diff options
Diffstat (limited to 'drivers/block')
-rw-r--r-- | drivers/block/floppy.c | 3 | ||||
-rw-r--r-- | drivers/block/nbd.c | 3 | ||||
-rw-r--r-- | drivers/block/null_blk.h | 17 | ||||
-rw-r--r-- | drivers/block/null_blk_main.c | 45 | ||||
-rw-r--r-- | drivers/block/null_blk_zoned.c | 34 | ||||
-rw-r--r-- | drivers/block/rbd.c | 235 | ||||
-rw-r--r-- | drivers/block/xen-blkback/blkback.c | 99 | ||||
-rw-r--r-- | drivers/block/xen-blkback/common.h | 14 | ||||
-rw-r--r-- | drivers/block/xen-blkfront.c | 110 |
9 files changed, 399 insertions, 161 deletions
diff --git a/drivers/block/floppy.c b/drivers/block/floppy.c index 48f622728ce6..f2b6f4da1034 100644 --- a/drivers/block/floppy.c +++ b/drivers/block/floppy.c @@ -3467,6 +3467,9 @@ static int fd_locked_ioctl(struct block_device *bdev, fmode_t mode, unsigned int (struct floppy_struct **)&outparam); if (ret) return ret; + memcpy(&inparam.g, outparam, + offsetof(struct floppy_struct, name)); + outparam = &inparam.g; break; case FDMSGON: UDP->flags |= FTD_MSG; diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c index 3863c00372bb..14a51254c3db 100644 --- a/drivers/block/nbd.c +++ b/drivers/block/nbd.c @@ -1239,6 +1239,9 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd, case NBD_SET_SOCK: return nbd_add_socket(nbd, arg, false); case NBD_SET_BLKSIZE: + if (!arg || !is_power_of_2(arg) || arg < 512 || + arg > PAGE_SIZE) + return -EINVAL; nbd_size_set(nbd, arg, div_s64(config->bytesize, arg)); return 0; diff --git a/drivers/block/null_blk.h b/drivers/block/null_blk.h index d81781f22dba..34e0030f0592 100644 --- a/drivers/block/null_blk.h +++ b/drivers/block/null_blk.h @@ -87,10 +87,10 @@ struct nullb { #ifdef CONFIG_BLK_DEV_ZONED int null_zone_init(struct nullb_device *dev); void null_zone_exit(struct nullb_device *dev); -blk_status_t null_zone_report(struct nullb *nullb, - struct nullb_cmd *cmd); -void null_zone_write(struct nullb_cmd *cmd); -void null_zone_reset(struct nullb_cmd *cmd); +blk_status_t null_zone_report(struct nullb *nullb, struct bio *bio); +void null_zone_write(struct nullb_cmd *cmd, sector_t sector, + unsigned int nr_sectors); +void null_zone_reset(struct nullb_cmd *cmd, sector_t sector); #else static inline int null_zone_init(struct nullb_device *dev) { @@ -98,11 +98,14 @@ static inline int null_zone_init(struct nullb_device *dev) } static inline void null_zone_exit(struct nullb_device *dev) {} static inline blk_status_t null_zone_report(struct nullb *nullb, - struct nullb_cmd *cmd) + struct bio *bio) { return BLK_STS_NOTSUPP; } -static inline void null_zone_write(struct nullb_cmd *cmd) {} -static inline void null_zone_reset(struct nullb_cmd *cmd) {} +static inline void null_zone_write(struct nullb_cmd *cmd, sector_t sector, + unsigned int nr_sectors) +{ +} +static inline void null_zone_reset(struct nullb_cmd *cmd, sector_t sector) {} #endif /* CONFIG_BLK_DEV_ZONED */ #endif /* __NULL_BLK_H */ diff --git a/drivers/block/null_blk_main.c b/drivers/block/null_blk_main.c index 6127e3ff7b4b..093b614d6524 100644 --- a/drivers/block/null_blk_main.c +++ b/drivers/block/null_blk_main.c @@ -1157,16 +1157,33 @@ static void null_restart_queue_async(struct nullb *nullb) } } +static bool cmd_report_zone(struct nullb *nullb, struct nullb_cmd *cmd) +{ + struct nullb_device *dev = cmd->nq->dev; + + if (dev->queue_mode == NULL_Q_BIO) { + if (bio_op(cmd->bio) == REQ_OP_ZONE_REPORT) { + cmd->error = null_zone_report(nullb, cmd->bio); + return true; + } + } else { + if (req_op(cmd->rq) == REQ_OP_ZONE_REPORT) { + cmd->error = null_zone_report(nullb, cmd->rq->bio); + return true; + } + } + + return false; +} + static blk_status_t null_handle_cmd(struct nullb_cmd *cmd) { struct nullb_device *dev = cmd->nq->dev; struct nullb *nullb = dev->nullb; int err = 0; - if (req_op(cmd->rq) == REQ_OP_ZONE_REPORT) { - cmd->error = null_zone_report(nullb, cmd); + if (cmd_report_zone(nullb, cmd)) goto out; - } if (test_bit(NULLB_DEV_FL_THROTTLED, &dev->flags)) { struct request *rq = cmd->rq; @@ -1234,10 +1251,24 @@ static blk_status_t null_handle_cmd(struct nullb_cmd *cmd) cmd->error = errno_to_blk_status(err); if (!cmd->error && dev->zoned) { - if (req_op(cmd->rq) == REQ_OP_WRITE) - null_zone_write(cmd); - else if (req_op(cmd->rq) == REQ_OP_ZONE_RESET) - null_zone_reset(cmd); + sector_t sector; + unsigned int nr_sectors; + int op; + + if (dev->queue_mode == NULL_Q_BIO) { + op = bio_op(cmd->bio); + sector = cmd->bio->bi_iter.bi_sector; + nr_sectors = cmd->bio->bi_iter.bi_size >> 9; + } else { + op = req_op(cmd->rq); + sector = blk_rq_pos(cmd->rq); + nr_sectors = blk_rq_sectors(cmd->rq); + } + + if (op == REQ_OP_WRITE) + null_zone_write(cmd, sector, nr_sectors); + else if (op == REQ_OP_ZONE_RESET) + null_zone_reset(cmd, sector); } out: /* Complete IO by inline, softirq or timer */ diff --git a/drivers/block/null_blk_zoned.c b/drivers/block/null_blk_zoned.c index a979ca00d7be..7c6b86d98700 100644 --- a/drivers/block/null_blk_zoned.c +++ b/drivers/block/null_blk_zoned.c @@ -48,8 +48,8 @@ void null_zone_exit(struct nullb_device *dev) kvfree(dev->zones); } -static void null_zone_fill_rq(struct nullb_device *dev, struct request *rq, - unsigned int zno, unsigned int nr_zones) +static void null_zone_fill_bio(struct nullb_device *dev, struct bio *bio, + unsigned int zno, unsigned int nr_zones) { struct blk_zone_report_hdr *hdr = NULL; struct bio_vec bvec; @@ -57,7 +57,7 @@ static void null_zone_fill_rq(struct nullb_device *dev, struct request *rq, void *addr; unsigned int zones_to_cpy; - bio_for_each_segment(bvec, rq->bio, iter) { + bio_for_each_segment(bvec, bio, iter) { addr = kmap_atomic(bvec.bv_page); zones_to_cpy = bvec.bv_len / sizeof(struct blk_zone); @@ -84,29 +84,24 @@ static void null_zone_fill_rq(struct nullb_device *dev, struct request *rq, } } -blk_status_t null_zone_report(struct nullb *nullb, - struct nullb_cmd *cmd) +blk_status_t null_zone_report(struct nullb *nullb, struct bio *bio) { struct nullb_device *dev = nullb->dev; - struct request *rq = cmd->rq; - unsigned int zno = null_zone_no(dev, blk_rq_pos(rq)); + unsigned int zno = null_zone_no(dev, bio->bi_iter.bi_sector); unsigned int nr_zones = dev->nr_zones - zno; - unsigned int max_zones = (blk_rq_bytes(rq) / - sizeof(struct blk_zone)) - 1; + unsigned int max_zones; + max_zones = (bio->bi_iter.bi_size / sizeof(struct blk_zone)) - 1; nr_zones = min_t(unsigned int, nr_zones, max_zones); - - null_zone_fill_rq(nullb->dev, rq, zno, nr_zones); + null_zone_fill_bio(nullb->dev, bio, zno, nr_zones); return BLK_STS_OK; } -void null_zone_write(struct nullb_cmd *cmd) +void null_zone_write(struct nullb_cmd *cmd, sector_t sector, + unsigned int nr_sectors) { struct nullb_device *dev = cmd->nq->dev; - struct request *rq = cmd->rq; - sector_t sector = blk_rq_pos(rq); - unsigned int rq_sectors = blk_rq_sectors(rq); unsigned int zno = null_zone_no(dev, sector); struct blk_zone *zone = &dev->zones[zno]; @@ -118,7 +113,7 @@ void null_zone_write(struct nullb_cmd *cmd) case BLK_ZONE_COND_EMPTY: case BLK_ZONE_COND_IMP_OPEN: /* Writes must be at the write pointer position */ - if (blk_rq_pos(rq) != zone->wp) { + if (sector != zone->wp) { cmd->error = BLK_STS_IOERR; break; } @@ -126,7 +121,7 @@ void null_zone_write(struct nullb_cmd *cmd) if (zone->cond == BLK_ZONE_COND_EMPTY) zone->cond = BLK_ZONE_COND_IMP_OPEN; - zone->wp += rq_sectors; + zone->wp += nr_sectors; if (zone->wp == zone->start + zone->len) zone->cond = BLK_ZONE_COND_FULL; break; @@ -137,11 +132,10 @@ void null_zone_write(struct nullb_cmd *cmd) } } -void null_zone_reset(struct nullb_cmd *cmd) +void null_zone_reset(struct nullb_cmd *cmd, sector_t sector) { struct nullb_device *dev = cmd->nq->dev; - struct request *rq = cmd->rq; - unsigned int zno = null_zone_no(dev, blk_rq_pos(rq)); + unsigned int zno = null_zone_no(dev, sector); struct blk_zone *zone = &dev->zones[zno]; zone->cond = BLK_ZONE_COND_EMPTY; diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 7915f3b03736..73ed5f3a862d 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -4207,11 +4207,13 @@ static ssize_t rbd_parent_show(struct device *dev, count += sprintf(&buf[count], "%s" "pool_id %llu\npool_name %s\n" + "pool_ns %s\n" "image_id %s\nimage_name %s\n" "snap_id %llu\nsnap_name %s\n" "overlap %llu\n", !count ? "" : "\n", /* first? */ spec->pool_id, spec->pool_name, + spec->pool_ns ?: "", spec->image_id, spec->image_name ?: "(unknown)", spec->snap_id, spec->snap_name, rbd_dev->parent_overlap); @@ -4584,47 +4586,177 @@ static int rbd_dev_v2_features(struct rbd_device *rbd_dev) &rbd_dev->header.features); } +struct parent_image_info { + u64 pool_id; + const char *pool_ns; + const char *image_id; + u64 snap_id; + + bool has_overlap; + u64 overlap; +}; + +/* + * The caller is responsible for @pii. + */ +static int decode_parent_image_spec(void **p, void *end, + struct parent_image_info *pii) +{ + u8 struct_v; + u32 struct_len; + int ret; + + ret = ceph_start_decoding(p, end, 1, "ParentImageSpec", + &struct_v, &struct_len); + if (ret) + return ret; + + ceph_decode_64_safe(p, end, pii->pool_id, e_inval); + pii->pool_ns = ceph_extract_encoded_string(p, end, NULL, GFP_KERNEL); + if (IS_ERR(pii->pool_ns)) { + ret = PTR_ERR(pii->pool_ns); + pii->pool_ns = NULL; + return ret; + } + pii->image_id = ceph_extract_encoded_string(p, end, NULL, GFP_KERNEL); + if (IS_ERR(pii->image_id)) { + ret = PTR_ERR(pii->image_id); + pii->image_id = NULL; + return ret; + } + ceph_decode_64_safe(p, end, pii->snap_id, e_inval); + return 0; + +e_inval: + return -EINVAL; +} + +static int __get_parent_info(struct rbd_device *rbd_dev, + struct page *req_page, + struct page *reply_page, + struct parent_image_info *pii) +{ + struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc; + size_t reply_len = PAGE_SIZE; + void *p, *end; + int ret; + + ret = ceph_osdc_call(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc, + "rbd", "parent_get", CEPH_OSD_FLAG_READ, + req_page, sizeof(u64), reply_page, &reply_len); + if (ret) + return ret == -EOPNOTSUPP ? 1 : ret; + + p = page_address(reply_page); + end = p + reply_len; + ret = decode_parent_image_spec(&p, end, pii); + if (ret) + return ret; + + ret = ceph_osdc_call(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc, + "rbd", "parent_overlap_get", CEPH_OSD_FLAG_READ, + req_page, sizeof(u64), reply_page, &reply_len); + if (ret) + return ret; + + p = page_address(reply_page); + end = p + reply_len; + ceph_decode_8_safe(&p, end, pii->has_overlap, e_inval); + if (pii->has_overlap) + ceph_decode_64_safe(&p, end, pii->overlap, e_inval); + + return 0; + +e_inval: + return -EINVAL; +} + +/* + * The caller is responsible for @pii. + */ +static int __get_parent_info_legacy(struct rbd_device *rbd_dev, + struct page *req_page, + struct page *reply_page, + struct parent_image_info *pii) +{ + struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc; + size_t reply_len = PAGE_SIZE; + void *p, *end; + int ret; + + ret = ceph_osdc_call(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc, + "rbd", "get_parent", CEPH_OSD_FLAG_READ, + req_page, sizeof(u64), reply_page, &reply_len); + if (ret) + return ret; + + p = page_address(reply_page); + end = p + reply_len; + ceph_decode_64_safe(&p, end, pii->pool_id, e_inval); + pii->image_id = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL); + if (IS_ERR(pii->image_id)) { + ret = PTR_ERR(pii->image_id); + pii->image_id = NULL; + return ret; + } + ceph_decode_64_safe(&p, end, pii->snap_id, e_inval); + pii->has_overlap = true; + ceph_decode_64_safe(&p, end, pii->overlap, e_inval); + + return 0; + +e_inval: + return -EINVAL; +} + +static int get_parent_info(struct rbd_device *rbd_dev, + struct parent_image_info *pii) +{ + struct page *req_page, *reply_page; + void *p; + int ret; + + req_page = alloc_page(GFP_KERNEL); + if (!req_page) + return -ENOMEM; + + reply_page = alloc_page(GFP_KERNEL); + if (!reply_page) { + __free_page(req_page); + return -ENOMEM; + } + + p = page_address(req_page); + ceph_encode_64(&p, rbd_dev->spec->snap_id); + ret = __get_parent_info(rbd_dev, req_page, reply_page, pii); + if (ret > 0) + ret = __get_parent_info_legacy(rbd_dev, req_page, reply_page, + pii); + + __free_page(req_page); + __free_page(reply_page); + return ret; +} + static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev) { struct rbd_spec *parent_spec; - size_t size; - void *reply_buf = NULL; - __le64 snapid; - void *p; - void *end; - u64 pool_id; - char *image_id; - u64 snap_id; - u64 overlap; + struct parent_image_info pii = { 0 }; int ret; parent_spec = rbd_spec_alloc(); if (!parent_spec) return -ENOMEM; - size = sizeof (__le64) + /* pool_id */ - sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX + /* image_id */ - sizeof (__le64) + /* snap_id */ - sizeof (__le64); /* overlap */ - reply_buf = kmalloc(size, GFP_KERNEL); - if (!reply_buf) { - ret = -ENOMEM; + ret = get_parent_info(rbd_dev, &pii); + if (ret) goto out_err; - } - snapid = cpu_to_le64(rbd_dev->spec->snap_id); - ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid, - &rbd_dev->header_oloc, "get_parent", - &snapid, sizeof(snapid), reply_buf, size); - dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret); - if (ret < 0) - goto out_err; + dout("%s pool_id %llu pool_ns %s image_id %s snap_id %llu has_overlap %d overlap %llu\n", + __func__, pii.pool_id, pii.pool_ns, pii.image_id, pii.snap_id, + pii.has_overlap, pii.overlap); - p = reply_buf; - end = reply_buf + ret; - ret = -ERANGE; - ceph_decode_64_safe(&p, end, pool_id, out_err); - if (pool_id == CEPH_NOPOOL) { + if (pii.pool_id == CEPH_NOPOOL || !pii.has_overlap) { /* * Either the parent never existed, or we have * record of it but the image got flattened so it no @@ -4633,6 +4765,10 @@ static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev) * overlap to 0. The effect of this is that all new * requests will be treated as if the image had no * parent. + * + * If !pii.has_overlap, the parent image spec is not + * applicable. It's there to avoid duplication in each + * snapshot record. */ if (rbd_dev->parent_overlap) { rbd_dev->parent_overlap = 0; @@ -4647,51 +4783,36 @@ static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev) /* The ceph file layout needs to fit pool id in 32 bits */ ret = -EIO; - if (pool_id > (u64)U32_MAX) { + if (pii.pool_id > (u64)U32_MAX) { rbd_warn(NULL, "parent pool id too large (%llu > %u)", - (unsigned long long)pool_id, U32_MAX); + (unsigned long long)pii.pool_id, U32_MAX); goto out_err; } - image_id = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL); - if (IS_ERR(image_id)) { - ret = PTR_ERR(image_id); - goto out_err; - } - ceph_decode_64_safe(&p, end, snap_id, out_err); - ceph_decode_64_safe(&p, end, overlap, out_err); - /* * The parent won't change (except when the clone is * flattened, already handled that). So we only need to * record the parent spec we have not already done so. */ if (!rbd_dev->parent_spec) { - parent_spec->pool_id = pool_id; - parent_spec->image_id = image_id; - parent_spec->snap_id = snap_id; - - /* TODO: support cloning across namespaces */ - if (rbd_dev->spec->pool_ns) { - parent_spec->pool_ns = kstrdup(rbd_dev->spec->pool_ns, - GFP_KERNEL); - if (!parent_spec->pool_ns) { - ret = -ENOMEM; - goto out_err; - } + parent_spec->pool_id = pii.pool_id; + if (pii.pool_ns && *pii.pool_ns) { + parent_spec->pool_ns = pii.pool_ns; + pii.pool_ns = NULL; } + parent_spec->image_id = pii.image_id; + pii.image_id = NULL; + parent_spec->snap_id = pii.snap_id; rbd_dev->parent_spec = parent_spec; parent_spec = NULL; /* rbd_dev now owns this */ - } else { - kfree(image_id); } /* * We always update the parent overlap. If it's zero we issue * a warning, as we will proceed as if there was no parent. */ - if (!overlap) { + if (!pii.overlap) { if (parent_spec) { /* refresh, careful to warn just once */ if (rbd_dev->parent_overlap) @@ -4702,14 +4823,14 @@ static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev) rbd_warn(rbd_dev, "clone is standalone (overlap 0)"); } } - rbd_dev->parent_overlap = overlap; + rbd_dev->parent_overlap = pii.overlap; out: ret = 0; out_err: - kfree(reply_buf); + kfree(pii.pool_ns); + kfree(pii.image_id); rbd_spec_put(parent_spec); - return ret; } diff --git a/drivers/block/xen-blkback/blkback.c b/drivers/block/xen-blkback/blkback.c index b55b245e8052..fd1e19f1a49f 100644 --- a/drivers/block/xen-blkback/blkback.c +++ b/drivers/block/xen-blkback/blkback.c @@ -84,6 +84,18 @@ MODULE_PARM_DESC(max_persistent_grants, "Maximum number of grants to map persistently"); /* + * How long a persistent grant is allowed to remain allocated without being in + * use. The time is in seconds, 0 means indefinitely long. + */ + +static unsigned int xen_blkif_pgrant_timeout = 60; +module_param_named(persistent_grant_unused_seconds, xen_blkif_pgrant_timeout, + uint, 0644); +MODULE_PARM_DESC(persistent_grant_unused_seconds, + "Time in seconds an unused persistent grant is allowed to " + "remain allocated. Default is 60, 0 means unlimited."); + +/* * Maximum number of rings/queues blkback supports, allow as many queues as there * are CPUs if user has not specified a value. */ @@ -123,6 +135,13 @@ module_param(log_stats, int, 0644); /* Number of free pages to remove on each call to gnttab_free_pages */ #define NUM_BATCH_FREE_PAGES 10 +static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) +{ + return xen_blkif_pgrant_timeout && + (jiffies - persistent_gnt->last_used >= + HZ * xen_blkif_pgrant_timeout); +} + static inline int get_free_page(struct xen_blkif_ring *ring, struct page **page) { unsigned long flags; @@ -236,8 +255,7 @@ static int add_persistent_gnt(struct xen_blkif_ring *ring, } } - bitmap_zero(persistent_gnt->flags, PERSISTENT_GNT_FLAGS_SIZE); - set_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags); + persistent_gnt->active = true; /* Add new node and rebalance tree. */ rb_link_node(&(persistent_gnt->node), parent, new); rb_insert_color(&(persistent_gnt->node), &ring->persistent_gnts); @@ -261,11 +279,11 @@ static struct persistent_gnt *get_persistent_gnt(struct xen_blkif_ring *ring, else if (gref > data->gnt) node = node->rb_right; else { - if(test_bit(PERSISTENT_GNT_ACTIVE, data->flags)) { + if (data->active) { pr_alert_ratelimited("requesting a grant already in use\n"); return NULL; } - set_bit(PERSISTENT_GNT_ACTIVE, data->flags); + data->active = true; atomic_inc(&ring->persistent_gnt_in_use); return data; } @@ -276,10 +294,10 @@ static struct persistent_gnt *get_persistent_gnt(struct xen_blkif_ring *ring, static void put_persistent_gnt(struct xen_blkif_ring *ring, struct persistent_gnt *persistent_gnt) { - if(!test_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags)) + if (!persistent_gnt->active) pr_alert_ratelimited("freeing a grant already unused\n"); - set_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags); - clear_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags); + persistent_gnt->last_used = jiffies; + persistent_gnt->active = false; atomic_dec(&ring->persistent_gnt_in_use); } @@ -371,26 +389,26 @@ static void purge_persistent_gnt(struct xen_blkif_ring *ring) struct persistent_gnt *persistent_gnt; struct rb_node *n; unsigned int num_clean, total; - bool scan_used = false, clean_used = false; + bool scan_used = false; struct rb_root *root; - if (ring->persistent_gnt_c < xen_blkif_max_pgrants || - (ring->persistent_gnt_c == xen_blkif_max_pgrants && - !ring->blkif->vbd.overflow_max_grants)) { - goto out; - } - if (work_busy(&ring->persistent_purge_work)) { pr_alert_ratelimited("Scheduled work from previous purge is still busy, cannot purge list\n"); goto out; } - num_clean = (xen_blkif_max_pgrants / 100) * LRU_PERCENT_CLEAN; - num_clean = ring->persistent_gnt_c - xen_blkif_max_pgrants + num_clean; - num_clean = min(ring->persistent_gnt_c, num_clean); - if ((num_clean == 0) || - (num_clean > (ring->persistent_gnt_c - atomic_read(&ring->persistent_gnt_in_use)))) - goto out; + if (ring->persistent_gnt_c < xen_blkif_max_pgrants || + (ring->persistent_gnt_c == xen_blkif_max_pgrants && + !ring->blkif->vbd.overflow_max_grants)) { + num_clean = 0; + } else { + num_clean = (xen_blkif_max_pgrants / 100) * LRU_PERCENT_CLEAN; + num_clean = ring->persistent_gnt_c - xen_blkif_max_pgrants + + num_clean; + num_clean = min(ring->persistent_gnt_c, num_clean); + pr_debug("Going to purge at least %u persistent grants\n", + num_clean); + } /* * At this point, we can assure that there will be no calls @@ -401,9 +419,7 @@ static void purge_persistent_gnt(struct xen_blkif_ring *ring) * number of grants. */ - total = num_clean; - - pr_debug("Going to purge %u persistent grants\n", num_clean); + total = 0; BUG_ON(!list_empty(&ring->persistent_purge_list)); root = &ring->persistent_gnts; @@ -412,46 +428,37 @@ purge_list: BUG_ON(persistent_gnt->handle == BLKBACK_INVALID_HANDLE); - if (clean_used) { - clear_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags); + if (persistent_gnt->active) continue; - } - - if (test_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags)) + if (!scan_used && !persistent_gnt_timeout(persistent_gnt)) continue; - if (!scan_used && - (test_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags))) + if (scan_used && total >= num_clean) continue; rb_erase(&persistent_gnt->node, root); list_add(&persistent_gnt->remove_node, &ring->persistent_purge_list); - if (--num_clean == 0) - goto finished; + total++; } /* - * If we get here it means we also need to start cleaning + * Check whether we also need to start cleaning * grants that were used since last purge in order to cope * with the requested num */ - if (!scan_used && !clean_used) { - pr_debug("Still missing %u purged frames\n", num_clean); + if (!scan_used && total < num_clean) { + pr_debug("Still missing %u purged frames\n", num_clean - total); scan_used = true; goto purge_list; } -finished: - if (!clean_used) { - pr_debug("Finished scanning for grants to clean, removing used flag\n"); - clean_used = true; - goto purge_list; - } - ring->persistent_gnt_c -= (total - num_clean); - ring->blkif->vbd.overflow_max_grants = 0; + if (total) { + ring->persistent_gnt_c -= total; + ring->blkif->vbd.overflow_max_grants = 0; - /* We can defer this work */ - schedule_work(&ring->persistent_purge_work); - pr_debug("Purged %u/%u\n", (total - num_clean), total); + /* We can defer this work */ + schedule_work(&ring->persistent_purge_work); + pr_debug("Purged %u/%u\n", num_clean, total); + } out: return; diff --git a/drivers/block/xen-blkback/common.h b/drivers/block/xen-blkback/common.h index ecb35fe8ca8d..1d3002d773f7 100644 --- a/drivers/block/xen-blkback/common.h +++ b/drivers/block/xen-blkback/common.h @@ -233,16 +233,6 @@ struct xen_vbd { struct backend_info; -/* Number of available flags */ -#define PERSISTENT_GNT_FLAGS_SIZE 2 -/* This persistent grant is currently in use */ -#define PERSISTENT_GNT_ACTIVE 0 -/* - * This persistent grant has been used, this flag is set when we remove the - * PERSISTENT_GNT_ACTIVE, to know that this grant has been used recently. - */ -#define PERSISTENT_GNT_WAS_ACTIVE 1 - /* Number of requests that we can fit in a ring */ #define XEN_BLKIF_REQS_PER_PAGE 32 @@ -250,7 +240,8 @@ struct persistent_gnt { struct page *page; grant_ref_t gnt; grant_handle_t handle; - DECLARE_BITMAP(flags, PERSISTENT_GNT_FLAGS_SIZE); + unsigned long last_used; + bool active; struct rb_node node; struct list_head remove_node; }; @@ -278,7 +269,6 @@ struct xen_blkif_ring { wait_queue_head_t pending_free_wq; /* Tree to store persistent grants. */ - spinlock_t pers_gnts_lock; struct rb_root persistent_gnts; unsigned int persistent_gnt_c; atomic_t persistent_gnt_in_use; diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c index 8986adab9bf5..429d20131c7e 100644 --- a/drivers/block/xen-blkfront.c +++ b/drivers/block/xen-blkfront.c @@ -46,6 +46,7 @@ #include <linux/scatterlist.h> #include <linux/bitmap.h> #include <linux/list.h> +#include <linux/workqueue.h> #include <xen/xen.h> #include <xen/xenbus.h> @@ -121,6 +122,8 @@ static inline struct blkif_req *blkif_req(struct request *rq) static DEFINE_MUTEX(blkfront_mutex); static const struct block_device_operations xlvbd_block_fops; +static struct delayed_work blkfront_work; +static LIST_HEAD(info_list); /* * Maximum number of segments in indirect requests, the actual value used by @@ -216,6 +219,7 @@ struct blkfront_info /* Save uncomplete reqs and bios for migration. */ struct list_head requests; struct bio_list bio_list; + struct list_head info_list; }; static unsigned int nr_minors; @@ -1759,6 +1763,12 @@ abort_transaction: return err; } +static void free_info(struct blkfront_info *info) +{ + list_del(&info->info_list); + kfree(info); +} + /* Common code used when first setting up, and when resuming. */ static int talk_to_blkback(struct xenbus_device *dev, struct blkfront_info *info) @@ -1880,7 +1890,10 @@ again: destroy_blkring: blkif_free(info, 0); - kfree(info); + mutex_lock(&blkfront_mutex); + free_info(info); + mutex_unlock(&blkfront_mutex); + dev_set_drvdata(&dev->dev, NULL); return err; @@ -1991,6 +2004,10 @@ static int blkfront_probe(struct xenbus_device *dev, info->handle = simple_strtoul(strrchr(dev->nodename, '/')+1, NULL, 0); dev_set_drvdata(&dev->dev, info); + mutex_lock(&blkfront_mutex); + list_add(&info->info_list, &info_list); + mutex_unlock(&blkfront_mutex); + return 0; } @@ -2301,6 +2318,12 @@ static void blkfront_gather_backend_features(struct blkfront_info *info) if (indirect_segments <= BLKIF_MAX_SEGMENTS_PER_REQUEST) indirect_segments = 0; info->max_indirect_segments = indirect_segments; + + if (info->feature_persistent) { + mutex_lock(&blkfront_mutex); + schedule_delayed_work(&blkfront_work, HZ * 10); + mutex_unlock(&blkfront_mutex); + } } /* @@ -2482,7 +2505,9 @@ static int blkfront_remove(struct xenbus_device *xbdev) mutex_unlock(&info->mutex); if (!bdev) { - kfree(info); + mutex_lock(&blkfront_mutex); + free_info(info); + mutex_unlock(&blkfront_mutex); return 0; } @@ -2502,7 +2527,9 @@ static int blkfront_remove(struct xenbus_device *xbdev) if (info && !bdev->bd_openers) { xlvbd_release_gendisk(info); disk->private_data = NULL; - kfree(info); + mutex_lock(&blkfront_mutex); + free_info(info); + mutex_unlock(&blkfront_mutex); } mutex_unlock(&bdev->bd_mutex); @@ -2585,7 +2612,7 @@ static void blkif_release(struct gendisk *disk, fmode_t mode) dev_info(disk_to_dev(bdev->bd_disk), "releasing disk\n"); xlvbd_release_gendisk(info); disk->private_data = NULL; - kfree(info); + free_info(info); } out: @@ -2618,6 +2645,61 @@ static struct xenbus_driver blkfront_driver = { .is_ready = blkfront_is_ready, }; +static void purge_persistent_grants(struct blkfront_info *info) +{ + unsigned int i; + unsigned long flags; + + for (i = 0; i < info->nr_rings; i++) { + struct blkfront_ring_info *rinfo = &info->rinfo[i]; + struct grant *gnt_list_entry, *tmp; + + spin_lock_irqsave(&rinfo->ring_lock, flags); + + if (rinfo->persistent_gnts_c == 0) { + spin_unlock_irqrestore(&rinfo->ring_lock, flags); + continue; + } + + list_for_each_entry_safe(gnt_list_entry, tmp, &rinfo->grants, + node) { + if (gnt_list_entry->gref == GRANT_INVALID_REF || + gnttab_query_foreign_access(gnt_list_entry->gref)) + continue; + + list_del(&gnt_list_entry->node); + gnttab_end_foreign_access(gnt_list_entry->gref, 0, 0UL); + rinfo->persistent_gnts_c--; + gnt_list_entry->gref = GRANT_INVALID_REF; + list_add_tail(&gnt_list_entry->node, &rinfo->grants); + } + + spin_unlock_irqrestore(&rinfo->ring_lock, flags); + } +} + +static void blkfront_delay_work(struct work_struct *work) +{ + struct blkfront_info *info; + bool need_schedule_work = false; + + mutex_lock(&blkfront_mutex); + + list_for_each_entry(info, &info_list, info_list) { + if (info->feature_persistent) { + need_schedule_work = true; + mutex_lock(&info->mutex); + purge_persistent_grants(info); + mutex_unlock(&info->mutex); + } + } + + if (need_schedule_work) + schedule_delayed_work(&blkfront_work, HZ * 10); + + mutex_unlock(&blkfront_mutex); +} + static int __init xlblk_init(void) { int ret; @@ -2626,6 +2708,15 @@ static int __init xlblk_init(void) if (!xen_domain()) return -ENODEV; + if (!xen_has_pv_disk_devices()) + return -ENODEV; + + if (register_blkdev(XENVBD_MAJOR, DEV_NAME)) { + pr_warn("xen_blk: can't get major %d with name %s\n", + XENVBD_MAJOR, DEV_NAME); + return -ENODEV; + } + if (xen_blkif_max_segments < BLKIF_MAX_SEGMENTS_PER_REQUEST) xen_blkif_max_segments = BLKIF_MAX_SEGMENTS_PER_REQUEST; @@ -2641,14 +2732,7 @@ static int __init xlblk_init(void) xen_blkif_max_queues = nr_cpus; } - if (!xen_has_pv_disk_devices()) - return -ENODEV; - - if (register_blkdev(XENVBD_MAJOR, DEV_NAME)) { - printk(KERN_WARNING "xen_blk: can't get major %d with name %s\n", - XENVBD_MAJOR, DEV_NAME); - return -ENODEV; - } + INIT_DELAYED_WORK(&blkfront_work, blkfront_delay_work); ret = xenbus_register_frontend(&blkfront_driver); if (ret) { @@ -2663,6 +2747,8 @@ module_init(xlblk_init); static void __exit xlblk_exit(void) { + cancel_delayed_work_sync(&blkfront_work); + xenbus_unregister_driver(&blkfront_driver); unregister_blkdev(XENVBD_MAJOR, DEV_NAME); kfree(minors); |