CACHED_MEMORY; + + ret = i915_gem_object_pin(obj, 4096); + if (ret != 0) { + goto err_unref; + } + + ring->status_page.gfx_addr = obj_priv->gtt_offset; + ring->status_page.page_addr = kmap(obj_priv->pages[0]); + if (ring->status_page.page_addr == NULL) { + memset(&dev_priv->hws_map, 0, sizeof(dev_priv->hws_map)); + goto err_unpin; + } + ring->status_page.obj = obj; + memset(ring->status_page.page_addr, 0, PAGE_SIZE); + + ring->setup_status_page(dev, ring); + DRM_DEBUG_DRIVER("%s hws offset: 0x%08x\n", + ring->name, ring->status_page.gfx_addr); + + return 0; + +err_unpin: + i915_gem_object_unpin(obj); +err_unref: + drm_gem_object_unreference(obj); +err: + return ret; +} + + +int intel_init_ring_buffer(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + int ret; + struct drm_i915_gem_object *obj_priv; + struct drm_gem_object *obj; + ring->dev = dev; + + if (I915_NEED_GFX_HWS(dev)) { + ret = init_status_page(dev, ring); + if (ret) + return ret; + } + + obj = i915_gem_alloc_object(dev, ring->size); + if (obj == NULL) { + DRM_ERROR("Failed to allocate ringbuffer\n"); + ret = -ENOMEM; + goto cleanup; + } + + ring->gem_object = obj; + + ret = i915_gem_object_pin(obj, ring->alignment); + if (ret != 0) { + drm_gem_object_unreference(obj); + goto cleanup; + } + + obj_priv = to_intel_bo(obj); + ring->map.size = ring->size; + ring->map.offset = dev->agp->base + obj_priv->gtt_offset; + ring->map.type = 0; + ring->map.flags = 0; + ring->map.mtrr = 0; + + drm_core_ioremap_wc(&ring->map, dev); + if (ring->map.handle == NULL) { + DRM_ERROR("Failed to map ringbuffer.\n"); + i915_gem_object_unpin(obj); + drm_gem_object_unreference(obj); + ret = -EINVAL; + goto cleanup; + } + + ring->virtual_start = ring->map.handle; + ret = ring->init(dev, ring); + if (ret != 0) { + intel_cleanup_ring_buffer(dev, ring); + return ret; + } + + if (!drm_core_check_feature(dev, DRIVER_MODESET)) + i915_kernel_lost_context(dev); + else { + ring->head = ring->get_head(dev, ring); + ring->tail = ring->get_tail(dev, ring); + ring->space = ring->head - (ring->tail + 8); + if (ring->space < 0) + ring->space += ring->size; + } + INIT_LIST_HEAD(&ring->active_list); + INIT_LIST_HEAD(&ring->request_list); + return ret; +cleanup: + cleanup_status_page(dev, ring); + return ret; +} + +void intel_cleanup_ring_buffer(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + if (ring->gem_object == NULL) + return; + + drm_core_ioremapfree(&ring->map, dev); + + i915_gem_object_unpin(ring->gem_object); + drm_gem_object_unreference(ring->gem_object); + ring->gem_object = NULL; + cleanup_status_page(dev, ring); +} + +int intel_wrap_ring_buffer(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + unsigned int *virt; + int rem; + rem = ring->size - ring->tail; + + if (ring->space < rem) { + int ret = intel_wait_ring_buffer(dev, ring, rem); + if (ret) + return ret; + } + + virt = (unsigned int *)(ring->virtual_start + ring->tail); + rem /= 4; + while (rem--) + *virt++ = MI_NOOP; + + ring->tail = 0; + ring->space = ring->head - 8; + + return 0; +} + +int intel_wait_ring_buffer(struct drm_device *dev, + struct intel_ring_buffer *ring, int n) +{ + unsigned long end; + + trace_i915_ring_wait_begin (dev); + end = jiffies + 3 * HZ; + do { + ring->head = ring->get_head(dev, ring); + ring->space = ring->head - (ring->tail + 8); + if (ring->space < 0) + ring->space += ring->size; + if (ring->space >= n) { + trace_i915_ring_wait_end (dev); + return 0; + } + + if (dev->primary->master) { + struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv; + if (master_priv->sarea_priv) + master_priv->sarea_priv->perf_boxes |= I915_BOX_WAIT; + } + + yield(); + } while (!time_after(jiffies, end)); + trace_i915_ring_wait_end (dev); + return -EBUSY; +} + +void intel_ring_begin(struct drm_device *dev, + struct intel_ring_buffer *ring, int num_dwords) +{ + int n = 4*num_dwords; + if (unlikely(ring->tail + n > ring->size)) + intel_wrap_ring_buffer(dev, ring); + if (unlikely(ring->space < n)) + intel_wait_ring_buffer(dev, ring, n); +} + +void intel_ring_emit(struct drm_device *dev, + struct intel_ring_buffer *ring, unsigned int data) +{ + unsigned int *virt = ring->virtual_start + ring->tail; + *virt = data; + ring->tail += 4; + ring->tail &= ring->size - 1; + ring->space -= 4; +} + +void intel_ring_advance(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + ring->advance_ring(dev, ring); +} + +void intel_fill_struct(struct drm_device *dev, + struct intel_ring_buffer *ring, + void *data, + unsigned int len) +{ + unsigned int *virt = ring->virtual_start + ring->tail; + BUG_ON((len&~(4-1)) != 0); + intel_ring_begin(dev, ring, len/4); + memcpy(virt, data, len); + ring->tail += len; + ring->tail &= ring->size - 1; + ring->space -= len; + intel_ring_advance(dev, ring); +} + +u32 intel_ring_get_seqno(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + u32 seqno; + seqno = ring->next_seqno; + + /* reserve 0 for non-seqno */ + if (++ring->next_seqno == 0) + ring->next_seqno = 1; + return seqno; +} + +struct intel_ring_buffer render_ring = { + .name = "render ring", + .regs = { + .ctl = PRB0_CTL, + .head = PRB0_HEAD, + .tail = PRB0_TAIL, + .start = PRB0_START + }, + .ring_flag = I915_EXEC_RENDER, + .size = 32 * PAGE_SIZE, + .alignment = PAGE_SIZE, + .virtual_start = NULL, + .dev = NULL, + .gem_object = NULL, + .head = 0, + .tail = 0, + .space = 0, + .next_seqno = 1, + .user_irq_refcount = 0, + .irq_gem_seqno = 0, + .waiting_gem_seqno = 0, + .setup_status_page = render_setup_status_page, + .init = init_render_ring, + .get_head = render_ring_get_head, + .get_tail = render_ring_get_tail, + .get_active_head = render_ring_get_active_head, + .advance_ring = render_ring_advance_ring, + .flush = render_ring_flush, + .add_request = render_ring_add_request, + .get_gem_seqno = render_ring_get_gem_seqno, + .user_irq_get = render_ring_get_user_irq, + .user_irq_put = render_ring_put_user_irq, + .dispatch_gem_execbuffer = render_ring_dispatch_gem_execbuffer, + .status_page = {NULL, 0, NULL}, + .map = {0,} +}; + +/* ring buffer for bit-stream decoder */ + +struct intel_ring_buffer bsd_ring = { + .name = "bsd ring", + .regs = { + .ctl = BSD_RING_CTL, + .head = BSD_RING_HEAD, + .tail = BSD_RING_TAIL, + .start = BSD_RING_START + }, + .ring_flag = I915_EXEC_BSD, + .size = 32 * PAGE_SIZE, + .alignment = PAGE_SIZE, + .virtual_start = NULL, + .dev = NULL, + .gem_object = NULL, + .head = 0, + .tail = 0, + .space = 0, + .next_seqno = 1, + .user_irq_refcount = 0, + .irq_gem_seqno = 0, + .waiting_gem_seqno = 0, + .setup_status_page = bsd_setup_status_page, + .init = init_bsd_ring, + .get_head = bsd_ring_get_head, + .get_tail = bsd_ring_get_tail, + .get_active_head = bsd_ring_get_active_head, + .advance_ring = bsd_ring_advance_ring, + .flush = bsd_ring_flush, + .add_request = bsd_ring_add_request, + .get_gem_seqno = bsd_ring_get_gem_seqno, + .user_irq_get = bsd_ring_get_user_irq, + .user_irq_put = bsd_ring_put_user_irq, + .dispatch_gem_execbuffer = bsd_ring_dispatch_gem_execbuffer, + .status_page = {NULL, 0, NULL}, + .map = {0,} +}; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h new file mode 100644 index 000000000000..d5568d3766de --- /dev/null +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -0,0 +1,124 @@ +#ifndef _INTEL_RINGBUFFER_H_ +#define _INTEL_RINGBUFFER_H_ + +struct intel_hw_status_page { + void *page_addr; + unsigned int gfx_addr; + struct drm_gem_object *obj; +}; + +struct drm_i915_gem_execbuffer2; +struct intel_ring_buffer { + const char *name; + struct ring_regs { + u32 ctl; + u32 head; + u32 tail; + u32 start; + } regs; + unsigned int ring_flag; + unsigned long size; + unsigned int alignment; + void *virtual_start; + struct drm_device *dev; + struct drm_gem_object *gem_object; + + unsigned int head; + unsigned int tail; + unsigned int space; + u32 next_seqno; + struct intel_hw_status_page status_page; + + u32 irq_gem_seqno; /* last seq seem at irq time */ + u32 waiting_gem_seqno; + int user_irq_refcount; + void (*user_irq_get)(struct drm_device *dev, + struct intel_ring_buffer *ring); + void (*user_irq_put)(struct drm_device *dev, + struct intel_ring_buffer *ring); + void (*setup_status_page)(struct drm_device *dev, + struct intel_ring_buffer *ring); + + int (*init)(struct drm_device *dev, + struct intel_ring_buffer *ring); + + unsigned int (*get_head)(struct drm_device *dev, + struct intel_ring_buffer *ring); + unsigned int (*get_tail)(struct drm_device *dev, + struct intel_ring_buffer *ring); + unsigned int (*get_active_head)(struct drm_device *dev, + struct intel_ring_buffer *ring); + void (*advance_ring)(struct drm_device *dev, + struct intel_ring_buffer *ring); + void (*flush)(struct drm_device *dev, + struct intel_ring_buffer *ring, + u32 invalidate_domains, + u32 flush_domains); + u32 (*add_request)(struct drm_device *dev, + struct intel_ring_buffer *ring, + struct drm_file *file_priv, + u32 flush_domains); + u32 (*get_gem_seqno)(struct drm_device *dev, + struct intel_ring_buffer *ring); + int (*dispatch_gem_execbuffer)(struct drm_device *dev, + struct intel_ring_buffer *ring, + struct drm_i915_gem_execbuffer2 *exec, + struct drm_clip_rect *cliprects, + uint64_t exec_offset); + + /** + * List of objects currently involved in rendering from the + * ringbuffer. + * + * Includes buffers having the contents of their GPU caches + * flushed, not necessarily primitives. last_rendering_seqno + * represents when the rendering involved will be completed. + * + * A reference is held on the buffer while on this list. + */ + struct list_head active_list; + + /** + * List of breadcrumbs associated with GPU requests currently + * outstanding. + */ + struct list_head request_list; + + wait_queue_head_t irq_queue; + drm_local_map_t map; +}; + +static inline u32 +intel_read_status_page(struct intel_ring_buffer *ring, + int reg) +{ + u32 *regs = ring->status_page.page_addr; + return regs[reg]; +} + +int intel_init_ring_buffer(struct drm_device *dev, + struct intel_ring_buffer *ring); +void intel_cleanup_ring_buffer(struct drm_device *dev, + struct intel_ring_buffer *ring); +int intel_wait_ring_buffer(struct drm_device *dev, + struct intel_ring_buffer *ring, int n); +int intel_wrap_ring_buffer(struct drm_device *dev, + struct intel_ring_buffer *ring); +void intel_ring_begin(struct drm_device *dev, + struct intel_ring_buffer *ring, int n); +void intel_ring_emit(struct drm_device *dev, + struct intel_ring_buffer *ring, u32 data); +void intel_fill_struct(struct drm_device *dev, + struct intel_ring_buffer *ring, + void *data, + unsigned int len); +void intel_ring_advance(struct drm_device *dev, + struct intel_ring_buffer *ring); + +u32 intel_ring_get_seqno(struct drm_device *dev, + struct intel_ring_buffer *ring); + +extern struct intel_ring_buffer render_ring; +extern struct intel_ring_buffer bsd_ring; + +#endif /* _INTEL_RINGBUFFER_H_ */ diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c index aba72c489a2f..76993ac16cc1 100644 --- a/drivers/gpu/drm/i915/intel_sdvo.c +++ b/drivers/gpu/drm/i915/intel_sdvo.c @@ -1479,7 +1479,7 @@ intel_find_analog_connector(struct drm_device *dev) intel_encoder = enc_to_intel_encoder(encoder); if (intel_encoder->type == INTEL_OUTPUT_ANALOG) { list_for_each_entry(connector, &dev->mode_config.connector_list, head) { - if (connector && encoder == intel_attached_encoder(connector)) + if (encoder == intel_attached_encoder(connector)) return connector; } } diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c index e13f6af0037a..d4bcca8a5133 100644 --- a/drivers/gpu/drm/nouveau/nouveau_acpi.c +++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c @@ -34,7 +34,7 @@ static struct nouveau_dsm_priv { bool dsm_detected; acpi_handle dhandle; - acpi_handle dsm_handle; + acpi_handle rom_handle; } nouveau_dsm_priv; static const char nouveau_dsm_muid[] = { @@ -107,9 +107,9 @@ static int nouveau_dsm_set_discrete_state(acpi_handle handle, enum vga_switchero static int nouveau_dsm_switchto(enum vga_switcheroo_client_id id) { if (id == VGA_SWITCHEROO_IGD) - return nouveau_dsm_switch_mux(nouveau_dsm_priv.dsm_handle, NOUVEAU_DSM_LED_STAMINA); + return nouveau_dsm_switch_mux(nouveau_dsm_priv.dhandle, NOUVEAU_DSM_LED_STAMINA); else - return nouveau_dsm_switch_mux(nouveau_dsm_priv.dsm_handle, NOUVEAU_DSM_LED_SPEED); + return nouveau_dsm_switch_mux(nouveau_dsm_priv.dhandle, NOUVEAU_DSM_LED_SPEED); } static int nouveau_dsm_power_state(enum vga_switcheroo_client_id id, @@ -118,7 +118,7 @@ static int nouveau_dsm_power_state(enum vga_switcheroo_client_id id, if (id == VGA_SWITCHEROO_IGD) return 0; - return nouveau_dsm_set_discrete_state(nouveau_dsm_priv.dsm_handle, state); + return nouveau_dsm_set_discrete_state(nouveau_dsm_priv.dhandle, state); } static int nouveau_dsm_init(void) @@ -151,18 +151,18 @@ static bool nouveau_dsm_pci_probe(struct pci_dev *pdev) dhandle = DEVICE_ACPI_HANDLE(&pdev->dev); if (!dhandle) return false; + status = acpi_get_handle(dhandle, "_DSM", &nvidia_handle); if (ACPI_FAILURE(status)) { return false; } - ret= nouveau_dsm(nvidia_handle, NOUVEAU_DSM_SUPPORTED, - NOUVEAU_DSM_SUPPORTED_FUNCTIONS, &result); + ret = nouveau_dsm(dhandle, NOUVEAU_DSM_SUPPORTED, + NOUVEAU_DSM_SUPPORTED_FUNCTIONS, &result); if (ret < 0) return false; nouveau_dsm_priv.dhandle = dhandle; - nouveau_dsm_priv.dsm_handle = nvidia_handle; return true; } @@ -173,6 +173,7 @@ static bool nouveau_dsm_detect(void) struct pci_dev *pdev = NULL; int has_dsm = 0; int vga_count = 0; + while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, pdev)) != NULL) { vga_count++; @@ -180,7 +181,7 @@ static bool nouveau_dsm_detect(void) } if (vga_count == 2 && has_dsm) { - acpi_get_name(nouveau_dsm_priv.dsm_handle, ACPI_FULL_PATHNAME, &buffer); + acpi_get_name(nouveau_dsm_priv.dhandle, ACPI_FULL_PATHNAME, &buffer); printk(KERN_INFO "VGA switcheroo: detected DSM switching method %s handle\n", acpi_method_name); nouveau_dsm_priv.dsm_detected = true; @@ -204,3 +205,57 @@ void nouveau_unregister_dsm_handler(void) { vga_switcheroo_unregister_handler(); } + +/* retrieve the ROM in 4k blocks */ +static int nouveau_rom_call(acpi_handle rom_handle, uint8_t *bios, + int offset, int len) +{ + acpi_status status; + union acpi_object rom_arg_elements[2], *obj; + struct acpi_object_list rom_arg; + struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL}; + + rom_arg.count = 2; + rom_arg.pointer = &rom_arg_elements[0]; + + rom_arg_elements[0].type = ACPI_TYPE_INTEGER; + rom_arg_elements[0].integer.value = offset; + + rom_arg_elements[1].type = ACPI_TYPE_INTEGER; + rom_arg_elements[1].integer.value = len; + + status = acpi_evaluate_object(rom_handle, NULL, &rom_arg, &buffer); + if (ACPI_FAILURE(status)) { + printk(KERN_INFO "failed to evaluate ROM got %s\n", acpi_format_exception(status)); + return -ENODEV; + } + obj = (union acpi_object *)buffer.pointer; + memcpy(bios+offset, obj->buffer.pointer, len); + kfree(buffer.pointer); + return len; +} + +bool nouveau_acpi_rom_supported(struct pci_dev *pdev) +{ + acpi_status status; + acpi_handle dhandle, rom_handle; + + if (!nouveau_dsm_priv.dsm_detected) + return false; + + dhandle = DEVICE_ACPI_HANDLE(&pdev->dev); + if (!dhandle) + return false; + + status = acpi_get_handle(dhandle, "_ROM", &rom_handle); + if (ACPI_FAILURE(status)) + return false; + + nouveau_dsm_priv.rom_handle = rom_handle; + return true; +} + +int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len) +{ + return nouveau_rom_call(nouveau_dsm_priv.rom_handle, bios, offset, len); +} diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c index e7e69ccce5c9..fc924b649195 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bios.c +++ b/drivers/gpu/drm/nouveau/nouveau_bios.c @@ -178,6 +178,25 @@ out: pci_disable_rom(dev->pdev); } +static void load_vbios_acpi(struct drm_device *dev, uint8_t *data) +{ + int i; + int ret; + int size = 64 * 1024; + + if (!nouveau_acpi_rom_supported(dev->pdev)) + return; + + for (i = 0; i < (size / ROM_BIOS_PAGE); i++) { + ret = nouveau_acpi_get_bios_chunk(data, + (i * ROM_BIOS_PAGE), + ROM_BIOS_PAGE); + if (ret <= 0) + break; + } + return; +} + struct methods { const char desc[8]; void (*loadbios)(struct drm_device *, uint8_t *); @@ -191,6 +210,7 @@ static struct methods nv04_methods[] = { }; static struct methods nv50_methods[] = { + { "ACPI", load_vbios_acpi, true }, { "PRAMIN", load_vbios_pramin, true }, { "PROM", load_vbios_prom, false }, { "PCIROM", load_vbios_pci, true }, @@ -814,7 +834,7 @@ init_i2c_device_find(struct drm_device *dev, int i2c_index) if (i2c_index == 0x81) i2c_index = (dcb->i2c_default_indices & 0xf0) >> 4; - if (i2c_index > DCB_MAX_NUM_I2C_ENTRIES) { + if (i2c_index >= DCB_MAX_NUM_I2C_ENTRIES) { NV_ERROR(dev, "invalid i2c_index 0x%x\n", i2c_index); return NULL; } @@ -2807,7 +2827,10 @@ init_gpio(struct nvbios *bios, uint16_t offset, struct init_exec *iexec) BIOSLOG(bios, "0x%04X: Entry: 0x%08X\n", offset, gpio->entry); - nv50_gpio_set(bios->dev, gpio->tag, gpio->state_default); + BIOSLOG(bios, "0x%04X: set gpio 0x%02x, state %d\n", + offset, gpio->tag, gpio->state_default); + if (bios->execute) + nv50_gpio_set(bios->dev, gpio->tag, gpio->state_default); /* The NVIDIA binary driver doesn't appear to actually do * any of this, my VBIOS does however. @@ -3897,7 +3920,8 @@ int nouveau_bios_parse_lvds_table(struct drm_device *dev, int pxclk, bool *dl, b static uint8_t * bios_output_config_match(struct drm_device *dev, struct dcb_entry *dcbent, - uint16_t record, int record_len, int record_nr) + uint16_t record, int record_len, int record_nr, + bool match_link) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nvbios *bios = &dev_priv->vbios; @@ -3905,12 +3929,28 @@ bios_output_config_match(struct drm_device *dev, struct dcb_entry *dcbent, uint16_t table; int i, v; + switch (dcbent->type) { + case OUTPUT_TMDS: + case OUTPUT_LVDS: + case OUTPUT_DP: + break; + default: + match_link = false; + break; + } + for (i = 0; i < record_nr; i++, record += record_len) { table = ROM16(bios->data[record]); if (!table) continue; entry = ROM32(bios->data[table]); + if (match_link) { + v = (entry & 0x00c00000) >> 22; + if (!(v & dcbent->sorconf.link)) + continue; + } + v = (entry & 0x000f0000) >> 16; if (!(v & dcbent->or)) continue; @@ -3952,7 +3992,7 @@ nouveau_bios_dp_table(struct drm_device *dev, struct dcb_entry *dcbent, *length = table[4]; return bios_output_config_match(dev, dcbent, bios->display.dp_table_ptr + table[1], - table[2], table[3]); + table[2], table[3], table[0] >= 0x21); } int @@ -4041,7 +4081,7 @@ nouveau_bios_run_display_table(struct drm_device *dev, struct dcb_entry *dcbent, dcbent->type, dcbent->location, dcbent->or); otable = bios_output_config_match(dev, dcbent, table[1] + bios->display.script_table_ptr, - table[2], table[3]); + table[2], table[3], table[0] >= 0x21); if (!otable) { NV_ERROR(dev, "Couldn't find matching output script table\n"); return 1; @@ -5533,12 +5573,6 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb, entry->bus = (conn >> 16) & 0xf; entry->location = (conn >> 20) & 0x3; entry->or = (conn >> 24) & 0xf; - /* - * Normal entries consist of a single bit, but dual link has the - * next most significant bit set too - */ - entry->duallink_possible = - ((1 << (ffs(entry->or) - 1)) * 3 == entry->or); switch (entry->type) { case OUTPUT_ANALOG: @@ -5622,6 +5656,16 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb, break; } + if (dcb->version < 0x40) { + /* Normal entries consist of a single bit, but dual link has + * the next most significant bit set too + */ + entry->duallink_possible = + ((1 << (ffs(entry->or) - 1)) * 3 == entry->or); + } else { + entry->duallink_possible = (entry->sorconf.link == 3); + } + /* unsure what DCB version introduces this, 3.0? */ if (conf & 0x100000) entry->i2c_upper_default = true; @@ -6205,6 +6249,30 @@ nouveau_bios_i2c_devices_takedown(struct drm_device *dev) nouveau_i2c_fini(dev, entry); } +static bool +nouveau_bios_posted(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + bool was_locked; + unsigned htotal; + + if (dev_priv->chipset >= NV_50) { + if (NVReadVgaCrtc(dev, 0, 0x00) == 0 && + NVReadVgaCrtc(dev, 0, 0x1a) == 0) + return false; + return true; + } + + was_locked = NVLockVgaCrtcs(dev, false); + htotal = NVReadVgaCrtc(dev, 0, 0x06); + htotal |= (NVReadVgaCrtc(dev, 0, 0x07) & 0x01) << 8; + htotal |= (NVReadVgaCrtc(dev, 0, 0x07) & 0x20) << 4; + htotal |= (NVReadVgaCrtc(dev, 0, 0x25) & 0x01) << 10; + htotal |= (NVReadVgaCrtc(dev, 0, 0x41) & 0x01) << 11; + NVLockVgaCrtcs(dev, was_locked); + return (htotal != 0); +} + int nouveau_bios_init(struct drm_device *dev) { @@ -6239,11 +6307,9 @@ nouveau_bios_init(struct drm_device *dev) bios->execute = false; /* ... unless card isn't POSTed already */ - if (dev_priv->card_type >= NV_10 && - NVReadVgaCrtc(dev, 0, 0x00) == 0 && - NVReadVgaCrtc(dev, 0, 0x1a) == 0) { + if (!nouveau_bios_posted(dev)) { NV_INFO(dev, "Adaptor not initialised\n"); - if (dev_priv->card_type < NV_50) { + if (dev_priv->card_type < NV_40) { NV_ERROR(dev, "Unable to POST this chipset\n"); return -ENODEV; } diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c index 266b0ff441af..149ed224c3cb 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.c +++ b/drivers/gpu/drm/nouveau/nouveau_connector.c @@ -432,24 +432,27 @@ nouveau_connector_set_property(struct drm_connector *connector, } static struct drm_display_mode * -nouveau_connector_native_mode(struct nouveau_connector *connector) +nouveau_connector_native_mode(struct drm_connector *connector) { - struct drm_device *dev = connector->base.dev; + struct drm_connector_helper_funcs *helper = connector->helper_private; + struct nouveau_connector *nv_connector = nouveau_connector(connector); + struct drm_device *dev = connector->dev; struct drm_display_mode *mode, *largest = NULL; int high_w = 0, high_h = 0, high_v = 0; - /* Use preferred mode if there is one.. */ - list_for_each_entry(mode, &connector->base.probed_modes, head) { + list_for_each_entry(mode, &nv_connector->base.probed_modes, head) { + if (helper->mode_valid(connector, mode) != MODE_OK) + continue; + + /* Use preferred mode if there is one.. */ if (mode->type & DRM_MODE_TYPE_PREFERRED) { NV_DEBUG_KMS(dev, "native mode from preferred\n"); return drm_mode_duplicate(dev, mode); } - } - /* Otherwise, take the resolution with the largest width, then height, - * then vertical refresh - */ - list_for_each_entry(mode, &connector->base.probed_modes, head) { + /* Otherwise, take the resolution with the largest width, then + * height, then vertical refresh + */ if (mode->hdisplay < high_w) continue; @@ -553,7 +556,7 @@ nouveau_connector_get_modes(struct drm_connector *connector) */ if (!nv_connector->native_mode) nv_connector->native_mode = - nouveau_connector_native_mode(nv_connector); + nouveau_connector_native_mode(connector); if (ret == 0 && nv_connector->native_mode) { struct drm_display_mode *mode; @@ -584,9 +587,9 @@ nouveau_connector_mode_valid(struct drm_connector *connector, switch (nv_encoder->dcb->type) { case OUTPUT_LVDS: - BUG_ON(!nv_connector->native_mode); - if (mode->hdisplay > nv_connector->native_mode->hdisplay || - mode->vdisplay > nv_connector->native_mode->vdisplay) + if (nv_connector->native_mode && + (mode->hdisplay > nv_connector->native_mode->hdisplay || + mode->vdisplay > nv_connector->native_mode->vdisplay)) return MODE_PANEL; min_clock = 0; @@ -594,8 +597,7 @@ nouveau_connector_mode_valid(struct drm_connector *connector, break; case OUTPUT_TMDS: if ((dev_priv->card_type >= NV_50 && !nouveau_duallink) || - (dev_priv->card_type < NV_50 && - !nv_encoder->dcb->duallink_possible)) + !nv_encoder->dcb->duallink_possible) max_clock = 165000; else max_clock = 330000; @@ -729,7 +731,7 @@ nouveau_connector_create_lvds(struct drm_device *dev, if (ret == 0) goto out; nv_connector->detected_encoder = nv_encoder; - nv_connector->native_mode = nouveau_connector_native_mode(nv_connector); + nv_connector->native_mode = nouveau_connector_native_mode(connector); list_for_each_entry_safe(mode, temp, &connector->probed_modes, head) drm_mode_remove(connector, mode); diff --git a/drivers/gpu/drm/nouveau/nouveau_crtc.h b/drivers/gpu/drm/nouveau/nouveau_crtc.h index 49fa7b2d257e..cb1ce2a09162 100644 --- a/drivers/gpu/drm/nouveau/nouveau_crtc.h +++ b/drivers/gpu/drm/nouveau/nouveau_crtc.h @@ -40,6 +40,8 @@ struct nouveau_crtc { int sharpness; int last_dpms; + int cursor_saved_x, cursor_saved_y; + struct { int cpp; bool blanked; diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.c b/drivers/gpu/drm/nouveau/nouveau_drv.c index f60a2b2ae445..0fd8e10dbbde 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drv.c +++ b/drivers/gpu/drm/nouveau/nouveau_drv.c @@ -175,6 +175,13 @@ nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state) nouveau_bo_unpin(nouveau_fb->nvbo); } + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); + + nouveau_bo_unmap(nv_crtc->cursor.nvbo); + nouveau_bo_unpin(nv_crtc->cursor.nvbo); + } + NV_INFO(dev, "Evicting buffers...\n"); ttm_bo_evict_mm(&dev_priv->ttm.bdev, TTM_PL_VRAM); @@ -314,12 +321,34 @@ nouveau_pci_resume(struct pci_dev *pdev) nouveau_bo_pin(nouveau_fb->nvbo, TTM_PL_FLAG_VRAM); } + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); + int ret; + + ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM); + if (!ret) + ret = nouveau_bo_map(nv_crtc->cursor.nvbo); + if (ret) + NV_ERROR(dev, "Could not pin/map cursor.\n"); + } + if (dev_priv->card_type < NV_50) { nv04_display_restore(dev); NVLockVgaCrtcs(dev, false); } else nv50_display_init(dev); + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); + + nv_crtc->cursor.set_offset(nv_crtc, + nv_crtc->cursor.nvbo->bo.offset - + dev_priv->vm_vram_base); + + nv_crtc->cursor.set_pos(nv_crtc, nv_crtc->cursor_saved_x, + nv_crtc->cursor_saved_y); + } + /* Force CLUT to get re-loaded during modeset */ list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h index 5b134438effe..c69719106489 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drv.h +++ b/drivers/gpu/drm/nouveau/nouveau_drv.h @@ -851,12 +851,17 @@ extern int nouveau_dma_init(struct nouveau_channel *); extern int nouveau_dma_wait(struct nouveau_channel *, int slots, int size); /* nouveau_acpi.c */ +#define ROM_BIOS_PAGE 4096 #if defined(CONFIG_ACPI) void nouveau_register_dsm_handler(void); void nouveau_unregister_dsm_handler(void); +int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len); +bool nouveau_acpi_rom_supported(struct pci_dev *pdev); #else static inline void nouveau_register_dsm_handler(void) {} static inline void nouveau_unregister_dsm_handler(void) {} +static inline bool nouveau_acpi_rom_supported(struct pci_dev *pdev) { return false; } +static inline int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len) { return -EINVAL; } #endif /* nouveau_backlight.c */ diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c index fd4a2df715e9..c9a4a0d2a115 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c +++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c @@ -377,6 +377,7 @@ int nouveau_fbcon_init(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_fbdev *nfbdev; + int ret; nfbdev = kzalloc(sizeof(struct nouveau_fbdev), GFP_KERNEL); if (!nfbdev) @@ -386,7 +387,12 @@ int nouveau_fbcon_init(struct drm_device *dev) dev_priv->nfbdev = nfbdev; nfbdev->helper.funcs = &nouveau_fbcon_helper_funcs; - drm_fb_helper_init(dev, &nfbdev->helper, 2, 4); + ret = drm_fb_helper_init(dev, &nfbdev->helper, 2, 4); + if (ret) { + kfree(nfbdev); + return ret; + } + drm_fb_helper_single_add_all_connectors(&nfbdev->helper); drm_fb_helper_initial_config(&nfbdev->helper, 32); return 0; diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c index 37c7bf8e8296..fb6b791506b2 100644 --- a/drivers/gpu/drm/nouveau/nouveau_mem.c +++ b/drivers/gpu/drm/nouveau/nouveau_mem.c @@ -541,7 +541,8 @@ nouveau_mem_detect(struct drm_device *dev) dev_priv->vram_size = nv_rd32(dev, NV04_FIFO_DATA); dev_priv->vram_size &= NV10_FIFO_DATA_RAM_AMOUNT_MB_MASK; if (dev_priv->chipset == 0xaa || dev_priv->chipset == 0xac) - dev_priv->vram_sys_base = nv_rd32(dev, 0x100e10) << 12; + dev_priv->vram_sys_base = nv_rd32(dev, 0x100e10); + dev_priv->vram_sys_base <<= 12; } NV_INFO(dev, "Detected %dMiB VRAM\n", (int)(dev_priv->vram_size >> 20)); diff --git a/drivers/gpu/drm/nouveau/nouveau_state.c b/drivers/gpu/drm/nouveau/nouveau_state.c index e632339c323e..b02a231d6937 100644 --- a/drivers/gpu/drm/nouveau/nouveau_state.c +++ b/drivers/gpu/drm/nouveau/nouveau_state.c @@ -376,12 +376,15 @@ out_err: static void nouveau_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_state state) { + struct drm_device *dev = pci_get_drvdata(pdev); pm_message_t pmm = { .event = PM_EVENT_SUSPEND }; if (state == VGA_SWITCHEROO_ON) { printk(KERN_ERR "VGA switcheroo: switched nouveau on\n"); nouveau_pci_resume(pdev); + drm_kms_helper_poll_enable(dev); } else { printk(KERN_ERR "VGA switcheroo: switched nouveau off\n"); + drm_kms_helper_poll_disable(dev); nouveau_pci_suspend(pdev, pmm); } } @@ -776,29 +779,24 @@ int nouveau_load(struct drm_device *dev, unsigned long flags) return ret; } - /* map larger RAMIN aperture on NV40 cards */ - dev_priv->ramin = NULL; + /* Map PRAMIN BAR, or on older cards, the aperture withing BAR0 */ if (dev_priv->card_type >= NV_40) { int ramin_bar = 2; if (pci_resource_len(dev->pdev, ramin_bar) == 0) ramin_bar = 3; dev_priv->ramin_size = pci_resource_len(dev->pdev, ramin_bar); - dev_priv->ramin = ioremap( - pci_resource_start(dev->pdev, ramin_bar), + dev_priv->ramin = + ioremap(pci_resource_start(dev->pdev, ramin_bar), dev_priv->ramin_size); if (!dev_priv->ramin) { - NV_ERROR(dev, "Failed to init RAMIN mapping, " - "limited instance memory available\n"); + NV_ERROR(dev, "Failed to PRAMIN BAR"); + return -ENOMEM; } - } - - /* On older cards (or if the above failed), create a map covering - * the BAR0 PRAMIN aperture */ - if (!dev_priv->ramin) { + } else { dev_priv->ramin_size = 1 * 1024 * 1024; dev_priv->ramin = ioremap(mmio_start_offs + NV_RAMIN, - dev_priv->ramin_size); + dev_priv->ramin_size); if (!dev_priv->ramin) { NV_ERROR(dev, "Failed to map BAR0 PRAMIN.\n"); return -ENOMEM; @@ -913,6 +911,9 @@ int nouveau_ioctl_getparam(struct drm_device *dev, void *data, case NOUVEAU_GETPARAM_VM_VRAM_BASE: getparam->value = dev_priv->vm_vram_base; break; + case NOUVEAU_GETPARAM_PTIMER_TIME: + getparam->value = dev_priv->engine.timer.read(dev); + break; case NOUVEAU_GETPARAM_GRAPH_UNITS: /* NV40 and NV50 versions are quite different, but register * address is the same. User is supposed to know the card diff --git a/drivers/gpu/drm/nouveau/nv04_cursor.c b/drivers/gpu/drm/nouveau/nv04_cursor.c index 89a91b9d8b25..aaf3de3bc816 100644 --- a/drivers/gpu/drm/nouveau/nv04_cursor.c +++ b/drivers/gpu/drm/nouveau/nv04_cursor.c @@ -20,6 +20,7 @@ nv04_cursor_hide(struct nouveau_crtc *nv_crtc, bool update) static void nv04_cursor_set_pos(struct nouveau_crtc *nv_crtc, int x, int y) { + nv_crtc->cursor_saved_x = x; nv_crtc->cursor_saved_y = y; NVWriteRAMDAC(nv_crtc->base.dev, nv_crtc->index, NV_PRAMDAC_CU_START_POS, XLATE(y, 0, NV_PRAMDAC_CU_START_POS_Y) | diff --git a/drivers/gpu/drm/nouveau/nv50_cursor.c b/drivers/gpu/drm/nouveau/nv50_cursor.c index 753e723adb3a..03ad7ab14f09 100644 --- a/drivers/gpu/drm/nouveau/nv50_cursor.c +++ b/drivers/gpu/drm/nouveau/nv50_cursor.c @@ -107,6 +107,7 @@ nv50_cursor_set_pos(struct nouveau_crtc *nv_crtc, int x, int y) { struct drm_device *dev = nv_crtc->base.dev; + nv_crtc->cursor_saved_x = x; nv_crtc->cursor_saved_y = y; nv_wr32(dev, NV50_PDISPLAY_CURSOR_USER_POS(nv_crtc->index), ((y & 0xFFFF) << 16) | (x & 0xFFFF)); /* Needed to make the cursor move. */ diff --git a/drivers/gpu/drm/nouveau/nv50_fb.c b/drivers/gpu/drm/nouveau/nv50_fb.c index a95e6941ba88..32611bd30e6d 100644 --- a/drivers/gpu/drm/nouveau/nv50_fb.c +++ b/drivers/gpu/drm/nouveau/nv50_fb.c @@ -6,10 +6,16 @@ int nv50_fb_init(struct drm_device *dev) { - /* This is needed to get meaningful information from 100c90 - * on traps. No idea what these values mean exactly. */ struct drm_nouveau_private *dev_priv = dev->dev_private; + /* Not a clue what this is exactly. Without pointing it at a + * scratch page, VRAM->GART blits with M2MF (as in DDX DFS) + * cause IOMMU "read from address 0" errors (rh#561267) + */ + nv_wr32(dev, 0x100c08, dev_priv->gart_info.sg_dummy_bus >> 8); + + /* This is needed to get meaningful information from 100c90 + * on traps. No idea what these values mean exactly. */ switch (dev_priv->chipset) { case 0x50: nv_wr32(dev, 0x100c90, 0x0707ff); diff --git a/drivers/gpu/drm/nouveau/nv50_gpio.c b/drivers/gpu/drm/nouveau/nv50_gpio.c index c61782b314e7..bb47ad737267 100644 --- a/drivers/gpu/drm/nouveau/nv50_gpio.c +++ b/drivers/gpu/drm/nouveau/nv50_gpio.c @@ -31,7 +31,7 @@ nv50_gpio_location(struct dcb_gpio_entry *gpio, uint32_t *reg, uint32_t *shift) { const uint32_t nv50_gpio_reg[4] = { 0xe104, 0xe108, 0xe280, 0xe284 }; - if (gpio->line > 32) + if (gpio->line >= 32) return -EINVAL; *reg = nv50_gpio_reg[gpio->line >> 3]; diff --git a/drivers/gpu/drm/nouveau/nv50_sor.c b/drivers/gpu/drm/nouveau/nv50_sor.c index b11eaf9c5c7c..812778db76ac 100644 --- a/drivers/gpu/drm/nouveau/nv50_sor.c +++ b/drivers/gpu/drm/nouveau/nv50_sor.c @@ -274,7 +274,6 @@ static const struct drm_encoder_funcs nv50_sor_encoder_funcs = { int nv50_sor_create(struct drm_device *dev, struct dcb_entry *entry) { - struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_encoder *nv_encoder = NULL; struct drm_encoder *encoder; bool dum; @@ -324,11 +323,7 @@ nv50_sor_create(struct drm_device *dev, struct dcb_entry *entry) int or = nv_encoder->or, link = !(entry->dpconf.sor.link & 1); uint32_t tmp; - if (dev_priv->chipset < 0x90 || - dev_priv->chipset == 0x92 || dev_priv->chipset == 0xa0) - tmp = nv_rd32(dev, NV50_PDISPLAY_SOR_MODE_CTRL_C(or)); - else - tmp = nv_rd32(dev, NV90_PDISPLAY_SOR_MODE_CTRL_C(or)); + tmp = nv_rd32(dev, 0x61c700 + (or * 0x800)); switch ((tmp & 0x00000f00) >> 8) { case 8: diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile index 3c91312dea9a..84b1f2729d43 100644 --- a/drivers/gpu/drm/radeon/Makefile +++ b/drivers/gpu/drm/radeon/Makefile @@ -33,6 +33,9 @@ $(obj)/rs600_reg_safe.h: $(src)/reg_srcs/rs600 $(obj)/mkregtable $(obj)/r600_reg_safe.h: $(src)/reg_srcs/r600 $(obj)/mkregtable $(call if_changed,mkregtable) +$(obj)/evergreen_reg_safe.h: $(src)/reg_srcs/evergreen $(obj)/mkregtable + $(call if_changed,mkregtable) + $(obj)/r100.o: $(obj)/r100_reg_safe.h $(obj)/rn50_reg_safe.h $(obj)/r200.o: $(obj)/r200_reg_safe.h @@ -47,6 +50,8 @@ $(obj)/rs600.o: $(obj)/rs600_reg_safe.h $(obj)/r600_cs.o: $(obj)/r600_reg_safe.h +$(obj)/evergreen_cs.o: $(obj)/evergreen_reg_safe.h + radeon-y := radeon_drv.o radeon_cp.o radeon_state.o radeon_mem.o \ radeon_irq.o r300_cmdbuf.o r600_cp.o # add KMS driver @@ -60,7 +65,7 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \ rs400.o rs600.o rs690.o rv515.o r520.o r600.o rv770.o radeon_test.o \ r200.o radeon_legacy_tv.o r600_cs.o r600_blit.o r600_blit_shaders.o \ r600_blit_kms.o radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o \ - evergreen.o + evergreen.o evergreen_cs.o radeon-$(CONFIG_COMPAT) += radeon_ioc32.o radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o diff --git a/drivers/gpu/drm/radeon/atombios.h b/drivers/gpu/drm/radeon/atombios.h index 2ebcb979dd7e..1bc72c3190a9 100644 --- a/drivers/gpu/drm/radeon/atombios.h +++ b/drivers/gpu/drm/radeon/atombios.h @@ -3780,7 +3780,7 @@ typedef struct _ATOM_ASIC_SS_ASSIGNMENT UCHAR ucReserved[2]; }ATOM_ASIC_SS_ASSIGNMENT; -//Define ucClockIndication, SW uses the IDs below to search if the SS is requried/enabled on a clock branch/signal type. +//Define ucClockIndication, SW uses the IDs below to search if the SS is required/enabled on a clock branch/signal type. //SS is not required or enabled if a match is not found. #define ASIC_INTERNAL_MEMORY_SS 1 #define ASIC_INTERNAL_ENGINE_SS 2 @@ -5967,7 +5967,7 @@ typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO UCHAR ucPadding; // For proper alignment and size. USHORT usVDDC; // For the 780, use: None, Low, High, Variable UCHAR ucMaxHTLinkWidth; // From SBIOS - {2, 4, 8, 16} - UCHAR ucMinHTLinkWidth; // From SBIOS - {2, 4, 8, 16}. Effective only if CDLW enabled. Minimum down stream width could be bigger as display BW requriement. + UCHAR ucMinHTLinkWidth; // From SBIOS - {2, 4, 8, 16}. Effective only if CDLW enabled. Minimum down stream width could be bigger as display BW requirement. USHORT usHTLinkFreq; // See definition ATOM_PPLIB_RS780_HTLINKFREQ_xxx or in MHz(>=200). ULONG ulFlags; } ATOM_PPLIB_RS780_CLOCK_INFO; diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index f3f2827017ef..8c2d6478a221 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -498,7 +498,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, if ((rdev->family == CHIP_RS600) || (rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740)) - pll->flags |= (RADEON_PLL_USE_FRAC_FB_DIV | + pll->flags |= (/*RADEON_PLL_USE_FRAC_FB_DIV |*/ RADEON_PLL_PREFER_CLOSEST_LOWER); if (ASIC_IS_DCE32(rdev) && mode->clock > 200000) /* range limits??? */ diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c index a4745e49ecf1..057192acdd36 100644 --- a/drivers/gpu/drm/radeon/evergreen.c +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -41,7 +41,18 @@ void evergreen_fini(struct radeon_device *rdev); void evergreen_pm_misc(struct radeon_device *rdev) { - + int req_ps_idx = rdev->pm.requested_power_state_index; + int req_cm_idx = rdev->pm.requested_clock_mode_index; + struct radeon_power_state *ps = &rdev->pm.power_state[req_ps_idx]; + struct radeon_voltage *voltage = &ps->clock_info[req_cm_idx].voltage; + + if ((voltage->type == VOLTAGE_SW) && voltage->voltage) { + if (voltage->voltage != rdev->pm.current_vddc) { + radeon_atom_set_voltage(rdev, voltage->voltage); + rdev->pm.current_vddc = voltage->voltage; + DRM_DEBUG("Setting: v: %d\n", voltage->voltage); + } + } } void evergreen_pm_prepare(struct radeon_device *rdev) @@ -596,7 +607,7 @@ static void evergreen_mc_program(struct radeon_device *rdev) WREG32(MC_VM_FB_LOCATION, tmp); WREG32(HDP_NONSURFACE_BASE, (rdev->mc.vram_start >> 8)); WREG32(HDP_NONSURFACE_INFO, (2 << 7)); - WREG32(HDP_NONSURFACE_SIZE, (rdev->mc.mc_vram_size - 1) | 0x3FF); + WREG32(HDP_NONSURFACE_SIZE, 0x3FFFFFFF); if (rdev->flags & RADEON_IS_AGP) { WREG32(MC_VM_AGP_TOP, rdev->mc.gtt_end >> 16); WREG32(MC_VM_AGP_BOT, rdev->mc.gtt_start >> 16); @@ -1211,11 +1222,11 @@ static void evergreen_gpu_init(struct radeon_device *rdev) ps_thread_count = 128; sq_thread_resource_mgmt = NUM_PS_THREADS(ps_thread_count); - sq_thread_resource_mgmt |= NUM_VS_THREADS(((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8; - sq_thread_resource_mgmt |= NUM_GS_THREADS(((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8; - sq_thread_resource_mgmt |= NUM_ES_THREADS(((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8; - sq_thread_resource_mgmt_2 = NUM_HS_THREADS(((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8; - sq_thread_resource_mgmt_2 |= NUM_LS_THREADS(((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8; + sq_thread_resource_mgmt |= NUM_VS_THREADS((((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8); + sq_thread_resource_mgmt |= NUM_GS_THREADS((((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8); + sq_thread_resource_mgmt |= NUM_ES_THREADS((((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8); + sq_thread_resource_mgmt_2 = NUM_HS_THREADS((((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8); + sq_thread_resource_mgmt_2 |= NUM_LS_THREADS((((rdev->config.evergreen.max_threads - ps_thread_count) / 6) / 8) * 8); sq_stack_resource_mgmt_1 = NUM_PS_STACK_ENTRIES((rdev->config.evergreen.max_stack_entries * 1) / 6); sq_stack_resource_mgmt_1 |= NUM_VS_STACK_ENTRIES((rdev->config.evergreen.max_stack_entries * 1) / 6); @@ -1249,6 +1260,9 @@ static void evergreen_gpu_init(struct radeon_device *rdev) WREG32(VGT_GS_VERTEX_REUSE, 16); WREG32(PA_SC_LINE_STIPPLE_STATE, 0); + WREG32(VGT_VERTEX_REUSE_BLOCK_CNTL, 14); + WREG32(VGT_OUT_DEALLOC_CNTL, 16); + WREG32(CB_PERF_CTR0_SEL_0, 0); WREG32(CB_PERF_CTR0_SEL_1, 0); WREG32(CB_PERF_CTR1_SEL_0, 0); @@ -1258,6 +1272,26 @@ static void evergreen_gpu_init(struct radeon_device *rdev) WREG32(CB_PERF_CTR3_SEL_0, 0); WREG32(CB_PERF_CTR3_SEL_1, 0); + /* clear render buffer base addresses */ + WREG32(CB_COLOR0_BASE, 0); + WREG32(CB_COLOR1_BASE, 0); + WREG32(CB_COLOR2_BASE, 0); + WREG32(CB_COLOR3_BASE, 0); + WREG32(CB_COLOR4_BASE, 0); + WREG32(CB_COLOR5_BASE, 0); + WREG32(CB_COLOR6_BASE, 0); + WREG32(CB_COLOR7_BASE, 0); + WREG32(CB_COLOR8_BASE, 0); + WREG32(CB_COLOR9_BASE, 0); + WREG32(CB_COLOR10_BASE, 0); + WREG32(CB_COLOR11_BASE, 0); + + /* set the shader const cache sizes to 0 */ + for (i = SQ_ALU_CONST_BUFFER_SIZE_PS_0; i < 0x28200; i += 4) + WREG32(i, 0); + for (i = SQ_ALU_CONST_BUFFER_SIZE_HS_0; i < 0x29000; i += 4) + WREG32(i, 0); + hdp_host_path_cntl = RREG32(HDP_HOST_PATH_CNTL); WREG32(HDP_HOST_PATH_CNTL, hdp_host_path_cntl); @@ -2148,7 +2182,7 @@ int evergreen_init(struct radeon_device *rdev) if (r) return r; - rdev->accel_working = false; + rdev->accel_working = true; r = evergreen_startup(rdev); if (r) { dev_err(rdev->dev, "disabling GPU acceleration\n"); diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c new file mode 100644 index 000000000000..010963d4570f --- /dev/null +++ b/drivers/gpu/drm/radeon/evergreen_cs.c @@ -0,0 +1,1356 @@ +/* + * Copyright 2010 Advanced Micro Devices, Inc. + * Copyright 2008 Red Hat Inc. + * Copyright 2009 Jerome Glisse. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Dave Airlie + * Alex Deucher + * Jerome Glisse + */ +#include "drmP.h" +#include "radeon.h" +#include "evergreend.h" +#include "evergreen_reg_safe.h" + +static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p, + struct radeon_cs_reloc **cs_reloc); + +struct evergreen_cs_track { + u32 group_size; + u32 nbanks; + u32 npipes; + /* value we track */ + u32 nsamples; + u32 cb_color_base_last[12]; + struct radeon_bo *cb_color_bo[12]; + u32 cb_color_bo_offset[12]; + struct radeon_bo *cb_color_fmask_bo[8]; + struct radeon_bo *cb_color_cmask_bo[8]; + u32 cb_color_info[12]; + u32 cb_color_view[12]; + u32 cb_color_pitch_idx[12]; + u32 cb_color_slice_idx[12]; + u32 cb_color_dim_idx[12]; + u32 cb_color_dim[12]; + u32 cb_color_pitch[12]; + u32 cb_color_slice[12]; + u32 cb_color_cmask_slice[8]; + u32 cb_color_fmask_slice[8]; + u32 cb_target_mask; + u32 cb_shader_mask; + u32 vgt_strmout_config; + u32 vgt_strmout_buffer_config; + u32 db_depth_control; + u32 db_depth_view; + u32 db_depth_size; + u32 db_depth_size_idx; + u32 db_z_info; + u32 db_z_idx; + u32 db_z_read_offset; + u32 db_z_write_offset; + struct radeon_bo *db_z_read_bo; + struct radeon_bo *db_z_write_bo; + u32 db_s_info; + u32 db_s_idx; + u32 db_s_read_offset; + u32 db_s_write_offset; + struct radeon_bo *db_s_read_bo; + struct radeon_bo *db_s_write_bo; +}; + +static void evergreen_cs_track_init(struct evergreen_cs_track *track) +{ + int i; + + for (i = 0; i < 8; i++) { + track->cb_color_fmask_bo[i] = NULL; + track->cb_color_cmask_bo[i] = NULL; + track->cb_color_cmask_slice[i] = 0; + track->cb_color_fmask_slice[i] = 0; + } + + for (i = 0; i < 12; i++) { + track->cb_color_base_last[i] = 0; + track->cb_color_bo[i] = NULL; + track->cb_color_bo_offset[i] = 0xFFFFFFFF; + track->cb_color_info[i] = 0; + track->cb_color_view[i] = 0; + track->cb_color_pitch_idx[i] = 0; + track->cb_color_slice_idx[i] = 0; + track->cb_color_dim[i] = 0; + track->cb_color_pitch[i] = 0; + track->cb_color_slice[i] = 0; + track->cb_color_dim[i] = 0; + } + track->cb_target_mask = 0xFFFFFFFF; + track->cb_shader_mask = 0xFFFFFFFF; + + track->db_depth_view = 0xFFFFC000; + track->db_depth_size = 0xFFFFFFFF; + track->db_depth_size_idx = 0; + track->db_depth_control = 0xFFFFFFFF; + track->db_z_info = 0xFFFFFFFF; + track->db_z_idx = 0xFFFFFFFF; + track->db_z_read_offset = 0xFFFFFFFF; + track->db_z_write_offset = 0xFFFFFFFF; + track->db_z_read_bo = NULL; + track->db_z_write_bo = NULL; + track->db_s_info = 0xFFFFFFFF; + track->db_s_idx = 0xFFFFFFFF; + track->db_s_read_offset = 0xFFFFFFFF; + track->db_s_write_offset = 0xFFFFFFFF; + track->db_s_read_bo = NULL; + track->db_s_write_bo = NULL; +} + +static inline int evergreen_cs_track_validate_cb(struct radeon_cs_parser *p, int i) +{ + /* XXX fill in */ + return 0; +} + +static int evergreen_cs_track_check(struct radeon_cs_parser *p) +{ + struct evergreen_cs_track *track = p->track; + + /* we don't support stream out buffer yet */ + if (track->vgt_strmout_config || track->vgt_strmout_buffer_config) { + dev_warn(p->dev, "this kernel doesn't support SMX output buffer\n"); + return -EINVAL; + } + + /* XXX fill in */ + return 0; +} + +/** + * evergreen_cs_packet_parse() - parse cp packet and point ib index to next packet + * @parser: parser structure holding parsing context. + * @pkt: where to store packet informations + * + * Assume that chunk_ib_index is properly set. Will return -EINVAL + * if packet is bigger than remaining ib size. or if packets is unknown. + **/ +int evergreen_cs_packet_parse(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt, + unsigned idx) +{ + struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx]; + uint32_t header; + + if (idx >= ib_chunk->length_dw) { + DRM_ERROR("Can not parse packet at %d after CS end %d !\n", + idx, ib_chunk->length_dw); + return -EINVAL; + } + header = radeon_get_ib_value(p, idx); + pkt->idx = idx; + pkt->type = CP_PACKET_GET_TYPE(header); + pkt->count = CP_PACKET_GET_COUNT(header); + pkt->one_reg_wr = 0; + switch (pkt->type) { + case PACKET_TYPE0: + pkt->reg = CP_PACKET0_GET_REG(header); + break; + case PACKET_TYPE3: + pkt->opcode = CP_PACKET3_GET_OPCODE(header); + break; + case PACKET_TYPE2: + pkt->count = -1; + break; + default: + DRM_ERROR("Unknown packet type %d at %d !\n", pkt->type, idx); + return -EINVAL; + } + if ((pkt->count + 1 + pkt->idx) >= ib_chunk->length_dw) { + DRM_ERROR("Packet (%d:%d:%d) end after CS buffer (%d) !\n", + pkt->idx, pkt->type, pkt->count, ib_chunk->length_dw); + return -EINVAL; + } + return 0; +} + +/** + * evergreen_cs_packet_next_reloc() - parse next packet which should be reloc packet3 + * @parser: parser structure holding parsing context. + * @data: pointer to relocation data + * @offset_start: starting offset + * @offset_mask: offset mask (to align start offset on) + * @reloc: reloc informations + * + * Check next packet is relocation packet3, do bo validation and compute + * GPU offset using the provided start. + **/ +static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p, + struct radeon_cs_reloc **cs_reloc) +{ + struct radeon_cs_chunk *relocs_chunk; + struct radeon_cs_packet p3reloc; + unsigned idx; + int r; + + if (p->chunk_relocs_idx == -1) { + DRM_ERROR("No relocation chunk !\n"); + return -EINVAL; + } + *cs_reloc = NULL; + relocs_chunk = &p->chunks[p->chunk_relocs_idx]; + r = evergreen_cs_packet_parse(p, &p3reloc, p->idx); + if (r) { + return r; + } + p->idx += p3reloc.count + 2; + if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) { + DRM_ERROR("No packet3 for relocation for packet at %d.\n", + p3reloc.idx); + return -EINVAL; + } + idx = radeon_get_ib_value(p, p3reloc.idx + 1); + if (idx >= relocs_chunk->length_dw) { + DRM_ERROR("Relocs at %d after relocations chunk end %d !\n", + idx, relocs_chunk->length_dw); + return -EINVAL; + } + /* FIXME: we assume reloc size is 4 dwords */ + *cs_reloc = p->relocs_ptr[(idx / 4)]; + return 0; +} + +/** + * evergreen_cs_packet_next_is_pkt3_nop() - test if next packet is packet3 nop for reloc + * @parser: parser structure holding parsing context. + * + * Check next packet is relocation packet3, do bo validation and compute + * GPU offset using the provided start. + **/ +static inline int evergreen_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p) +{ + struct radeon_cs_packet p3reloc; + int r; + + r = evergreen_cs_packet_parse(p, &p3reloc, p->idx); + if (r) { + return 0; + } + if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) { + return 0; + } + return 1; +} + +/** + * evergreen_cs_packet_next_vline() - parse userspace VLINE packet + * @parser: parser structure holding parsing context. + * + * Userspace sends a special sequence for VLINE waits. + * PACKET0 - VLINE_START_END + value + * PACKET3 - WAIT_REG_MEM poll vline status reg + * RELOC (P3) - crtc_id in reloc. + * + * This function parses this and relocates the VLINE START END + * and WAIT_REG_MEM packets to the correct crtc. + * It also detects a switched off crtc and nulls out the + * wait in that case. + */ +static int evergreen_cs_packet_parse_vline(struct radeon_cs_parser *p) +{ + struct drm_mode_object *obj; + struct drm_crtc *crtc; + struct radeon_crtc *radeon_crtc; + struct radeon_cs_packet p3reloc, wait_reg_mem; + int crtc_id; + int r; + uint32_t header, h_idx, reg, wait_reg_mem_info; + volatile uint32_t *ib; + + ib = p->ib->ptr; + + /* parse the WAIT_REG_MEM */ + r = evergreen_cs_packet_parse(p, &wait_reg_mem, p->idx); + if (r) + return r; + + /* check its a WAIT_REG_MEM */ + if (wait_reg_mem.type != PACKET_TYPE3 || + wait_reg_mem.opcode != PACKET3_WAIT_REG_MEM) { + DRM_ERROR("vline wait missing WAIT_REG_MEM segment\n"); + r = -EINVAL; + return r; + } + + wait_reg_mem_info = radeon_get_ib_value(p, wait_reg_mem.idx + 1); + /* bit 4 is reg (0) or mem (1) */ + if (wait_reg_mem_info & 0x10) { + DRM_ERROR("vline WAIT_REG_MEM waiting on MEM rather than REG\n"); + r = -EINVAL; + return r; + } + /* waiting for value to be equal */ + if ((wait_reg_mem_info & 0x7) != 0x3) { + DRM_ERROR("vline WAIT_REG_MEM function not equal\n"); + r = -EINVAL; + return r; + } + if ((radeon_get_ib_value(p, wait_reg_mem.idx + 2) << 2) != EVERGREEN_VLINE_STATUS) { + DRM_ERROR("vline WAIT_REG_MEM bad reg\n"); + r = -EINVAL; + return r; + } + + if (radeon_get_ib_value(p, wait_reg_mem.idx + 5) != EVERGREEN_VLINE_STAT) { + DRM_ERROR("vline WAIT_REG_MEM bad bit mask\n"); + r = -EINVAL; + return r; + } + + /* jump over the NOP */ + r = evergreen_cs_packet_parse(p, &p3reloc, p->idx + wait_reg_mem.count + 2); + if (r) + return r; + + h_idx = p->idx - 2; + p->idx += wait_reg_mem.count + 2; + p->idx += p3reloc.count + 2; + + header = radeon_get_ib_value(p, h_idx); + crtc_id = radeon_get_ib_value(p, h_idx + 2 + 7 + 1); + reg = CP_PACKET0_GET_REG(header); + mutex_lock(&p->rdev->ddev->mode_config.mutex); + obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC); + if (!obj) { + DRM_ERROR("cannot find crtc %d\n", crtc_id); + r = -EINVAL; + goto out; + } + crtc = obj_to_crtc(obj); + radeon_crtc = to_radeon_crtc(crtc); + crtc_id = radeon_crtc->crtc_id; + + if (!crtc->enabled) { + /* if the CRTC isn't enabled - we need to nop out the WAIT_REG_MEM */ + ib[h_idx + 2] = PACKET2(0); + ib[h_idx + 3] = PACKET2(0); + ib[h_idx + 4] = PACKET2(0); + ib[h_idx + 5] = PACKET2(0); + ib[h_idx + 6] = PACKET2(0); + ib[h_idx + 7] = PACKET2(0); + ib[h_idx + 8] = PACKET2(0); + } else { + switch (reg) { + case EVERGREEN_VLINE_START_END: + header &= ~R600_CP_PACKET0_REG_MASK; + header |= (EVERGREEN_VLINE_START_END + radeon_crtc->crtc_offset) >> 2; + ib[h_idx] = header; + ib[h_idx + 4] = (EVERGREEN_VLINE_STATUS + radeon_crtc->crtc_offset) >> 2; + break; + default: + DRM_ERROR("unknown crtc reloc\n"); + r = -EINVAL; + goto out; + } + } +out: + mutex_unlock(&p->rdev->ddev->mode_config.mutex); + return r; +} + +static int evergreen_packet0_check(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt, + unsigned idx, unsigned reg) +{ + int r; + + switch (reg) { + case EVERGREEN_VLINE_START_END: + r = evergreen_cs_packet_parse_vline(p); + if (r) { + DRM_ERROR("No reloc for ib[%d]=0x%04X\n", + idx, reg); + return r; + } + break; + default: + printk(KERN_ERR "Forbidden register 0x%04X in cs at %d\n", + reg, idx); + return -EINVAL; + } + return 0; +} + +static int evergreen_cs_parse_packet0(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt) +{ + unsigned reg, i; + unsigned idx; + int r; + + idx = pkt->idx + 1; + reg = pkt->reg; + for (i = 0; i <= pkt->count; i++, idx++, reg += 4) { + r = evergreen_packet0_check(p, pkt, idx, reg); + if (r) { + return r; + } + } + return 0; +} + +/** + * evergreen_cs_check_reg() - check if register is authorized or not + * @parser: parser structure holding parsing context + * @reg: register we are testing + * @idx: index into the cs buffer + * + * This function will test against evergreen_reg_safe_bm and return 0 + * if register is safe. If register is not flag as safe this function + * will test it against a list of register needind special handling. + */ +static inline int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) +{ + struct evergreen_cs_track *track = (struct evergreen_cs_track *)p->track; + struct radeon_cs_reloc *reloc; + u32 last_reg = ARRAY_SIZE(evergreen_reg_safe_bm); + u32 m, i, tmp, *ib; + int r; + + i = (reg >> 7); + if (i > last_reg) { + dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); + return -EINVAL; + } + m = 1 << ((reg >> 2) & 31); + if (!(evergreen_reg_safe_bm[i] & m)) + return 0; + ib = p->ib->ptr; + switch (reg) { + /* force following reg to 0 in an attemp to disable out buffer + * which will need us to better understand how it works to perform + * security check on it (Jerome) + */ + case SQ_ESGS_RING_SIZE: + case SQ_GSVS_RING_SIZE: + case SQ_ESTMP_RING_SIZE: + case SQ_GSTMP_RING_SIZE: + case SQ_HSTMP_RING_SIZE: + case SQ_LSTMP_RING_SIZE: + case SQ_PSTMP_RING_SIZE: + case SQ_VSTMP_RING_SIZE: + case SQ_ESGS_RING_ITEMSIZE: + case SQ_ESTMP_RING_ITEMSIZE: + case SQ_GSTMP_RING_ITEMSIZE: + case SQ_GSVS_RING_ITEMSIZE: + case SQ_GS_VERT_ITEMSIZE: + case SQ_GS_VERT_ITEMSIZE_1: + case SQ_GS_VERT_ITEMSIZE_2: + case SQ_GS_VERT_ITEMSIZE_3: + case SQ_GSVS_RING_OFFSET_1: + case SQ_GSVS_RING_OFFSET_2: + case SQ_GSVS_RING_OFFSET_3: + case SQ_HSTMP_RING_ITEMSIZE: + case SQ_LSTMP_RING_ITEMSIZE: + case SQ_PSTMP_RING_ITEMSIZE: + case SQ_VSTMP_RING_ITEMSIZE: + case VGT_TF_RING_SIZE: + /* get value to populate the IB don't remove */ + tmp =radeon_get_ib_value(p, idx); + ib[idx] = 0; + break; + case DB_DEPTH_CONTROL: + track->db_depth_control = radeon_get_ib_value(p, idx); + break; + case DB_Z_INFO: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + track->db_z_info = radeon_get_ib_value(p, idx); + ib[idx] &= ~Z_ARRAY_MODE(0xf); + track->db_z_info &= ~Z_ARRAY_MODE(0xf); + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + ib[idx] |= Z_ARRAY_MODE(ARRAY_2D_TILED_THIN1); + track->db_z_info |= Z_ARRAY_MODE(ARRAY_2D_TILED_THIN1); + } else { + ib[idx] |= Z_ARRAY_MODE(ARRAY_1D_TILED_THIN1); + track->db_z_info |= Z_ARRAY_MODE(ARRAY_1D_TILED_THIN1); + } + break; + case DB_STENCIL_INFO: + track->db_s_info = radeon_get_ib_value(p, idx); + break; + case DB_DEPTH_VIEW: + track->db_depth_view = radeon_get_ib_value(p, idx); + break; + case DB_DEPTH_SIZE: + track->db_depth_size = radeon_get_ib_value(p, idx); + track->db_depth_size_idx = idx; + break; + case DB_Z_READ_BASE: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + track->db_z_read_offset = radeon_get_ib_value(p, idx); + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->db_z_read_bo = reloc->robj; + break; + case DB_Z_WRITE_BASE: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + track->db_z_write_offset = radeon_get_ib_value(p, idx); + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->db_z_write_bo = reloc->robj; + break; + case DB_STENCIL_READ_BASE: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + track->db_s_read_offset = radeon_get_ib_value(p, idx); + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->db_s_read_bo = reloc->robj; + break; + case DB_STENCIL_WRITE_BASE: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + track->db_s_write_offset = radeon_get_ib_value(p, idx); + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->db_s_write_bo = reloc->robj; + break; + case VGT_STRMOUT_CONFIG: + track->vgt_strmout_config = radeon_get_ib_value(p, idx); + break; + case VGT_STRMOUT_BUFFER_CONFIG: + track->vgt_strmout_buffer_config = radeon_get_ib_value(p, idx); + break; + case CB_TARGET_MASK: + track->cb_target_mask = radeon_get_ib_value(p, idx); + break; + case CB_SHADER_MASK: + track->cb_shader_mask = radeon_get_ib_value(p, idx); + break; + case PA_SC_AA_CONFIG: + tmp = radeon_get_ib_value(p, idx) & MSAA_NUM_SAMPLES_MASK; + track->nsamples = 1 << tmp; + break; + case CB_COLOR0_VIEW: + case CB_COLOR1_VIEW: + case CB_COLOR2_VIEW: + case CB_COLOR3_VIEW: + case CB_COLOR4_VIEW: + case CB_COLOR5_VIEW: + case CB_COLOR6_VIEW: + case CB_COLOR7_VIEW: + tmp = (reg - CB_COLOR0_VIEW) / 0x3c; + track->cb_color_view[tmp] = radeon_get_ib_value(p, idx); + break; + case CB_COLOR8_VIEW: + case CB_COLOR9_VIEW: + case CB_COLOR10_VIEW: + case CB_COLOR11_VIEW: + tmp = ((reg - CB_COLOR8_VIEW) / 0x1c) + 8; + track->cb_color_view[tmp] = radeon_get_ib_value(p, idx); + break; + case CB_COLOR0_INFO: + case CB_COLOR1_INFO: + case CB_COLOR2_INFO: + case CB_COLOR3_INFO: + case CB_COLOR4_INFO: + case CB_COLOR5_INFO: + case CB_COLOR6_INFO: + case CB_COLOR7_INFO: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + tmp = (reg - CB_COLOR0_INFO) / 0x3c; + track->cb_color_info[tmp] = radeon_get_ib_value(p, idx); + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + ib[idx] |= CB_ARRAY_MODE(ARRAY_2D_TILED_THIN1); + track->cb_color_info[tmp] |= CB_ARRAY_MODE(ARRAY_2D_TILED_THIN1); + } else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { + ib[idx] |= CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1); + track->cb_color_info[tmp] |= CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1); + } + break; + case CB_COLOR8_INFO: + case CB_COLOR9_INFO: + case CB_COLOR10_INFO: + case CB_COLOR11_INFO: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + tmp = ((reg - CB_COLOR8_INFO) / 0x1c) + 8; + track->cb_color_info[tmp] = radeon_get_ib_value(p, idx); + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + ib[idx] |= CB_ARRAY_MODE(ARRAY_2D_TILED_THIN1); + track->cb_color_info[tmp] |= CB_ARRAY_MODE(ARRAY_2D_TILED_THIN1); + } else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { + ib[idx] |= CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1); + track->cb_color_info[tmp] |= CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1); + } + break; + case CB_COLOR0_PITCH: + case CB_COLOR1_PITCH: + case CB_COLOR2_PITCH: + case CB_COLOR3_PITCH: + case CB_COLOR4_PITCH: + case CB_COLOR5_PITCH: + case CB_COLOR6_PITCH: + case CB_COLOR7_PITCH: + tmp = (reg - CB_COLOR0_PITCH) / 0x3c; + track->cb_color_pitch[tmp] = radeon_get_ib_value(p, idx); + track->cb_color_pitch_idx[tmp] = idx; + break; + case CB_COLOR8_PITCH: + case CB_COLOR9_PITCH: + case CB_COLOR10_PITCH: + case CB_COLOR11_PITCH: + tmp = ((reg - CB_COLOR8_PITCH) / 0x1c) + 8; + track->cb_color_pitch[tmp] = radeon_get_ib_value(p, idx); + track->cb_color_pitch_idx[tmp] = idx; + break; + case CB_COLOR0_SLICE: + case CB_COLOR1_SLICE: + case CB_COLOR2_SLICE: + case CB_COLOR3_SLICE: + case CB_COLOR4_SLICE: + case CB_COLOR5_SLICE: + case CB_COLOR6_SLICE: + case CB_COLOR7_SLICE: + tmp = (reg - CB_COLOR0_SLICE) / 0x3c; + track->cb_color_slice[tmp] = radeon_get_ib_value(p, idx); + track->cb_color_slice_idx[tmp] = idx; + break; + case CB_COLOR8_SLICE: + case CB_COLOR9_SLICE: + case CB_COLOR10_SLICE: + case CB_COLOR11_SLICE: + tmp = ((reg - CB_COLOR8_SLICE) / 0x1c) + 8; + track->cb_color_slice[tmp] = radeon_get_ib_value(p, idx); + track->cb_color_slice_idx[tmp] = idx; + break; + case CB_COLOR0_ATTRIB: + case CB_COLOR1_ATTRIB: + case CB_COLOR2_ATTRIB: + case CB_COLOR3_ATTRIB: + case CB_COLOR4_ATTRIB: + case CB_COLOR5_ATTRIB: + case CB_COLOR6_ATTRIB: + case CB_COLOR7_ATTRIB: + case CB_COLOR8_ATTRIB: + case CB_COLOR9_ATTRIB: + case CB_COLOR10_ATTRIB: + case CB_COLOR11_ATTRIB: + break; + case CB_COLOR0_DIM: + case CB_COLOR1_DIM: + case CB_COLOR2_DIM: + case CB_COLOR3_DIM: + case CB_COLOR4_DIM: + case CB_COLOR5_DIM: + case CB_COLOR6_DIM: + case CB_COLOR7_DIM: + tmp = (reg - CB_COLOR0_DIM) / 0x3c; + track->cb_color_dim[tmp] = radeon_get_ib_value(p, idx); + track->cb_color_dim_idx[tmp] = idx; + break; + case CB_COLOR8_DIM: + case CB_COLOR9_DIM: + case CB_COLOR10_DIM: + case CB_COLOR11_DIM: + tmp = ((reg - CB_COLOR8_DIM) / 0x1c) + 8; + track->cb_color_dim[tmp] = radeon_get_ib_value(p, idx); + track->cb_color_dim_idx[tmp] = idx; + break; + case CB_COLOR0_FMASK: + case CB_COLOR1_FMASK: + case CB_COLOR2_FMASK: + case CB_COLOR3_FMASK: + case CB_COLOR4_FMASK: + case CB_COLOR5_FMASK: + case CB_COLOR6_FMASK: + case CB_COLOR7_FMASK: + tmp = (reg - CB_COLOR0_FMASK) / 0x3c; + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg); + return -EINVAL; + } + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->cb_color_fmask_bo[tmp] = reloc->robj; + break; + case CB_COLOR0_CMASK: + case CB_COLOR1_CMASK: + case CB_COLOR2_CMASK: + case CB_COLOR3_CMASK: + case CB_COLOR4_CMASK: + case CB_COLOR5_CMASK: + case CB_COLOR6_CMASK: + case CB_COLOR7_CMASK: + tmp = (reg - CB_COLOR0_CMASK) / 0x3c; + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg); + return -EINVAL; + } + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->cb_color_cmask_bo[tmp] = reloc->robj; + break; + case CB_COLOR0_FMASK_SLICE: + case CB_COLOR1_FMASK_SLICE: + case CB_COLOR2_FMASK_SLICE: + case CB_COLOR3_FMASK_SLICE: + case CB_COLOR4_FMASK_SLICE: + case CB_COLOR5_FMASK_SLICE: + case CB_COLOR6_FMASK_SLICE: + case CB_COLOR7_FMASK_SLICE: + tmp = (reg - CB_COLOR0_FMASK_SLICE) / 0x3c; + track->cb_color_fmask_slice[tmp] = radeon_get_ib_value(p, idx); + break; + case CB_COLOR0_CMASK_SLICE: + case CB_COLOR1_CMASK_SLICE: + case CB_COLOR2_CMASK_SLICE: + case CB_COLOR3_CMASK_SLICE: + case CB_COLOR4_CMASK_SLICE: + case CB_COLOR5_CMASK_SLICE: + case CB_COLOR6_CMASK_SLICE: + case CB_COLOR7_CMASK_SLICE: + tmp = (reg - CB_COLOR0_CMASK_SLICE) / 0x3c; + track->cb_color_cmask_slice[tmp] = radeon_get_ib_value(p, idx); + break; + case CB_COLOR0_BASE: + case CB_COLOR1_BASE: + case CB_COLOR2_BASE: + case CB_COLOR3_BASE: + case CB_COLOR4_BASE: + case CB_COLOR5_BASE: + case CB_COLOR6_BASE: + case CB_COLOR7_BASE: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + tmp = (reg - CB_COLOR0_BASE) / 0x3c; + track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx); + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->cb_color_base_last[tmp] = ib[idx]; + track->cb_color_bo[tmp] = reloc->robj; + break; + case CB_COLOR8_BASE: + case CB_COLOR9_BASE: + case CB_COLOR10_BASE: + case CB_COLOR11_BASE: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + tmp = ((reg - CB_COLOR8_BASE) / 0x1c) + 8; + track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx); + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->cb_color_base_last[tmp] = ib[idx]; + track->cb_color_bo[tmp] = reloc->robj; + break; + case CB_IMMED0_BASE: + case CB_IMMED1_BASE: + case CB_IMMED2_BASE: + case CB_IMMED3_BASE: + case CB_IMMED4_BASE: + case CB_IMMED5_BASE: + case CB_IMMED6_BASE: + case CB_IMMED7_BASE: + case CB_IMMED8_BASE: + case CB_IMMED9_BASE: + case CB_IMMED10_BASE: + case CB_IMMED11_BASE: + case DB_HTILE_DATA_BASE: + case SQ_PGM_START_FS: + case SQ_PGM_START_ES: + case SQ_PGM_START_VS: + case SQ_PGM_START_GS: + case SQ_PGM_START_PS: + case SQ_PGM_START_HS: + case SQ_PGM_START_LS: + case GDS_ADDR_BASE: + case SQ_CONST_MEM_BASE: + case SQ_ALU_CONST_CACHE_GS_0: + case SQ_ALU_CONST_CACHE_GS_1: + case SQ_ALU_CONST_CACHE_GS_2: + case SQ_ALU_CONST_CACHE_GS_3: + case SQ_ALU_CONST_CACHE_GS_4: + case SQ_ALU_CONST_CACHE_GS_5: + case SQ_ALU_CONST_CACHE_GS_6: + case SQ_ALU_CONST_CACHE_GS_7: + case SQ_ALU_CONST_CACHE_GS_8: + case SQ_ALU_CONST_CACHE_GS_9: + case SQ_ALU_CONST_CACHE_GS_10: + case SQ_ALU_CONST_CACHE_GS_11: + case SQ_ALU_CONST_CACHE_GS_12: + case SQ_ALU_CONST_CACHE_GS_13: + case SQ_ALU_CONST_CACHE_GS_14: + case SQ_ALU_CONST_CACHE_GS_15: + case SQ_ALU_CONST_CACHE_PS_0: + case SQ_ALU_CONST_CACHE_PS_1: + case SQ_ALU_CONST_CACHE_PS_2: + case SQ_ALU_CONST_CACHE_PS_3: + case SQ_ALU_CONST_CACHE_PS_4: + case SQ_ALU_CONST_CACHE_PS_5: + case SQ_ALU_CONST_CACHE_PS_6: + case SQ_ALU_CONST_CACHE_PS_7: + case SQ_ALU_CONST_CACHE_PS_8: + case SQ_ALU_CONST_CACHE_PS_9: + case SQ_ALU_CONST_CACHE_PS_10: + case SQ_ALU_CONST_CACHE_PS_11: + case SQ_ALU_CONST_CACHE_PS_12: + case SQ_ALU_CONST_CACHE_PS_13: + case SQ_ALU_CONST_CACHE_PS_14: + case SQ_ALU_CONST_CACHE_PS_15: + case SQ_ALU_CONST_CACHE_VS_0: + case SQ_ALU_CONST_CACHE_VS_1: + case SQ_ALU_CONST_CACHE_VS_2: + case SQ_ALU_CONST_CACHE_VS_3: + case SQ_ALU_CONST_CACHE_VS_4: + case SQ_ALU_CONST_CACHE_VS_5: + case SQ_ALU_CONST_CACHE_VS_6: + case SQ_ALU_CONST_CACHE_VS_7: + case SQ_ALU_CONST_CACHE_VS_8: + case SQ_ALU_CONST_CACHE_VS_9: + case SQ_ALU_CONST_CACHE_VS_10: + case SQ_ALU_CONST_CACHE_VS_11: + case SQ_ALU_CONST_CACHE_VS_12: + case SQ_ALU_CONST_CACHE_VS_13: + case SQ_ALU_CONST_CACHE_VS_14: + case SQ_ALU_CONST_CACHE_VS_15: + case SQ_ALU_CONST_CACHE_HS_0: + case SQ_ALU_CONST_CACHE_HS_1: + case SQ_ALU_CONST_CACHE_HS_2: + case SQ_ALU_CONST_CACHE_HS_3: + case SQ_ALU_CONST_CACHE_HS_4: + case SQ_ALU_CONST_CACHE_HS_5: + case SQ_ALU_CONST_CACHE_HS_6: + case SQ_ALU_CONST_CACHE_HS_7: + case SQ_ALU_CONST_CACHE_HS_8: + case SQ_ALU_CONST_CACHE_HS_9: + case SQ_ALU_CONST_CACHE_HS_10: + case SQ_ALU_CONST_CACHE_HS_11: + case SQ_ALU_CONST_CACHE_HS_12: + case SQ_ALU_CONST_CACHE_HS_13: + case SQ_ALU_CONST_CACHE_HS_14: + case SQ_ALU_CONST_CACHE_HS_15: + case SQ_ALU_CONST_CACHE_LS_0: + case SQ_ALU_CONST_CACHE_LS_1: + case SQ_ALU_CONST_CACHE_LS_2: + case SQ_ALU_CONST_CACHE_LS_3: + case SQ_ALU_CONST_CACHE_LS_4: + case SQ_ALU_CONST_CACHE_LS_5: + case SQ_ALU_CONST_CACHE_LS_6: + case SQ_ALU_CONST_CACHE_LS_7: + case SQ_ALU_CONST_CACHE_LS_8: + case SQ_ALU_CONST_CACHE_LS_9: + case SQ_ALU_CONST_CACHE_LS_10: + case SQ_ALU_CONST_CACHE_LS_11: + case SQ_ALU_CONST_CACHE_LS_12: + case SQ_ALU_CONST_CACHE_LS_13: + case SQ_ALU_CONST_CACHE_LS_14: + case SQ_ALU_CONST_CACHE_LS_15: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + break; + default: + dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); + return -EINVAL; + } + return 0; +} + +/** + * evergreen_check_texture_resource() - check if register is authorized or not + * @p: parser structure holding parsing context + * @idx: index into the cs buffer + * @texture: texture's bo structure + * @mipmap: mipmap's bo structure + * + * This function will check that the resource has valid field and that + * the texture and mipmap bo object are big enough to cover this resource. + */ +static inline int evergreen_check_texture_resource(struct radeon_cs_parser *p, u32 idx, + struct radeon_bo *texture, + struct radeon_bo *mipmap) +{ + /* XXX fill in */ + return 0; +} + +static int evergreen_packet3_check(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt) +{ + struct radeon_cs_reloc *reloc; + struct evergreen_cs_track *track; + volatile u32 *ib; + unsigned idx; + unsigned i; + unsigned start_reg, end_reg, reg; + int r; + u32 idx_value; + + track = (struct evergreen_cs_track *)p->track; + ib = p->ib->ptr; + idx = pkt->idx + 1; + idx_value = radeon_get_ib_value(p, idx); + + switch (pkt->opcode) { + case PACKET3_CONTEXT_CONTROL: + if (pkt->count != 1) { + DRM_ERROR("bad CONTEXT_CONTROL\n"); + return -EINVAL; + } + break; + case PACKET3_INDEX_TYPE: + case PACKET3_NUM_INSTANCES: + case PACKET3_CLEAR_STATE: + if (pkt->count) { + DRM_ERROR("bad INDEX_TYPE/NUM_INSTANCES/CLEAR_STATE\n"); + return -EINVAL; + } + break; + case PACKET3_INDEX_BASE: + if (pkt->count != 1) { + DRM_ERROR("bad INDEX_BASE\n"); + return -EINVAL; + } + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad INDEX_BASE\n"); + return -EINVAL; + } + ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); + ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + r = evergreen_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); + return r; + } + break; + case PACKET3_DRAW_INDEX: + if (pkt->count != 3) { + DRM_ERROR("bad DRAW_INDEX\n"); + return -EINVAL; + } + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad DRAW_INDEX\n"); + return -EINVAL; + } + ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); + ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + r = evergreen_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); + return r; + } + break; + case PACKET3_DRAW_INDEX_2: + if (pkt->count != 4) { + DRM_ERROR("bad DRAW_INDEX_2\n"); + return -EINVAL; + } + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad DRAW_INDEX_2\n"); + return -EINVAL; + } + ib[idx+1] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); + ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + r = evergreen_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); + return r; + } + break; + case PACKET3_DRAW_INDEX_AUTO: + if (pkt->count != 1) { + DRM_ERROR("bad DRAW_INDEX_AUTO\n"); + return -EINVAL; + } + r = evergreen_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream %d\n", __func__, __LINE__, idx); + return r; + } + break; + case PACKET3_DRAW_INDEX_MULTI_AUTO: + if (pkt->count != 2) { + DRM_ERROR("bad DRAW_INDEX_MULTI_AUTO\n"); + return -EINVAL; + } + r = evergreen_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream %d\n", __func__, __LINE__, idx); + return r; + } + break; + case PACKET3_DRAW_INDEX_IMMD: + if (pkt->count < 2) { + DRM_ERROR("bad DRAW_INDEX_IMMD\n"); + return -EINVAL; + } + r = evergreen_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); + return r; + } + break; + case PACKET3_DRAW_INDEX_OFFSET: + if (pkt->count != 2) { + DRM_ERROR("bad DRAW_INDEX_OFFSET\n"); + return -EINVAL; + } + r = evergreen_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); + return r; + } + break; + case PACKET3_DRAW_INDEX_OFFSET_2: + if (pkt->count != 3) { + DRM_ERROR("bad DRAW_INDEX_OFFSET_2\n"); + return -EINVAL; + } + r = evergreen_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); + return r; + } + break; + case PACKET3_WAIT_REG_MEM: + if (pkt->count != 5) { + DRM_ERROR("bad WAIT_REG_MEM\n"); + return -EINVAL; + } + /* bit 4 is reg (0) or mem (1) */ + if (idx_value & 0x10) { + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad WAIT_REG_MEM\n"); + return -EINVAL; + } + ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); + ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + } + break; + case PACKET3_SURFACE_SYNC: + if (pkt->count != 3) { + DRM_ERROR("bad SURFACE_SYNC\n"); + return -EINVAL; + } + /* 0xffffffff/0x0 is flush all cache flag */ + if (radeon_get_ib_value(p, idx + 1) != 0xffffffff || + radeon_get_ib_value(p, idx + 2) != 0) { + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad SURFACE_SYNC\n"); + return -EINVAL; + } + ib[idx+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + } + break; + case PACKET3_EVENT_WRITE: + if (pkt->count != 2 && pkt->count != 0) { + DRM_ERROR("bad EVENT_WRITE\n"); + return -EINVAL; + } + if (pkt->count) { + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad EVENT_WRITE\n"); + return -EINVAL; + } + ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); + ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + } + break; + case PACKET3_EVENT_WRITE_EOP: + if (pkt->count != 4) { + DRM_ERROR("bad EVENT_WRITE_EOP\n"); + return -EINVAL; + } + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad EVENT_WRITE_EOP\n"); + return -EINVAL; + } + ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); + ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + break; + case PACKET3_EVENT_WRITE_EOS: + if (pkt->count != 3) { + DRM_ERROR("bad EVENT_WRITE_EOS\n"); + return -EINVAL; + } + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad EVENT_WRITE_EOS\n"); + return -EINVAL; + } + ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); + ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + break; + case PACKET3_SET_CONFIG_REG: + start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_START; + end_reg = 4 * pkt->count + start_reg - 4; + if ((start_reg < PACKET3_SET_CONFIG_REG_START) || + (start_reg >= PACKET3_SET_CONFIG_REG_END) || + (end_reg >= PACKET3_SET_CONFIG_REG_END)) { + DRM_ERROR("bad PACKET3_SET_CONFIG_REG\n"); + return -EINVAL; + } + for (i = 0; i < pkt->count; i++) { + reg = start_reg + (4 * i); + r = evergreen_cs_check_reg(p, reg, idx+1+i); + if (r) + return r; + } + break; + case PACKET3_SET_CONTEXT_REG: + start_reg = (idx_value << 2) + PACKET3_SET_CONTEXT_REG_START; + end_reg = 4 * pkt->count + start_reg - 4; + if ((start_reg < PACKET3_SET_CONTEXT_REG_START) || + (start_reg >= PACKET3_SET_CONTEXT_REG_END) || + (end_reg >= PACKET3_SET_CONTEXT_REG_END)) { + DRM_ERROR("bad PACKET3_SET_CONTEXT_REG\n"); + return -EINVAL; + } + for (i = 0; i < pkt->count; i++) { + reg = start_reg + (4 * i); + r = evergreen_cs_check_reg(p, reg, idx+1+i); + if (r) + return r; + } + break; + case PACKET3_SET_RESOURCE: + if (pkt->count % 8) { + DRM_ERROR("bad SET_RESOURCE\n"); + return -EINVAL; + } + start_reg = (idx_value << 2) + PACKET3_SET_RESOURCE_START; + end_reg = 4 * pkt->count + start_reg - 4; + if ((start_reg < PACKET3_SET_RESOURCE_START) || + (start_reg >= PACKET3_SET_RESOURCE_END) || + (end_reg >= PACKET3_SET_RESOURCE_END)) { + DRM_ERROR("bad SET_RESOURCE\n"); + return -EINVAL; + } + for (i = 0; i < (pkt->count / 8); i++) { + struct radeon_bo *texture, *mipmap; + u32 size, offset; + + switch (G__SQ_CONSTANT_TYPE(radeon_get_ib_value(p, idx+1+(i*8)+7))) { + case SQ_TEX_VTX_VALID_TEXTURE: + /* tex base */ + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad SET_RESOURCE (tex)\n"); + return -EINVAL; + } + ib[idx+1+(i*8)+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + ib[idx+1+(i*8)+1] |= TEX_ARRAY_MODE(ARRAY_2D_TILED_THIN1); + else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + ib[idx+1+(i*8)+1] |= TEX_ARRAY_MODE(ARRAY_1D_TILED_THIN1); + texture = reloc->robj; + /* tex mip base */ + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad SET_RESOURCE (tex)\n"); + return -EINVAL; + } + ib[idx+1+(i*8)+3] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + mipmap = reloc->robj; + r = evergreen_check_texture_resource(p, idx+1+(i*8), + texture, mipmap); + if (r) + return r; + break; + case SQ_TEX_VTX_VALID_BUFFER: + /* vtx base */ + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad SET_RESOURCE (vtx)\n"); + return -EINVAL; + } + offset = radeon_get_ib_value(p, idx+1+(i*8)+0); + size = radeon_get_ib_value(p, idx+1+(i*8)+1); + if (p->rdev && (size + offset) > radeon_bo_size(reloc->robj)) { + /* force size to size of the buffer */ + dev_warn(p->dev, "vbo resource seems too big for the bo\n"); + ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj); + } + ib[idx+1+(i*8)+0] += (u32)((reloc->lobj.gpu_offset) & 0xffffffff); + ib[idx+1+(i*8)+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + break; + case SQ_TEX_VTX_INVALID_TEXTURE: + case SQ_TEX_VTX_INVALID_BUFFER: + default: + DRM_ERROR("bad SET_RESOURCE\n"); + return -EINVAL; + } + } + break; + case PACKET3_SET_ALU_CONST: + /* XXX fix me ALU const buffers only */ + break; + case PACKET3_SET_BOOL_CONST: + start_reg = (idx_value << 2) + PACKET3_SET_BOOL_CONST_START; + end_reg = 4 * pkt->count + start_reg - 4; + if ((start_reg < PACKET3_SET_BOOL_CONST_START) || + (start_reg >= PACKET3_SET_BOOL_CONST_END) || + (end_reg >= PACKET3_SET_BOOL_CONST_END)) { + DRM_ERROR("bad SET_BOOL_CONST\n"); + return -EINVAL; + } + break; + case PACKET3_SET_LOOP_CONST: + start_reg = (idx_value << 2) + PACKET3_SET_LOOP_CONST_START; + end_reg = 4 * pkt->count + start_reg - 4; + if ((start_reg < PACKET3_SET_LOOP_CONST_START) || + (start_reg >= PACKET3_SET_LOOP_CONST_END) || + (end_reg >= PACKET3_SET_LOOP_CONST_END)) { + DRM_ERROR("bad SET_LOOP_CONST\n"); + return -EINVAL; + } + break; + case PACKET3_SET_CTL_CONST: + start_reg = (idx_value << 2) + PACKET3_SET_CTL_CONST_START; + end_reg = 4 * pkt->count + start_reg - 4; + if ((start_reg < PACKET3_SET_CTL_CONST_START) || + (start_reg >= PACKET3_SET_CTL_CONST_END) || + (end_reg >= PACKET3_SET_CTL_CONST_END)) { + DRM_ERROR("bad SET_CTL_CONST\n"); + return -EINVAL; + } + break; + case PACKET3_SET_SAMPLER: + if (pkt->count % 3) { + DRM_ERROR("bad SET_SAMPLER\n"); + return -EINVAL; + } + start_reg = (idx_value << 2) + PACKET3_SET_SAMPLER_START; + end_reg = 4 * pkt->count + start_reg - 4; + if ((start_reg < PACKET3_SET_SAMPLER_START) || + (start_reg >= PACKET3_SET_SAMPLER_END) || + (end_reg >= PACKET3_SET_SAMPLER_END)) { + DRM_ERROR("bad SET_SAMPLER\n"); + return -EINVAL; + } + break; + case PACKET3_NOP: + break; + default: + DRM_ERROR("Packet3 opcode %x not supported\n", pkt->opcode); + return -EINVAL; + } + return 0; +} + +int evergreen_cs_parse(struct radeon_cs_parser *p) +{ + struct radeon_cs_packet pkt; + struct evergreen_cs_track *track; + int r; + + if (p->track == NULL) { + /* initialize tracker, we are in kms */ + track = kzalloc(sizeof(*track), GFP_KERNEL); + if (track == NULL) + return -ENOMEM; + evergreen_cs_track_init(track); + track->npipes = p->rdev->config.evergreen.tiling_npipes; + track->nbanks = p->rdev->config.evergreen.tiling_nbanks; + track->group_size = p->rdev->config.evergreen.tiling_group_size; + p->track = track; + } + do { + r = evergreen_cs_packet_parse(p, &pkt, p->idx); + if (r) { + kfree(p->track); + p->track = NULL; + return r; + } + p->idx += pkt.count + 2; + switch (pkt.type) { + case PACKET_TYPE0: + r = evergreen_cs_parse_packet0(p, &pkt); + break; + case PACKET_TYPE2: + break; + case PACKET_TYPE3: + r = evergreen_packet3_check(p, &pkt); + break; + default: + DRM_ERROR("Unknown packet type %d !\n", pkt.type); + kfree(p->track); + p->track = NULL; + return -EINVAL; + } + if (r) { + kfree(p->track); + p->track = NULL; + return r; + } + } while (p->idx < p->chunks[p->chunk_ib_idx].length_dw); +#if 0 + for (r = 0; r < p->ib->length_dw; r++) { + printk(KERN_INFO "%05d 0x%08X\n", r, p->ib->ptr[r]); + mdelay(1); + } +#endif + kfree(p->track); + p->track = NULL; + return 0; +} + diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h index af86af836f13..e028c1cd9d9b 100644 --- a/drivers/gpu/drm/radeon/evergreen_reg.h +++ b/drivers/gpu/drm/radeon/evergreen_reg.h @@ -151,6 +151,9 @@ #define EVERGREEN_DATA_FORMAT 0x6b00 # define EVERGREEN_INTERLEAVE_EN (1 << 0) #define EVERGREEN_DESKTOP_HEIGHT 0x6b04 +#define EVERGREEN_VLINE_START_END 0x6b08 +#define EVERGREEN_VLINE_STATUS 0x6bb8 +# define EVERGREEN_VLINE_STAT (1 << 12) #define EVERGREEN_VIEWPORT_START 0x6d70 #define EVERGREEN_VIEWPORT_SIZE 0x6d74 diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h index 93e9e17ad54a..a1cd621780e2 100644 --- a/drivers/gpu/drm/radeon/evergreend.h +++ b/drivers/gpu/drm/radeon/evergreend.h @@ -218,6 +218,8 @@ #define CLIP_VTX_REORDER_ENA (1 << 0) #define NUM_CLIP_SEQ(x) ((x) << 1) #define PA_SC_AA_CONFIG 0x28C04 +#define MSAA_NUM_SAMPLES_SHIFT 0 +#define MSAA_NUM_SAMPLES_MASK 0x3 #define PA_SC_CLIPRECT_RULE 0x2820C #define PA_SC_EDGERULE 0x28230 #define PA_SC_FIFO_SIZE 0x8BCC @@ -553,4 +555,469 @@ # define DC_HPDx_RX_INT_TIMER(x) ((x) << 16) # define DC_HPDx_EN (1 << 28) +/* + * PM4 + */ +#define PACKET_TYPE0 0 +#define PACKET_TYPE1 1 +#define PACKET_TYPE2 2 +#define PACKET_TYPE3 3 + +#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3) +#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF) +#define CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2) +#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF) +#define PACKET0(reg, n) ((PACKET_TYPE0 << 30) | \ + (((reg) >> 2) & 0xFFFF) | \ + ((n) & 0x3FFF) << 16) +#define CP_PACKET2 0x80000000 +#define PACKET2_PAD_SHIFT 0 +#define PACKET2_PAD_MASK (0x3fffffff << 0) + +#define PACKET2(v) (CP_PACKET2 | REG_SET(PACKET2_PAD, (v))) + +#define PACKET3(op, n) ((PACKET_TYPE3 << 30) | \ + (((op) & 0xFF) << 8) | \ + ((n) & 0x3FFF) << 16) + +/* Packet 3 types */ +#define PACKET3_NOP 0x10 +#define PACKET3_SET_BASE 0x11 +#define PACKET3_CLEAR_STATE 0x12 +#define PACKET3_INDIRECT_BUFFER_SIZE 0x13 +#define PACKET3_DISPATCH_DIRECT 0x15 +#define PACKET3_DISPATCH_INDIRECT 0x16 +#define PACKET3_INDIRECT_BUFFER_END 0x17 +#define PACKET3_SET_PREDICATION 0x20 +#define PACKET3_REG_RMW 0x21 +#define PACKET3_COND_EXEC 0x22 +#define PACKET3_PRED_EXEC 0x23 +#define PACKET3_DRAW_INDIRECT 0x24 +#define PACKET3_DRAW_INDEX_INDIRECT 0x25 +#define PACKET3_INDEX_BASE 0x26 +#define PACKET3_DRAW_INDEX_2 0x27 +#define PACKET3_CONTEXT_CONTROL 0x28 +#define PACKET3_DRAW_INDEX_OFFSET 0x29 +#define PACKET3_INDEX_TYPE 0x2A +#define PACKET3_DRAW_INDEX 0x2B +#define PACKET3_DRAW_INDEX_AUTO 0x2D +#define PACKET3_DRAW_INDEX_IMMD 0x2E +#define PACKET3_NUM_INSTANCES 0x2F +#define PACKET3_DRAW_INDEX_MULTI_AUTO 0x30 +#define PACKET3_STRMOUT_BUFFER_UPDATE 0x34 +#define PACKET3_DRAW_INDEX_OFFSET_2 0x35 +#define PACKET3_DRAW_INDEX_MULTI_ELEMENT 0x36 +#define PACKET3_MEM_SEMAPHORE 0x39 +#define PACKET3_MPEG_INDEX 0x3A +#define PACKET3_WAIT_REG_MEM 0x3C +#define PACKET3_MEM_WRITE 0x3D +#define PACKET3_INDIRECT_BUFFER 0x32 +#define PACKET3_SURFACE_SYNC 0x43 +# define PACKET3_CB0_DEST_BASE_ENA (1 << 6) +# define PACKET3_CB1_DEST_BASE_ENA (1 << 7) +# define PACKET3_CB2_DEST_BASE_ENA (1 << 8) +# define PACKET3_CB3_DEST_BASE_ENA (1 << 9) +# define PACKET3_CB4_DEST_BASE_ENA (1 << 10) +# define PACKET3_CB5_DEST_BASE_ENA (1 << 11) +# define PACKET3_CB6_DEST_BASE_ENA (1 << 12) +# define PACKET3_CB7_DEST_BASE_ENA (1 << 13) +# define PACKET3_DB_DEST_BASE_ENA (1 << 14) +# define PACKET3_CB8_DEST_BASE_ENA (1 << 15) +# define PACKET3_CB9_DEST_BASE_ENA (1 << 16) +# define PACKET3_CB10_DEST_BASE_ENA (1 << 17) +# define PACKET3_CB11_DEST_BASE_ENA (1 << 17) +# define PACKET3_FULL_CACHE_ENA (1 << 20) +# define PACKET3_TC_ACTION_ENA (1 << 23) +# define PACKET3_VC_ACTION_ENA (1 << 24) +# define PACKET3_CB_ACTION_ENA (1 << 25) +# define PACKET3_DB_ACTION_ENA (1 << 26) +# define PACKET3_SH_ACTION_ENA (1 << 27) +# define PACKET3_SMX_ACTION_ENA (1 << 28) +#define PACKET3_ME_INITIALIZE 0x44 +#define PACKET3_ME_INITIALIZE_DEVICE_ID(x) ((x) << 16) +#define PACKET3_COND_WRITE 0x45 +#define PACKET3_EVENT_WRITE 0x46 +#define PACKET3_EVENT_WRITE_EOP 0x47 +#define PACKET3_EVENT_WRITE_EOS 0x48 +#define PACKET3_PREAMBLE_CNTL 0x4A +#define PACKET3_RB_OFFSET 0x4B +#define PACKET3_ALU_PS_CONST_BUFFER_COPY 0x4C +#define PACKET3_ALU_VS_CONST_BUFFER_COPY 0x4D +#define PACKET3_ALU_PS_CONST_UPDATE 0x4E +#define PACKET3_ALU_VS_CONST_UPDATE 0x4F +#define PACKET3_ONE_REG_WRITE 0x57 +#define PACKET3_SET_CONFIG_REG 0x68 +#define PACKET3_SET_CONFIG_REG_START 0x00008000 +#define PACKET3_SET_CONFIG_REG_END 0x0000ac00 +#define PACKET3_SET_CONTEXT_REG 0x69 +#define PACKET3_SET_CONTEXT_REG_START 0x00028000 +#define PACKET3_SET_CONTEXT_REG_END 0x00029000 +#define PACKET3_SET_ALU_CONST 0x6A +/* alu const buffers only; no reg file */ +#define PACKET3_SET_BOOL_CONST 0x6B +#define PACKET3_SET_BOOL_CONST_START 0x0003a500 +#define PACKET3_SET_BOOL_CONST_END 0x0003a518 +#define PACKET3_SET_LOOP_CONST 0x6C +#define PACKET3_SET_LOOP_CONST_START 0x0003a200 +#define PACKET3_SET_LOOP_CONST_END 0x0003a500 +#define PACKET3_SET_RESOURCE 0x6D +#define PACKET3_SET_RESOURCE_START 0x00030000 +#define PACKET3_SET_RESOURCE_END 0x00038000 +#define PACKET3_SET_SAMPLER 0x6E +#define PACKET3_SET_SAMPLER_START 0x0003c000 +#define PACKET3_SET_SAMPLER_END 0x0003c600 +#define PACKET3_SET_CTL_CONST 0x6F +#define PACKET3_SET_CTL_CONST_START 0x0003cff0 +#define PACKET3_SET_CTL_CONST_END 0x0003ff0c +#define PACKET3_SET_RESOURCE_OFFSET 0x70 +#define PACKET3_SET_ALU_CONST_VS 0x71 +#define PACKET3_SET_ALU_CONST_DI 0x72 +#define PACKET3_SET_CONTEXT_REG_INDIRECT 0x73 +#define PACKET3_SET_RESOURCE_INDIRECT 0x74 +#define PACKET3_SET_APPEND_CNT 0x75 + +#define SQ_RESOURCE_CONSTANT_WORD7_0 0x3001c +#define S__SQ_CONSTANT_TYPE(x) (((x) & 3) << 30) +#define G__SQ_CONSTANT_TYPE(x) (((x) >> 30) & 3) +#define SQ_TEX_VTX_INVALID_TEXTURE 0x0 +#define SQ_TEX_VTX_INVALID_BUFFER 0x1 +#define SQ_TEX_VTX_VALID_TEXTURE 0x2 +#define SQ_TEX_VTX_VALID_BUFFER 0x3 + +#define SQ_CONST_MEM_BASE 0x8df8 + +#define SQ_ESGS_RING_SIZE 0x8c44 +#define SQ_GSVS_RING_SIZE 0x8c4c +#define SQ_ESTMP_RING_SIZE 0x8c54 +#define SQ_GSTMP_RING_SIZE 0x8c5c +#define SQ_VSTMP_RING_SIZE 0x8c64 +#define SQ_PSTMP_RING_SIZE 0x8c6c +#define SQ_LSTMP_RING_SIZE 0x8e14 +#define SQ_HSTMP_RING_SIZE 0x8e1c +#define VGT_TF_RING_SIZE 0x8988 + +#define SQ_ESGS_RING_ITEMSIZE 0x28900 +#define SQ_GSVS_RING_ITEMSIZE 0x28904 +#define SQ_ESTMP_RING_ITEMSIZE 0x28908 +#define SQ_GSTMP_RING_ITEMSIZE 0x2890c +#define SQ_VSTMP_RING_ITEMSIZE 0x28910 +#define SQ_PSTMP_RING_ITEMSIZE 0x28914 +#define SQ_LSTMP_RING_ITEMSIZE 0x28830 +#define SQ_HSTMP_RING_ITEMSIZE 0x28834 + +#define SQ_GS_VERT_ITEMSIZE 0x2891c +#define SQ_GS_VERT_ITEMSIZE_1 0x28920 +#define SQ_GS_VERT_ITEMSIZE_2 0x28924 +#define SQ_GS_VERT_ITEMSIZE_3 0x28928 +#define SQ_GSVS_RING_OFFSET_1 0x2892c +#define SQ_GSVS_RING_OFFSET_2 0x28930 +#define SQ_GSVS_RING_OFFSET_3 0x28934 + +#define SQ_ALU_CONST_BUFFER_SIZE_PS_0 0x28140 +#define SQ_ALU_CONST_BUFFER_SIZE_HS_0 0x28f80 + +#define SQ_ALU_CONST_CACHE_PS_0 0x28940 +#define SQ_ALU_CONST_CACHE_PS_1 0x28944 +#define SQ_ALU_CONST_CACHE_PS_2 0x28948 +#define SQ_ALU_CONST_CACHE_PS_3 0x2894c +#define SQ_ALU_CONST_CACHE_PS_4 0x28950 +#define SQ_ALU_CONST_CACHE_PS_5 0x28954 +#define SQ_ALU_CONST_CACHE_PS_6 0x28958 +#define SQ_ALU_CONST_CACHE_PS_7 0x2895c +#define SQ_ALU_CONST_CACHE_PS_8 0x28960 +#define SQ_ALU_CONST_CACHE_PS_9 0x28964 +#define SQ_ALU_CONST_CACHE_PS_10 0x28968 +#define SQ_ALU_CONST_CACHE_PS_11 0x2896c +#define SQ_ALU_CONST_CACHE_PS_12 0x28970 +#define SQ_ALU_CONST_CACHE_PS_13 0x28974 +#define SQ_ALU_CONST_CACHE_PS_14 0x28978 +#define SQ_ALU_CONST_CACHE_PS_15 0x2897c +#define SQ_ALU_CONST_CACHE_VS_0 0x28980 +#define SQ_ALU_CONST_CACHE_VS_1 0x28984 +#define SQ_ALU_CONST_CACHE_VS_2 0x28988 +#define SQ_ALU_CONST_CACHE_VS_3 0x2898c +#define SQ_ALU_CONST_CACHE_VS_4 0x28990 +#define SQ_ALU_CONST_CACHE_VS_5 0x28994 +#define SQ_ALU_CONST_CACHE_VS_6 0x28998 +#define SQ_ALU_CONST_CACHE_VS_7 0x2899c +#define SQ_ALU_CONST_CACHE_VS_8 0x289a0 +#define SQ_ALU_CONST_CACHE_VS_9 0x289a4 +#define SQ_ALU_CONST_CACHE_VS_10 0x289a8 +#define SQ_ALU_CONST_CACHE_VS_11 0x289ac +#define SQ_ALU_CONST_CACHE_VS_12 0x289b0 +#define SQ_ALU_CONST_CACHE_VS_13 0x289b4 +#define SQ_ALU_CONST_CACHE_VS_14 0x289b8 +#define SQ_ALU_CONST_CACHE_VS_15 0x289bc +#define SQ_ALU_CONST_CACHE_GS_0 0x289c0 +#define SQ_ALU_CONST_CACHE_GS_1 0x289c4 +#define SQ_ALU_CONST_CACHE_GS_2 0x289c8 +#define SQ_ALU_CONST_CACHE_GS_3 0x289cc +#define SQ_ALU_CONST_CACHE_GS_4 0x289d0 +#define SQ_ALU_CONST_CACHE_GS_5 0x289d4 +#define SQ_ALU_CONST_CACHE_GS_6 0x289d8 +#define SQ_ALU_CONST_CACHE_GS_7 0x289dc +#define SQ_ALU_CONST_CACHE_GS_8 0x289e0 +#define SQ_ALU_CONST_CACHE_GS_9 0x289e4 +#define SQ_ALU_CONST_CACHE_GS_10 0x289e8 +#define SQ_ALU_CONST_CACHE_GS_11 0x289ec +#define SQ_ALU_CONST_CACHE_GS_12 0x289f0 +#define SQ_ALU_CONST_CACHE_GS_13 0x289f4 +#define SQ_ALU_CONST_CACHE_GS_14 0x289f8 +#define SQ_ALU_CONST_CACHE_GS_15 0x289fc +#define SQ_ALU_CONST_CACHE_HS_0 0x28f00 +#define SQ_ALU_CONST_CACHE_HS_1 0x28f04 +#define SQ_ALU_CONST_CACHE_HS_2 0x28f08 +#define SQ_ALU_CONST_CACHE_HS_3 0x28f0c +#define SQ_ALU_CONST_CACHE_HS_4 0x28f10 +#define SQ_ALU_CONST_CACHE_HS_5 0x28f14 +#define SQ_ALU_CONST_CACHE_HS_6 0x28f18 +#define SQ_ALU_CONST_CACHE_HS_7 0x28f1c +#define SQ_ALU_CONST_CACHE_HS_8 0x28f20 +#define SQ_ALU_CONST_CACHE_HS_9 0x28f24 +#define SQ_ALU_CONST_CACHE_HS_10 0x28f28 +#define SQ_ALU_CONST_CACHE_HS_11 0x28f2c +#define SQ_ALU_CONST_CACHE_HS_12 0x28f30 +#define SQ_ALU_CONST_CACHE_HS_13 0x28f34 +#define SQ_ALU_CONST_CACHE_HS_14 0x28f38 +#define SQ_ALU_CONST_CACHE_HS_15 0x28f3c +#define SQ_ALU_CONST_CACHE_LS_0 0x28f40 +#define SQ_ALU_CONST_CACHE_LS_1 0x28f44 +#define SQ_ALU_CONST_CACHE_LS_2 0x28f48 +#define SQ_ALU_CONST_CACHE_LS_3 0x28f4c +#define SQ_ALU_CONST_CACHE_LS_4 0x28f50 +#define SQ_ALU_CONST_CACHE_LS_5 0x28f54 +#define SQ_ALU_CONST_CACHE_LS_6 0x28f58 +#define SQ_ALU_CONST_CACHE_LS_7 0x28f5c +#define SQ_ALU_CONST_CACHE_LS_8 0x28f60 +#define SQ_ALU_CONST_CACHE_LS_9 0x28f64 +#define SQ_ALU_CONST_CACHE_LS_10 0x28f68 +#define SQ_ALU_CONST_CACHE_LS_11 0x28f6c +#define SQ_ALU_CONST_CACHE_LS_12 0x28f70 +#define SQ_ALU_CONST_CACHE_LS_13 0x28f74 +#define SQ_ALU_CONST_CACHE_LS_14 0x28f78 +#define SQ_ALU_CONST_CACHE_LS_15 0x28f7c + +#define DB_DEPTH_CONTROL 0x28800 +#define DB_DEPTH_VIEW 0x28008 +#define DB_HTILE_DATA_BASE 0x28014 +#define DB_Z_INFO 0x28040 +# define Z_ARRAY_MODE(x) ((x) << 4) +#define DB_STENCIL_INFO 0x28044 +#define DB_Z_READ_BASE 0x28048 +#define DB_STENCIL_READ_BASE 0x2804c +#define DB_Z_WRITE_BASE 0x28050 +#define DB_STENCIL_WRITE_BASE 0x28054 +#define DB_DEPTH_SIZE 0x28058 + +#define SQ_PGM_START_PS 0x28840 +#define SQ_PGM_START_VS 0x2885c +#define SQ_PGM_START_GS 0x28874 +#define SQ_PGM_START_ES 0x2888c +#define SQ_PGM_START_FS 0x288a4 +#define SQ_PGM_START_HS 0x288b8 +#define SQ_PGM_START_LS 0x288d0 + +#define VGT_STRMOUT_CONFIG 0x28b94 +#define VGT_STRMOUT_BUFFER_CONFIG 0x28b98 + +#define CB_TARGET_MASK 0x28238 +#define CB_SHADER_MASK 0x2823c + +#define GDS_ADDR_BASE 0x28720 + +#define CB_IMMED0_BASE 0x28b9c +#define CB_IMMED1_BASE 0x28ba0 +#define CB_IMMED2_BASE 0x28ba4 +#define CB_IMMED3_BASE 0x28ba8 +#define CB_IMMED4_BASE 0x28bac +#define CB_IMMED5_BASE 0x28bb0 +#define CB_IMMED6_BASE 0x28bb4 +#define CB_IMMED7_BASE 0x28bb8 +#define CB_IMMED8_BASE 0x28bbc +#define CB_IMMED9_BASE 0x28bc0 +#define CB_IMMED10_BASE 0x28bc4 +#define CB_IMMED11_BASE 0x28bc8 + +/* all 12 CB blocks have these regs */ +#define CB_COLOR0_BASE 0x28c60 +#define CB_COLOR0_PITCH 0x28c64 +#define CB_COLOR0_SLICE 0x28c68 +#define CB_COLOR0_VIEW 0x28c6c +#define CB_COLOR0_INFO 0x28c70 +# define CB_ARRAY_MODE(x) ((x) << 8) +# define ARRAY_LINEAR_GENERAL 0 +# define ARRAY_LINEAR_ALIGNED 1 +# define ARRAY_1D_TILED_THIN1 2 +# define ARRAY_2D_TILED_THIN1 4 +#define CB_COLOR0_ATTRIB 0x28c74 +#define CB_COLOR0_DIM 0x28c78 +/* only CB0-7 blocks have these regs */ +#define CB_COLOR0_CMASK 0x28c7c +#define CB_COLOR0_CMASK_SLICE 0x28c80 +#define CB_COLOR0_FMASK 0x28c84 +#define CB_COLOR0_FMASK_SLICE 0x28c88 +#define CB_COLOR0_CLEAR_WORD0 0x28c8c +#define CB_COLOR0_CLEAR_WORD1 0x28c90 +#define CB_COLOR0_CLEAR_WORD2 0x28c94 +#define CB_COLOR0_CLEAR_WORD3 0x28c98 + +#define CB_COLOR1_BASE 0x28c9c +#define CB_COLOR2_BASE 0x28cd8 +#define CB_COLOR3_BASE 0x28d14 +#define CB_COLOR4_BASE 0x28d50 +#define CB_COLOR5_BASE 0x28d8c +#define CB_COLOR6_BASE 0x28dc8 +#define CB_COLOR7_BASE 0x28e04 +#define CB_COLOR8_BASE 0x28e40 +#define CB_COLOR9_BASE 0x28e5c +#define CB_COLOR10_BASE 0x28e78 +#define CB_COLOR11_BASE 0x28e94 + +#define CB_COLOR1_PITCH 0x28ca0 +#define CB_COLOR2_PITCH 0x28cdc +#define CB_COLOR3_PITCH 0x28d18 +#define CB_COLOR4_PITCH 0x28d54 +#define CB_COLOR5_PITCH 0x28d90 +#define CB_COLOR6_PITCH 0x28dcc +#define CB_COLOR7_PITCH 0x28e08 +#define CB_COLOR8_PITCH 0x28e44 +#define CB_COLOR9_PITCH 0x28e60 +#define CB_COLOR10_PITCH 0x28e7c +#define CB_COLOR11_PITCH 0x28e98 + +#define CB_COLOR1_SLICE 0x28ca4 +#define CB_COLOR2_SLICE 0x28ce0 +#define CB_COLOR3_SLICE 0x28d1c +#define CB_COLOR4_SLICE 0x28d58 +#define CB_COLOR5_SLICE 0x28d94 +#define CB_COLOR6_SLICE 0x28dd0 +#define CB_COLOR7_SLICE 0x28e0c +#define CB_COLOR8_SLICE 0x28e48 +#define CB_COLOR9_SLICE 0x28e64 +#define CB_COLOR10_SLICE 0x28e80 +#define CB_COLOR11_SLICE 0x28e9c + +#define CB_COLOR1_VIEW 0x28ca8 +#define CB_COLOR2_VIEW 0x28ce4 +#define CB_COLOR3_VIEW 0x28d20 +#define CB_COLOR4_VIEW 0x28d5c +#define CB_COLOR5_VIEW 0x28d98 +#define CB_COLOR6_VIEW 0x28dd4 +#define CB_COLOR7_VIEW 0x28e10 +#define CB_COLOR8_VIEW 0x28e4c +#define CB_COLOR9_VIEW 0x28e68 +#define CB_COLOR10_VIEW 0x28e84 +#define CB_COLOR11_VIEW 0x28ea0 + +#define CB_COLOR1_INFO 0x28cac +#define CB_COLOR2_INFO 0x28ce8 +#define CB_COLOR3_INFO 0x28d24 +#define CB_COLOR4_INFO 0x28d60 +#define CB_COLOR5_INFO 0x28d9c +#define CB_COLOR6_INFO 0x28dd8 +#define CB_COLOR7_INFO 0x28e14 +#define CB_COLOR8_INFO 0x28e50 +#define CB_COLOR9_INFO 0x28e6c +#define CB_COLOR10_INFO 0x28e88 +#define CB_COLOR11_INFO 0x28ea4 + +#define CB_COLOR1_ATTRIB 0x28cb0 +#define CB_COLOR2_ATTRIB 0x28cec +#define CB_COLOR3_ATTRIB 0x28d28 +#define CB_COLOR4_ATTRIB 0x28d64 +#define CB_COLOR5_ATTRIB 0x28da0 +#define CB_COLOR6_ATTRIB 0x28ddc +#define CB_COLOR7_ATTRIB 0x28e18 +#define CB_COLOR8_ATTRIB 0x28e54 +#define CB_COLOR9_ATTRIB 0x28e70 +#define CB_COLOR10_ATTRIB 0x28e8c +#define CB_COLOR11_ATTRIB 0x28ea8 + +#define CB_COLOR1_DIM 0x28cb4 +#define CB_COLOR2_DIM 0x28cf0 +#define CB_COLOR3_DIM 0x28d2c +#define CB_COLOR4_DIM 0x28d68 +#define CB_COLOR5_DIM 0x28da4 +#define CB_COLOR6_DIM 0x28de0 +#define CB_COLOR7_DIM 0x28e1c +#define CB_COLOR8_DIM 0x28e58 +#define CB_COLOR9_DIM 0x28e74 +#define CB_COLOR10_DIM 0x28e90 +#define CB_COLOR11_DIM 0x28eac + +#define CB_COLOR1_CMASK 0x28cb8 +#define CB_COLOR2_CMASK 0x28cf4 +#define CB_COLOR3_CMASK 0x28d30 +#define CB_COLOR4_CMASK 0x28d6c +#define CB_COLOR5_CMASK 0x28da8 +#define CB_COLOR6_CMASK 0x28de4 +#define CB_COLOR7_CMASK 0x28e20 + +#define CB_COLOR1_CMASK_SLICE 0x28cbc +#define CB_COLOR2_CMASK_SLICE 0x28cf8 +#define CB_COLOR3_CMASK_SLICE 0x28d34 +#define CB_COLOR4_CMASK_SLICE 0x28d70 +#define CB_COLOR5_CMASK_SLICE 0x28dac +#define CB_COLOR6_CMASK_SLICE 0x28de8 +#define CB_COLOR7_CMASK_SLICE 0x28e24 + +#define CB_COLOR1_FMASK 0x28cc0 +#define CB_COLOR2_FMASK 0x28cfc +#define CB_COLOR3_FMASK 0x28d38 +#define CB_COLOR4_FMASK 0x28d74 +#define CB_COLOR5_FMASK 0x28db0 +#define CB_COLOR6_FMASK 0x28dec +#define CB_COLOR7_FMASK 0x28e28 + +#define CB_COLOR1_FMASK_SLICE 0x28cc4 +#define CB_COLOR2_FMASK_SLICE 0x28d00 +#define CB_COLOR3_FMASK_SLICE 0x28d3c +#define CB_COLOR4_FMASK_SLICE 0x28d78 +#define CB_COLOR5_FMASK_SLICE 0x28db4 +#define CB_COLOR6_FMASK_SLICE 0x28df0 +#define CB_COLOR7_FMASK_SLICE 0x28e2c + +#define CB_COLOR1_CLEAR_WORD0 0x28cc8 +#define CB_COLOR2_CLEAR_WORD0 0x28d04 +#define CB_COLOR3_CLEAR_WORD0 0x28d40 +#define CB_COLOR4_CLEAR_WORD0 0x28d7c +#define CB_COLOR5_CLEAR_WORD0 0x28db8 +#define CB_COLOR6_CLEAR_WORD0 0x28df4 +#define CB_COLOR7_CLEAR_WORD0 0x28e30 + +#define CB_COLOR1_CLEAR_WORD1 0x28ccc +#define CB_COLOR2_CLEAR_WORD1 0x28d08 +#define CB_COLOR3_CLEAR_WORD1 0x28d44 +#define CB_COLOR4_CLEAR_WORD1 0x28d80 +#define CB_COLOR5_CLEAR_WORD1 0x28dbc +#define CB_COLOR6_CLEAR_WORD1 0x28df8 +#define CB_COLOR7_CLEAR_WORD1 0x28e34 + +#define CB_COLOR1_CLEAR_WORD2 0x28cd0 +#define CB_COLOR2_CLEAR_WORD2 0x28d0c +#define CB_COLOR3_CLEAR_WORD2 0x28d48 +#define CB_COLOR4_CLEAR_WORD2 0x28d84 +#define CB_COLOR5_CLEAR_WORD2 0x28dc0 +#define CB_COLOR6_CLEAR_WORD2 0x28dfc +#define CB_COLOR7_CLEAR_WORD2 0x28e38 + +#define CB_COLOR1_CLEAR_WORD3 0x28cd4 +#define CB_COLOR2_CLEAR_WORD3 0x28d10 +#define CB_COLOR3_CLEAR_WORD3 0x28d4c +#define CB_COLOR4_CLEAR_WORD3 0x28d88 +#define CB_COLOR5_CLEAR_WORD3 0x28dc4 +#define CB_COLOR6_CLEAR_WORD3 0x28e00 +#define CB_COLOR7_CLEAR_WORD3 0x28e3c + +#define SQ_TEX_RESOURCE_WORD0_0 0x30000 +#define SQ_TEX_RESOURCE_WORD1_0 0x30004 +# define TEX_ARRAY_MODE(x) ((x) << 28) +#define SQ_TEX_RESOURCE_WORD2_0 0x30008 +#define SQ_TEX_RESOURCE_WORD3_0 0x3000C +#define SQ_TEX_RESOURCE_WORD4_0 0x30010 +#define SQ_TEX_RESOURCE_WORD5_0 0x30014 +#define SQ_TEX_RESOURCE_WORD6_0 0x30018 +#define SQ_TEX_RESOURCE_WORD7_0 0x3001c + + #endif diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index c485c2cec4da..ab37717a5d39 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -162,6 +162,11 @@ void r100_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; + /* mid sh */ + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 0; /* high sh */ rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; @@ -172,6 +177,11 @@ void r100_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; + /* mid mh */ + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 0; /* high mh */ rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; @@ -1618,6 +1628,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, case RADEON_TXFORMAT_RGB332: case RADEON_TXFORMAT_Y8: track->textures[i].cpp = 1; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case RADEON_TXFORMAT_AI88: case RADEON_TXFORMAT_ARGB1555: @@ -1629,12 +1640,14 @@ static int r100_packet0_check(struct radeon_cs_parser *p, case RADEON_TXFORMAT_LDUDV655: case RADEON_TXFORMAT_DUDV88: track->textures[i].cpp = 2; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case RADEON_TXFORMAT_ARGB8888: case RADEON_TXFORMAT_RGBA8888: case RADEON_TXFORMAT_SHADOW32: case RADEON_TXFORMAT_LDUDUV8888: track->textures[i].cpp = 4; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case RADEON_TXFORMAT_DXT1: track->textures[i].cpp = 1; @@ -2594,12 +2607,6 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg, int surf_index = reg * 16; int flags = 0; - /* r100/r200 divide by 16 */ - if (rdev->family < CHIP_R300) - flags = pitch / 16; - else - flags = pitch / 8; - if (rdev->family <= CHIP_RS200) { if ((tiling_flags & (RADEON_TILING_MACRO|RADEON_TILING_MICRO)) == (RADEON_TILING_MACRO|RADEON_TILING_MICRO)) @@ -2623,6 +2630,20 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg, if (tiling_flags & RADEON_TILING_SWAP_32BIT) flags |= RADEON_SURF_AP0_SWP_32BPP | RADEON_SURF_AP1_SWP_32BPP; + /* when we aren't tiling the pitch seems to needs to be furtherdivided down. - tested on power5 + rn50 server */ + if (tiling_flags & (RADEON_TILING_SWAP_16BIT | RADEON_TILING_SWAP_32BIT)) { + if (!(tiling_flags & (RADEON_TILING_MACRO | RADEON_TILING_MICRO))) + if (ASIC_IS_RN50(rdev)) + pitch /= 16; + } + + /* r100/r200 divide by 16 */ + if (rdev->family < CHIP_R300) + flags |= pitch / 16; + else + flags |= pitch / 8; + + DRM_DEBUG("writing surface %d %d %x %x\n", reg, flags, offset, offset+obj_size-1); WREG32(RADEON_SURFACE0_INFO + surf_index, flags); WREG32(RADEON_SURFACE0_LOWER_BOUND + surf_index, offset); @@ -3137,33 +3158,6 @@ static inline void r100_cs_track_texture_print(struct r100_cs_track_texture *t) DRM_ERROR("compress format %d\n", t->compress_format); } -static int r100_cs_track_cube(struct radeon_device *rdev, - struct r100_cs_track *track, unsigned idx) -{ - unsigned face, w, h; - struct radeon_bo *cube_robj; - unsigned long size; - - for (face = 0; face < 5; face++) { - cube_robj = track->textures[idx].cube_info[face].robj; - w = track->textures[idx].cube_info[face].width; - h = track->textures[idx].cube_info[face].height; - - size = w * h; - size *= track->textures[idx].cpp; - - size += track->textures[idx].cube_info[face].offset; - - if (size > radeon_bo_size(cube_robj)) { - DRM_ERROR("Cube texture offset greater than object size %lu %lu\n", - size, radeon_bo_size(cube_robj)); - r100_cs_track_texture_print(&track->textures[idx]); - return -1; - } - } - return 0; -} - static int r100_track_compress_size(int compress_format, int w, int h) { int block_width, block_height, block_bytes; @@ -3194,6 +3188,37 @@ static int r100_track_compress_size(int compress_format, int w, int h) return sz; } +static int r100_cs_track_cube(struct radeon_device *rdev, + struct r100_cs_track *track, unsigned idx) +{ + unsigned face, w, h; + struct radeon_bo *cube_robj; + unsigned long size; + unsigned compress_format = track->textures[idx].compress_format; + + for (face = 0; face < 5; face++) { + cube_robj = track->textures[idx].cube_info[face].robj; + w = track->textures[idx].cube_info[face].width; + h = track->textures[idx].cube_info[face].height; + + if (compress_format) { + size = r100_track_compress_size(compress_format, w, h); + } else + size = w * h; + size *= track->textures[idx].cpp; + + size += track->textures[idx].cube_info[face].offset; + + if (size > radeon_bo_size(cube_robj)) { + DRM_ERROR("Cube texture offset greater than object size %lu %lu\n", + size, radeon_bo_size(cube_robj)); + r100_cs_track_texture_print(&track->textures[idx]); + return -1; + } + } + return 0; +} + static int r100_cs_track_texture_check(struct radeon_device *rdev, struct r100_cs_track *track) { diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c index 85617c311212..0266d72e0a4c 100644 --- a/drivers/gpu/drm/radeon/r200.c +++ b/drivers/gpu/drm/radeon/r200.c @@ -415,6 +415,8 @@ int r200_packet0_check(struct radeon_cs_parser *p, /* 2D, 3D, CUBE */ switch (tmp) { case 0: + case 3: + case 4: case 5: case 6: case 7: @@ -450,6 +452,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, case R200_TXFORMAT_RGB332: case R200_TXFORMAT_Y8: track->textures[i].cpp = 1; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case R200_TXFORMAT_AI88: case R200_TXFORMAT_ARGB1555: @@ -461,6 +464,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, case R200_TXFORMAT_DVDU88: case R200_TXFORMAT_AVYU4444: track->textures[i].cpp = 2; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case R200_TXFORMAT_ARGB8888: case R200_TXFORMAT_RGBA8888: @@ -468,6 +472,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, case R200_TXFORMAT_BGR111110: case R200_TXFORMAT_LDVDU8888: track->textures[i].cpp = 4; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case R200_TXFORMAT_DXT1: track->textures[i].cpp = 1; diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index b2f9efe2897c..7e81db5eb804 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -881,6 +881,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, case R300_TX_FORMAT_Y4X4: case R300_TX_FORMAT_Z3Y3X2: track->textures[i].cpp = 1; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case R300_TX_FORMAT_X16: case R300_TX_FORMAT_Y8X8: @@ -892,6 +893,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, case R300_TX_FORMAT_B8G8_B8G8: case R300_TX_FORMAT_G8R8_G8B8: track->textures[i].cpp = 2; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case R300_TX_FORMAT_Y16X16: case R300_TX_FORMAT_Z11Y11X10: @@ -902,14 +904,17 @@ static int r300_packet0_check(struct radeon_cs_parser *p, case R300_TX_FORMAT_FL_I32: case 0x1e: track->textures[i].cpp = 4; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case R300_TX_FORMAT_W16Z16Y16X16: case R300_TX_FORMAT_FL_R16G16B16A16: case R300_TX_FORMAT_FL_I32A32: track->textures[i].cpp = 8; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case R300_TX_FORMAT_FL_R32G32B32A32: track->textures[i].cpp = 16; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case R300_TX_FORMAT_DXT1: track->textures[i].cpp = 1; diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c index 4415a5ee5871..e6c89142bb4d 100644 --- a/drivers/gpu/drm/radeon/r420.c +++ b/drivers/gpu/drm/radeon/r420.c @@ -45,9 +45,14 @@ void r420_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_on_cm_idx = 0; /* low sh */ rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_ps_idx = 0; - rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = 1; + rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; + /* mid sh */ + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = 1; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 0; /* high sh */ rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; @@ -58,6 +63,11 @@ void r420_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; + /* mid mh */ + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 0; /* high mh */ rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index 4959619f8851..a73a6e17588d 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -130,9 +130,14 @@ void r600_pm_get_dynpm_state(struct radeon_device *rdev) break; } } - } else - rdev->pm.requested_power_state_index = - rdev->pm.current_power_state_index - 1; + } else { + if (rdev->pm.current_power_state_index == 0) + rdev->pm.requested_power_state_index = + rdev->pm.num_power_states - 1; + else + rdev->pm.requested_power_state_index = + rdev->pm.current_power_state_index - 1; + } } rdev->pm.requested_clock_mode_index = 0; /* don't use the power state if crtcs are active and no display flag is set */ @@ -291,6 +296,11 @@ void rs780_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; + /* mid sh */ + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 0; /* high sh */ rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = 1; @@ -301,6 +311,11 @@ void rs780_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; + /* mid mh */ + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 0; /* high mh */ rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = 1; @@ -317,6 +332,11 @@ void rs780_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = 1; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; + /* mid sh */ + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = 1; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = 1; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 0; /* high sh */ rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = 1; rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = 2; @@ -327,6 +347,11 @@ void rs780_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = 1; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; + /* mid mh */ + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = 1; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = 1; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 0; /* high mh */ rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = 1; rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = 2; @@ -343,6 +368,11 @@ void rs780_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = 2; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; + /* mid sh */ + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = 2; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = 2; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 0; /* high sh */ rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = 2; rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = 3; @@ -353,6 +383,11 @@ void rs780_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; + /* mid mh */ + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = 2; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 0; /* high mh */ rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = 2; rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = 3; @@ -375,6 +410,11 @@ void r600_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; + /* mid sh */ + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 0; /* high sh */ rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = rdev->pm.default_power_state_index; rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; @@ -385,6 +425,11 @@ void r600_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; + /* mid mh */ + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 0; /* high mh */ rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = rdev->pm.default_power_state_index; rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; @@ -401,7 +446,12 @@ void r600_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_ps_idx = 1; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = 1; rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; - rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 1; + rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; + /* mid sh */ + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = 1; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = 1; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 1; /* high sh */ rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = 1; rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = 1; @@ -411,7 +461,12 @@ void r600_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_ps_idx = 2; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = 2; rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; - rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 1; + rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; + /* low mh */ + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = 2; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = 2; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 1; /* high mh */ rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = 2; rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = 2; @@ -430,14 +485,30 @@ void r600_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 0); rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; - rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 1; + rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; } else { rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_ps_idx = r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0); rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0); rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; - rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 1; + rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; + } + /* mid sh */ + if (rdev->flags & RADEON_IS_MOBILITY) { + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = + r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 0); + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = + r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 0); + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 1; + } else { + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = + r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0); + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = + r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0); + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 1; } /* high sh */ rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = @@ -453,14 +524,30 @@ void r600_pm_init_profile(struct radeon_device *rdev) rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 1); rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; - rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 2; + rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; } else { rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_ps_idx = r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1); rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1); rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; - rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 1; + rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; + } + /* mid mh */ + if (rdev->flags & RADEON_IS_MOBILITY) { + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = + r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 1); + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = + r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 1); + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 1; + } else { + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = + r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1); + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = + r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1); + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 1; } /* high mh */ rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = @@ -475,7 +562,18 @@ void r600_pm_init_profile(struct radeon_device *rdev) void r600_pm_misc(struct radeon_device *rdev) { + int req_ps_idx = rdev->pm.requested_power_state_index; + int req_cm_idx = rdev->pm.requested_clock_mode_index; + struct radeon_power_state *ps = &rdev->pm.power_state[req_ps_idx]; + struct radeon_voltage *voltage = &ps->clock_info[req_cm_idx].voltage; + if ((voltage->type == VOLTAGE_SW) && voltage->voltage) { + if (voltage->voltage != rdev->pm.current_vddc) { + radeon_atom_set_voltage(rdev, voltage->voltage); + rdev->pm.current_vddc = voltage->voltage; + DRM_DEBUG("Setting: v: %d\n", voltage->voltage); + } + } } bool r600_gui_idle(struct radeon_device *rdev) @@ -1004,7 +1102,7 @@ static void r600_mc_program(struct radeon_device *rdev) WREG32(MC_VM_FB_LOCATION, tmp); WREG32(HDP_NONSURFACE_BASE, (rdev->mc.vram_start >> 8)); WREG32(HDP_NONSURFACE_INFO, (2 << 7)); - WREG32(HDP_NONSURFACE_SIZE, rdev->mc.mc_vram_size | 0x3FF); + WREG32(HDP_NONSURFACE_SIZE, 0x3FFFFFFF); if (rdev->flags & RADEON_IS_AGP) { WREG32(MC_VM_AGP_TOP, rdev->mc.gtt_end >> 22); WREG32(MC_VM_AGP_BOT, rdev->mc.gtt_start >> 22); @@ -1126,8 +1224,10 @@ int r600_mc_init(struct radeon_device *rdev) rdev->mc.visible_vram_size = rdev->mc.aper_size; r600_vram_gtt_location(rdev, &rdev->mc); - if (rdev->flags & RADEON_IS_IGP) + if (rdev->flags & RADEON_IS_IGP) { + rs690_pm_info(rdev); rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev); + } radeon_update_bandwidth_info(rdev); return 0; } diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 669feb689bfc..ab61aaa887bb 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -176,6 +176,8 @@ void radeon_pm_suspend(struct radeon_device *rdev); void radeon_pm_resume(struct radeon_device *rdev); void radeon_combios_get_power_modes(struct radeon_device *rdev); void radeon_atombios_get_power_modes(struct radeon_device *rdev); +void radeon_atom_set_voltage(struct radeon_device *rdev, u16 level); +void rs690_pm_info(struct radeon_device *rdev); /* * Fences. @@ -618,7 +620,8 @@ enum radeon_dynpm_state { DYNPM_STATE_DISABLED, DYNPM_STATE_MINIMUM, DYNPM_STATE_PAUSED, - DYNPM_STATE_ACTIVE + DYNPM_STATE_ACTIVE, + DYNPM_STATE_SUSPENDED, }; enum radeon_dynpm_action { DYNPM_ACTION_NONE, @@ -647,15 +650,18 @@ enum radeon_pm_profile_type { PM_PROFILE_DEFAULT, PM_PROFILE_AUTO, PM_PROFILE_LOW, + PM_PROFILE_MID, PM_PROFILE_HIGH, }; #define PM_PROFILE_DEFAULT_IDX 0 #define PM_PROFILE_LOW_SH_IDX 1 -#define PM_PROFILE_HIGH_SH_IDX 2 -#define PM_PROFILE_LOW_MH_IDX 3 -#define PM_PROFILE_HIGH_MH_IDX 4 -#define PM_PROFILE_MAX 5 +#define PM_PROFILE_MID_SH_IDX 2 +#define PM_PROFILE_HIGH_SH_IDX 3 +#define PM_PROFILE_LOW_MH_IDX 4 +#define PM_PROFILE_MID_MH_IDX 5 +#define PM_PROFILE_HIGH_MH_IDX 6 +#define PM_PROFILE_MAX 7 struct radeon_pm_profile { int dpms_off_ps_idx; @@ -744,6 +750,7 @@ struct radeon_pm { int default_power_state_index; u32 current_sclk; u32 current_mclk; + u32 current_vddc; struct radeon_i2c_chan *i2c_bus; /* selected pm method */ enum radeon_pm_method pm_method; diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c index e57df08d4aeb..646f96f97c77 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.c +++ b/drivers/gpu/drm/radeon/radeon_asic.c @@ -724,8 +724,8 @@ static struct radeon_asic evergreen_asic = { .irq_set = &evergreen_irq_set, .irq_process = &evergreen_irq_process, .get_vblank_counter = &evergreen_get_vblank_counter, - .fence_ring_emit = NULL, - .cs_parse = NULL, + .fence_ring_emit = &r600_fence_ring_emit, + .cs_parse = &evergreen_cs_parse, .copy_blit = NULL, .copy_dma = NULL, .copy = NULL, @@ -780,6 +780,13 @@ int radeon_asic_init(struct radeon_device *rdev) case CHIP_R423: case CHIP_RV410: rdev->asic = &r420_asic; + /* handle macs */ + if (rdev->bios == NULL) { + rdev->asic->get_engine_clock = &radeon_legacy_get_engine_clock; + rdev->asic->set_engine_clock = &radeon_legacy_set_engine_clock; + rdev->asic->get_memory_clock = &radeon_legacy_get_memory_clock; + rdev->asic->set_memory_clock = NULL; + } break; case CHIP_RS400: case CHIP_RS480: diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h index 5c40a3dfaca2..c0bbaa64157a 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.h +++ b/drivers/gpu/drm/radeon/radeon_asic.h @@ -314,6 +314,7 @@ void evergreen_hpd_set_polarity(struct radeon_device *rdev, u32 evergreen_get_vblank_counter(struct radeon_device *rdev, int crtc); int evergreen_irq_set(struct radeon_device *rdev); int evergreen_irq_process(struct radeon_device *rdev); +extern int evergreen_cs_parse(struct radeon_cs_parser *p); extern void evergreen_pm_misc(struct radeon_device *rdev); extern void evergreen_pm_prepare(struct radeon_device *rdev); extern void evergreen_pm_finish(struct radeon_device *rdev); diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c index 24ea683f7cf5..99bd8a9c56b3 100644 --- a/drivers/gpu/drm/radeon/radeon_atombios.c +++ b/drivers/gpu/drm/radeon/radeon_atombios.c @@ -1538,7 +1538,8 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev) rdev->pm.power_state[state_index].pcie_lanes = power_info->info.asPowerPlayInfo[i].ucNumPciELanes; misc = le32_to_cpu(power_info->info.asPowerPlayInfo[i].ulMiscInfo); - if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) { + if ((misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) || + (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_ACTIVE_HIGH)) { rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_GPIO; rdev->pm.power_state[state_index].clock_info[0].voltage.gpio = @@ -1605,7 +1606,8 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev) power_info->info_2.asPowerPlayInfo[i].ucNumPciELanes; misc = le32_to_cpu(power_info->info_2.asPowerPlayInfo[i].ulMiscInfo); misc2 = le32_to_cpu(power_info->info_2.asPowerPlayInfo[i].ulMiscInfo2); - if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) { + if ((misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) || + (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_ACTIVE_HIGH)) { rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_GPIO; rdev->pm.power_state[state_index].clock_info[0].voltage.gpio = @@ -1679,7 +1681,8 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev) power_info->info_3.asPowerPlayInfo[i].ucNumPciELanes; misc = le32_to_cpu(power_info->info_3.asPowerPlayInfo[i].ulMiscInfo); misc2 = le32_to_cpu(power_info->info_3.asPowerPlayInfo[i].ulMiscInfo2); - if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) { + if ((misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) || + (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_ACTIVE_HIGH)) { rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_GPIO; rdev->pm.power_state[state_index].clock_info[0].voltage.gpio = @@ -1755,9 +1758,22 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev) rdev->pm.power_state[state_index].misc2 = 0; } } else { + int fw_index = GetIndexIntoMasterTable(DATA, FirmwareInfo); + uint8_t fw_frev, fw_crev; + uint16_t fw_data_offset, vddc = 0; + union firmware_info *firmware_info; + ATOM_PPLIB_THERMALCONTROLLER *controller = &power_info->info_4.sThermalController; + + if (atom_parse_data_header(mode_info->atom_context, fw_index, NULL, + &fw_frev, &fw_crev, &fw_data_offset)) { + firmware_info = + (union firmware_info *)(mode_info->atom_context->bios + + fw_data_offset); + vddc = firmware_info->info_14.usBootUpVDDCVoltage; + } + /* add the i2c bus for thermal/fan chip */ /* no support for internal controller yet */ - ATOM_PPLIB_THERMALCONTROLLER *controller = &power_info->info_4.sThermalController; if (controller->ucType > 0) { if ((controller->ucType == ATOM_PP_THERMALCONTROLLER_RV6xx) || (controller->ucType == ATOM_PP_THERMALCONTROLLER_RV770) || @@ -1817,10 +1833,7 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev) /* skip invalid modes */ if (rdev->pm.power_state[state_index].clock_info[mode_index].sclk == 0) continue; - rdev->pm.power_state[state_index].clock_info[mode_index].voltage.type = - VOLTAGE_SW; - rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage = - clock_info->usVDDC; + /* voltage works differently on IGPs */ mode_index++; } else if (ASIC_IS_DCE4(rdev)) { struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO *clock_info = @@ -1904,6 +1917,16 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev) rdev->pm.default_power_state_index = state_index; rdev->pm.power_state[state_index].default_clock_mode = &rdev->pm.power_state[state_index].clock_info[mode_index - 1]; + /* patch the table values with the default slck/mclk from firmware info */ + for (j = 0; j < mode_index; j++) { + rdev->pm.power_state[state_index].clock_info[j].mclk = + rdev->clock.default_mclk; + rdev->pm.power_state[state_index].clock_info[j].sclk = + rdev->clock.default_sclk; + if (vddc) + rdev->pm.power_state[state_index].clock_info[j].voltage.voltage = + vddc; + } } state_index++; } @@ -1943,6 +1966,7 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev) rdev->pm.current_power_state_index = rdev->pm.default_power_state_index; rdev->pm.current_clock_mode_index = 0; + rdev->pm.current_vddc = rdev->pm.power_state[rdev->pm.default_power_state_index].clock_info[0].voltage.voltage; } void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable) @@ -1998,6 +2022,42 @@ void radeon_atom_set_memory_clock(struct radeon_device *rdev, atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } +union set_voltage { + struct _SET_VOLTAGE_PS_ALLOCATION alloc; + struct _SET_VOLTAGE_PARAMETERS v1; + struct _SET_VOLTAGE_PARAMETERS_V2 v2; +}; + +void radeon_atom_set_voltage(struct radeon_device *rdev, u16 level) +{ + union set_voltage args; + int index = GetIndexIntoMasterTable(COMMAND, SetVoltage); + u8 frev, crev, volt_index = level; + + if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev)) + return; + + switch (crev) { + case 1: + args.v1.ucVoltageType = SET_VOLTAGE_TYPE_ASIC_VDDC; + args.v1.ucVoltageMode = SET_ASIC_VOLTAGE_MODE_ALL_SOURCE; + args.v1.ucVoltageIndex = volt_index; + break; + case 2: + args.v2.ucVoltageType = SET_VOLTAGE_TYPE_ASIC_VDDC; + args.v2.ucVoltageMode = SET_ASIC_VOLTAGE_MODE_SET_VOLTAGE; + args.v2.usVoltageLevel = cpu_to_le16(level); + break; + default: + DRM_ERROR("Unknown table version %d, %d\n", frev, crev); + return; + } + + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); +} + + + void radeon_atom_initialize_bios_scratch_regs(struct drm_device *dev) { struct radeon_device *rdev = dev->dev_private; diff --git a/drivers/gpu/drm/radeon/radeon_bios.c b/drivers/gpu/drm/radeon/radeon_bios.c index 91f5b5a29a9f..654787ec43f4 100644 --- a/drivers/gpu/drm/radeon/radeon_bios.c +++ b/drivers/gpu/drm/radeon/radeon_bios.c @@ -48,6 +48,10 @@ static bool igp_read_bios_from_vram(struct radeon_device *rdev) resource_size_t vram_base; resource_size_t size = 256 * 1024; /* ??? */ + if (!(rdev->flags & RADEON_IS_IGP)) + if (!radeon_card_posted(rdev)) + return false; + rdev->bios = NULL; vram_base = pci_resource_start(rdev->pdev, 0); bios = ioremap(vram_base, size); diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c index 7b5e10d3e9c9..d1c1d8dd93ce 100644 --- a/drivers/gpu/drm/radeon/radeon_combios.c +++ b/drivers/gpu/drm/radeon/radeon_combios.c @@ -1411,6 +1411,11 @@ bool radeon_get_legacy_connector_info_from_table(struct drm_device *dev) rdev->mode_info.connector_table = CT_IMAC_G5_ISIGHT; } else #endif /* CONFIG_PPC_PMAC */ +#ifdef CONFIG_PPC64 + if (ASIC_IS_RN50(rdev)) + rdev->mode_info.connector_table = CT_RN50_POWER; + else +#endif rdev->mode_info.connector_table = CT_GENERIC; } @@ -1853,6 +1858,33 @@ bool radeon_get_legacy_connector_info_from_table(struct drm_device *dev) CONNECTOR_OBJECT_ID_SVIDEO, &hpd); break; + case CT_RN50_POWER: + DRM_INFO("Connector Table: %d (rn50-power)\n", + rdev->mode_info.connector_table); + /* VGA - primary dac */ + ddc_i2c = combios_setup_i2c_bus(rdev, RADEON_GPIO_VGA_DDC); + hpd.hpd = RADEON_HPD_NONE; + radeon_add_legacy_encoder(dev, + radeon_get_encoder_id(dev, + ATOM_DEVICE_CRT1_SUPPORT, + 1), + ATOM_DEVICE_CRT1_SUPPORT); + radeon_add_legacy_connector(dev, 0, ATOM_DEVICE_CRT1_SUPPORT, + DRM_MODE_CONNECTOR_VGA, &ddc_i2c, + CONNECTOR_OBJECT_ID_VGA, + &hpd); + ddc_i2c = combios_setup_i2c_bus(rdev, RADEON_GPIO_CRT2_DDC); + hpd.hpd = RADEON_HPD_NONE; + radeon_add_legacy_encoder(dev, + radeon_get_encoder_id(dev, + ATOM_DEVICE_CRT2_SUPPORT, + 2), + ATOM_DEVICE_CRT2_SUPPORT); + radeon_add_legacy_connector(dev, 1, ATOM_DEVICE_CRT2_SUPPORT, + DRM_MODE_CONNECTOR_VGA, &ddc_i2c, + CONNECTOR_OBJECT_ID_VGA, + &hpd); + break; default: DRM_INFO("Connector table: %d (invalid)\n", rdev->mode_info.connector_table); @@ -1906,15 +1938,6 @@ static bool radeon_apply_legacy_quirks(struct drm_device *dev, return false; } - /* Some RV100 cards with 2 VGA ports show up with DVI+VGA */ - if (dev->pdev->device == 0x5159 && - dev->pdev->subsystem_vendor == 0x1002 && - dev->pdev->subsystem_device == 0x013a) { - if (*legacy_connector == CONNECTOR_DVI_I_LEGACY) - *legacy_connector = CONNECTOR_CRT_LEGACY; - - } - /* X300 card with extra non-existent DVI port */ if (dev->pdev->device == 0x5B60 && dev->pdev->subsystem_vendor == 0x17af && @@ -2026,6 +2049,7 @@ bool radeon_get_legacy_connector_info_from_bios(struct drm_device *dev) combios_setup_i2c_bus(rdev, RADEON_GPIO_CRT2_DDC); break; default: + ddc_i2c.valid = false; break; } @@ -2339,6 +2363,7 @@ bool radeon_get_legacy_connector_info_from_bios(struct drm_device *dev) if (RBIOS8(tv_info + 6) == 'T') { if (radeon_apply_legacy_tv_quirks(dev)) { hpd.hpd = RADEON_HPD_NONE; + ddc_i2c.valid = false; radeon_add_legacy_encoder(dev, radeon_get_encoder_id (dev, @@ -2454,7 +2479,12 @@ default_mode: rdev->pm.power_state[state_index].clock_info[0].mclk = rdev->clock.default_mclk; rdev->pm.power_state[state_index].clock_info[0].sclk = rdev->clock.default_sclk; rdev->pm.power_state[state_index].default_clock_mode = &rdev->pm.power_state[state_index].clock_info[0]; - rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_NONE; + if ((state_index > 0) && + (rdev->pm.power_state[0].clock_info[0].voltage.type == VOLTAGE_GPIO)) + rdev->pm.power_state[state_index].clock_info[0].voltage = + rdev->pm.power_state[0].clock_info[0].voltage; + else + rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_NONE; rdev->pm.power_state[state_index].pcie_lanes = 16; rdev->pm.power_state[state_index].flags = 0; rdev->pm.default_power_state_index = state_index; @@ -3012,6 +3042,14 @@ void radeon_combios_asic_init(struct drm_device *dev) combios_write_ram_size(dev); } + /* quirk for rs4xx HP nx6125 laptop to make it resume + * - it hangs on resume inside the dynclk 1 table. + */ + if (rdev->family == CHIP_RS480 && + rdev->pdev->subsystem_vendor == 0x103c && + rdev->pdev->subsystem_device == 0x308b) + return; + /* DYN CLK 1 */ table = combios_get_table_offset(dev, COMBIOS_DYN_CLK_1_TABLE); if (table) diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c index 0c7ccc6961a3..f58f8bd8f77b 100644 --- a/drivers/gpu/drm/radeon/radeon_connectors.c +++ b/drivers/gpu/drm/radeon/radeon_connectors.c @@ -785,7 +785,9 @@ static enum drm_connector_status radeon_dvi_detect(struct drm_connector *connect if (connector == list_connector) continue; list_radeon_connector = to_radeon_connector(list_connector); - if (radeon_connector->devices == list_radeon_connector->devices) { + if (list_radeon_connector->shared_ddc && + (list_radeon_connector->ddc_bus->rec.i2c_id == + radeon_connector->ddc_bus->rec.i2c_id)) { if (drm_detect_hdmi_monitor(radeon_connector->edid)) { if (connector->connector_type == DRM_MODE_CONNECTOR_DVID) { kfree(radeon_connector->edid); diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c index b7023fff89eb..4eb67c0e0996 100644 --- a/drivers/gpu/drm/radeon/radeon_cursor.c +++ b/drivers/gpu/drm/radeon/radeon_cursor.c @@ -194,7 +194,7 @@ unpin: fail: drm_gem_object_unreference_unlocked(obj); - return 0; + return ret; } int radeon_crtc_cursor_move(struct drm_crtc *crtc, diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index 2a897a7ca26f..37533bec1f25 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -546,8 +546,10 @@ static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero /* don't suspend or resume card normally */ rdev->powered_down = false; radeon_resume_kms(dev); + drm_kms_helper_poll_enable(dev); } else { printk(KERN_INFO "radeon: switched off\n"); + drm_kms_helper_poll_disable(dev); radeon_suspend_kms(dev, pmm); /* don't suspend or resume card normally */ rdev->powered_down = true; @@ -711,6 +713,7 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) { struct radeon_device *rdev; struct drm_crtc *crtc; + struct drm_connector *connector; int r; if (dev == NULL || dev->dev_private == NULL) { @@ -723,6 +726,12 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) if (rdev->powered_down) return 0; + + /* turn off display hw */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF); + } + /* unpin the front buffers */ list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { struct radeon_framebuffer *rfb = to_radeon_framebuffer(crtc->fb); @@ -770,6 +779,7 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) int radeon_resume_kms(struct drm_device *dev) { + struct drm_connector *connector; struct radeon_device *rdev = dev->dev_private; if (rdev->powered_down) @@ -788,6 +798,12 @@ int radeon_resume_kms(struct drm_device *dev) radeon_resume(rdev); radeon_pm_resume(rdev); radeon_restore_bios_scratch_regs(rdev); + + /* turn on display hw */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + drm_helper_connector_dpms(connector, DRM_MODE_DPMS_ON); + } + radeon_fbdev_set_suspend(rdev, 0); release_console_sem(); diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index 1006549d1570..8154cdf796e4 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -284,8 +284,7 @@ static const char *connector_names[15] = { "eDP", }; -static const char *hpd_names[7] = { - "NONE", +static const char *hpd_names[6] = { "HPD1", "HPD2", "HPD3", diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index 683e281b4092..ed0ceb3fc40a 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -45,9 +45,10 @@ * - 2.2.0 - add r6xx/r7xx const buffer support * - 2.3.0 - add MSPOS + 3D texture + r500 VAP regs * - 2.4.0 - add crtc id query + * - 2.5.0 - add get accel 2 to work around ddx breakage for evergreen */ #define KMS_DRIVER_MAJOR 2 -#define KMS_DRIVER_MINOR 4 +#define KMS_DRIVER_MINOR 5 #define KMS_DRIVER_PATCHLEVEL 0 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags); int radeon_driver_unload_kms(struct drm_device *dev); @@ -213,6 +214,7 @@ static struct drm_driver driver_old = { .mmap = drm_mmap, .poll = drm_poll, .fasync = drm_fasync, + .read = drm_read, #ifdef CONFIG_COMPAT .compat_ioctl = radeon_compat_ioctl, #endif @@ -301,6 +303,7 @@ static struct drm_driver kms_driver = { .mmap = radeon_mmap, .poll = drm_poll, .fasync = drm_fasync, + .read = drm_read, #ifdef CONFIG_COMPAT .compat_ioctl = radeon_kms_compat_ioctl, #endif diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c index 1ebb100015b7..e0b30b264c28 100644 --- a/drivers/gpu/drm/radeon/radeon_encoders.c +++ b/drivers/gpu/drm/radeon/radeon_encoders.c @@ -1072,6 +1072,8 @@ radeon_atom_encoder_dpms(struct drm_encoder *encoder, int mode) if (is_dig) { switch (mode) { case DRM_MODE_DPMS_ON: + if (!ASIC_IS_DCE4(rdev)) + atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE_OUTPUT, 0, 0); if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_DP) { struct drm_connector *connector = radeon_get_connector_for_encoder(encoder); @@ -1079,8 +1081,6 @@ radeon_atom_encoder_dpms(struct drm_encoder *encoder, int mode) if (ASIC_IS_DCE4(rdev)) atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_VIDEO_ON); } - if (!ASIC_IS_DCE4(rdev)) - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE_OUTPUT, 0, 0); break; case DRM_MODE_DPMS_STANDBY: case DRM_MODE_DPMS_SUSPEND: diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c index e192acfbf0cd..dc1634bb0c11 100644 --- a/drivers/gpu/drm/radeon/radeon_fb.c +++ b/drivers/gpu/drm/radeon/radeon_fb.c @@ -363,6 +363,7 @@ int radeon_fbdev_init(struct radeon_device *rdev) { struct radeon_fbdev *rfbdev; int bpp_sel = 32; + int ret; /* select 8 bpp console on RN50 or 16MB cards */ if (ASIC_IS_RN50(rdev) || rdev->mc.real_vram_size <= (32*1024*1024)) @@ -376,9 +377,14 @@ int radeon_fbdev_init(struct radeon_device *rdev) rdev->mode_info.rfbdev = rfbdev; rfbdev->helper.funcs = &radeon_fb_helper_funcs; - drm_fb_helper_init(rdev->ddev, &rfbdev->helper, - rdev->num_crtc, - RADEONFB_CONN_LIMIT); + ret = drm_fb_helper_init(rdev->ddev, &rfbdev->helper, + rdev->num_crtc, + RADEONFB_CONN_LIMIT); + if (ret) { + kfree(rfbdev); + return ret; + } + drm_fb_helper_single_add_all_connectors(&rfbdev->helper); drm_fb_helper_initial_config(&rfbdev->helper, bpp_sel); return 0; diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 04068352ccd2..6a70c0dc7f92 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -118,7 +118,11 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) value = rdev->num_z_pipes; break; case RADEON_INFO_ACCEL_WORKING: - value = rdev->accel_working; + /* xf86-video-ati 6.13.0 relies on this being false for evergreen */ + if ((rdev->family >= CHIP_CEDAR) && (rdev->family <= CHIP_HEMLOCK)) + value = false; + else + value = rdev->accel_working; break; case RADEON_INFO_CRTC_FROM_ID: for (i = 0, found = 0; i < rdev->num_crtc; i++) { @@ -134,6 +138,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) return -EINVAL; } break; + case RADEON_INFO_ACCEL_WORKING2: + value = rdev->accel_working; + break; default: DRM_DEBUG("Invalid request %d\n", info->request); return -EINVAL; diff --git a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c index 5a13b3eeef19..bad77f40a9da 100644 --- a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c +++ b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c @@ -928,16 +928,14 @@ static void radeon_legacy_tv_dac_mode_set(struct drm_encoder *encoder, if (ASIC_IS_R300(rdev)) { gpiopad_a = RREG32(RADEON_GPIOPAD_A) | 1; disp_output_cntl = RREG32(RADEON_DISP_OUTPUT_CNTL); - } - - if (rdev->family == CHIP_R200 || ASIC_IS_R300(rdev)) - disp_tv_out_cntl = RREG32(RADEON_DISP_TV_OUT_CNTL); - else + } else if (rdev->family != CHIP_R200) disp_hw_debug = RREG32(RADEON_DISP_HW_DEBUG); - - if (rdev->family == CHIP_R200) + else if (rdev->family == CHIP_R200) fp2_gen_cntl = RREG32(RADEON_FP2_GEN_CNTL); + if (rdev->family >= CHIP_R200) + disp_tv_out_cntl = RREG32(RADEON_DISP_TV_OUT_CNTL); + if (is_tv) { uint32_t dac_cntl; @@ -1002,15 +1000,13 @@ static void radeon_legacy_tv_dac_mode_set(struct drm_encoder *encoder, if (ASIC_IS_R300(rdev)) { WREG32_P(RADEON_GPIOPAD_A, gpiopad_a, ~1); WREG32(RADEON_DISP_OUTPUT_CNTL, disp_output_cntl); - } + } else if (rdev->family != CHIP_R200) + WREG32(RADEON_DISP_HW_DEBUG, disp_hw_debug); + else if (rdev->family == CHIP_R200) + WREG32(RADEON_FP2_GEN_CNTL, fp2_gen_cntl); if (rdev->family >= CHIP_R200) WREG32(RADEON_DISP_TV_OUT_CNTL, disp_tv_out_cntl); - else - WREG32(RADEON_DISP_HW_DEBUG, disp_hw_debug); - - if (rdev->family == CHIP_R200) - WREG32(RADEON_FP2_GEN_CNTL, fp2_gen_cntl); if (is_tv) radeon_legacy_tv_mode_set(encoder, mode, adjusted_mode); @@ -1168,6 +1164,17 @@ static enum drm_connector_status radeon_legacy_tv_dac_detect(struct drm_encoder struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct radeon_encoder_tv_dac *tv_dac = radeon_encoder->enc_priv; bool color = true; + struct drm_crtc *crtc; + + /* find out if crtc2 is in use or if this encoder is using it */ + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + if ((radeon_crtc->crtc_id == 1) && crtc->enabled) { + if (encoder->crtc != crtc) { + return connector_status_disconnected; + } + } + } if (connector->connector_type == DRM_MODE_CONNECTOR_SVIDEO || connector->connector_type == DRM_MODE_CONNECTOR_Composite || diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h index 67358baf28b2..95696aa57ac8 100644 --- a/drivers/gpu/drm/radeon/radeon_mode.h +++ b/drivers/gpu/drm/radeon/radeon_mode.h @@ -206,6 +206,7 @@ enum radeon_connector_table { CT_MINI_INTERNAL, CT_IMAC_G5_ISIGHT, CT_EMAC, + CT_RN50_POWER, }; enum radeon_dvo_chip { diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index a8d162c6f829..115d26b762cc 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -33,6 +33,14 @@ #define RADEON_WAIT_VBLANK_TIMEOUT 200 #define RADEON_WAIT_IDLE_TIMEOUT 200 +static const char *radeon_pm_state_type_name[5] = { + "Default", + "Powersave", + "Battery", + "Balanced", + "Performance", +}; + static void radeon_dynpm_idle_work_handler(struct work_struct *work); static int radeon_debugfs_pm_init(struct radeon_device *rdev); static bool radeon_pm_in_vbl(struct radeon_device *rdev); @@ -84,9 +92,9 @@ static void radeon_pm_update_profile(struct radeon_device *rdev) rdev->pm.profile_index = PM_PROFILE_HIGH_SH_IDX; } else { if (rdev->pm.active_crtc_count > 1) - rdev->pm.profile_index = PM_PROFILE_LOW_MH_IDX; + rdev->pm.profile_index = PM_PROFILE_MID_MH_IDX; else - rdev->pm.profile_index = PM_PROFILE_LOW_SH_IDX; + rdev->pm.profile_index = PM_PROFILE_MID_SH_IDX; } break; case PM_PROFILE_LOW: @@ -95,6 +103,12 @@ static void radeon_pm_update_profile(struct radeon_device *rdev) else rdev->pm.profile_index = PM_PROFILE_LOW_SH_IDX; break; + case PM_PROFILE_MID: + if (rdev->pm.active_crtc_count > 1) + rdev->pm.profile_index = PM_PROFILE_MID_MH_IDX; + else + rdev->pm.profile_index = PM_PROFILE_MID_SH_IDX; + break; case PM_PROFILE_HIGH: if (rdev->pm.active_crtc_count > 1) rdev->pm.profile_index = PM_PROFILE_HIGH_MH_IDX; @@ -127,15 +141,6 @@ static void radeon_unmap_vram_bos(struct radeon_device *rdev) if (bo->tbo.mem.mem_type == TTM_PL_VRAM) ttm_bo_unmap_virtual(&bo->tbo); } - - if (rdev->gart.table.vram.robj) - ttm_bo_unmap_virtual(&rdev->gart.table.vram.robj->tbo); - - if (rdev->stollen_vga_memory) - ttm_bo_unmap_virtual(&rdev->stollen_vga_memory->tbo); - - if (rdev->r600_blit.shader_obj) - ttm_bo_unmap_virtual(&rdev->r600_blit.shader_obj->tbo); } static void radeon_sync_with_vblank(struct radeon_device *rdev) @@ -151,6 +156,7 @@ static void radeon_sync_with_vblank(struct radeon_device *rdev) static void radeon_set_power_state(struct radeon_device *rdev) { u32 sclk, mclk; + bool misc_after = false; if ((rdev->pm.requested_clock_mode_index == rdev->pm.current_clock_mode_index) && (rdev->pm.requested_power_state_index == rdev->pm.current_power_state_index)) @@ -167,55 +173,47 @@ static void radeon_set_power_state(struct radeon_device *rdev) if (mclk > rdev->clock.default_mclk) mclk = rdev->clock.default_mclk; - /* voltage, pcie lanes, etc.*/ - radeon_pm_misc(rdev); + /* upvolt before raising clocks, downvolt after lowering clocks */ + if (sclk < rdev->pm.current_sclk) + misc_after = true; - if (rdev->pm.pm_method == PM_METHOD_DYNPM) { - radeon_sync_with_vblank(rdev); + radeon_sync_with_vblank(rdev); + if (rdev->pm.pm_method == PM_METHOD_DYNPM) { if (!radeon_pm_in_vbl(rdev)) return; + } - radeon_pm_prepare(rdev); - /* set engine clock */ - if (sclk != rdev->pm.current_sclk) { - radeon_pm_debug_check_in_vbl(rdev, false); - radeon_set_engine_clock(rdev, sclk); - radeon_pm_debug_check_in_vbl(rdev, true); - rdev->pm.current_sclk = sclk; - DRM_DEBUG("Setting: e: %d\n", sclk); - } + radeon_pm_prepare(rdev); - /* set memory clock */ - if (rdev->asic->set_memory_clock && (mclk != rdev->pm.current_mclk)) { - radeon_pm_debug_check_in_vbl(rdev, false); - radeon_set_memory_clock(rdev, mclk); - radeon_pm_debug_check_in_vbl(rdev, true); - rdev->pm.current_mclk = mclk; - DRM_DEBUG("Setting: m: %d\n", mclk); - } - radeon_pm_finish(rdev); - } else { - /* set engine clock */ - if (sclk != rdev->pm.current_sclk) { - radeon_sync_with_vblank(rdev); - radeon_pm_prepare(rdev); - radeon_set_engine_clock(rdev, sclk); - radeon_pm_finish(rdev); - rdev->pm.current_sclk = sclk; - DRM_DEBUG("Setting: e: %d\n", sclk); - } - /* set memory clock */ - if (rdev->asic->set_memory_clock && (mclk != rdev->pm.current_mclk)) { - radeon_sync_with_vblank(rdev); - radeon_pm_prepare(rdev); - radeon_set_memory_clock(rdev, mclk); - radeon_pm_finish(rdev); - rdev->pm.current_mclk = mclk; - DRM_DEBUG("Setting: m: %d\n", mclk); - } + if (!misc_after) + /* voltage, pcie lanes, etc.*/ + radeon_pm_misc(rdev); + + /* set engine clock */ + if (sclk != rdev->pm.current_sclk) { + radeon_pm_debug_check_in_vbl(rdev, false); + radeon_set_engine_clock(rdev, sclk); + radeon_pm_debug_check_in_vbl(rdev, true); + rdev->pm.current_sclk = sclk; + DRM_DEBUG("Setting: e: %d\n", sclk); } + /* set memory clock */ + if (rdev->asic->set_memory_clock && (mclk != rdev->pm.current_mclk)) { + radeon_pm_debug_check_in_vbl(rdev, false); + radeon_set_memory_clock(rdev, mclk); + radeon_pm_debug_check_in_vbl(rdev, true); + rdev->pm.current_mclk = mclk; + DRM_DEBUG("Setting: m: %d\n", mclk); + } + + if (misc_after) + /* voltage, pcie lanes, etc.*/ + radeon_pm_misc(rdev); + + radeon_pm_finish(rdev); + rdev->pm.current_power_state_index = rdev->pm.requested_power_state_index; rdev->pm.current_clock_mode_index = rdev->pm.requested_clock_mode_index; } else @@ -288,6 +286,42 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev) mutex_unlock(&rdev->ddev->struct_mutex); } +static void radeon_pm_print_states(struct radeon_device *rdev) +{ + int i, j; + struct radeon_power_state *power_state; + struct radeon_pm_clock_info *clock_info; + + DRM_DEBUG("%d Power State(s)\n", rdev->pm.num_power_states); + for (i = 0; i < rdev->pm.num_power_states; i++) { + power_state = &rdev->pm.power_state[i]; + DRM_DEBUG("State %d: %s\n", i, + radeon_pm_state_type_name[power_state->type]); + if (i == rdev->pm.default_power_state_index) + DRM_DEBUG("\tDefault"); + if ((rdev->flags & RADEON_IS_PCIE) && !(rdev->flags & RADEON_IS_IGP)) + DRM_DEBUG("\t%d PCIE Lanes\n", power_state->pcie_lanes); + if (power_state->flags & RADEON_PM_STATE_SINGLE_DISPLAY_ONLY) + DRM_DEBUG("\tSingle display only\n"); + DRM_DEBUG("\t%d Clock Mode(s)\n", power_state->num_clock_modes); + for (j = 0; j < power_state->num_clock_modes; j++) { + clock_info = &(power_state->clock_info[j]); + if (rdev->flags & RADEON_IS_IGP) + DRM_DEBUG("\t\t%d e: %d%s\n", + j, + clock_info->sclk * 10, + clock_info->flags & RADEON_PM_MODE_NO_DISPLAY ? "\tNo display only" : ""); + else + DRM_DEBUG("\t\t%d e: %d\tm: %d\tv: %d%s\n", + j, + clock_info->sclk * 10, + clock_info->mclk * 10, + clock_info->voltage.voltage, + clock_info->flags & RADEON_PM_MODE_NO_DISPLAY ? "\tNo display only" : ""); + } + } +} + static ssize_t radeon_get_pm_profile(struct device *dev, struct device_attribute *attr, char *buf) @@ -318,6 +352,8 @@ static ssize_t radeon_set_pm_profile(struct device *dev, rdev->pm.profile = PM_PROFILE_AUTO; else if (strncmp("low", buf, strlen("low")) == 0) rdev->pm.profile = PM_PROFILE_LOW; + else if (strncmp("mid", buf, strlen("mid")) == 0) + rdev->pm.profile = PM_PROFILE_MID; else if (strncmp("high", buf, strlen("high")) == 0) rdev->pm.profile = PM_PROFILE_HIGH; else { @@ -361,13 +397,20 @@ static ssize_t radeon_set_pm_method(struct device *dev, rdev->pm.dynpm_planned_action = DYNPM_ACTION_DEFAULT; mutex_unlock(&rdev->pm.mutex); } else if (strncmp("profile", buf, strlen("profile")) == 0) { + bool flush_wq = false; + mutex_lock(&rdev->pm.mutex); - rdev->pm.pm_method = PM_METHOD_PROFILE; + if (rdev->pm.pm_method == PM_METHOD_DYNPM) { + cancel_delayed_work(&rdev->pm.dynpm_idle_work); + flush_wq = true; + } /* disable dynpm */ rdev->pm.dynpm_state = DYNPM_STATE_DISABLED; rdev->pm.dynpm_planned_action = DYNPM_ACTION_NONE; - cancel_delayed_work(&rdev->pm.dynpm_idle_work); + rdev->pm.pm_method = PM_METHOD_PROFILE; mutex_unlock(&rdev->pm.mutex); + if (flush_wq) + flush_workqueue(rdev->wq); } else { DRM_ERROR("invalid power method!\n"); goto fail; @@ -382,17 +425,36 @@ static DEVICE_ATTR(power_method, S_IRUGO | S_IWUSR, radeon_get_pm_method, radeon void radeon_pm_suspend(struct radeon_device *rdev) { + bool flush_wq = false; + mutex_lock(&rdev->pm.mutex); - cancel_delayed_work(&rdev->pm.dynpm_idle_work); - rdev->pm.current_power_state_index = -1; - rdev->pm.current_clock_mode_index = -1; - rdev->pm.current_sclk = 0; - rdev->pm.current_mclk = 0; + if (rdev->pm.pm_method == PM_METHOD_DYNPM) { + cancel_delayed_work(&rdev->pm.dynpm_idle_work); + if (rdev->pm.dynpm_state == DYNPM_STATE_ACTIVE) + rdev->pm.dynpm_state = DYNPM_STATE_SUSPENDED; + flush_wq = true; + } mutex_unlock(&rdev->pm.mutex); + if (flush_wq) + flush_workqueue(rdev->wq); } void radeon_pm_resume(struct radeon_device *rdev) { + /* asic init will reset the default power state */ + mutex_lock(&rdev->pm.mutex); + rdev->pm.current_power_state_index = rdev->pm.default_power_state_index; + rdev->pm.current_clock_mode_index = 0; + rdev->pm.current_sclk = rdev->clock.default_sclk; + rdev->pm.current_mclk = rdev->clock.default_mclk; + rdev->pm.current_vddc = rdev->pm.power_state[rdev->pm.default_power_state_index].clock_info[0].voltage.voltage; + if (rdev->pm.pm_method == PM_METHOD_DYNPM + && rdev->pm.dynpm_state == DYNPM_STATE_SUSPENDED) { + rdev->pm.dynpm_state = DYNPM_STATE_ACTIVE; + queue_delayed_work(rdev->wq, &rdev->pm.dynpm_idle_work, + msecs_to_jiffies(RADEON_IDLE_LOOP_MS)); + } + mutex_unlock(&rdev->pm.mutex); radeon_pm_compute_clocks(rdev); } @@ -401,32 +463,24 @@ int radeon_pm_init(struct radeon_device *rdev) int ret; /* default to profile method */ rdev->pm.pm_method = PM_METHOD_PROFILE; + rdev->pm.profile = PM_PROFILE_DEFAULT; rdev->pm.dynpm_state = DYNPM_STATE_DISABLED; rdev->pm.dynpm_planned_action = DYNPM_ACTION_NONE; rdev->pm.dynpm_can_upclock = true; rdev->pm.dynpm_can_downclock = true; - rdev->pm.current_sclk = 0; - rdev->pm.current_mclk = 0; + rdev->pm.current_sclk = rdev->clock.default_sclk; + rdev->pm.current_mclk = rdev->clock.default_mclk; if (rdev->bios) { if (rdev->is_atom_bios) radeon_atombios_get_power_modes(rdev); else radeon_combios_get_power_modes(rdev); + radeon_pm_print_states(rdev); radeon_pm_init_profile(rdev); - rdev->pm.current_power_state_index = -1; - rdev->pm.current_clock_mode_index = -1; } if (rdev->pm.num_power_states > 1) { - if (rdev->pm.pm_method == PM_METHOD_PROFILE) { - mutex_lock(&rdev->pm.mutex); - rdev->pm.profile = PM_PROFILE_DEFAULT; - radeon_pm_update_profile(rdev); - radeon_pm_set_clocks(rdev); - mutex_unlock(&rdev->pm.mutex); - } - /* where's the best place to put these? */ ret = device_create_file(rdev->dev, &dev_attr_power_profile); if (ret) @@ -454,6 +508,8 @@ int radeon_pm_init(struct radeon_device *rdev) void radeon_pm_fini(struct radeon_device *rdev) { if (rdev->pm.num_power_states > 1) { + bool flush_wq = false; + mutex_lock(&rdev->pm.mutex); if (rdev->pm.pm_method == PM_METHOD_PROFILE) { rdev->pm.profile = PM_PROFILE_DEFAULT; @@ -461,13 +517,16 @@ void radeon_pm_fini(struct radeon_device *rdev) radeon_pm_set_clocks(rdev); } else if (rdev->pm.pm_method == PM_METHOD_DYNPM) { /* cancel work */ - cancel_delayed_work_sync(&rdev->pm.dynpm_idle_work); + cancel_delayed_work(&rdev->pm.dynpm_idle_work); + flush_wq = true; /* reset default clocks */ rdev->pm.dynpm_state = DYNPM_STATE_DISABLED; rdev->pm.dynpm_planned_action = DYNPM_ACTION_DEFAULT; radeon_pm_set_clocks(rdev); } mutex_unlock(&rdev->pm.mutex); + if (flush_wq) + flush_workqueue(rdev->wq); device_remove_file(rdev->dev, &dev_attr_power_profile); device_remove_file(rdev->dev, &dev_attr_power_method); @@ -688,12 +747,12 @@ static void radeon_dynpm_idle_work_handler(struct work_struct *work) radeon_pm_get_dynpm_state(rdev); radeon_pm_set_clocks(rdev); } + + queue_delayed_work(rdev->wq, &rdev->pm.dynpm_idle_work, + msecs_to_jiffies(RADEON_IDLE_LOOP_MS)); } mutex_unlock(&rdev->pm.mutex); ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched); - - queue_delayed_work(rdev->wq, &rdev->pm.dynpm_idle_work, - msecs_to_jiffies(RADEON_IDLE_LOOP_MS)); } /* @@ -712,6 +771,8 @@ static int radeon_debugfs_pm_info(struct seq_file *m, void *data) seq_printf(m, "default memory clock: %u0 kHz\n", rdev->clock.default_mclk); if (rdev->asic->get_memory_clock) seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev)); + if (rdev->pm.current_vddc) + seq_printf(m, "voltage: %u mV\n", rdev->pm.current_vddc); if (rdev->asic->get_pcie_lanes) seq_printf(m, "PCIE lanes: %d\n", radeon_get_pcie_lanes(rdev)); diff --git a/drivers/gpu/drm/radeon/radeon_state.c b/drivers/gpu/drm/radeon/radeon_state.c index 11ce94c668e7..b3ba44c0a818 100644 --- a/drivers/gpu/drm/radeon/radeon_state.c +++ b/drivers/gpu/drm/radeon/radeon_state.c @@ -424,7 +424,7 @@ static __inline__ int radeon_check_and_fixup_packet3(drm_radeon_private_t * if ((*cmd & RADEON_GMC_SRC_PITCH_OFFSET_CNTL) && (*cmd & RADEON_GMC_DST_PITCH_OFFSET_CNTL)) { u32 *cmd3 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 3); - offset = *cmd << 10; + offset = *cmd3 << 10; if (radeon_check_and_fixup_offset (dev_priv, file_priv, &offset)) { DRM_ERROR("Invalid second packet offset\n"); @@ -2896,9 +2896,12 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, return rv; rv = drm_buffer_copy_from_user(cmdbuf->buffer, buffer, cmdbuf->bufsz); - if (rv) + if (rv) { + drm_buffer_free(cmdbuf->buffer); return rv; - } + } + } else + goto done; orig_nbox = cmdbuf->nbox; @@ -2906,8 +2909,7 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, int temp; temp = r300_do_cp_cmdbuf(dev, file_priv, cmdbuf); - if (cmdbuf->bufsz != 0) - drm_buffer_free(cmdbuf->buffer); + drm_buffer_free(cmdbuf->buffer); return temp; } @@ -3013,16 +3015,15 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, } } - if (cmdbuf->bufsz != 0) - drm_buffer_free(cmdbuf->buffer); + drm_buffer_free(cmdbuf->buffer); + done: DRM_DEBUG("DONE\n"); COMMIT_RING(); return 0; err: - if (cmdbuf->bufsz != 0) - drm_buffer_free(cmdbuf->buffer); + drm_buffer_free(cmdbuf->buffer); return -EINVAL; } diff --git a/drivers/gpu/drm/radeon/reg_srcs/evergreen b/drivers/gpu/drm/radeon/reg_srcs/evergreen new file mode 100644 index 000000000000..f78fd592544d --- /dev/null +++ b/drivers/gpu/drm/radeon/reg_srcs/evergreen @@ -0,0 +1,611 @@ +evergreen 0x9400 +0x00008040 WAIT_UNTIL +0x00008044 WAIT_UNTIL_POLL_CNTL +0x00008048 WAIT_UNTIL_POLL_MASK +0x0000804c WAIT_UNTIL_POLL_REFDATA +0x000088B0 VGT_VTX_VECT_EJECT_REG +0x000088C4 VGT_CACHE_INVALIDATION +0x000088D4 VGT_GS_VERTEX_REUSE +0x00008958 VGT_PRIMITIVE_TYPE +0x0000895C VGT_INDEX_TYPE +0x00008970 VGT_NUM_INDICES +0x00008974 VGT_NUM_INSTANCES +0x00008990 VGT_COMPUTE_DIM_X +0x00008994 VGT_COMPUTE_DIM_Y +0x00008998 VGT_COMPUTE_DIM_Z +0x0000899C VGT_COMPUTE_START_X +0x000089A0 VGT_COMPUTE_START_Y +0x000089A4 VGT_COMPUTE_START_Z +0x000089AC VGT_COMPUTE_THREAD_GOURP_SIZE +0x00008A14 PA_CL_ENHANCE +0x00008A60 PA_SC_LINE_STIPPLE_VALUE +0x00008B10 PA_SC_LINE_STIPPLE_STATE +0x00008BF0 PA_SC_ENHANCE +0x00008D8C SQ_DYN_GPR_CNTL_PS_FLUSH_REQ +0x00008C00 SQ_CONFIG +0x00008C04 SQ_GPR_RESOURCE_MGMT_1 +0x00008C08 SQ_GPR_RESOURCE_MGMT_2 +0x00008C0C SQ_GPR_RESOURCE_MGMT_3 +0x00008C10 SQ_GLOBAL_GPR_RESOURCE_MGMT_1 +0x00008C14 SQ_GLOBAL_GPR_RESOURCE_MGMT_2 +0x00008C18 SQ_THREAD_RESOURCE_MGMT +0x00008C1C SQ_THREAD_RESOURCE_MGMT_2 +0x00008C20 SQ_STACK_RESOURCE_MGMT_1 +0x00008C24 SQ_STACK_RESOURCE_MGMT_2 +0x00008C28 SQ_STACK_RESOURCE_MGMT_3 +0x00008DF8 SQ_CONST_MEM_BASE +0x00008E48 SQ_EX_ALLOC_TABLE_SLOTS +0x00009100 SPI_CONFIG_CNTL +0x0000913C SPI_CONFIG_CNTL_1 +0x00009700 VC_CNTL +0x00009714 VC_ENHANCE +0x00009830 DB_DEBUG +0x00009834 DB_DEBUG2 +0x00009838 DB_DEBUG3 +0x0000983C DB_DEBUG4 +0x00009854 DB_WATERMARKS +0x0000A400 TD_PS_BORDER_COLOR_INDEX +0x0000A404 TD_PS_BORDER_COLOR_RED +0x0000A408 TD_PS_BORDER_COLOR_GREEN +0x0000A40C TD_PS_BORDER_COLOR_BLUE +0x0000A410 TD_PS_BORDER_COLOR_ALPHA +0x0000A414 TD_VS_BORDER_COLOR_INDEX +0x0000A418 TD_VS_BORDER_COLOR_RED +0x0000A41C TD_VS_BORDER_COLOR_GREEN +0x0000A420 TD_VS_BORDER_COLOR_BLUE +0x0000A424 TD_VS_BORDER_COLOR_ALPHA +0x0000A428 TD_GS_BORDER_COLOR_INDEX +0x0000A42C TD_GS_BORDER_COLOR_RED +0x0000A430 TD_GS_BORDER_COLOR_GREEN +0x0000A434 TD_GS_BORDER_COLOR_BLUE +0x0000A438 TD_GS_BORDER_COLOR_ALPHA +0x0000A43C TD_HS_BORDER_COLOR_INDEX +0x0000A440 TD_HS_BORDER_COLOR_RED +0x0000A444 TD_HS_BORDER_COLOR_GREEN +0x0000A448 TD_HS_BORDER_COLOR_BLUE +0x0000A44C TD_HS_BORDER_COLOR_ALPHA +0x0000A450 TD_LS_BORDER_COLOR_INDEX +0x0000A454 TD_LS_BORDER_COLOR_RED +0x0000A458 TD_LS_BORDER_COLOR_GREEN +0x0000A45C TD_LS_BORDER_COLOR_BLUE +0x0000A460 TD_LS_BORDER_COLOR_ALPHA +0x0000A464 TD_CS_BORDER_COLOR_INDEX +0x0000A468 TD_CS_BORDER_COLOR_RED +0x0000A46C TD_CS_BORDER_COLOR_GREEN +0x0000A470 TD_CS_BORDER_COLOR_BLUE +0x0000A474 TD_CS_BORDER_COLOR_ALPHA +0x00028000 DB_RENDER_CONTROL +0x00028004 DB_COUNT_CONTROL +0x0002800C DB_RENDER_OVERRIDE +0x00028010 DB_RENDER_OVERRIDE2 +0x00028028 DB_STENCIL_CLEAR +0x0002802C DB_DEPTH_CLEAR +0x00028030 PA_SC_SCREEN_SCISSOR_TL +0x00028034 PA_SC_SCREEN_SCISSOR_BR +0x0002805C DB_DEPTH_SLICE +0x00028140 SQ_ALU_CONST_BUFFER_SIZE_PS_0 +0x00028144 SQ_ALU_CONST_BUFFER_SIZE_PS_1 +0x00028148 SQ_ALU_CONST_BUFFER_SIZE_PS_2 +0x0002814C SQ_ALU_CONST_BUFFER_SIZE_PS_3 +0x00028150 SQ_ALU_CONST_BUFFER_SIZE_PS_4 +0x00028154 SQ_ALU_CONST_BUFFER_SIZE_PS_5 +0x00028158 SQ_ALU_CONST_BUFFER_SIZE_PS_6 +0x0002815C SQ_ALU_CONST_BUFFER_SIZE_PS_7 +0x00028160 SQ_ALU_CONST_BUFFER_SIZE_PS_8 +0x00028164 SQ_ALU_CONST_BUFFER_SIZE_PS_9 +0x00028168 SQ_ALU_CONST_BUFFER_SIZE_PS_10 +0x0002816C SQ_ALU_CONST_BUFFER_SIZE_PS_11 +0x00028170 SQ_ALU_CONST_BUFFER_SIZE_PS_12 +0x00028174 SQ_ALU_CONST_BUFFER_SIZE_PS_13 +0x00028178 SQ_ALU_CONST_BUFFER_SIZE_PS_14 +0x0002817C SQ_ALU_CONST_BUFFER_SIZE_PS_15 +0x00028180 SQ_ALU_CONST_BUFFER_SIZE_VS_0 +0x00028184 SQ_ALU_CONST_BUFFER_SIZE_VS_1 +0x00028188 SQ_ALU_CONST_BUFFER_SIZE_VS_2 +0x0002818C SQ_ALU_CONST_BUFFER_SIZE_VS_3 +0x00028190 SQ_ALU_CONST_BUFFER_SIZE_VS_4 +0x00028194 SQ_ALU_CONST_BUFFER_SIZE_VS_5 +0x00028198 SQ_ALU_CONST_BUFFER_SIZE_VS_6 +0x0002819C SQ_ALU_CONST_BUFFER_SIZE_VS_7 +0x000281A0 SQ_ALU_CONST_BUFFER_SIZE_VS_8 +0x000281A4 SQ_ALU_CONST_BUFFER_SIZE_VS_9 +0x000281A8 SQ_ALU_CONST_BUFFER_SIZE_VS_10 +0x000281AC SQ_ALU_CONST_BUFFER_SIZE_VS_11 +0x000281B0 SQ_ALU_CONST_BUFFER_SIZE_VS_12 +0x000281B4 SQ_ALU_CONST_BUFFER_SIZE_VS_13 +0x000281B8 SQ_ALU_CONST_BUFFER_SIZE_VS_14 +0x000281BC SQ_ALU_CONST_BUFFER_SIZE_VS_15 +0x000281C0 SQ_ALU_CONST_BUFFER_SIZE_GS_0 +0x000281C4 SQ_ALU_CONST_BUFFER_SIZE_GS_1 +0x000281C8 SQ_ALU_CONST_BUFFER_SIZE_GS_2 +0x000281CC SQ_ALU_CONST_BUFFER_SIZE_GS_3 +0x000281D0 SQ_ALU_CONST_BUFFER_SIZE_GS_4 +0x000281D4 SQ_ALU_CONST_BUFFER_SIZE_GS_5 +0x000281D8 SQ_ALU_CONST_BUFFER_SIZE_GS_6 +0x000281DC SQ_ALU_CONST_BUFFER_SIZE_GS_7 +0x000281E0 SQ_ALU_CONST_BUFFER_SIZE_GS_8 +0x000281E4 SQ_ALU_CONST_BUFFER_SIZE_GS_9 +0x000281E8 SQ_ALU_CONST_BUFFER_SIZE_GS_10 +0x000281EC SQ_ALU_CONST_BUFFER_SIZE_GS_11 +0x000281F0 SQ_ALU_CONST_BUFFER_SIZE_GS_12 +0x000281F4 SQ_ALU_CONST_BUFFER_SIZE_GS_13 +0x000281F8 SQ_ALU_CONST_BUFFER_SIZE_GS_14 +0x000281FC SQ_ALU_CONST_BUFFER_SIZE_GS_15 +0x00028200 PA_SC_WINDOW_OFFSET +0x00028204 PA_SC_WINDOW_SCISSOR_TL +0x00028208 PA_SC_WINDOW_SCISSOR_BR +0x0002820C PA_SC_CLIPRECT_RULE +0x00028210 PA_SC_CLIPRECT_0_TL +0x00028214 PA_SC_CLIPRECT_0_BR +0x00028218 PA_SC_CLIPRECT_1_TL +0x0002821C PA_SC_CLIPRECT_1_BR +0x00028220 PA_SC_CLIPRECT_2_TL +0x00028224 PA_SC_CLIPRECT_2_BR +0x00028228 PA_SC_CLIPRECT_3_TL +0x0002822C PA_SC_CLIPRECT_3_BR +0x00028230 PA_SC_EDGERULE +0x00028234 PA_SU_HARDWARE_SCREEN_OFFSET +0x00028240 PA_SC_GENERIC_SCISSOR_TL +0x00028244 PA_SC_GENERIC_SCISSOR_BR +0x00028250 PA_SC_VPORT_SCISSOR_0_TL +0x00028254 PA_SC_VPORT_SCISSOR_0_BR +0x00028258 PA_SC_VPORT_SCISSOR_1_TL +0x0002825C PA_SC_VPORT_SCISSOR_1_BR +0x00028260 PA_SC_VPORT_SCISSOR_2_TL +0x00028264 PA_SC_VPORT_SCISSOR_2_BR +0x00028268 PA_SC_VPORT_SCISSOR_3_TL +0x0002826C PA_SC_VPORT_SCISSOR_3_BR +0x00028270 PA_SC_VPORT_SCISSOR_4_TL +0x00028274 PA_SC_VPORT_SCISSOR_4_BR +0x00028278 PA_SC_VPORT_SCISSOR_5_TL +0x0002827C PA_SC_VPORT_SCISSOR_5_BR +0x00028280 PA_SC_VPORT_SCISSOR_6_TL +0x00028284 PA_SC_VPORT_SCISSOR_6_BR +0x00028288 PA_SC_VPORT_SCISSOR_7_TL +0x0002828C PA_SC_VPORT_SCISSOR_7_BR +0x00028290 PA_SC_VPORT_SCISSOR_8_TL +0x00028294 PA_SC_VPORT_SCISSOR_8_BR +0x00028298 PA_SC_VPORT_SCISSOR_9_TL +0x0002829C PA_SC_VPORT_SCISSOR_9_BR +0x000282A0 PA_SC_VPORT_SCISSOR_10_TL +0x000282A4 PA_SC_VPORT_SCISSOR_10_BR +0x000282A8 PA_SC_VPORT_SCISSOR_11_TL +0x000282AC PA_SC_VPORT_SCISSOR_11_BR +0x000282B0 PA_SC_VPORT_SCISSOR_12_TL +0x000282B4 PA_SC_VPORT_SCISSOR_12_BR +0x000282B8 PA_SC_VPORT_SCISSOR_13_TL +0x000282BC PA_SC_VPORT_SCISSOR_13_BR +0x000282C0 PA_SC_VPORT_SCISSOR_14_TL +0x000282C4 PA_SC_VPORT_SCISSOR_14_BR +0x000282C8 PA_SC_VPORT_SCISSOR_15_TL +0x000282CC PA_SC_VPORT_SCISSOR_15_BR +0x000282D0 PA_SC_VPORT_ZMIN_0 +0x000282D4 PA_SC_VPORT_ZMAX_0 +0x000282D8 PA_SC_VPORT_ZMIN_1 +0x000282DC PA_SC_VPORT_ZMAX_1 +0x000282E0 PA_SC_VPORT_ZMIN_2 +0x000282E4 PA_SC_VPORT_ZMAX_2 +0x000282E8 PA_SC_VPORT_ZMIN_3 +0x000282EC PA_SC_VPORT_ZMAX_3 +0x000282F0 PA_SC_VPORT_ZMIN_4 +0x000282F4 PA_SC_VPORT_ZMAX_4 +0x000282F8 PA_SC_VPORT_ZMIN_5 +0x000282FC PA_SC_VPORT_ZMAX_5 +0x00028300 PA_SC_VPORT_ZMIN_6 +0x00028304 PA_SC_VPORT_ZMAX_6 +0x00028308 PA_SC_VPORT_ZMIN_7 +0x0002830C PA_SC_VPORT_ZMAX_7 +0x00028310 PA_SC_VPORT_ZMIN_8 +0x00028314 PA_SC_VPORT_ZMAX_8 +0x00028318 PA_SC_VPORT_ZMIN_9 +0x0002831C PA_SC_VPORT_ZMAX_9 +0x00028320 PA_SC_VPORT_ZMIN_10 +0x00028324 PA_SC_VPORT_ZMAX_10 +0x00028328 PA_SC_VPORT_ZMIN_11 +0x0002832C PA_SC_VPORT_ZMAX_11 +0x00028330 PA_SC_VPORT_ZMIN_12 +0x00028334 PA_SC_VPORT_ZMAX_12 +0x00028338 PA_SC_VPORT_ZMIN_13 +0x0002833C PA_SC_VPORT_ZMAX_13 +0x00028340 PA_SC_VPORT_ZMIN_14 +0x00028344 PA_SC_VPORT_ZMAX_14 +0x00028348 PA_SC_VPORT_ZMIN_15 +0x0002834C PA_SC_VPORT_ZMAX_15 +0x00028350 SX_MISC +0x00028380 SQ_VTX_SEMANTIC_0 +0x00028384 SQ_VTX_SEMANTIC_1 +0x00028388 SQ_VTX_SEMANTIC_2 +0x0002838C SQ_VTX_SEMANTIC_3 +0x00028390 SQ_VTX_SEMANTIC_4 +0x00028394 SQ_VTX_SEMANTIC_5 +0x00028398 SQ_VTX_SEMANTIC_6 +0x0002839C SQ_VTX_SEMANTIC_7 +0x000283A0 SQ_VTX_SEMANTIC_8 +0x000283A4 SQ_VTX_SEMANTIC_9 +0x000283A8 SQ_VTX_SEMANTIC_10 +0x000283AC SQ_VTX_SEMANTIC_11 +0x000283B0 SQ_VTX_SEMANTIC_12 +0x000283B4 SQ_VTX_SEMANTIC_13 +0x000283B8 SQ_VTX_SEMANTIC_14 +0x000283BC SQ_VTX_SEMANTIC_15 +0x000283C0 SQ_VTX_SEMANTIC_16 +0x000283C4 SQ_VTX_SEMANTIC_17 +0x000283C8 SQ_VTX_SEMANTIC_18 +0x000283CC SQ_VTX_SEMANTIC_19 +0x000283D0 SQ_VTX_SEMANTIC_20 +0x000283D4 SQ_VTX_SEMANTIC_21 +0x000283D8 SQ_VTX_SEMANTIC_22 +0x000283DC SQ_VTX_SEMANTIC_23 +0x000283E0 SQ_VTX_SEMANTIC_24 +0x000283E4 SQ_VTX_SEMANTIC_25 +0x000283E8 SQ_VTX_SEMANTIC_26 +0x000283EC SQ_VTX_SEMANTIC_27 +0x000283F0 SQ_VTX_SEMANTIC_28 +0x000283F4 SQ_VTX_SEMANTIC_29 +0x000283F8 SQ_VTX_SEMANTIC_30 +0x000283FC SQ_VTX_SEMANTIC_31 +0x00028400 VGT_MAX_VTX_INDX +0x00028404 VGT_MIN_VTX_INDX +0x00028408 VGT_INDX_OFFSET +0x0002840C VGT_MULTI_PRIM_IB_RESET_INDX +0x00028410 SX_ALPHA_TEST_CONTROL +0x00028414 CB_BLEND_RED +0x00028418 CB_BLEND_GREEN +0x0002841C CB_BLEND_BLUE +0x00028420 CB_BLEND_ALPHA +0x00028430 DB_STENCILREFMASK +0x00028434 DB_STENCILREFMASK_BF +0x00028438 SX_ALPHA_REF +0x0002843C PA_CL_VPORT_XSCALE_0 +0x00028440 PA_CL_VPORT_XOFFSET_0 +0x00028444 PA_CL_VPORT_YSCALE_0 +0x00028448 PA_CL_VPORT_YOFFSET_0 +0x0002844C PA_CL_VPORT_ZSCALE_0 +0x00028450 PA_CL_VPORT_ZOFFSET_0 +0x00028454 PA_CL_VPORT_XSCALE_1 +0x00028458 PA_CL_VPORT_XOFFSET_1 +0x0002845C PA_CL_VPORT_YSCALE_1 +0x00028460 PA_CL_VPORT_YOFFSET_1 +0x00028464 PA_CL_VPORT_ZSCALE_1 +0x00028468 PA_CL_VPORT_ZOFFSET_1 +0x0002846C PA_CL_VPORT_XSCALE_2 +0x00028470 PA_CL_VPORT_XOFFSET_2 +0x00028474 PA_CL_VPORT_YSCALE_2 +0x00028478 PA_CL_VPORT_YOFFSET_2 +0x0002847C PA_CL_VPORT_ZSCALE_2 +0x00028480 PA_CL_VPORT_ZOFFSET_2 +0x00028484 PA_CL_VPORT_XSCALE_3 +0x00028488 PA_CL_VPORT_XOFFSET_3 +0x0002848C PA_CL_VPORT_YSCALE_3 +0x00028490 PA_CL_VPORT_YOFFSET_3 +0x00028494 PA_CL_VPORT_ZSCALE_3 +0x00028498 PA_CL_VPORT_ZOFFSET_3 +0x0002849C PA_CL_VPORT_XSCALE_4 +0x000284A0 PA_CL_VPORT_XOFFSET_4 +0x000284A4 PA_CL_VPORT_YSCALE_4 +0x000284A8 PA_CL_VPORT_YOFFSET_4 +0x000284AC PA_CL_VPORT_ZSCALE_4 +0x000284B0 PA_CL_VPORT_ZOFFSET_4 +0x000284B4 PA_CL_VPORT_XSCALE_5 +0x000284B8 PA_CL_VPORT_XOFFSET_5 +0x000284BC PA_CL_VPORT_YSCALE_5 +0x000284C0 PA_CL_VPORT_YOFFSET_5 +0x000284C4 PA_CL_VPORT_ZSCALE_5 +0x000284C8 PA_CL_VPORT_ZOFFSET_5 +0x000284CC PA_CL_VPORT_XSCALE_6 +0x000284D0 PA_CL_VPORT_XOFFSET_6 +0x000284D4 PA_CL_VPORT_YSCALE_6 +0x000284D8 PA_CL_VPORT_YOFFSET_6 +0x000284DC PA_CL_VPORT_ZSCALE_6 +0x000284E0 PA_CL_VPORT_ZOFFSET_6 +0x000284E4 PA_CL_VPORT_XSCALE_7 +0x000284E8 PA_CL_VPORT_XOFFSET_7 +0x000284EC PA_CL_VPORT_YSCALE_7 +0x000284F0 PA_CL_VPORT_YOFFSET_7 +0x000284F4 PA_CL_VPORT_ZSCALE_7 +0x000284F8 PA_CL_VPORT_ZOFFSET_7 +0x000284FC PA_CL_VPORT_XSCALE_8 +0x00028500 PA_CL_VPORT_XOFFSET_8 +0x00028504 PA_CL_VPORT_YSCALE_8 +0x00028508 PA_CL_VPORT_YOFFSET_8 +0x0002850C PA_CL_VPORT_ZSCALE_8 +0x00028510 PA_CL_VPORT_ZOFFSET_8 +0x00028514 PA_CL_VPORT_XSCALE_9 +0x00028518 PA_CL_VPORT_XOFFSET_9 +0x0002851C PA_CL_VPORT_YSCALE_9 +0x00028520 PA_CL_VPORT_YOFFSET_9 +0x00028524 PA_CL_VPORT_ZSCALE_9 +0x00028528 PA_CL_VPORT_ZOFFSET_9 +0x0002852C PA_CL_VPORT_XSCALE_10 +0x00028530 PA_CL_VPORT_XOFFSET_10 +0x00028534 PA_CL_VPORT_YSCALE_10 +0x00028538 PA_CL_VPORT_YOFFSET_10 +0x0002853C PA_CL_VPORT_ZSCALE_10 +0x00028540 PA_CL_VPORT_ZOFFSET_10 +0x00028544 PA_CL_VPORT_XSCALE_11 +0x00028548 PA_CL_VPORT_XOFFSET_11 +0x0002854C PA_CL_VPORT_YSCALE_11 +0x00028550 PA_CL_VPORT_YOFFSET_11 +0x00028554 PA_CL_VPORT_ZSCALE_11 +0x00028558 PA_CL_VPORT_ZOFFSET_11 +0x0002855C PA_CL_VPORT_XSCALE_12 +0x00028560 PA_CL_VPORT_XOFFSET_12 +0x00028564 PA_CL_VPORT_YSCALE_12 +0x00028568 PA_CL_VPORT_YOFFSET_12 +0x0002856C PA_CL_VPORT_ZSCALE_12 +0x00028570 PA_CL_VPORT_ZOFFSET_12 +0x00028574 PA_CL_VPORT_XSCALE_13 +0x00028578 PA_CL_VPORT_XOFFSET_13 +0x0002857C PA_CL_VPORT_YSCALE_13 +0x00028580 PA_CL_VPORT_YOFFSET_13 +0x00028584 PA_CL_VPORT_ZSCALE_13 +0x00028588 PA_CL_VPORT_ZOFFSET_13 +0x0002858C PA_CL_VPORT_XSCALE_14 +0x00028590 PA_CL_VPORT_XOFFSET_14 +0x00028594 PA_CL_VPORT_YSCALE_14 +0x00028598 PA_CL_VPORT_YOFFSET_14 +0x0002859C PA_CL_VPORT_ZSCALE_14 +0x000285A0 PA_CL_VPORT_ZOFFSET_14 +0x000285A4 PA_CL_VPORT_XSCALE_15 +0x000285A8 PA_CL_VPORT_XOFFSET_15 +0x000285AC PA_CL_VPORT_YSCALE_15 +0x000285B0 PA_CL_VPORT_YOFFSET_15 +0x000285B4 PA_CL_VPORT_ZSCALE_15 +0x000285B8 PA_CL_VPORT_ZOFFSET_15 +0x000285BC PA_CL_UCP_0_X +0x000285C0 PA_CL_UCP_0_Y +0x000285C4 PA_CL_UCP_0_Z +0x000285C8 PA_CL_UCP_0_W +0x000285CC PA_CL_UCP_1_X +0x000285D0 PA_CL_UCP_1_Y +0x000285D4 PA_CL_UCP_1_Z +0x000285D8 PA_CL_UCP_1_W +0x000285DC PA_CL_UCP_2_X +0x000285E0 PA_CL_UCP_2_Y +0x000285E4 PA_CL_UCP_2_Z +0x000285E8 PA_CL_UCP_2_W +0x000285EC PA_CL_UCP_3_X +0x000285F0 PA_CL_UCP_3_Y +0x000285F4 PA_CL_UCP_3_Z +0x000285F8 PA_CL_UCP_3_W +0x000285FC PA_CL_UCP_4_X +0x00028600 PA_CL_UCP_4_Y +0x00028604 PA_CL_UCP_4_Z +0x00028608 PA_CL_UCP_4_W +0x0002860C PA_CL_UCP_5_X +0x00028610 PA_CL_UCP_5_Y +0x00028614 PA_CL_UCP_5_Z +0x00028618 PA_CL_UCP_5_W +0x0002861C SPI_VS_OUT_ID_0 +0x00028620 SPI_VS_OUT_ID_1 +0x00028624 SPI_VS_OUT_ID_2 +0x00028628 SPI_VS_OUT_ID_3 +0x0002862C SPI_VS_OUT_ID_4 +0x00028630 SPI_VS_OUT_ID_5 +0x00028634 SPI_VS_OUT_ID_6 +0x00028638 SPI_VS_OUT_ID_7 +0x0002863C SPI_VS_OUT_ID_8 +0x00028640 SPI_VS_OUT_ID_9 +0x00028644 SPI_PS_INPUT_CNTL_0 +0x00028648 SPI_PS_INPUT_CNTL_1 +0x0002864C SPI_PS_INPUT_CNTL_2 +0x00028650 SPI_PS_INPUT_CNTL_3 +0x00028654 SPI_PS_INPUT_CNTL_4 +0x00028658 SPI_PS_INPUT_CNTL_5 +0x0002865C SPI_PS_INPUT_CNTL_6 +0x00028660 SPI_PS_INPUT_CNTL_7 +0x00028664 SPI_PS_INPUT_CNTL_8 +0x00028668 SPI_PS_INPUT_CNTL_9 +0x0002866C SPI_PS_INPUT_CNTL_10 +0x00028670 SPI_PS_INPUT_CNTL_11 +0x00028674 SPI_PS_INPUT_CNTL_12 +0x00028678 SPI_PS_INPUT_CNTL_13 +0x0002867C SPI_PS_INPUT_CNTL_14 +0x00028680 SPI_PS_INPUT_CNTL_15 +0x00028684 SPI_PS_INPUT_CNTL_16 +0x00028688 SPI_PS_INPUT_CNTL_17 +0x0002868C SPI_PS_INPUT_CNTL_18 +0x00028690 SPI_PS_INPUT_CNTL_19 +0x00028694 SPI_PS_INPUT_CNTL_20 +0x00028698 SPI_PS_INPUT_CNTL_21 +0x0002869C SPI_PS_INPUT_CNTL_22 +0x000286A0 SPI_PS_INPUT_CNTL_23 +0x000286A4 SPI_PS_INPUT_CNTL_24 +0x000286A8 SPI_PS_INPUT_CNTL_25 +0x000286AC SPI_PS_INPUT_CNTL_26 +0x000286B0 SPI_PS_INPUT_CNTL_27 +0x000286B4 SPI_PS_INPUT_CNTL_28 +0x000286B8 SPI_PS_INPUT_CNTL_29 +0x000286BC SPI_PS_INPUT_CNTL_30 +0x000286C0 SPI_PS_INPUT_CNTL_31 +0x000286C4 SPI_VS_OUT_CONFIG +0x000286C8 SPI_THREAD_GROUPING +0x000286CC SPI_PS_IN_CONTROL_0 +0x000286D0 SPI_PS_IN_CONTROL_1 +0x000286D4 SPI_INTERP_CONTROL_0 +0x000286D8 SPI_INPUT_Z +0x000286DC SPI_FOG_CNTL +0x000286E0 SPI_BARYC_CNTL +0x000286E4 SPI_PS_IN_CONTROL_2 +0x000286E8 SPI_COMPUTE_INPUT_CNTL +0x000286EC SPI_COMPUTE_NUM_THREAD_X +0x000286F0 SPI_COMPUTE_NUM_THREAD_Y +0x000286F4 SPI_COMPUTE_NUM_THREAD_Z +0x000286F8 GDS_ADDR_SIZE +0x00028780 CB_BLEND0_CONTROL +0x00028784 CB_BLEND1_CONTROL +0x00028788 CB_BLEND2_CONTROL +0x0002878C CB_BLEND3_CONTROL +0x00028790 CB_BLEND4_CONTROL +0x00028794 CB_BLEND5_CONTROL +0x00028798 CB_BLEND6_CONTROL +0x0002879C CB_BLEND7_CONTROL +0x000287CC CS_COPY_STATE +0x000287D0 GFX_COPY_STATE +0x000287D4 PA_CL_POINT_X_RAD +0x000287D8 PA_CL_POINT_Y_RAD +0x000287DC PA_CL_POINT_SIZE +0x000287E0 PA_CL_POINT_CULL_RAD +0x00028808 CB_COLOR_CONTROL +0x0002880C DB_SHADER_CONTROL +0x00028810 PA_CL_CLIP_CNTL +0x00028814 PA_SU_SC_MODE_CNTL +0x00028818 PA_CL_VTE_CNTL +0x0002881C PA_CL_VS_OUT_CNTL +0x00028820 PA_CL_NANINF_CNTL +0x00028824 PA_SU_LINE_STIPPLE_CNTL +0x00028828 PA_SU_LINE_STIPPLE_SCALE +0x0002882C PA_SU_PRIM_FILTER_CNTL +0x00028838 SQ_DYN_GPR_RESOURCE_LIMIT_1 +0x00028844 SQ_PGM_RESOURCES_PS +0x00028848 SQ_PGM_RESOURCES_2_PS +0x0002884C SQ_PGM_EXPORTS_PS +0x00028860 SQ_PGM_RESOURCES_VS +0x00028864 SQ_PGM_RESOURCES_2_VS +0x00028878 SQ_PGM_RESOURCES_GS +0x0002887C SQ_PGM_RESOURCES_2_GS +0x00028890 SQ_PGM_RESOURCES_ES +0x00028894 SQ_PGM_RESOURCES_2_ES +0x000288A8 SQ_PGM_RESOURCES_FS +0x000288BC SQ_PGM_RESOURCES_HS +0x000288C0 SQ_PGM_RESOURCES_2_HS +0x000288D4 SQ_PGM_RESOURCES_LS +0x000288D8 SQ_PGM_RESOURCES_2_LS +0x000288E8 SQ_LDS_ALLOC +0x000288EC SQ_LDS_ALLOC_PS +0x000288F0 SQ_VTX_SEMANTIC_CLEAR +0x00028A00 PA_SU_POINT_SIZE +0x00028A04 PA_SU_POINT_MINMAX +0x00028A08 PA_SU_LINE_CNTL +0x00028A0C PA_SC_LINE_STIPPLE +0x00028A10 VGT_OUTPUT_PATH_CNTL +0x00028A14 VGT_HOS_CNTL +0x00028A18 VGT_HOS_MAX_TESS_LEVEL +0x00028A1C VGT_HOS_MIN_TESS_LEVEL +0x00028A20 VGT_HOS_REUSE_DEPTH +0x00028A24 VGT_GROUP_PRIM_TYPE +0x00028A28 VGT_GROUP_FIRST_DECR +0x00028A2C VGT_GROUP_DECR +0x00028A30 VGT_GROUP_VECT_0_CNTL +0x00028A34 VGT_GROUP_VECT_1_CNTL +0x00028A38 VGT_GROUP_VECT_0_FMT_CNTL +0x00028A3C VGT_GROUP_VECT_1_FMT_CNTL +0x00028A40 VGT_GS_MODE +0x00028A48 PA_SC_MODE_CNTL_0 +0x00028A4C PA_SC_MODE_CNTL_1 +0x00028A50 VGT_ENHANCE +0x00028A54 VGT_GS_PER_ES +0x00028A58 VGT_ES_PER_GS +0x00028A5C VGT_GS_PER_VS +0x00028A6C VGT_GS_OUT_PRIM_TYPE +0x00028A84 VGT_PRIMITIVEID_EN +0x00028A94 VGT_MULTI_PRIM_IB_RESET_EN +0x00028AA0 VGT_INSTANCE_STEP_RATE_0 +0x00028AA4 VGT_INSTANCE_STEP_RATE_1 +0x00028AB4 VGT_REUSE_OFF +0x00028AB8 VGT_VTX_CNT_EN +0x00028ABC DB_HTILE_SURFACE +0x00028AC0 DB_SRESULTS_COMPARE_STATE0 +0x00028AC4 DB_SRESULTS_COMPARE_STATE1 +0x00028AC8 DB_PRELOAD_CONTROL +0x00028B38 VGT_GS_MAX_VERT_OUT +0x00028B54 VGT_SHADER_STAGES_EN +0x00028B58 VGT_LS_HS_CONFIG +0x00028B5C VGT_LS_SIZE +0x00028B60 VGT_HS_SIZE +0x00028B64 VGT_LS_HS_ALLOC +0x00028B68 VGT_HS_PATCH_CONST +0x00028B6C VGT_TF_PARAM +0x00028B70 DB_ALPHA_TO_MASK +0x00028B74 VGT_DISPATCH_INITIATOR +0x00028B78 PA_SU_POLY_OFFSET_DB_FMT_CNTL +0x00028B7C PA_SU_POLY_OFFSET_CLAMP +0x00028B80 PA_SU_POLY_OFFSET_FRONT_SCALE +0x00028B84 PA_SU_POLY_OFFSET_FRONT_OFFSET +0x00028B88 PA_SU_POLY_OFFSET_BACK_SCALE +0x00028B8C PA_SU_POLY_OFFSET_BACK_OFFSET +0x00028B74 VGT_GS_INSTANCE_CNT +0x00028C00 PA_SC_LINE_CNTL +0x00028C08 PA_SU_VTX_CNTL +0x00028C0C PA_CL_GB_VERT_CLIP_ADJ +0x00028C10 PA_CL_GB_VERT_DISC_ADJ +0x00028C14 PA_CL_GB_HORZ_CLIP_ADJ +0x00028C18 PA_CL_GB_HORZ_DISC_ADJ +0x00028C1C PA_SC_AA_SAMPLE_LOCS_0 +0x00028C20 PA_SC_AA_SAMPLE_LOCS_1 +0x00028C24 PA_SC_AA_SAMPLE_LOCS_2 +0x00028C28 PA_SC_AA_SAMPLE_LOCS_3 +0x00028C2C PA_SC_AA_SAMPLE_LOCS_4 +0x00028C30 PA_SC_AA_SAMPLE_LOCS_5 +0x00028C34 PA_SC_AA_SAMPLE_LOCS_6 +0x00028C38 PA_SC_AA_SAMPLE_LOCS_7 +0x00028C3C PA_SC_AA_MASK +0x00028C8C CB_COLOR0_CLEAR_WORD0 +0x00028C90 CB_COLOR0_CLEAR_WORD1 +0x00028C94 CB_COLOR0_CLEAR_WORD2 +0x00028C98 CB_COLOR0_CLEAR_WORD3 +0x00028CC8 CB_COLOR1_CLEAR_WORD0 +0x00028CCC CB_COLOR1_CLEAR_WORD1 +0x00028CD0 CB_COLOR1_CLEAR_WORD2 +0x00028CD4 CB_COLOR1_CLEAR_WORD3 +0x00028D04 CB_COLOR2_CLEAR_WORD0 +0x00028D08 CB_COLOR2_CLEAR_WORD1 +0x00028D0C CB_COLOR2_CLEAR_WORD2 +0x00028D10 CB_COLOR2_CLEAR_WORD3 +0x00028D40 CB_COLOR3_CLEAR_WORD0 +0x00028D44 CB_COLOR3_CLEAR_WORD1 +0x00028D48 CB_COLOR3_CLEAR_WORD2 +0x00028D4C CB_COLOR3_CLEAR_WORD3 +0x00028D7C CB_COLOR4_CLEAR_WORD0 +0x00028D80 CB_COLOR4_CLEAR_WORD1 +0x00028D84 CB_COLOR4_CLEAR_WORD2 +0x00028D88 CB_COLOR4_CLEAR_WORD3 +0x00028DB8 CB_COLOR5_CLEAR_WORD0 +0x00028DBC CB_COLOR5_CLEAR_WORD1 +0x00028DC0 CB_COLOR5_CLEAR_WORD2 +0x00028DC4 CB_COLOR5_CLEAR_WORD3 +0x00028DF4 CB_COLOR6_CLEAR_WORD0 +0x00028DF8 CB_COLOR6_CLEAR_WORD1 +0x00028DFC CB_COLOR6_CLEAR_WORD2 +0x00028E00 CB_COLOR6_CLEAR_WORD3 +0x00028E30 CB_COLOR7_CLEAR_WORD0 +0x00028E34 CB_COLOR7_CLEAR_WORD1 +0x00028E38 CB_COLOR7_CLEAR_WORD2 +0x00028E3C CB_COLOR7_CLEAR_WORD3 +0x00028F80 SQ_ALU_CONST_BUFFER_SIZE_HS_0 +0x00028F84 SQ_ALU_CONST_BUFFER_SIZE_HS_1 +0x00028F88 SQ_ALU_CONST_BUFFER_SIZE_HS_2 +0x00028F8C SQ_ALU_CONST_BUFFER_SIZE_HS_3 +0x00028F90 SQ_ALU_CONST_BUFFER_SIZE_HS_4 +0x00028F94 SQ_ALU_CONST_BUFFER_SIZE_HS_5 +0x00028F98 SQ_ALU_CONST_BUFFER_SIZE_HS_6 +0x00028F9C SQ_ALU_CONST_BUFFER_SIZE_HS_7 +0x00028FA0 SQ_ALU_CONST_BUFFER_SIZE_HS_8 +0x00028FA4 SQ_ALU_CONST_BUFFER_SIZE_HS_9 +0x00028FA8 SQ_ALU_CONST_BUFFER_SIZE_HS_10 +0x00028FAC SQ_ALU_CONST_BUFFER_SIZE_HS_11 +0x00028FB0 SQ_ALU_CONST_BUFFER_SIZE_HS_12 +0x00028FB4 SQ_ALU_CONST_BUFFER_SIZE_HS_13 +0x00028FB8 SQ_ALU_CONST_BUFFER_SIZE_HS_14 +0x00028FBC SQ_ALU_CONST_BUFFER_SIZE_HS_15 +0x00028FC0 SQ_ALU_CONST_BUFFER_SIZE_LS_0 +0x00028FC4 SQ_ALU_CONST_BUFFER_SIZE_LS_1 +0x00028FC8 SQ_ALU_CONST_BUFFER_SIZE_LS_2 +0x00028FCC SQ_ALU_CONST_BUFFER_SIZE_LS_3 +0x00028FD0 SQ_ALU_CONST_BUFFER_SIZE_LS_4 +0x00028FD4 SQ_ALU_CONST_BUFFER_SIZE_LS_5 +0x00028FD8 SQ_ALU_CONST_BUFFER_SIZE_LS_6 +0x00028FDC SQ_ALU_CONST_BUFFER_SIZE_LS_7 +0x00028FE0 SQ_ALU_CONST_BUFFER_SIZE_LS_8 +0x00028FE4 SQ_ALU_CONST_BUFFER_SIZE_LS_9 +0x00028FE8 SQ_ALU_CONST_BUFFER_SIZE_LS_10 +0x00028FEC SQ_ALU_CONST_BUFFER_SIZE_LS_11 +0x00028FF0 SQ_ALU_CONST_BUFFER_SIZE_LS_12 +0x00028FF4 SQ_ALU_CONST_BUFFER_SIZE_LS_13 +0x00028FF8 SQ_ALU_CONST_BUFFER_SIZE_LS_14 +0x00028FFC SQ_ALU_CONST_BUFFER_SIZE_LS_15 +0x0003CFF0 SQ_VTX_BASE_VTX_LOC +0x0003CFF4 SQ_VTX_START_INST_LOC +0x0003FF00 SQ_TEX_SAMPLER_CLEAR +0x0003FF04 SQ_TEX_RESOURCE_CLEAR +0x0003FF08 SQ_LOOP_BOOL_CLEAR diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c index 340c7611f2ac..5ce3ccc7a423 100644 --- a/drivers/gpu/drm/radeon/rs600.c +++ b/drivers/gpu/drm/radeon/rs600.c @@ -74,7 +74,8 @@ void rs600_pm_misc(struct radeon_device *rdev) if (voltage->delay) udelay(voltage->delay); } - } + } else if (voltage->type == VOLTAGE_VDDC) + radeon_atom_set_voltage(rdev, voltage->vddc_id); dyn_pwrmgt_sclk_length = RREG32_PLL(DYN_PWRMGT_SCLK_LENGTH); dyn_pwrmgt_sclk_length &= ~REDUCED_POWER_SCLK_HILEN(0xf); diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c index a18ba98885f3..5fea094ed8cb 100644 --- a/drivers/gpu/drm/radeon/rs690.c +++ b/drivers/gpu/drm/radeon/rs690.c @@ -79,7 +79,13 @@ void rs690_pm_info(struct radeon_device *rdev) tmp.full = dfixed_const(100); rdev->pm.igp_sideport_mclk.full = dfixed_const(info->info.ulBootUpMemoryClock); rdev->pm.igp_sideport_mclk.full = dfixed_div(rdev->pm.igp_sideport_mclk, tmp); - rdev->pm.igp_system_mclk.full = dfixed_const(le16_to_cpu(info->info.usK8MemoryClock)); + if (info->info.usK8MemoryClock) + rdev->pm.igp_system_mclk.full = dfixed_const(le16_to_cpu(info->info.usK8MemoryClock)); + else if (rdev->clock.default_mclk) { + rdev->pm.igp_system_mclk.full = dfixed_const(rdev->clock.default_mclk); + rdev->pm.igp_system_mclk.full = dfixed_div(rdev->pm.igp_system_mclk, tmp); + } else + rdev->pm.igp_system_mclk.full = dfixed_const(400); rdev->pm.igp_ht_link_clk.full = dfixed_const(le16_to_cpu(info->info.usFSBClock)); rdev->pm.igp_ht_link_width.full = dfixed_const(info->info.ucHTLinkWidth); break; @@ -87,34 +93,31 @@ void rs690_pm_info(struct radeon_device *rdev) tmp.full = dfixed_const(100); rdev->pm.igp_sideport_mclk.full = dfixed_const(info->info_v2.ulBootUpSidePortClock); rdev->pm.igp_sideport_mclk.full = dfixed_div(rdev->pm.igp_sideport_mclk, tmp); - rdev->pm.igp_system_mclk.full = dfixed_const(info->info_v2.ulBootUpUMAClock); + if (info->info_v2.ulBootUpUMAClock) + rdev->pm.igp_system_mclk.full = dfixed_const(info->info_v2.ulBootUpUMAClock); + else if (rdev->clock.default_mclk) + rdev->pm.igp_system_mclk.full = dfixed_const(rdev->clock.default_mclk); + else + rdev->pm.igp_system_mclk.full = dfixed_const(66700); rdev->pm.igp_system_mclk.full = dfixed_div(rdev->pm.igp_system_mclk, tmp); rdev->pm.igp_ht_link_clk.full = dfixed_const(info->info_v2.ulHTLinkFreq); rdev->pm.igp_ht_link_clk.full = dfixed_div(rdev->pm.igp_ht_link_clk, tmp); rdev->pm.igp_ht_link_width.full = dfixed_const(le16_to_cpu(info->info_v2.usMinHTLinkWidth)); break; default: - tmp.full = dfixed_const(100); /* We assume the slower possible clock ie worst case */ - /* DDR 333Mhz */ - rdev->pm.igp_sideport_mclk.full = dfixed_const(333); - /* FIXME: system clock ? */ - rdev->pm.igp_system_mclk.full = dfixed_const(100); - rdev->pm.igp_system_mclk.full = dfixed_div(rdev->pm.igp_system_mclk, tmp); - rdev->pm.igp_ht_link_clk.full = dfixed_const(200); + rdev->pm.igp_sideport_mclk.full = dfixed_const(200); + rdev->pm.igp_system_mclk.full = dfixed_const(200); + rdev->pm.igp_ht_link_clk.full = dfixed_const(1000); rdev->pm.igp_ht_link_width.full = dfixed_const(8); DRM_ERROR("No integrated system info for your GPU, using safe default\n"); break; } } else { - tmp.full = dfixed_const(100); /* We assume the slower possible clock ie worst case */ - /* DDR 333Mhz */ - rdev->pm.igp_sideport_mclk.full = dfixed_const(333); - /* FIXME: system clock ? */ - rdev->pm.igp_system_mclk.full = dfixed_const(100); - rdev->pm.igp_system_mclk.full = dfixed_div(rdev->pm.igp_system_mclk, tmp); - rdev->pm.igp_ht_link_clk.full = dfixed_const(200); + rdev->pm.igp_sideport_mclk.full = dfixed_const(200); + rdev->pm.igp_system_mclk.full = dfixed_const(200); + rdev->pm.igp_ht_link_clk.full = dfixed_const(1000); rdev->pm.igp_ht_link_width.full = dfixed_const(8); DRM_ERROR("No integrated system info for your GPU, using safe default\n"); } @@ -228,10 +231,6 @@ void rs690_crtc_bandwidth_compute(struct radeon_device *rdev, fixed20_12 a, b, c; fixed20_12 pclk, request_fifo_depth, tolerable_latency, estimated_width; fixed20_12 consumption_time, line_time, chunk_time, read_delay_latency; - /* FIXME: detect IGP with sideport memory, i don't think there is any - * such product available - */ - bool sideport = false; if (!crtc->base.enabled) { /* FIXME: wouldn't it better to set priority mark to maximum */ @@ -300,7 +299,7 @@ void rs690_crtc_bandwidth_compute(struct radeon_device *rdev, /* Maximun bandwidth is the minimun bandwidth of all component */ rdev->pm.max_bandwidth = rdev->pm.core_bandwidth; - if (sideport) { + if (rdev->mc.igp_sideport_enabled) { if (rdev->pm.max_bandwidth.full > rdev->pm.sideport_bandwidth.full && rdev->pm.sideport_bandwidth.full) rdev->pm.max_bandwidth = rdev->pm.sideport_bandwidth; diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c index 5c7f0b97c6aa..6a7bf1091971 100644 --- a/drivers/gpu/drm/radeon/rv770.c +++ b/drivers/gpu/drm/radeon/rv770.c @@ -44,7 +44,18 @@ void rv770_fini(struct radeon_device *rdev); void rv770_pm_misc(struct radeon_device *rdev) { - + int req_ps_idx = rdev->pm.requested_power_state_index; + int req_cm_idx = rdev->pm.requested_clock_mode_index; + struct radeon_power_state *ps = &rdev->pm.power_state[req_ps_idx]; + struct radeon_voltage *voltage = &ps->clock_info[req_cm_idx].voltage; + + if ((voltage->type == VOLTAGE_SW) && voltage->voltage) { + if (voltage->voltage != rdev->pm.current_vddc) { + radeon_atom_set_voltage(rdev, voltage->voltage); + rdev->pm.current_vddc = voltage->voltage; + DRM_DEBUG("Setting: v: %d\n", voltage->voltage); + } + } } /* @@ -213,7 +224,7 @@ static void rv770_mc_program(struct radeon_device *rdev) WREG32(MC_VM_FB_LOCATION, tmp); WREG32(HDP_NONSURFACE_BASE, (rdev->mc.vram_start >> 8)); WREG32(HDP_NONSURFACE_INFO, (2 << 7)); - WREG32(HDP_NONSURFACE_SIZE, (rdev->mc.mc_vram_size - 1) | 0x3FF); + WREG32(HDP_NONSURFACE_SIZE, 0x3FFFFFFF); if (rdev->flags & RADEON_IS_AGP) { WREG32(MC_VM_AGP_TOP, rdev->mc.gtt_end >> 16); WREG32(MC_VM_AGP_BOT, rdev->mc.gtt_start >> 16); diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 4c2299299ab2..555ebb12ace8 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -1718,40 +1718,12 @@ int ttm_bo_wait(struct ttm_buffer_object *bo, } EXPORT_SYMBOL(ttm_bo_wait); -void ttm_bo_unblock_reservation(struct ttm_buffer_object *bo) -{ - atomic_set(&bo->reserved, 0); - wake_up_all(&bo->event_queue); -} - -int ttm_bo_block_reservation(struct ttm_buffer_object *bo, bool interruptible, - bool no_wait) -{ - int ret; - - while (unlikely(atomic_cmpxchg(&bo->reserved, 0, 1) != 0)) { - if (no_wait) - return -EBUSY; - else if (interruptible) { - ret = wait_event_interruptible - (bo->event_queue, atomic_read(&bo->reserved) == 0); - if (unlikely(ret != 0)) - return ret; - } else { - wait_event(bo->event_queue, - atomic_read(&bo->reserved) == 0); - } - } - return 0; -} - int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait) { int ret = 0; /* - * Using ttm_bo_reserve instead of ttm_bo_block_reservation - * makes sure the lru lists are updated. + * Using ttm_bo_reserve makes sure the lru lists are updated. */ ret = ttm_bo_reserve(bo, true, no_wait, false, 0); diff --git a/drivers/gpu/drm/ttm/ttm_lock.c b/drivers/gpu/drm/ttm/ttm_lock.c index 3d172ef04ee1..de41e55a944a 100644 --- a/drivers/gpu/drm/ttm/ttm_lock.c +++ b/drivers/gpu/drm/ttm/ttm_lock.c @@ -204,7 +204,6 @@ static int __ttm_vt_unlock(struct ttm_lock *lock) lock->flags &= ~TTM_VT_LOCK; wake_up_all(&lock->queue); spin_unlock(&lock->lock); - printk(KERN_INFO TTM_PFX "vt unlock.\n"); return ret; } @@ -265,10 +264,8 @@ int ttm_vt_lock(struct ttm_lock *lock, ttm_lock_type, &ttm_vt_lock_remove, NULL); if (ret) (void)__ttm_vt_unlock(lock); - else { + else lock->vt_holder = tfile; - printk(KERN_INFO TTM_PFX "vt lock.\n"); - } return ret; } diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c index 0d9a42c2394f..b1d67dc973dc 100644 --- a/drivers/gpu/drm/ttm/ttm_page_alloc.c +++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c @@ -77,7 +77,7 @@ struct ttm_page_pool { /** * Limits for the pool. They are handled without locks because only place where * they may change is in sysfs store. They won't have immediate effect anyway - * so forcing serialiazation to access them is pointless. + * so forcing serialization to access them is pointless. */ struct ttm_pool_opts { @@ -104,7 +104,6 @@ struct ttm_pool_opts { struct ttm_pool_manager { struct kobject kobj; struct shrinker mm_shrink; - atomic_t page_alloc_inited; struct ttm_pool_opts options; union { @@ -142,7 +141,7 @@ static void ttm_pool_kobj_release(struct kobject *kobj) { struct ttm_pool_manager *m = container_of(kobj, struct ttm_pool_manager, kobj); - (void)m; + kfree(m); } static ssize_t ttm_pool_store(struct kobject *kobj, @@ -165,16 +164,18 @@ static ssize_t ttm_pool_store(struct kobject *kobj, m->options.small = val; else if (attr == &ttm_page_pool_alloc_size) { if (val > NUM_PAGES_TO_ALLOC*8) { - printk(KERN_ERR "[ttm] Setting allocation size to %lu " - "is not allowed. Recomended size is " - "%lu\n", - NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 7), - NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10)); + printk(KERN_ERR TTM_PFX + "Setting allocation size to %lu " + "is not allowed. Recommended size is " + "%lu\n", + NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 7), + NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10)); return size; } else if (val > NUM_PAGES_TO_ALLOC) { - printk(KERN_WARNING "[ttm] Setting allocation size to " - "larger than %lu is not recomended.\n", - NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10)); + printk(KERN_WARNING TTM_PFX + "Setting allocation size to " + "larger than %lu is not recommended.\n", + NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10)); } m->options.alloc_size = val; } @@ -212,9 +213,7 @@ static struct kobj_type ttm_pool_kobj_type = { .default_attrs = ttm_pool_attrs, }; -static struct ttm_pool_manager _manager = { - .page_alloc_inited = ATOMIC_INIT(0) -}; +static struct ttm_pool_manager *_manager; #ifndef CONFIG_X86 static int set_pages_array_wb(struct page **pages, int addrinarray) @@ -269,7 +268,7 @@ static struct ttm_page_pool *ttm_get_pool(int flags, if (flags & TTM_PAGE_FLAG_DMA32) pool_index |= 0x2; - return &_manager.pools[pool_index]; + return &_manager->pools[pool_index]; } /* set memory back to wb and free the pages. */ @@ -277,7 +276,7 @@ static void ttm_pages_put(struct page *pages[], unsigned npages) { unsigned i; if (set_pages_array_wb(pages, npages)) - printk(KERN_ERR "[ttm] Failed to set %d pages to wb!\n", + printk(KERN_ERR TTM_PFX "Failed to set %d pages to wb!\n", npages); for (i = 0; i < npages; ++i) __free_page(pages[i]); @@ -313,7 +312,8 @@ static int ttm_page_pool_free(struct ttm_page_pool *pool, unsigned nr_free) pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), GFP_KERNEL); if (!pages_to_free) { - printk(KERN_ERR "Failed to allocate memory for pool free operation.\n"); + printk(KERN_ERR TTM_PFX + "Failed to allocate memory for pool free operation.\n"); return 0; } @@ -384,13 +384,13 @@ static int ttm_pool_get_num_unused_pages(void) unsigned i; int total = 0; for (i = 0; i < NUM_POOLS; ++i) - total += _manager.pools[i].npages; + total += _manager->pools[i].npages; return total; } /** - * Calback for mm to request pool to reduce number of page held. + * Callback for mm to request pool to reduce number of page held. */ static int ttm_pool_mm_shrink(int shrink_pages, gfp_t gfp_mask) { @@ -405,7 +405,7 @@ static int ttm_pool_mm_shrink(int shrink_pages, gfp_t gfp_mask) unsigned nr_free = shrink_pages; if (shrink_pages == 0) break; - pool = &_manager.pools[(i + pool_offset)%NUM_POOLS]; + pool = &_manager->pools[(i + pool_offset)%NUM_POOLS]; shrink_pages = ttm_page_pool_free(pool, nr_free); } /* return estimated number of unused pages in pool */ @@ -433,14 +433,16 @@ static int ttm_set_pages_caching(struct page **pages, case tt_uncached: r = set_pages_array_uc(pages, cpages); if (r) - printk(KERN_ERR "[ttm] Failed to set %d pages to uc!\n", - cpages); + printk(KERN_ERR TTM_PFX + "Failed to set %d pages to uc!\n", + cpages); break; case tt_wc: r = set_pages_array_wc(pages, cpages); if (r) - printk(KERN_ERR "[ttm] Failed to set %d pages to wc!\n", - cpages); + printk(KERN_ERR TTM_PFX + "Failed to set %d pages to wc!\n", + cpages); break; default: break; @@ -458,7 +460,7 @@ static void ttm_handle_caching_state_failure(struct list_head *pages, struct page **failed_pages, unsigned cpages) { unsigned i; - /* Failed pages has to be reed */ + /* Failed pages have to be freed */ for (i = 0; i < cpages; ++i) { list_del(&failed_pages[i]->lru); __free_page(failed_pages[i]); @@ -485,7 +487,8 @@ static int ttm_alloc_new_pages(struct list_head *pages, int gfp_flags, caching_array = kmalloc(max_cpages*sizeof(struct page *), GFP_KERNEL); if (!caching_array) { - printk(KERN_ERR "[ttm] unable to allocate table for new pages."); + printk(KERN_ERR TTM_PFX + "Unable to allocate table for new pages."); return -ENOMEM; } @@ -493,12 +496,13 @@ static int ttm_alloc_new_pages(struct list_head *pages, int gfp_flags, p = alloc_page(gfp_flags); if (!p) { - printk(KERN_ERR "[ttm] unable to get page %u\n", i); + printk(KERN_ERR TTM_PFX "Unable to get page %u.\n", i); /* store already allocated pages in the pool after * setting the caching state */ if (cpages) { - r = ttm_set_pages_caching(caching_array, cstate, cpages); + r = ttm_set_pages_caching(caching_array, + cstate, cpages); if (r) ttm_handle_caching_state_failure(pages, ttm_flags, cstate, @@ -569,10 +573,10 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool, /* If allocation request is small and there is not enough * pages in pool we fill the pool first */ - if (count < _manager.options.small + if (count < _manager->options.small && count > pool->npages) { struct list_head new_pages; - unsigned alloc_size = _manager.options.alloc_size; + unsigned alloc_size = _manager->options.alloc_size; /** * Can't change page caching if in irqsave context. We have to @@ -590,7 +594,8 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool, ++pool->nrefills; pool->npages += alloc_size; } else { - printk(KERN_ERR "[ttm] Failed to fill pool (%p).", pool); + printk(KERN_ERR TTM_PFX + "Failed to fill pool (%p).", pool); /* If we have any pages left put them to the pool. */ list_for_each_entry(p, &pool->list, lru) { ++cpages; @@ -659,7 +664,7 @@ int ttm_get_pages(struct list_head *pages, int flags, { struct ttm_page_pool *pool = ttm_get_pool(flags, cstate); struct page *p = NULL; - int gfp_flags = 0; + int gfp_flags = GFP_USER; int r; /* set zero flag for page allocation if required */ @@ -671,13 +676,14 @@ int ttm_get_pages(struct list_head *pages, int flags, if (flags & TTM_PAGE_FLAG_DMA32) gfp_flags |= GFP_DMA32; else - gfp_flags |= __GFP_HIGHMEM; + gfp_flags |= GFP_HIGHUSER; for (r = 0; r < count; ++r) { p = alloc_page(gfp_flags); if (!p) { - printk(KERN_ERR "[ttm] unable to allocate page."); + printk(KERN_ERR TTM_PFX + "Unable to allocate page."); return -ENOMEM; } @@ -709,8 +715,9 @@ int ttm_get_pages(struct list_head *pages, int flags, if (r) { /* If there is any pages in the list put them back to * the pool. */ - printk(KERN_ERR "[ttm] Failed to allocate extra pages " - "for large request."); + printk(KERN_ERR TTM_PFX + "Failed to allocate extra pages " + "for large request."); ttm_put_pages(pages, 0, flags, cstate); return r; } @@ -749,8 +756,8 @@ void ttm_put_pages(struct list_head *pages, unsigned page_count, int flags, pool->npages += page_count; /* Check that we don't go over the pool limit */ page_count = 0; - if (pool->npages > _manager.options.max_size) { - page_count = pool->npages - _manager.options.max_size; + if (pool->npages > _manager->options.max_size) { + page_count = pool->npages - _manager->options.max_size; /* free at least NUM_PAGES_TO_ALLOC number of pages * to reduce calls to set_memory_wb */ if (page_count < NUM_PAGES_TO_ALLOC) @@ -775,33 +782,36 @@ static void ttm_page_pool_init_locked(struct ttm_page_pool *pool, int flags, int ttm_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages) { int ret; - if (atomic_add_return(1, &_manager.page_alloc_inited) > 1) - return 0; - printk(KERN_INFO "[ttm] Initializing pool allocator.\n"); + WARN_ON(_manager); + + printk(KERN_INFO TTM_PFX "Initializing pool allocator.\n"); - ttm_page_pool_init_locked(&_manager.wc_pool, GFP_HIGHUSER, "wc"); + _manager = kzalloc(sizeof(*_manager), GFP_KERNEL); - ttm_page_pool_init_locked(&_manager.uc_pool, GFP_HIGHUSER, "uc"); + ttm_page_pool_init_locked(&_manager->wc_pool, GFP_HIGHUSER, "wc"); - ttm_page_pool_init_locked(&_manager.wc_pool_dma32, GFP_USER | GFP_DMA32, - "wc dma"); + ttm_page_pool_init_locked(&_manager->uc_pool, GFP_HIGHUSER, "uc"); - ttm_page_pool_init_locked(&_manager.uc_pool_dma32, GFP_USER | GFP_DMA32, - "uc dma"); + ttm_page_pool_init_locked(&_manager->wc_pool_dma32, + GFP_USER | GFP_DMA32, "wc dma"); - _manager.options.max_size = max_pages; - _manager.options.small = SMALL_ALLOCATION; - _manager.options.alloc_size = NUM_PAGES_TO_ALLOC; + ttm_page_pool_init_locked(&_manager->uc_pool_dma32, + GFP_USER | GFP_DMA32, "uc dma"); - kobject_init(&_manager.kobj, &ttm_pool_kobj_type); - ret = kobject_add(&_manager.kobj, &glob->kobj, "pool"); + _manager->options.max_size = max_pages; + _manager->options.small = SMALL_ALLOCATION; + _manager->options.alloc_size = NUM_PAGES_TO_ALLOC; + + ret = kobject_init_and_add(&_manager->kobj, &ttm_pool_kobj_type, + &glob->kobj, "pool"); if (unlikely(ret != 0)) { - kobject_put(&_manager.kobj); + kobject_put(&_manager->kobj); + _manager = NULL; return ret; } - ttm_pool_mm_shrink_init(&_manager); + ttm_pool_mm_shrink_init(_manager); return 0; } @@ -810,16 +820,14 @@ void ttm_page_alloc_fini() { int i; - if (atomic_sub_return(1, &_manager.page_alloc_inited) > 0) - return; - - printk(KERN_INFO "[ttm] Finilizing pool allocator.\n"); - ttm_pool_mm_shrink_fini(&_manager); + printk(KERN_INFO TTM_PFX "Finalizing pool allocator.\n"); + ttm_pool_mm_shrink_fini(_manager); for (i = 0; i < NUM_POOLS; ++i) - ttm_page_pool_free(&_manager.pools[i], FREE_ALL_PAGES); + ttm_page_pool_free(&_manager->pools[i], FREE_ALL_PAGES); - kobject_put(&_manager.kobj); + kobject_put(&_manager->kobj); + _manager = NULL; } int ttm_page_alloc_debugfs(struct seq_file *m, void *data) @@ -827,14 +835,14 @@ int ttm_page_alloc_debugfs(struct seq_file *m, void *data) struct ttm_page_pool *p; unsigned i; char *h[] = {"pool", "refills", "pages freed", "size"}; - if (atomic_read(&_manager.page_alloc_inited) == 0) { + if (!_manager) { seq_printf(m, "No pool allocator running.\n"); return 0; } seq_printf(m, "%6s %12s %13s %8s\n", h[0], h[1], h[2], h[3]); for (i = 0; i < NUM_POOLS; ++i) { - p = &_manager.pools[i]; + p = &_manager->pools[i]; seq_printf(m, "%6s %12ld %13ld %8d\n", p->name, p->nrefills, diff --git a/drivers/gpu/drm/vmwgfx/Makefile b/drivers/gpu/drm/vmwgfx/Makefile index 1a3cb6816d1c..4505e17df3f5 100644 --- a/drivers/gpu/drm/vmwgfx/Makefile +++ b/drivers/gpu/drm/vmwgfx/Makefile @@ -4,6 +4,6 @@ ccflags-y := -Iinclude/drm vmwgfx-y := vmwgfx_execbuf.o vmwgfx_gmr.o vmwgfx_kms.o vmwgfx_drv.o \ vmwgfx_fb.o vmwgfx_ioctl.o vmwgfx_resource.o vmwgfx_buffer.o \ vmwgfx_fifo.o vmwgfx_irq.o vmwgfx_ldu.o vmwgfx_ttm_glue.o \ - vmwgfx_overlay.o + vmwgfx_overlay.o vmwgfx_fence.o obj-$(CONFIG_DRM_VMWGFX) := vmwgfx.o diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index f7f248dbff58..9dd395b90216 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -88,6 +88,9 @@ #define DRM_IOCTL_VMW_FENCE_WAIT \ DRM_IOWR(DRM_COMMAND_BASE + DRM_VMW_FENCE_WAIT, \ struct drm_vmw_fence_wait_arg) +#define DRM_IOCTL_VMW_UPDATE_LAYOUT \ + DRM_IOWR(DRM_COMMAND_BASE + DRM_VMW_UPDATE_LAYOUT, \ + struct drm_vmw_update_layout_arg) /** @@ -135,7 +138,9 @@ static struct drm_ioctl_desc vmw_ioctls[] = { VMW_IOCTL_DEF(DRM_IOCTL_VMW_FIFO_DEBUG, vmw_fifo_debug_ioctl, DRM_AUTH | DRM_ROOT_ONLY | DRM_MASTER | DRM_UNLOCKED), VMW_IOCTL_DEF(DRM_IOCTL_VMW_FENCE_WAIT, vmw_fence_wait_ioctl, - DRM_AUTH | DRM_UNLOCKED) + DRM_AUTH | DRM_UNLOCKED), + VMW_IOCTL_DEF(DRM_IOCTL_VMW_UPDATE_LAYOUT, vmw_kms_update_layout_ioctl, + DRM_MASTER | DRM_CONTROL_ALLOW | DRM_UNLOCKED) }; static struct pci_device_id vmw_pci_id_list[] = { @@ -318,6 +323,15 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) goto out_err3; } + /* Need mmio memory to check for fifo pitchlock cap. */ + if (!(dev_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY) && + !(dev_priv->capabilities & SVGA_CAP_PITCHLOCK) && + !vmw_fifo_have_pitchlock(dev_priv)) { + ret = -ENOSYS; + DRM_ERROR("Hardware has no pitchlock\n"); + goto out_err4; + } + dev_priv->tdev = ttm_object_device_init (dev_priv->mem_global_ref.object, 12); @@ -399,8 +413,6 @@ static int vmw_driver_unload(struct drm_device *dev) { struct vmw_private *dev_priv = vmw_priv(dev); - DRM_INFO(VMWGFX_DRIVER_NAME " unload.\n"); - unregister_pm_notifier(&dev_priv->pm_nb); vmw_fb_close(dev_priv); @@ -546,7 +558,6 @@ static int vmw_master_create(struct drm_device *dev, { struct vmw_master *vmaster; - DRM_INFO("Master create.\n"); vmaster = kzalloc(sizeof(*vmaster), GFP_KERNEL); if (unlikely(vmaster == NULL)) return -ENOMEM; @@ -563,7 +574,6 @@ static void vmw_master_destroy(struct drm_device *dev, { struct vmw_master *vmaster = vmw_master(master); - DRM_INFO("Master destroy.\n"); master->driver_priv = NULL; kfree(vmaster); } @@ -579,8 +589,6 @@ static int vmw_master_set(struct drm_device *dev, struct vmw_master *vmaster = vmw_master(file_priv->master); int ret = 0; - DRM_INFO("Master set.\n"); - if (active) { BUG_ON(active != &dev_priv->fbdev_master); ret = ttm_vt_lock(&active->lock, false, vmw_fp->tfile); @@ -622,8 +630,6 @@ static void vmw_master_drop(struct drm_device *dev, struct vmw_master *vmaster = vmw_master(file_priv->master); int ret; - DRM_INFO("Master drop.\n"); - /** * Make sure the master doesn't disappear while we have * it locked. diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index 356dc935ec13..eaad52095339 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -41,12 +41,13 @@ #define VMWGFX_DRIVER_DATE "20100209" #define VMWGFX_DRIVER_MAJOR 1 -#define VMWGFX_DRIVER_MINOR 0 +#define VMWGFX_DRIVER_MINOR 2 #define VMWGFX_DRIVER_PATCHLEVEL 0 #define VMWGFX_FILE_PAGE_OFFSET 0x00100000 #define VMWGFX_FIFO_STATIC_SIZE (1024*1024) #define VMWGFX_MAX_RELOCATIONS 2048 #define VMWGFX_MAX_GMRS 2048 +#define VMWGFX_MAX_DISPLAYS 16 struct vmw_fpriv { struct drm_master *locked_master; @@ -102,6 +103,13 @@ struct vmw_surface { struct vmw_cursor_snooper snooper; }; +struct vmw_fence_queue { + struct list_head head; + struct timespec lag; + struct timespec lag_time; + spinlock_t lock; +}; + struct vmw_fifo_state { unsigned long reserved_size; __le32 *dynamic_buffer; @@ -115,6 +123,7 @@ struct vmw_fifo_state { uint32_t capabilities; struct mutex fifo_mutex; struct rw_semaphore rwsem; + struct vmw_fence_queue fence_queue; }; struct vmw_relocation { @@ -144,6 +153,14 @@ struct vmw_master { struct ttm_lock lock; }; +struct vmw_vga_topology_state { + uint32_t width; + uint32_t height; + uint32_t primary; + uint32_t pos_x; + uint32_t pos_y; +}; + struct vmw_private { struct ttm_bo_device bdev; struct ttm_bo_global_ref bo_global_ref; @@ -171,14 +188,19 @@ struct vmw_private { * VGA registers. */ + struct vmw_vga_topology_state vga_save[VMWGFX_MAX_DISPLAYS]; uint32_t vga_width; uint32_t vga_height; uint32_t vga_depth; uint32_t vga_bpp; uint32_t vga_pseudo; uint32_t vga_red_mask; - uint32_t vga_blue_mask; uint32_t vga_green_mask; + uint32_t vga_blue_mask; + uint32_t vga_bpl; + uint32_t vga_pitchlock; + + uint32_t num_displays; /* * Framebuffer info. @@ -393,6 +415,7 @@ extern int vmw_fifo_send_fence(struct vmw_private *dev_priv, extern void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason); extern int vmw_fifo_mmap(struct file *filp, struct vm_area_struct *vma); extern bool vmw_fifo_have_3d(struct vmw_private *dev_priv); +extern bool vmw_fifo_have_pitchlock(struct vmw_private *dev_priv); /** * TTM glue - vmwgfx_ttm_glue.c @@ -441,6 +464,23 @@ extern int vmw_fallback_wait(struct vmw_private *dev_priv, uint32_t sequence, bool interruptible, unsigned long timeout); +extern void vmw_update_sequence(struct vmw_private *dev_priv, + struct vmw_fifo_state *fifo_state); + + +/** + * Rudimentary fence objects currently used only for throttling - + * vmwgfx_fence.c + */ + +extern void vmw_fence_queue_init(struct vmw_fence_queue *queue); +extern void vmw_fence_queue_takedown(struct vmw_fence_queue *queue); +extern int vmw_fence_push(struct vmw_fence_queue *queue, + uint32_t sequence); +extern int vmw_fence_pull(struct vmw_fence_queue *queue, + uint32_t signaled_sequence); +extern int vmw_wait_lag(struct vmw_private *dev_priv, + struct vmw_fence_queue *queue, uint32_t us); /** * Kernel framebuffer - vmwgfx_fb.c @@ -466,6 +506,11 @@ void vmw_kms_cursor_snoop(struct vmw_surface *srf, struct ttm_object_file *tfile, struct ttm_buffer_object *bo, SVGA3dCmdHeader *header); +void vmw_kms_write_svga(struct vmw_private *vmw_priv, + unsigned width, unsigned height, unsigned pitch, + unsigned bbp, unsigned depth); +int vmw_kms_update_layout_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); /** * Overlay control - vmwgfx_overlay.c diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c index dbd36b8910cf..8e396850513c 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c @@ -644,6 +644,7 @@ int vmw_execbuf_ioctl(struct drm_device *dev, void *data, ret = copy_from_user(cmd, user_cmd, arg->command_size); if (unlikely(ret != 0)) { + ret = -EFAULT; DRM_ERROR("Failed copying commands.\n"); goto out_commit; } @@ -669,6 +670,15 @@ int vmw_execbuf_ioctl(struct drm_device *dev, void *data, goto out_err; vmw_apply_relocations(sw_context); + + if (arg->throttle_us) { + ret = vmw_wait_lag(dev_priv, &dev_priv->fifo.fence_queue, + arg->throttle_us); + + if (unlikely(ret != 0)) + goto out_err; + } + vmw_fifo_commit(dev_priv, arg->command_size); ret = vmw_fifo_send_fence(dev_priv, &sequence); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c index 7421aaad8d09..b0866f04ec76 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c @@ -132,16 +132,14 @@ static int vmw_fb_check_var(struct fb_var_screeninfo *var, return -EINVAL; } - /* without multimon its hard to resize */ - if (!(vmw_priv->capabilities & SVGA_CAP_MULTIMON) && - (var->xres != par->max_width || - var->yres != par->max_height)) { - DRM_ERROR("Tried to resize, but we don't have multimon\n"); + if (!(vmw_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY) && + (var->xoffset != 0 || var->yoffset != 0)) { + DRM_ERROR("Can not handle panning without display topology\n"); return -EINVAL; } - if (var->xres > par->max_width || - var->yres > par->max_height) { + if ((var->xoffset + var->xres) > par->max_width || + (var->yoffset + var->yres) > par->max_height) { DRM_ERROR("Requested geom can not fit in framebuffer\n"); return -EINVAL; } @@ -154,27 +152,11 @@ static int vmw_fb_set_par(struct fb_info *info) struct vmw_fb_par *par = info->par; struct vmw_private *vmw_priv = par->vmw_priv; - if (vmw_priv->capabilities & SVGA_CAP_MULTIMON) { - vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY, true); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_X, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_Y, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_WIDTH, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_HEIGHT, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID); - - vmw_write(vmw_priv, SVGA_REG_ENABLE, 1); - vmw_write(vmw_priv, SVGA_REG_WIDTH, par->max_width); - vmw_write(vmw_priv, SVGA_REG_HEIGHT, par->max_height); - vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, par->bpp); - vmw_write(vmw_priv, SVGA_REG_DEPTH, par->depth); - vmw_write(vmw_priv, SVGA_REG_RED_MASK, 0x00ff0000); - vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, 0x0000ff00); - vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, 0x000000ff); - + vmw_kms_write_svga(vmw_priv, info->var.xres, info->var.yres, + info->fix.line_length, + par->bpp, par->depth); + if (vmw_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY) { /* TODO check if pitch and offset changes */ - vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1); vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, 0); vmw_write(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY, true); @@ -183,13 +165,13 @@ static int vmw_fb_set_par(struct fb_info *info) vmw_write(vmw_priv, SVGA_REG_DISPLAY_WIDTH, info->var.xres); vmw_write(vmw_priv, SVGA_REG_DISPLAY_HEIGHT, info->var.yres); vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID); - } else { - vmw_write(vmw_priv, SVGA_REG_WIDTH, info->var.xres); - vmw_write(vmw_priv, SVGA_REG_HEIGHT, info->var.yres); - - /* TODO check if pitch and offset changes */ } + /* This is really helpful since if this fails the user + * can probably not see anything on the screen. + */ + WARN_ON(vmw_read(vmw_priv, SVGA_REG_FB_OFFSET) != 0); + return 0; } @@ -416,48 +398,23 @@ int vmw_fb_init(struct vmw_private *vmw_priv) unsigned fb_bbp, fb_depth, fb_offset, fb_pitch, fb_size; int ret; + /* XXX These shouldn't be hardcoded. */ initial_width = 800; initial_height = 600; fb_bbp = 32; fb_depth = 24; - if (vmw_priv->capabilities & SVGA_CAP_MULTIMON) { - fb_width = min(vmw_priv->fb_max_width, (unsigned)2048); - fb_height = min(vmw_priv->fb_max_height, (unsigned)2048); - } else { - fb_width = min(vmw_priv->fb_max_width, initial_width); - fb_height = min(vmw_priv->fb_max_height, initial_height); - } + /* XXX As shouldn't these be as well. */ + fb_width = min(vmw_priv->fb_max_width, (unsigned)2048); + fb_height = min(vmw_priv->fb_max_height, (unsigned)2048); initial_width = min(fb_width, initial_width); initial_height = min(fb_height, initial_height); - vmw_write(vmw_priv, SVGA_REG_WIDTH, fb_width); - vmw_write(vmw_priv, SVGA_REG_HEIGHT, fb_height); - vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, fb_bbp); - vmw_write(vmw_priv, SVGA_REG_DEPTH, fb_depth); - vmw_write(vmw_priv, SVGA_REG_RED_MASK, 0x00ff0000); - vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, 0x0000ff00); - vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, 0x000000ff); - - fb_size = vmw_read(vmw_priv, SVGA_REG_FB_SIZE); + fb_pitch = fb_width * fb_bbp / 8; + fb_size = fb_pitch * fb_height; fb_offset = vmw_read(vmw_priv, SVGA_REG_FB_OFFSET); - fb_pitch = vmw_read(vmw_priv, SVGA_REG_BYTES_PER_LINE); - - DRM_DEBUG("width %u\n", vmw_read(vmw_priv, SVGA_REG_MAX_WIDTH)); - DRM_DEBUG("height %u\n", vmw_read(vmw_priv, SVGA_REG_MAX_HEIGHT)); - DRM_DEBUG("width %u\n", vmw_read(vmw_priv, SVGA_REG_WIDTH)); - DRM_DEBUG("height %u\n", vmw_read(vmw_priv, SVGA_REG_HEIGHT)); - DRM_DEBUG("bpp %u\n", vmw_read(vmw_priv, SVGA_REG_BITS_PER_PIXEL)); - DRM_DEBUG("depth %u\n", vmw_read(vmw_priv, SVGA_REG_DEPTH)); - DRM_DEBUG("bpl %u\n", vmw_read(vmw_priv, SVGA_REG_BYTES_PER_LINE)); - DRM_DEBUG("r mask %08x\n", vmw_read(vmw_priv, SVGA_REG_RED_MASK)); - DRM_DEBUG("g mask %08x\n", vmw_read(vmw_priv, SVGA_REG_GREEN_MASK)); - DRM_DEBUG("b mask %08x\n", vmw_read(vmw_priv, SVGA_REG_BLUE_MASK)); - DRM_DEBUG("fb_offset 0x%08x\n", fb_offset); - DRM_DEBUG("fb_pitch %u\n", fb_pitch); - DRM_DEBUG("fb_size %u kiB\n", fb_size / 1024); info = framebuffer_alloc(sizeof(*par), device); if (!info) @@ -659,6 +616,10 @@ int vmw_dmabuf_to_start_of_vram(struct vmw_private *vmw_priv, goto err_unlock; ret = ttm_bo_validate(bo, &ne_placement, false, false, false); + + /* Could probably bug on */ + WARN_ON(bo->offset != 0); + ttm_bo_unreserve(bo); err_unlock: ttm_write_unlock(&vmw_priv->active_master->lock); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c new file mode 100644 index 000000000000..61eacc1b5ca3 --- /dev/null +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c @@ -0,0 +1,173 @@ +/************************************************************************** + * + * Copyright (C) 2010 VMware, Inc., Palo Alto, CA., USA + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#include "vmwgfx_drv.h" + +struct vmw_fence { + struct list_head head; + uint32_t sequence; + struct timespec submitted; +}; + +void vmw_fence_queue_init(struct vmw_fence_queue *queue) +{ + INIT_LIST_HEAD(&queue->head); + queue->lag = ns_to_timespec(0); + getrawmonotonic(&queue->lag_time); + spin_lock_init(&queue->lock); +} + +void vmw_fence_queue_takedown(struct vmw_fence_queue *queue) +{ + struct vmw_fence *fence, *next; + + spin_lock(&queue->lock); + list_for_each_entry_safe(fence, next, &queue->head, head) { + kfree(fence); + } + spin_unlock(&queue->lock); +} + +int vmw_fence_push(struct vmw_fence_queue *queue, + uint32_t sequence) +{ + struct vmw_fence *fence = kmalloc(sizeof(*fence), GFP_KERNEL); + + if (unlikely(!fence)) + return -ENOMEM; + + fence->sequence = sequence; + getrawmonotonic(&fence->submitted); + spin_lock(&queue->lock); + list_add_tail(&fence->head, &queue->head); + spin_unlock(&queue->lock); + + return 0; +} + +int vmw_fence_pull(struct vmw_fence_queue *queue, + uint32_t signaled_sequence) +{ + struct vmw_fence *fence, *next; + struct timespec now; + bool updated = false; + + spin_lock(&queue->lock); + getrawmonotonic(&now); + + if (list_empty(&queue->head)) { + queue->lag = ns_to_timespec(0); + queue->lag_time = now; + updated = true; + goto out_unlock; + } + + list_for_each_entry_safe(fence, next, &queue->head, head) { + if (signaled_sequence - fence->sequence > (1 << 30)) + continue; + + queue->lag = timespec_sub(now, fence->submitted); + queue->lag_time = now; + updated = true; + list_del(&fence->head); + kfree(fence); + } + +out_unlock: + spin_unlock(&queue->lock); + + return (updated) ? 0 : -EBUSY; +} + +static struct timespec vmw_timespec_add(struct timespec t1, + struct timespec t2) +{ + t1.tv_sec += t2.tv_sec; + t1.tv_nsec += t2.tv_nsec; + if (t1.tv_nsec >= 1000000000L) { + t1.tv_sec += 1; + t1.tv_nsec -= 1000000000L; + } + + return t1; +} + +static struct timespec vmw_fifo_lag(struct vmw_fence_queue *queue) +{ + struct timespec now; + + spin_lock(&queue->lock); + getrawmonotonic(&now); + queue->lag = vmw_timespec_add(queue->lag, + timespec_sub(now, queue->lag_time)); + queue->lag_time = now; + spin_unlock(&queue->lock); + return queue->lag; +} + + +static bool vmw_lag_lt(struct vmw_fence_queue *queue, + uint32_t us) +{ + struct timespec lag, cond; + + cond = ns_to_timespec((s64) us * 1000); + lag = vmw_fifo_lag(queue); + return (timespec_compare(&lag, &cond) < 1); +} + +int vmw_wait_lag(struct vmw_private *dev_priv, + struct vmw_fence_queue *queue, uint32_t us) +{ + struct vmw_fence *fence; + uint32_t sequence; + int ret; + + while (!vmw_lag_lt(queue, us)) { + spin_lock(&queue->lock); + if (list_empty(&queue->head)) + sequence = atomic_read(&dev_priv->fence_seq); + else { + fence = list_first_entry(&queue->head, + struct vmw_fence, head); + sequence = fence->sequence; + } + spin_unlock(&queue->lock); + + ret = vmw_wait_fence(dev_priv, false, sequence, true, + 3*HZ); + + if (unlikely(ret != 0)) + return ret; + + (void) vmw_fence_pull(queue, sequence); + } + return 0; +} + + diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c index 39d43a01d846..e6a1eb7ea954 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c @@ -34,6 +34,9 @@ bool vmw_fifo_have_3d(struct vmw_private *dev_priv) __le32 __iomem *fifo_mem = dev_priv->mmio_virt; uint32_t fifo_min, hwversion; + if (!(dev_priv->capabilities & SVGA_CAP_EXTENDED_FIFO)) + return false; + fifo_min = ioread32(fifo_mem + SVGA_FIFO_MIN); if (fifo_min <= SVGA_FIFO_3D_HWVERSION * sizeof(unsigned int)) return false; @@ -48,6 +51,21 @@ bool vmw_fifo_have_3d(struct vmw_private *dev_priv) return true; } +bool vmw_fifo_have_pitchlock(struct vmw_private *dev_priv) +{ + __le32 __iomem *fifo_mem = dev_priv->mmio_virt; + uint32_t caps; + + if (!(dev_priv->capabilities & SVGA_CAP_EXTENDED_FIFO)) + return false; + + caps = ioread32(fifo_mem + SVGA_FIFO_CAPABILITIES); + if (caps & SVGA_FIFO_CAP_PITCHLOCK) + return true; + + return false; +} + int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo) { __le32 __iomem *fifo_mem = dev_priv->mmio_virt; @@ -120,7 +138,7 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo) atomic_set(&dev_priv->fence_seq, dev_priv->last_read_sequence); iowrite32(dev_priv->last_read_sequence, fifo_mem + SVGA_FIFO_FENCE); - + vmw_fence_queue_init(&fifo->fence_queue); return vmw_fifo_send_fence(dev_priv, &dummy); out_err: vfree(fifo->static_buffer); @@ -159,6 +177,7 @@ void vmw_fifo_release(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo) dev_priv->enable_state); mutex_unlock(&dev_priv->hw_mutex); + vmw_fence_queue_takedown(&fifo->fence_queue); if (likely(fifo->last_buffer != NULL)) { vfree(fifo->last_buffer); @@ -484,6 +503,8 @@ int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence) fifo_state->last_buffer_add = true; vmw_fifo_commit(dev_priv, bytes); fifo_state->last_buffer_add = false; + (void) vmw_fence_push(&fifo_state->fence_queue, *sequence); + vmw_update_sequence(dev_priv, fifo_state); out_err: return ret; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c index 4d7cb5393860..e92298a6a383 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c @@ -64,22 +64,33 @@ static bool vmw_fifo_idle(struct vmw_private *dev_priv, uint32_t sequence) return (busy == 0); } +void vmw_update_sequence(struct vmw_private *dev_priv, + struct vmw_fifo_state *fifo_state) +{ + __le32 __iomem *fifo_mem = dev_priv->mmio_virt; + + uint32_t sequence = ioread32(fifo_mem + SVGA_FIFO_FENCE); + + if (dev_priv->last_read_sequence != sequence) { + dev_priv->last_read_sequence = sequence; + vmw_fence_pull(&fifo_state->fence_queue, sequence); + } +} bool vmw_fence_signaled(struct vmw_private *dev_priv, uint32_t sequence) { - __le32 __iomem *fifo_mem = dev_priv->mmio_virt; struct vmw_fifo_state *fifo_state; bool ret; if (likely(dev_priv->last_read_sequence - sequence < VMW_FENCE_WRAP)) return true; - dev_priv->last_read_sequence = ioread32(fifo_mem + SVGA_FIFO_FENCE); + fifo_state = &dev_priv->fifo; + vmw_update_sequence(dev_priv, fifo_state); if (likely(dev_priv->last_read_sequence - sequence < VMW_FENCE_WRAP)) return true; - fifo_state = &dev_priv->fifo; if (!(fifo_state->capabilities & SVGA_FIFO_CAP_FENCE) && vmw_fifo_idle(dev_priv, sequence)) return true; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c index bbc7c4c30bc7..f1d626112415 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c @@ -30,6 +30,8 @@ /* Might need a hrtimer here? */ #define VMWGFX_PRESENT_RATE ((HZ / 60 > 0) ? HZ / 60 : 1) +static int vmw_surface_dmabuf_pin(struct vmw_framebuffer *vfb); +static int vmw_surface_dmabuf_unpin(struct vmw_framebuffer *vfb); void vmw_display_unit_cleanup(struct vmw_display_unit *du) { @@ -326,6 +328,7 @@ int vmw_framebuffer_create_handle(struct drm_framebuffer *fb, struct vmw_framebuffer_surface { struct vmw_framebuffer base; struct vmw_surface *surface; + struct vmw_dma_buffer *buffer; struct delayed_work d_work; struct mutex work_lock; bool present_fs; @@ -500,8 +503,8 @@ int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv, vfbs->base.base.depth = 24; vfbs->base.base.width = width; vfbs->base.base.height = height; - vfbs->base.pin = NULL; - vfbs->base.unpin = NULL; + vfbs->base.pin = &vmw_surface_dmabuf_pin; + vfbs->base.unpin = &vmw_surface_dmabuf_unpin; vfbs->surface = surface; mutex_init(&vfbs->work_lock); INIT_DELAYED_WORK(&vfbs->d_work, &vmw_framebuffer_present_fs_callback); @@ -589,6 +592,40 @@ static struct drm_framebuffer_funcs vmw_framebuffer_dmabuf_funcs = { .create_handle = vmw_framebuffer_create_handle, }; +static int vmw_surface_dmabuf_pin(struct vmw_framebuffer *vfb) +{ + struct vmw_private *dev_priv = vmw_priv(vfb->base.dev); + struct vmw_framebuffer_surface *vfbs = + vmw_framebuffer_to_vfbs(&vfb->base); + unsigned long size = vfbs->base.base.pitch * vfbs->base.base.height; + int ret; + + vfbs->buffer = kzalloc(sizeof(*vfbs->buffer), GFP_KERNEL); + if (unlikely(vfbs->buffer == NULL)) + return -ENOMEM; + + vmw_overlay_pause_all(dev_priv); + ret = vmw_dmabuf_init(dev_priv, vfbs->buffer, size, + &vmw_vram_ne_placement, + false, &vmw_dmabuf_bo_free); + vmw_overlay_resume_all(dev_priv); + + return ret; +} + +static int vmw_surface_dmabuf_unpin(struct vmw_framebuffer *vfb) +{ + struct ttm_buffer_object *bo; + struct vmw_framebuffer_surface *vfbs = + vmw_framebuffer_to_vfbs(&vfb->base); + + bo = &vfbs->buffer->base; + ttm_bo_unref(&bo); + vfbs->buffer = NULL; + + return 0; +} + static int vmw_framebuffer_dmabuf_pin(struct vmw_framebuffer *vfb) { struct vmw_private *dev_priv = vmw_priv(vfb->base.dev); @@ -596,33 +633,15 @@ static int vmw_framebuffer_dmabuf_pin(struct vmw_framebuffer *vfb) vmw_framebuffer_to_vfbd(&vfb->base); int ret; + vmw_overlay_pause_all(dev_priv); ret = vmw_dmabuf_to_start_of_vram(dev_priv, vfbd->buffer); - if (dev_priv->capabilities & SVGA_CAP_MULTIMON) { - vmw_write(dev_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1); - vmw_write(dev_priv, SVGA_REG_DISPLAY_ID, 0); - vmw_write(dev_priv, SVGA_REG_DISPLAY_IS_PRIMARY, true); - vmw_write(dev_priv, SVGA_REG_DISPLAY_POSITION_X, 0); - vmw_write(dev_priv, SVGA_REG_DISPLAY_POSITION_Y, 0); - vmw_write(dev_priv, SVGA_REG_DISPLAY_WIDTH, 0); - vmw_write(dev_priv, SVGA_REG_DISPLAY_HEIGHT, 0); - vmw_write(dev_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID); - - vmw_write(dev_priv, SVGA_REG_ENABLE, 1); - vmw_write(dev_priv, SVGA_REG_WIDTH, vfb->base.width); - vmw_write(dev_priv, SVGA_REG_HEIGHT, vfb->base.height); - vmw_write(dev_priv, SVGA_REG_BITS_PER_PIXEL, vfb->base.bits_per_pixel); - vmw_write(dev_priv, SVGA_REG_DEPTH, vfb->base.depth); - vmw_write(dev_priv, SVGA_REG_RED_MASK, 0x00ff0000); - vmw_write(dev_priv, SVGA_REG_GREEN_MASK, 0x0000ff00); - vmw_write(dev_priv, SVGA_REG_BLUE_MASK, 0x000000ff); - } else - WARN_ON(true); - vmw_overlay_resume_all(dev_priv); + WARN_ON(ret != 0); + return 0; } @@ -668,7 +687,7 @@ int vmw_kms_new_framebuffer_dmabuf(struct vmw_private *dev_priv, /* XXX get the first 3 from the surface info */ vfbd->base.base.bits_per_pixel = 32; - vfbd->base.base.pitch = width * 32 / 4; + vfbd->base.base.pitch = width * vfbd->base.base.bits_per_pixel / 8; vfbd->base.base.depth = 24; vfbd->base.base.width = width; vfbd->base.base.height = height; @@ -765,8 +784,9 @@ int vmw_kms_init(struct vmw_private *dev_priv) dev->mode_config.funcs = &vmw_kms_funcs; dev->mode_config.min_width = 1; dev->mode_config.min_height = 1; - dev->mode_config.max_width = dev_priv->fb_max_width; - dev->mode_config.max_height = dev_priv->fb_max_height; + /* assumed largest fb size */ + dev->mode_config.max_width = 8192; + dev->mode_config.max_height = 8192; ret = vmw_kms_init_legacy_display_system(dev_priv); @@ -826,49 +846,140 @@ out: return ret; } +void vmw_kms_write_svga(struct vmw_private *vmw_priv, + unsigned width, unsigned height, unsigned pitch, + unsigned bbp, unsigned depth) +{ + if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK) + vmw_write(vmw_priv, SVGA_REG_PITCHLOCK, pitch); + else if (vmw_fifo_have_pitchlock(vmw_priv)) + iowrite32(pitch, vmw_priv->mmio_virt + SVGA_FIFO_PITCHLOCK); + vmw_write(vmw_priv, SVGA_REG_WIDTH, width); + vmw_write(vmw_priv, SVGA_REG_HEIGHT, height); + vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, bbp); + vmw_write(vmw_priv, SVGA_REG_DEPTH, depth); + vmw_write(vmw_priv, SVGA_REG_RED_MASK, 0x00ff0000); + vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, 0x0000ff00); + vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, 0x000000ff); +} + int vmw_kms_save_vga(struct vmw_private *vmw_priv) { - /* - * setup a single multimon monitor with the size - * of 0x0, this stops the UI from resizing when we - * change the framebuffer size - */ - if (vmw_priv->capabilities & SVGA_CAP_MULTIMON) { - vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY, true); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_X, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_Y, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_WIDTH, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_HEIGHT, 0); - vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID); - } + struct vmw_vga_topology_state *save; + uint32_t i; vmw_priv->vga_width = vmw_read(vmw_priv, SVGA_REG_WIDTH); vmw_priv->vga_height = vmw_read(vmw_priv, SVGA_REG_HEIGHT); - vmw_priv->vga_bpp = vmw_read(vmw_priv, SVGA_REG_BITS_PER_PIXEL); vmw_priv->vga_depth = vmw_read(vmw_priv, SVGA_REG_DEPTH); + vmw_priv->vga_bpp = vmw_read(vmw_priv, SVGA_REG_BITS_PER_PIXEL); vmw_priv->vga_pseudo = vmw_read(vmw_priv, SVGA_REG_PSEUDOCOLOR); vmw_priv->vga_red_mask = vmw_read(vmw_priv, SVGA_REG_RED_MASK); - vmw_priv->vga_green_mask = vmw_read(vmw_priv, SVGA_REG_GREEN_MASK); vmw_priv->vga_blue_mask = vmw_read(vmw_priv, SVGA_REG_BLUE_MASK); + vmw_priv->vga_green_mask = vmw_read(vmw_priv, SVGA_REG_GREEN_MASK); + if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK) + vmw_priv->vga_pitchlock = + vmw_read(vmw_priv, SVGA_REG_PITCHLOCK); + else if (vmw_fifo_have_pitchlock(vmw_priv)) + vmw_priv->vga_pitchlock = ioread32(vmw_priv->mmio_virt + + SVGA_FIFO_PITCHLOCK); + + if (!(vmw_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY)) + return 0; + vmw_priv->num_displays = vmw_read(vmw_priv, + SVGA_REG_NUM_GUEST_DISPLAYS); + + for (i = 0; i < vmw_priv->num_displays; ++i) { + save = &vmw_priv->vga_save[i]; + vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, i); + save->primary = vmw_read(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY); + save->pos_x = vmw_read(vmw_priv, SVGA_REG_DISPLAY_POSITION_X); + save->pos_y = vmw_read(vmw_priv, SVGA_REG_DISPLAY_POSITION_Y); + save->width = vmw_read(vmw_priv, SVGA_REG_DISPLAY_WIDTH); + save->height = vmw_read(vmw_priv, SVGA_REG_DISPLAY_HEIGHT); + vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID); + } return 0; } int vmw_kms_restore_vga(struct vmw_private *vmw_priv) { + struct vmw_vga_topology_state *save; + uint32_t i; + vmw_write(vmw_priv, SVGA_REG_WIDTH, vmw_priv->vga_width); vmw_write(vmw_priv, SVGA_REG_HEIGHT, vmw_priv->vga_height); - vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, vmw_priv->vga_bpp); vmw_write(vmw_priv, SVGA_REG_DEPTH, vmw_priv->vga_depth); + vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, vmw_priv->vga_bpp); vmw_write(vmw_priv, SVGA_REG_PSEUDOCOLOR, vmw_priv->vga_pseudo); vmw_write(vmw_priv, SVGA_REG_RED_MASK, vmw_priv->vga_red_mask); vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, vmw_priv->vga_green_mask); vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, vmw_priv->vga_blue_mask); + if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK) + vmw_write(vmw_priv, SVGA_REG_PITCHLOCK, + vmw_priv->vga_pitchlock); + else if (vmw_fifo_have_pitchlock(vmw_priv)) + iowrite32(vmw_priv->vga_pitchlock, + vmw_priv->mmio_virt + SVGA_FIFO_PITCHLOCK); + + if (!(vmw_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY)) + return 0; - /* TODO check for multimon */ - vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 0); + for (i = 0; i < vmw_priv->num_displays; ++i) { + save = &vmw_priv->vga_save[i]; + vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, i); + vmw_write(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY, save->primary); + vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_X, save->pos_x); + vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_Y, save->pos_y); + vmw_write(vmw_priv, SVGA_REG_DISPLAY_WIDTH, save->width); + vmw_write(vmw_priv, SVGA_REG_DISPLAY_HEIGHT, save->height); + vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID); + } return 0; } + +int vmw_kms_update_layout_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct vmw_private *dev_priv = vmw_priv(dev); + struct drm_vmw_update_layout_arg *arg = + (struct drm_vmw_update_layout_arg *)data; + struct vmw_master *vmaster = vmw_master(file_priv->master); + void __user *user_rects; + struct drm_vmw_rect *rects; + unsigned rects_size; + int ret; + + ret = ttm_read_lock(&vmaster->lock, true); + if (unlikely(ret != 0)) + return ret; + + if (!arg->num_outputs) { + struct drm_vmw_rect def_rect = {0, 0, 800, 600}; + vmw_kms_ldu_update_layout(dev_priv, 1, &def_rect); + goto out_unlock; + } + + rects_size = arg->num_outputs * sizeof(struct drm_vmw_rect); + rects = kzalloc(rects_size, GFP_KERNEL); + if (unlikely(!rects)) { + ret = -ENOMEM; + goto out_unlock; + } + + user_rects = (void __user *)(unsigned long)arg->rects; + ret = copy_from_user(rects, user_rects, rects_size); + if (unlikely(ret != 0)) { + DRM_ERROR("Failed to get rects.\n"); + goto out_free; + } + + vmw_kms_ldu_update_layout(dev_priv, arg->num_outputs, rects); + +out_free: + kfree(rects); +out_unlock: + ttm_read_unlock(&vmaster->lock); + return ret; +} diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h index 8b95249f0531..8a398a0339b6 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h @@ -94,9 +94,11 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv, int vmw_du_crtc_cursor_move(struct drm_crtc *crtc, int x, int y); /* - * Legacy display unit functions - vmwgfx_ldu.h + * Legacy display unit functions - vmwgfx_ldu.c */ int vmw_kms_init_legacy_display_system(struct vmw_private *dev_priv); int vmw_kms_close_legacy_display_system(struct vmw_private *dev_priv); +int vmw_kms_ldu_update_layout(struct vmw_private *dev_priv, unsigned num, + struct drm_vmw_rect *rects); #endif diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c index 90891593bf6c..cfaf690a5b2f 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c @@ -38,6 +38,7 @@ struct vmw_legacy_display { struct list_head active; unsigned num_active; + unsigned last_num_active; struct vmw_framebuffer *fb; }; @@ -48,9 +49,12 @@ struct vmw_legacy_display { struct vmw_legacy_display_unit { struct vmw_display_unit base; - struct list_head active; + unsigned pref_width; + unsigned pref_height; + bool pref_active; + struct drm_display_mode *pref_mode; - unsigned unit; + struct list_head active; }; static void vmw_ldu_destroy(struct vmw_legacy_display_unit *ldu) @@ -88,23 +92,44 @@ static int vmw_ldu_commit_list(struct vmw_private *dev_priv) { struct vmw_legacy_display *lds = dev_priv->ldu_priv; struct vmw_legacy_display_unit *entry; - struct drm_crtc *crtc; + struct drm_framebuffer *fb = NULL; + struct drm_crtc *crtc = NULL; int i = 0; - /* to stop the screen from changing size on resize */ - vmw_write(dev_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 0); - for (i = 0; i < lds->num_active; i++) { - vmw_write(dev_priv, SVGA_REG_DISPLAY_ID, i); - vmw_write(dev_priv, SVGA_REG_DISPLAY_IS_PRIMARY, !i); - vmw_write(dev_priv, SVGA_REG_DISPLAY_POSITION_X, 0); - vmw_write(dev_priv, SVGA_REG_DISPLAY_POSITION_Y, 0); - vmw_write(dev_priv, SVGA_REG_DISPLAY_WIDTH, 0); - vmw_write(dev_priv, SVGA_REG_DISPLAY_HEIGHT, 0); - vmw_write(dev_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID); + /* If there is no display topology the host just assumes + * that the guest will set the same layout as the host. + */ + if (!(dev_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY)) { + int w = 0, h = 0; + list_for_each_entry(entry, &lds->active, active) { + crtc = &entry->base.crtc; + w = max(w, crtc->x + crtc->mode.hdisplay); + h = max(h, crtc->y + crtc->mode.vdisplay); + i++; + } + + if (crtc == NULL) + return 0; + fb = entry->base.crtc.fb; + + vmw_kms_write_svga(dev_priv, w, h, fb->pitch, + fb->bits_per_pixel, fb->depth); + + return 0; } - /* Now set the mode */ - vmw_write(dev_priv, SVGA_REG_NUM_GUEST_DISPLAYS, lds->num_active); + if (!list_empty(&lds->active)) { + entry = list_entry(lds->active.next, typeof(*entry), active); + fb = entry->base.crtc.fb; + + vmw_kms_write_svga(dev_priv, fb->width, fb->height, fb->pitch, + fb->bits_per_pixel, fb->depth); + } + + /* Make sure we always show something. */ + vmw_write(dev_priv, SVGA_REG_NUM_GUEST_DISPLAYS, + lds->num_active ? lds->num_active : 1); + i = 0; list_for_each_entry(entry, &lds->active, active) { crtc = &entry->base.crtc; @@ -120,6 +145,10 @@ static int vmw_ldu_commit_list(struct vmw_private *dev_priv) i++; } + BUG_ON(i != lds->num_active); + + lds->last_num_active = lds->num_active; + return 0; } @@ -130,6 +159,7 @@ static int vmw_ldu_del_active(struct vmw_private *vmw_priv, if (list_empty(&ldu->active)) return 0; + /* Must init otherwise list_empty(&ldu->active) will not work. */ list_del_init(&ldu->active); if (--(ld->num_active) == 0) { BUG_ON(!ld->fb); @@ -149,24 +179,29 @@ static int vmw_ldu_add_active(struct vmw_private *vmw_priv, struct vmw_legacy_display_unit *entry; struct list_head *at; + BUG_ON(!ld->num_active && ld->fb); + if (vfb != ld->fb) { + if (ld->fb && ld->fb->unpin) + ld->fb->unpin(ld->fb); + if (vfb->pin) + vfb->pin(vfb); + ld->fb = vfb; + } + if (!list_empty(&ldu->active)) return 0; at = &ld->active; list_for_each_entry(entry, &ld->active, active) { - if (entry->unit > ldu->unit) + if (entry->base.unit > ldu->base.unit) break; at = &entry->active; } list_add(&ldu->active, at); - if (ld->num_active++ == 0) { - BUG_ON(ld->fb); - if (vfb->pin) - vfb->pin(vfb); - ld->fb = vfb; - } + + ld->num_active++; return 0; } @@ -208,6 +243,8 @@ static int vmw_ldu_crtc_set_config(struct drm_mode_set *set) /* ldu only supports one fb active at the time */ if (dev_priv->ldu_priv->fb && vfb && + !(dev_priv->ldu_priv->num_active == 1 && + !list_empty(&ldu->active)) && dev_priv->ldu_priv->fb != vfb) { DRM_ERROR("Multiple framebuffers not supported\n"); return -EINVAL; @@ -300,8 +337,7 @@ static void vmw_ldu_connector_restore(struct drm_connector *connector) static enum drm_connector_status vmw_ldu_connector_detect(struct drm_connector *connector) { - /* XXX vmwctrl should control connection status */ - if (vmw_connector_to_ldu(connector)->base.unit == 0) + if (vmw_connector_to_ldu(connector)->pref_active) return connector_status_connected; return connector_status_disconnected; } @@ -312,10 +348,9 @@ static struct drm_display_mode vmw_ldu_connector_builtin[] = { 752, 800, 0, 480, 489, 492, 525, 0, DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 800x600@60Hz */ - { DRM_MODE("800x600", - DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED, - 40000, 800, 840, 968, 1056, 0, 600, 601, 605, 628, - 0, DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 40000, 800, 840, + 968, 1056, 0, 600, 601, 605, 628, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 1024x768@60Hz */ { DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 65000, 1024, 1048, 1184, 1344, 0, 768, 771, 777, 806, 0, @@ -387,10 +422,34 @@ static struct drm_display_mode vmw_ldu_connector_builtin[] = { static int vmw_ldu_connector_fill_modes(struct drm_connector *connector, uint32_t max_width, uint32_t max_height) { + struct vmw_legacy_display_unit *ldu = vmw_connector_to_ldu(connector); struct drm_device *dev = connector->dev; struct drm_display_mode *mode = NULL; + struct drm_display_mode prefmode = { DRM_MODE("preferred", + DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) + }; int i; + /* Add preferred mode */ + { + mode = drm_mode_duplicate(dev, &prefmode); + if (!mode) + return 0; + mode->hdisplay = ldu->pref_width; + mode->vdisplay = ldu->pref_height; + mode->vrefresh = drm_mode_vrefresh(mode); + drm_mode_probed_add(connector, mode); + + if (ldu->pref_mode) { + list_del_init(&ldu->pref_mode->head); + drm_mode_destroy(dev, ldu->pref_mode); + } + + ldu->pref_mode = mode; + } + for (i = 0; vmw_ldu_connector_builtin[i].type != 0; i++) { if (vmw_ldu_connector_builtin[i].hdisplay > max_width || vmw_ldu_connector_builtin[i].vdisplay > max_height) @@ -443,18 +502,21 @@ static int vmw_ldu_init(struct vmw_private *dev_priv, unsigned unit) if (!ldu) return -ENOMEM; - ldu->unit = unit; + ldu->base.unit = unit; crtc = &ldu->base.crtc; encoder = &ldu->base.encoder; connector = &ldu->base.connector; + INIT_LIST_HEAD(&ldu->active); + + ldu->pref_active = (unit == 0); + ldu->pref_width = 800; + ldu->pref_height = 600; + ldu->pref_mode = NULL; + drm_connector_init(dev, connector, &vmw_legacy_connector_funcs, DRM_MODE_CONNECTOR_LVDS); - /* Initial status */ - if (unit == 0) - connector->status = connector_status_connected; - else - connector->status = connector_status_disconnected; + connector->status = vmw_ldu_connector_detect(connector); drm_encoder_init(dev, encoder, &vmw_legacy_encoder_funcs, DRM_MODE_ENCODER_LVDS); @@ -462,8 +524,6 @@ static int vmw_ldu_init(struct vmw_private *dev_priv, unsigned unit) encoder->possible_crtcs = (1 << unit); encoder->possible_clones = 0; - INIT_LIST_HEAD(&ldu->active); - drm_crtc_init(dev, crtc, &vmw_legacy_crtc_funcs); drm_connector_attach_property(connector, @@ -487,18 +547,22 @@ int vmw_kms_init_legacy_display_system(struct vmw_private *dev_priv) INIT_LIST_HEAD(&dev_priv->ldu_priv->active); dev_priv->ldu_priv->num_active = 0; + dev_priv->ldu_priv->last_num_active = 0; dev_priv->ldu_priv->fb = NULL; drm_mode_create_dirty_info_property(dev_priv->dev); vmw_ldu_init(dev_priv, 0); - vmw_ldu_init(dev_priv, 1); - vmw_ldu_init(dev_priv, 2); - vmw_ldu_init(dev_priv, 3); - vmw_ldu_init(dev_priv, 4); - vmw_ldu_init(dev_priv, 5); - vmw_ldu_init(dev_priv, 6); - vmw_ldu_init(dev_priv, 7); + /* for old hardware without multimon only enable one display */ + if (dev_priv->capabilities & SVGA_CAP_MULTIMON) { + vmw_ldu_init(dev_priv, 1); + vmw_ldu_init(dev_priv, 2); + vmw_ldu_init(dev_priv, 3); + vmw_ldu_init(dev_priv, 4); + vmw_ldu_init(dev_priv, 5); + vmw_ldu_init(dev_priv, 6); + vmw_ldu_init(dev_priv, 7); + } return 0; } @@ -514,3 +578,42 @@ int vmw_kms_close_legacy_display_system(struct vmw_private *dev_priv) return 0; } + +int vmw_kms_ldu_update_layout(struct vmw_private *dev_priv, unsigned num, + struct drm_vmw_rect *rects) +{ + struct drm_device *dev = dev_priv->dev; + struct vmw_legacy_display_unit *ldu; + struct drm_connector *con; + int i; + + mutex_lock(&dev->mode_config.mutex); + +#if 0 + DRM_INFO("%s: new layout ", __func__); + for (i = 0; i < (int)num; i++) + DRM_INFO("(%i, %i %ux%u) ", rects[i].x, rects[i].y, + rects[i].w, rects[i].h); + DRM_INFO("\n"); +#else + (void)i; +#endif + + list_for_each_entry(con, &dev->mode_config.connector_list, head) { + ldu = vmw_connector_to_ldu(con); + if (num > ldu->base.unit) { + ldu->pref_width = rects[ldu->base.unit].w; + ldu->pref_height = rects[ldu->base.unit].h; + ldu->pref_active = true; + } else { + ldu->pref_width = 800; + ldu->pref_height = 600; + ldu->pref_active = false; + } + con->status = vmw_ldu_connector_detect(con); + } + + mutex_unlock(&dev->mode_config.mutex); + + return 0; +} diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c index ad566c85b075..df2036ed18d5 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c @@ -358,6 +358,8 @@ static int vmw_overlay_update_stream(struct vmw_private *dev_priv, if (stream->buf != buf) stream->buf = vmw_dmabuf_reference(buf); stream->saved = *arg; + /* stream is no longer stopped/paused */ + stream->paused = false; return 0; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c index f8fbbc67a406..8612378b131e 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c @@ -597,8 +597,10 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data, ret = copy_from_user(srf->sizes, user_sizes, srf->num_sizes * sizeof(*srf->sizes)); - if (unlikely(ret != 0)) + if (unlikely(ret != 0)) { + ret = -EFAULT; goto out_err1; + } if (srf->scanout && srf->num_sizes == 1 && @@ -697,9 +699,11 @@ int vmw_surface_reference_ioctl(struct drm_device *dev, void *data, if (user_sizes) ret = copy_to_user(user_sizes, srf->sizes, srf->num_sizes * sizeof(*srf->sizes)); - if (unlikely(ret != 0)) + if (unlikely(ret != 0)) { DRM_ERROR("copy_to_user failed %p %u\n", user_sizes, srf->num_sizes); + ret = -EFAULT; + } out_bad_resource: out_no_reference: ttm_base_object_unref(&base); |
