From 81b73dd92b97423b8f5324a59044da478c04f4c4 Mon Sep 17 00:00:00 2001 From: Haren Myneni Date: Thu, 27 Jul 2006 14:29:00 -0700 Subject: [POWERPC] Fix might-sleep warning on removing cpus Noticing the following might_sleep warning (dump_stack()) during kdump testing when CONFIG_DEBUG_SPINLOCK_SLEEP is enabled. All secondary CPUs will be calling rtas_set_indicator with interrupts disabled to remove them from global interrupt queue. BUG: sleeping function called from invalid context at arch/powerpc/kernel/rtas.c:463 in_atomic():1, irqs_disabled():1 Call Trace: [C00000000FFFB970] [C000000000010234] .show_stack+0x68/0x1b0 (unreliable) [C00000000FFFBA10] [C000000000059354] .__might_sleep+0xd8/0xf4 [C00000000FFFBA90] [C00000000001D1BC] .rtas_busy_delay+0x20/0x5c [C00000000FFFBB20] [C00000000001D8A8] .rtas_set_indicator+0x6c/0xcc [C00000000FFFBBC0] [C000000000048BF4] .xics_teardown_cpu+0x118/0x134 [C00000000FFFBC40] [C00000000004539C] .pseries_kexec_cpu_down_xics+0x74/0x8c [C00000000FFFBCC0] [C00000000002DF08] .crash_ipi_callback+0x15c/0x188 [C00000000FFFBD50] [C0000000000296EC] .smp_message_recv+0x84/0xdc [C00000000FFFBDC0] [C000000000048E08] .xics_ipi_dispatch+0xf0/0x130 [C00000000FFFBE50] [C00000000009EF10] .handle_IRQ_event+0x7c/0xf8 [C00000000FFFBF00] [C0000000000A0A14] .handle_percpu_irq+0x90/0x10c [C00000000FFFBF90] [C00000000002659C] .call_handle_irq+0x1c/0x2c [C00000000058B9C0] [C00000000000CA10] .do_IRQ+0xf4/0x1a4 [C00000000058BA50] [C0000000000044EC] hardware_interrupt_entry+0xc/0x10 --- Exception: 501 at .plpar_hcall_norets+0x14/0x1c LR = .pseries_dedicated_idle_sleep+0x190/0x1d4 [C00000000058BD40] [C00000000058BDE0] 0xc00000000058bde0 (unreliable) [C00000000058BDF0] [C00000000001270C] .cpu_idle+0x10c/0x1e0 [C00000000058BE70] [C000000000009274] .rest_init+0x44/0x5c To fix this issue, rtas_set_indicator_fast() is added so that will not wait for RTAS 'busy' delay and this new function is used for kdump (in xics_teardown_cpu()) and for CPU hotplug ( xics_migrate_irqs_away() and xics_setup_cpu()). Note that the platform architecture spec says that set-indicator on the indicator we're using here is not permitted to return the busy or extended busy status codes. Signed-off-by: Haren Myneni Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/rtas.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/rtas.c b/arch/powerpc/kernel/rtas.c index 4a4cb5598402..77f1e06d208d 100644 --- a/arch/powerpc/kernel/rtas.c +++ b/arch/powerpc/kernel/rtas.c @@ -569,6 +569,27 @@ int rtas_set_indicator(int indicator, int index, int new_value) } EXPORT_SYMBOL(rtas_set_indicator); +/* + * Ignoring RTAS extended delay + */ +int rtas_set_indicator_fast(int indicator, int index, int new_value) +{ + int rc; + int token = rtas_token("set-indicator"); + + if (token == RTAS_UNKNOWN_SERVICE) + return -ENOENT; + + rc = rtas_call(token, 3, 1, NULL, indicator, index, new_value); + + WARN_ON(rc == -2 || (rc >= 9900 && rc <= 9905)); + + if (rc < 0) + return rtas_error_rc(rc); + + return rc; +} + void rtas_restart(char *cmd) { if (rtas_flash_term_hook) -- cgit v1.2.1 From 012c437d03cb299814e58ac8d574f7510f5989a5 Mon Sep 17 00:00:00 2001 From: Horms Date: Sun, 13 Aug 2006 23:24:22 -0700 Subject: [PATCH] Change panic_on_oops message to "Fatal exception" Previously the message was "Fatal exception: panic_on_oops", as introduced in a recent patch whith removed a somewhat dangerous call to ssleep() in the panic_on_oops path. However, Paul Mackerras suggested that this was somewhat confusing, leadind people to believe that it was panic_on_oops that was the root cause of the fatal exception. On his suggestion, this patch changes the message to simply "Fatal exception". A suitable oops message should already have been displayed. Signed-off-by: Simon Horman Cc: Paul Mackerras Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman --- arch/powerpc/kernel/traps.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c index 2105767fcc57..05682a24b210 100644 --- a/arch/powerpc/kernel/traps.c +++ b/arch/powerpc/kernel/traps.c @@ -151,7 +151,7 @@ int die(const char *str, struct pt_regs *regs, long err) panic("Fatal exception in interrupt"); if (panic_on_oops) - panic("Fatal exception: panic_on_oops"); + panic("Fatal exception"); do_exit(err); -- cgit v1.2.1 From eac8392f9591b359847b8a6abb75ca60967bcecd Mon Sep 17 00:00:00 2001 From: David Wilder Date: Thu, 29 Jun 2006 15:17:30 -0700 Subject: [POWERPC] Make secondary CPUs call into kdump on reset exception In the case of a system hang, the user will invoke soft-reset to initiate the kdump boot. If xmon is enabled, the CPU(s) enter into the xmon debugger. Unfortunately, the secondary CPU(s) will return to the hung state when they exit from the debugger (returned from die() -> system_reset_exception()). This causes a problem in kdump since the hung CPU(s) will not respond to the IPI sent from kdump. This patch fixes the issue by calling crash_kexec_secondary() directly from system_reset_exception() without returning to the previous state. These secondary CPUs wait 5ms until the kdump boot is started by the primary CPU. In the case we exited from the debugger to "recover" (command 'x' in xmon) the primary and the secondary CPUs will all return from die() -> system_reset_exception() ->crash_kexec_secondary() wait 5ms, then return to the previous state. A kdump boot is not started in this case. Signed-off-by: Haren Myneni Signed-off-by: David Wilder Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/traps.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c index 2105767fcc57..5147175f921c 100644 --- a/arch/powerpc/kernel/traps.c +++ b/arch/powerpc/kernel/traps.c @@ -211,6 +211,19 @@ void system_reset_exception(struct pt_regs *regs) die("System Reset", regs, SIGABRT); + /* + * Some CPUs when released from the debugger will execute this path. + * These CPUs entered the debugger via a soft-reset. If the CPU was + * hung before entering the debugger it will return to the hung + * state when exiting this function. This causes a problem in + * kdump since the hung CPU(s) will not respond to the IPI sent + * from kdump. To prevent the problem we call crash_kexec_secondary() + * here. If a kdump had not been initiated or we exit the debugger + * with the "exit and recover" command (x) crash_kexec_secondary() + * will return after 5ms and the CPU returns to its previous state. + */ + crash_kexec_secondary(regs); + /* Must die if the interrupt is not recoverable */ if (!(regs->msr & MSR_RI)) panic("Unrecoverable System Reset"); -- cgit v1.2.1 From 47585d8f5dea10dea49c948f1fb13ef7632409c7 Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Wed, 5 Jul 2006 14:39:42 +1000 Subject: [POWERPC] Move some kexec logic into machine_kexec.c We're missing a few functions for kexec to compile on 32-bit. There's nothing really 64-bit specific about the 64-bit versions, so make them generic rather than adding empty definitions for 32-bit. Signed-off-by: Michael Ellerman Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/machine_kexec.c | 56 ++++++++++++++++++++++++++++++++++ arch/powerpc/kernel/machine_kexec_64.c | 56 ---------------------------------- 2 files changed, 56 insertions(+), 56 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/machine_kexec.c b/arch/powerpc/kernel/machine_kexec.c index a81ca1b841ec..e60a0c544d63 100644 --- a/arch/powerpc/kernel/machine_kexec.c +++ b/arch/powerpc/kernel/machine_kexec.c @@ -13,6 +13,7 @@ #include #include #include +#include void machine_crash_shutdown(struct pt_regs *regs) { @@ -59,3 +60,58 @@ NORET_TYPE void machine_kexec(struct kimage *image) } for(;;); } + +static int __init early_parse_crashk(char *p) +{ + unsigned long size; + + if (!p) + return 1; + + size = memparse(p, &p); + + if (*p == '@') + crashk_res.start = memparse(p + 1, &p); + else + crashk_res.start = KDUMP_KERNELBASE; + + crashk_res.end = crashk_res.start + size - 1; + + return 0; +} +early_param("crashkernel", early_parse_crashk); + +void __init reserve_crashkernel(void) +{ + unsigned long size; + + if (crashk_res.start == 0) + return; + + /* We might have got these values via the command line or the + * device tree, either way sanitise them now. */ + + size = crashk_res.end - crashk_res.start + 1; + + if (crashk_res.start != KDUMP_KERNELBASE) + printk("Crash kernel location must be 0x%x\n", + KDUMP_KERNELBASE); + + crashk_res.start = KDUMP_KERNELBASE; + size = PAGE_ALIGN(size); + crashk_res.end = crashk_res.start + size - 1; + + /* Crash kernel trumps memory limit */ + if (memory_limit && memory_limit <= crashk_res.end) { + memory_limit = crashk_res.end + 1; + printk("Adjusted memory limit for crashkernel, now 0x%lx\n", + memory_limit); + } + + lmb_reserve(crashk_res.start, size); +} + +int overlaps_crashkernel(unsigned long start, unsigned long size) +{ + return (start + size) > crashk_res.start && start <= crashk_res.end; +} diff --git a/arch/powerpc/kernel/machine_kexec_64.c b/arch/powerpc/kernel/machine_kexec_64.c index b438d45a068c..b6db19d61435 100644 --- a/arch/powerpc/kernel/machine_kexec_64.c +++ b/arch/powerpc/kernel/machine_kexec_64.c @@ -21,7 +21,6 @@ #include #include #include -#include #include #include /* _end */ #include @@ -385,58 +384,3 @@ static int __init kexec_setup(void) return 0; } __initcall(kexec_setup); - -static int __init early_parse_crashk(char *p) -{ - unsigned long size; - - if (!p) - return 1; - - size = memparse(p, &p); - - if (*p == '@') - crashk_res.start = memparse(p + 1, &p); - else - crashk_res.start = KDUMP_KERNELBASE; - - crashk_res.end = crashk_res.start + size - 1; - - return 0; -} -early_param("crashkernel", early_parse_crashk); - -void __init reserve_crashkernel(void) -{ - unsigned long size; - - if (crashk_res.start == 0) - return; - - /* We might have got these values via the command line or the - * device tree, either way sanitise them now. */ - - size = crashk_res.end - crashk_res.start + 1; - - if (crashk_res.start != KDUMP_KERNELBASE) - printk("Crash kernel location must be 0x%x\n", - KDUMP_KERNELBASE); - - crashk_res.start = KDUMP_KERNELBASE; - size = PAGE_ALIGN(size); - crashk_res.end = crashk_res.start + size - 1; - - /* Crash kernel trumps memory limit */ - if (memory_limit && memory_limit <= crashk_res.end) { - memory_limit = crashk_res.end + 1; - printk("Adjusted memory limit for crashkernel, now 0x%lx\n", - memory_limit); - } - - lmb_reserve(crashk_res.start, size); -} - -int overlaps_crashkernel(unsigned long start, unsigned long size) -{ - return (start + size) > crashk_res.start && start <= crashk_res.end; -} -- cgit v1.2.1 From b6f35b4966e0ae59cec45e5292b100698d12dc5f Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Wed, 5 Jul 2006 14:39:43 +1000 Subject: [POWERPC] Make crash.c work on 32-bit and 64-bit To compile kexec on 32-bit we need a few more bits and pieces. Rather than add empty definitions, we can make crash.c work on 32-bit, with only a couple of kludges. Signed-off-by: Michael Ellerman Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/Makefile | 4 ++-- arch/powerpc/kernel/crash.c | 11 +++++++++++ arch/powerpc/kernel/machine_kexec_64.c | 1 - arch/powerpc/kernel/traps.c | 3 --- 4 files changed, 13 insertions(+), 6 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile index 814f242aeb8c..956c2e5564b7 100644 --- a/arch/powerpc/kernel/Makefile +++ b/arch/powerpc/kernel/Makefile @@ -67,9 +67,9 @@ pci64-$(CONFIG_PPC64) += pci_64.o pci_dn.o pci_iommu.o \ pci_direct_iommu.o iomap.o pci32-$(CONFIG_PPC32) := pci_32.o obj-$(CONFIG_PCI) += $(pci64-y) $(pci32-y) -kexec-$(CONFIG_PPC64) := machine_kexec_64.o crash.o +kexec-$(CONFIG_PPC64) := machine_kexec_64.o kexec-$(CONFIG_PPC32) := machine_kexec_32.o -obj-$(CONFIG_KEXEC) += machine_kexec.o $(kexec-y) +obj-$(CONFIG_KEXEC) += machine_kexec.o crash.o $(kexec-y) ifeq ($(CONFIG_PPC_ISERIES),y) $(obj)/head_64.o: $(obj)/lparmap.s diff --git a/arch/powerpc/kernel/crash.c b/arch/powerpc/kernel/crash.c index 358cecdc6aef..f04c18e08b8b 100644 --- a/arch/powerpc/kernel/crash.c +++ b/arch/powerpc/kernel/crash.c @@ -44,6 +44,7 @@ /* This keeps a track of which one is crashing cpu. */ int crashing_cpu = -1; static cpumask_t cpus_in_crash = CPU_MASK_NONE; +cpumask_t cpus_in_sr = CPU_MASK_NONE; static u32 *append_elf_note(u32 *buf, char *name, unsigned type, void *data, size_t data_len) @@ -139,7 +140,13 @@ void crash_ipi_callback(struct pt_regs *regs) if (ppc_md.kexec_cpu_down) ppc_md.kexec_cpu_down(1, 1); + +#ifdef CONFIG_PPC64 kexec_smp_wait(); +#else + for (;;); /* FIXME */ +#endif + /* NOTREACHED */ } @@ -255,7 +262,11 @@ static void crash_kexec_prepare_cpus(int cpu) * * do this if kexec in setup.c ? */ +#ifdef CONFIG_PPC64 smp_release_cpus(); +#else + /* FIXME */ +#endif } void crash_kexec_secondary(struct pt_regs *regs) diff --git a/arch/powerpc/kernel/machine_kexec_64.c b/arch/powerpc/kernel/machine_kexec_64.c index b6db19d61435..be58985c7681 100644 --- a/arch/powerpc/kernel/machine_kexec_64.c +++ b/arch/powerpc/kernel/machine_kexec_64.c @@ -10,7 +10,6 @@ */ -#include #include #include #include diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c index 5147175f921c..125761aaa40b 100644 --- a/arch/powerpc/kernel/traps.c +++ b/arch/powerpc/kernel/traps.c @@ -55,9 +55,6 @@ #ifdef CONFIG_PPC64 /* XXX */ #define _IO_BASE pci_io_base -#ifdef CONFIG_KEXEC -cpumask_t cpus_in_sr = CPU_MASK_NONE; -#endif #endif #ifdef CONFIG_DEBUGGER -- cgit v1.2.1 From 83db3dde2604d48e30a468c82da163dd1a8602b3 Mon Sep 17 00:00:00 2001 From: Ananth N Mavinakayanahalli Date: Fri, 11 Aug 2006 17:01:34 +0530 Subject: [POWERPC] kprobes: Fix possible system crash during out-of-line single-stepping - On archs that have no-exec support, we vmalloc() a executable scratch area of PAGE_SIZE and divide it up into an array of slots of maximum instruction size for that arch - On a kprobe registration, the original instruction is copied to the first available free slot, so if multiple kprobes are registered, chances are, they get contiguous slots - On POWER4, due to not having coherent icaches, we could hit a situation where a probe that is registered on one processor, is hit immediately on another. This second processor could have fetched the stream of text from the out-of-line single-stepping area *before* the probe registration completed, possibly due to an earlier (and a different) kprobe hit and hence would see stale data at the slot. Executing such an arbitrary instruction lead to a problem as reported in LTC bugzilla 23555. The correct solution is to call flush_icache_range() as soon as the instruction is copied for out-of-line single-stepping, so the correct instruction is seen on all processors. Thanks to Will Schmidt who tracked this down. Signed-off-by: Ananth N Mavinakayanahalli Acked-by: Will Schmidt Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/kprobes.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/kprobes.c b/arch/powerpc/kernel/kprobes.c index 9f0898c89759..cd65c367b8b6 100644 --- a/arch/powerpc/kernel/kprobes.c +++ b/arch/powerpc/kernel/kprobes.c @@ -61,6 +61,8 @@ int __kprobes arch_prepare_kprobe(struct kprobe *p) if (!ret) { memcpy(p->ainsn.insn, p->addr, MAX_INSN_SIZE * sizeof(kprobe_opcode_t)); p->opcode = *p->addr; + flush_icache_range((unsigned long)p->ainsn.insn, + (unsigned long)p->ainsn.insn + sizeof(kprobe_opcode_t)); } return ret; -- cgit v1.2.1 From e5c14ce118ab6a96afb2af5c4ec9727b03ab1936 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Thu, 17 Aug 2006 03:59:27 +0200 Subject: [POWERPC] Fix irq radix tree remapping typo The code for using the radix tree for reverse mapping of interrupts has a typo that causes it to create incorrect mappings if the software and hardware numbers happen to be different. This would, among others, cause the IDE interrupt to fail on js20's. This fixes it. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/irq.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c index 01bdae35cb55..7ee685433319 100644 --- a/arch/powerpc/kernel/irq.c +++ b/arch/powerpc/kernel/irq.c @@ -693,7 +693,7 @@ unsigned int irq_radix_revmap(struct irq_host *host, /* If not there, try to insert it */ virq = irq_find_mapping(host, hwirq); if (virq != NO_IRQ) - radix_tree_insert(tree, virq, &irq_map[virq]); + radix_tree_insert(tree, hwirq, &irq_map[virq]); bail: spin_unlock_irqrestore(&irq_big_lock, flags); return virq; -- cgit v1.2.1 From 9a936a2e0526089194159eae31238e36b1c19e74 Mon Sep 17 00:00:00 2001 From: Olof Johansson Date: Sun, 6 Aug 2006 20:51:25 -0500 Subject: [POWERPC] powerpc: Clear HID0 attention enable on PPC970 at boot time Clear HID0[en_attn] at CPU init time on PPC970. Closes CVE-2006-4093. Signed-off-by: Olof Johansson Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/cpu_setup_power4.S | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/cpu_setup_power4.S b/arch/powerpc/kernel/cpu_setup_power4.S index f69af2c5d7b3..76e97aa71c45 100644 --- a/arch/powerpc/kernel/cpu_setup_power4.S +++ b/arch/powerpc/kernel/cpu_setup_power4.S @@ -76,6 +76,8 @@ _GLOBAL(__setup_cpu_ppc970) mfspr r0,SPRN_HID0 li r11,5 /* clear DOZE and SLEEP */ rldimi r0,r11,52,8 /* set NAP and DPM */ + li r11,0 + rldimi r0,r11,32,31 /* clear EN_ATTN */ mtspr SPRN_HID0,r0 mfspr r0,SPRN_HID0 mfspr r0,SPRN_HID0 -- cgit v1.2.1 From 5db9fa9593e2ff69f2b95f9d59229dc4faaa564d Mon Sep 17 00:00:00 2001 From: Nathan Lynch Date: Tue, 22 Aug 2006 20:36:05 -0500 Subject: [POWERPC] Fix gettimeofday inaccuracies There are two problems in the powerpc gettimeofday code which can cause incorrect results to be returned. The first is that there is a race between do_gettimeofday and the timer interrupt: 1. do_gettimeofday does get_tb() 2. decrementer exception on boot cpu which runs timer_recalc_offset, which also samples the timebase and updates the do_gtod structure with a greater timebase value. 3. do_gettimeofday calls __do_gettimeofday, which leads to the negative result from tb_val - temp_varp->tb_orig_stamp. The second is caused by taking the boot cpu offline, which can cause the value of tb_last_jiffy to be increased past the currently available timebase, causing the same underflow as above. [paulus@samba.org - define and use data_barrier() instead of mb().] Signed-off-by: Nathan Lynch Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/time.c | 25 +++++++++++++++++-------- 1 file changed, 17 insertions(+), 8 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 774c0a3c5019..18e59e43d2b3 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -417,7 +417,7 @@ static __inline__ void timer_check_rtc(void) /* * This version of gettimeofday has microsecond resolution. */ -static inline void __do_gettimeofday(struct timeval *tv, u64 tb_val) +static inline void __do_gettimeofday(struct timeval *tv) { unsigned long sec, usec; u64 tb_ticks, xsec; @@ -431,7 +431,12 @@ static inline void __do_gettimeofday(struct timeval *tv, u64 tb_val) * without a divide (and in fact, without a multiply) */ temp_varp = do_gtod.varp; - tb_ticks = tb_val - temp_varp->tb_orig_stamp; + + /* Sampling the time base must be done after loading + * do_gtod.varp in order to avoid racing with update_gtod. + */ + data_barrier(temp_varp); + tb_ticks = get_tb() - temp_varp->tb_orig_stamp; temp_tb_to_xs = temp_varp->tb_to_xs; temp_stamp_xsec = temp_varp->stamp_xsec; xsec = temp_stamp_xsec + mulhdu(tb_ticks, temp_tb_to_xs); @@ -464,7 +469,7 @@ void do_gettimeofday(struct timeval *tv) tv->tv_usec = usec; return; } - __do_gettimeofday(tv, get_tb()); + __do_gettimeofday(tv); } EXPORT_SYMBOL(do_gettimeofday); @@ -650,6 +655,7 @@ void timer_interrupt(struct pt_regs * regs) int next_dec; int cpu = smp_processor_id(); unsigned long ticks; + u64 tb_next_jiffy; #ifdef CONFIG_PPC32 if (atomic_read(&ppc_n_lost_interrupts) != 0) @@ -691,11 +697,14 @@ void timer_interrupt(struct pt_regs * regs) continue; write_seqlock(&xtime_lock); - tb_last_jiffy += tb_ticks_per_jiffy; - tb_last_stamp = per_cpu(last_jiffy, cpu); - do_timer(regs); - timer_recalc_offset(tb_last_jiffy); - timer_check_rtc(); + tb_next_jiffy = tb_last_jiffy + tb_ticks_per_jiffy; + if (per_cpu(last_jiffy, cpu) >= tb_next_jiffy) { + tb_last_jiffy = tb_next_jiffy; + tb_last_stamp = per_cpu(last_jiffy, cpu); + do_timer(regs); + timer_recalc_offset(tb_last_jiffy); + timer_check_rtc(); + } write_sequnlock(&xtime_lock); } -- cgit v1.2.1 From be9633e96a9e33ab97776bb6874244fa4db6987d Mon Sep 17 00:00:00 2001 From: Zang Roy-r61911 Date: Wed, 23 Aug 2006 10:20:27 +0800 Subject: [POWERPC] Pass UPIO_TSI flag to 8259 serial driver The patch passes the UPIO_TSI flag to general 8259 serial driver Signed-off-by: Roy Zang Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/legacy_serial.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/legacy_serial.c b/arch/powerpc/kernel/legacy_serial.c index 359ab89748e0..40a39291861f 100644 --- a/arch/powerpc/kernel/legacy_serial.c +++ b/arch/powerpc/kernel/legacy_serial.c @@ -115,6 +115,7 @@ static int __init add_legacy_soc_port(struct device_node *np, u64 addr; u32 *addrp; upf_t flags = UPF_BOOT_AUTOCONF | UPF_SKIP_TEST | UPF_SHARE_IRQ; + struct device_node *tsi = of_get_parent(np); /* We only support ports that have a clock frequency properly * encoded in the device-tree. @@ -134,7 +135,10 @@ static int __init add_legacy_soc_port(struct device_node *np, /* Add port, irq will be dealt with later. We passed a translated * IO port value. It will be fixed up later along with the irq */ - return add_legacy_port(np, -1, UPIO_MEM, addr, addr, NO_IRQ, flags, 0); + if (tsi && !strcmp(tsi->type, "tsi-bridge")) + return add_legacy_port(np, -1, UPIO_TSI, addr, addr, NO_IRQ, flags, 0); + else + return add_legacy_port(np, -1, UPIO_MEM, addr, addr, NO_IRQ, flags, 0); } static int __init add_legacy_isa_port(struct device_node *np, @@ -464,7 +468,7 @@ static int __init serial_dev_init(void) fixup_port_irq(i, np, port); if (port->iotype == UPIO_PORT) fixup_port_pio(i, np, port); - if (port->iotype == UPIO_MEM) + if ((port->iotype == UPIO_MEM) || (port->iotype == UPIO_TSI)) fixup_port_mmio(i, np, port); } -- cgit v1.2.1 From 87589f08beaec3e1f8a3af0c72406c845f706821 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Wed, 23 Aug 2006 16:58:39 +1000 Subject: [POWERPC] Correct masks used in emulating some instructions When we get an illegal instruction exception, we check to see whether the instruction is one that we emulate for the user program. Some of the masks we use in checking whether the offending instruction is one we care about didn't have the top bit set, which is the MSB of the major opcode. Thus some undefined opcodes could get emulated as other (defined but unimplemented) instructions. This corrects the masks. Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/traps.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c index 125761aaa40b..4d0b4e74d579 100644 --- a/arch/powerpc/kernel/traps.c +++ b/arch/powerpc/kernel/traps.c @@ -585,14 +585,14 @@ static void parse_fpe(struct pt_regs *regs) #define INST_MFSPR_PVR_MASK 0xfc1fffff #define INST_DCBA 0x7c0005ec -#define INST_DCBA_MASK 0x7c0007fe +#define INST_DCBA_MASK 0xfc0007fe #define INST_MCRXR 0x7c000400 -#define INST_MCRXR_MASK 0x7c0007fe +#define INST_MCRXR_MASK 0xfc0007fe #define INST_STRING 0x7c00042a -#define INST_STRING_MASK 0x7c0007fe -#define INST_STRING_GEN_MASK 0x7c00067e +#define INST_STRING_MASK 0xfc0007fe +#define INST_STRING_GEN_MASK 0xfc00067e #define INST_LSWI 0x7c0004aa #define INST_LSWX 0x7c00042a #define INST_STSWI 0x7c0005aa -- cgit v1.2.1 From d55c4a76f26160482158cd43788dcfc96a320a4f Mon Sep 17 00:00:00 2001 From: Li Yang Date: Wed, 23 Aug 2006 14:13:08 +0800 Subject: [POWERPC] Fix compile problem without CONFIG_PCI Compile fails without defining CONFIG_PCI. The patch fix this. [paulus@samba.org: Moved of_irq_pci_swizzle so we only need one #ifdef] Signed-off-by: Li Yang Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/prom_parse.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/prom_parse.c b/arch/powerpc/kernel/prom_parse.c index 6a7e997c401d..11052c212ad5 100644 --- a/arch/powerpc/kernel/prom_parse.c +++ b/arch/powerpc/kernel/prom_parse.c @@ -598,11 +598,6 @@ static struct device_node *of_irq_find_parent(struct device_node *child) return p; } -static u8 of_irq_pci_swizzle(u8 slot, u8 pin) -{ - return (((pin - 1) + slot) % 4) + 1; -} - /* This doesn't need to be called if you don't have any special workaround * flags to pass */ @@ -891,6 +886,12 @@ int of_irq_map_one(struct device_node *device, int index, struct of_irq *out_irq } EXPORT_SYMBOL_GPL(of_irq_map_one); +#ifdef CONFIG_PCI +static u8 of_irq_pci_swizzle(u8 slot, u8 pin) +{ + return (((pin - 1) + slot) % 4) + 1; +} + int of_irq_map_pci(struct pci_dev *pdev, struct of_irq *out_irq) { struct device_node *dn, *ppnode; @@ -967,4 +968,4 @@ int of_irq_map_pci(struct pci_dev *pdev, struct of_irq *out_irq) return of_irq_map_raw(ppnode, &lspec, laddr, out_irq); } EXPORT_SYMBOL_GPL(of_irq_map_pci); - +#endif /* CONFIG_PCI */ -- cgit v1.2.1 From af07ac276af6eec5111a6567cb7eaf4d222bd332 Mon Sep 17 00:00:00 2001 From: Matt Porter Date: Fri, 4 Aug 2006 11:41:51 -0500 Subject: [POWERPC] Remove flush_dcache_all export Removes the flush_dcache_all export for non coherent platforms. We removed the last in-kernel user of this years ago in arch/ppc so it no longer serves a purpose. Plus, it breaks the build at the moment. Signed-off-by: Matt Porter Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/ppc_ksyms.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/ppc_ksyms.c b/arch/powerpc/kernel/ppc_ksyms.c index f6a05f090b25..39d3bfcabcd2 100644 --- a/arch/powerpc/kernel/ppc_ksyms.c +++ b/arch/powerpc/kernel/ppc_ksyms.c @@ -126,10 +126,6 @@ EXPORT_SYMBOL(pci_bus_mem_base_phys); EXPORT_SYMBOL(pci_bus_to_hose); #endif /* CONFIG_PCI */ -#ifdef CONFIG_NOT_COHERENT_CACHE -EXPORT_SYMBOL(flush_dcache_all); -#endif - EXPORT_SYMBOL(start_thread); EXPORT_SYMBOL(kernel_thread); -- cgit v1.2.1 From 8ec8f2e85c6b88b4a1641eb3902275bcf2c6d60a Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Mon, 28 Aug 2006 11:17:37 +1000 Subject: [POWERPC] Fix performance regression in IRQ radix tree locking When reworking the powerpc irq code, I figured out that we were using the radix tree in a racy way. As a temporary fix, I put a spinlock in there. However, this can have a significant impact on performances. This patch reworks that to use a smarter technique based on the fact that what we need is in fact a rwlock with extremely rare writers (thus optimized for the read path). Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/irq.c | 84 +++++++++++++++++++++++++++++++++++++---------- 1 file changed, 66 insertions(+), 18 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c index 7ee685433319..12c5971d6565 100644 --- a/arch/powerpc/kernel/irq.c +++ b/arch/powerpc/kernel/irq.c @@ -322,7 +322,8 @@ EXPORT_SYMBOL(do_softirq); static LIST_HEAD(irq_hosts); static spinlock_t irq_big_lock = SPIN_LOCK_UNLOCKED; - +static DEFINE_PER_CPU(unsigned int, irq_radix_reader); +static unsigned int irq_radix_writer; struct irq_map_entry irq_map[NR_IRQS]; static unsigned int irq_virq_count = NR_IRQS; static struct irq_host *irq_default_host; @@ -455,6 +456,58 @@ void irq_set_virq_count(unsigned int count) irq_virq_count = count; } +/* radix tree not lockless safe ! we use a brlock-type mecanism + * for now, until we can use a lockless radix tree + */ +static void irq_radix_wrlock(unsigned long *flags) +{ + unsigned int cpu, ok; + + spin_lock_irqsave(&irq_big_lock, *flags); + irq_radix_writer = 1; + smp_mb(); + do { + barrier(); + ok = 1; + for_each_possible_cpu(cpu) { + if (per_cpu(irq_radix_reader, cpu)) { + ok = 0; + break; + } + } + if (!ok) + cpu_relax(); + } while(!ok); +} + +static void irq_radix_wrunlock(unsigned long flags) +{ + smp_wmb(); + irq_radix_writer = 0; + spin_unlock_irqrestore(&irq_big_lock, flags); +} + +static void irq_radix_rdlock(unsigned long *flags) +{ + local_irq_save(*flags); + __get_cpu_var(irq_radix_reader) = 1; + smp_mb(); + if (likely(irq_radix_writer == 0)) + return; + __get_cpu_var(irq_radix_reader) = 0; + smp_wmb(); + spin_lock(&irq_big_lock); + __get_cpu_var(irq_radix_reader) = 1; + spin_unlock(&irq_big_lock); +} + +static void irq_radix_rdunlock(unsigned long flags) +{ + __get_cpu_var(irq_radix_reader) = 0; + local_irq_restore(flags); +} + + unsigned int irq_create_mapping(struct irq_host *host, irq_hw_number_t hwirq) { @@ -604,13 +657,9 @@ void irq_dispose_mapping(unsigned int virq) /* Check if radix tree allocated yet */ if (host->revmap_data.tree.gfp_mask == 0) break; - /* XXX radix tree not safe ! remove lock whem it becomes safe - * and use some RCU sync to make sure everything is ok before we - * can re-use that map entry - */ - spin_lock_irqsave(&irq_big_lock, flags); + irq_radix_wrlock(&flags); radix_tree_delete(&host->revmap_data.tree, hwirq); - spin_unlock_irqrestore(&irq_big_lock, flags); + irq_radix_wrunlock(flags); break; } @@ -677,25 +726,24 @@ unsigned int irq_radix_revmap(struct irq_host *host, if (tree->gfp_mask == 0) return irq_find_mapping(host, hwirq); - /* XXX Current radix trees are NOT SMP safe !!! Remove that lock - * when that is fixed (when Nick's patch gets in - */ - spin_lock_irqsave(&irq_big_lock, flags); - /* Now try to resolve */ + irq_radix_rdlock(&flags); ptr = radix_tree_lookup(tree, hwirq); + irq_radix_rdunlock(flags); + /* Found it, return */ if (ptr) { virq = ptr - irq_map; - goto bail; + return virq; } /* If not there, try to insert it */ virq = irq_find_mapping(host, hwirq); - if (virq != NO_IRQ) + if (virq != NO_IRQ) { + irq_radix_wrlock(&flags); radix_tree_insert(tree, hwirq, &irq_map[virq]); - bail: - spin_unlock_irqrestore(&irq_big_lock, flags); + irq_radix_wrunlock(flags); + } return virq; } @@ -806,12 +854,12 @@ static int irq_late_init(void) struct irq_host *h; unsigned long flags; - spin_lock_irqsave(&irq_big_lock, flags); + irq_radix_wrlock(&flags); list_for_each_entry(h, &irq_hosts, link) { if (h->revmap_type == IRQ_HOST_MAP_TREE) INIT_RADIX_TREE(&h->revmap_data.tree, GFP_ATOMIC); } - spin_unlock_irqrestore(&irq_big_lock, flags); + irq_radix_wrunlock(flags); return 0; } -- cgit v1.2.1 From 006b64de607f895de2ba1e21d3179cddf059128f Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Fri, 25 Aug 2006 14:46:23 +1000 Subject: [POWERPC] Make OF irq map code detect more error cases Device-tree bugs on js20 with some versions of SLOF were causing the interrupt for IDE to not be parsed correctly and fail to boot. This patch adds a bit more sanity checking to the parser to detect some of those errors and fail instead of returning bogus information. The powerpc PCI code can then trigger a fallback that works on those machines. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/pci_64.c | 11 +++++++++-- arch/powerpc/kernel/prom_parse.c | 17 +++++++++++++---- 2 files changed, 22 insertions(+), 6 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/pci_64.c b/arch/powerpc/kernel/pci_64.c index 2fce7738e9e2..138134c8c17d 100644 --- a/arch/powerpc/kernel/pci_64.c +++ b/arch/powerpc/kernel/pci_64.c @@ -1289,6 +1289,9 @@ int pci_read_irq_line(struct pci_dev *pci_dev) DBG("Try to map irq for %s...\n", pci_name(pci_dev)); +#ifdef DEBUG + memset(&oirq, 0xff, sizeof(oirq)); +#endif /* Try to get a mapping from the device-tree */ if (of_irq_map_pci(pci_dev, &oirq)) { u8 line, pin; @@ -1314,8 +1317,9 @@ int pci_read_irq_line(struct pci_dev *pci_dev) if (virq != NO_IRQ) set_irq_type(virq, IRQ_TYPE_LEVEL_LOW); } else { - DBG(" -> got one, spec %d cells (0x%08x...) on %s\n", - oirq.size, oirq.specifier[0], oirq.controller->full_name); + DBG(" -> got one, spec %d cells (0x%08x 0x%08x...) on %s\n", + oirq.size, oirq.specifier[0], oirq.specifier[1], + oirq.controller->full_name); virq = irq_create_of_mapping(oirq.controller, oirq.specifier, oirq.size); @@ -1324,6 +1328,9 @@ int pci_read_irq_line(struct pci_dev *pci_dev) DBG(" -> failed to map !\n"); return -1; } + + DBG(" -> mapped to linux irq %d\n", virq); + pci_dev->irq = virq; pci_write_config_byte(pci_dev, PCI_INTERRUPT_LINE, virq); diff --git a/arch/powerpc/kernel/prom_parse.c b/arch/powerpc/kernel/prom_parse.c index 11052c212ad5..a10825a5dfe6 100644 --- a/arch/powerpc/kernel/prom_parse.c +++ b/arch/powerpc/kernel/prom_parse.c @@ -639,14 +639,17 @@ void of_irq_map_init(unsigned int flags) } -int of_irq_map_raw(struct device_node *parent, u32 *intspec, u32 *addr, - struct of_irq *out_irq) +int of_irq_map_raw(struct device_node *parent, u32 *intspec, u32 ointsize, + u32 *addr, struct of_irq *out_irq) { struct device_node *ipar, *tnode, *old = NULL, *newpar = NULL; u32 *tmp, *imap, *imask; u32 intsize = 1, addrsize, newintsize = 0, newaddrsize = 0; int imaplen, match, i; + DBG("of_irq_map_raw: par=%s,intspec=[0x%08x 0x%08x...],ointsize=%d\n", + parent->full_name, intspec[0], intspec[1], ointsize); + ipar = of_node_get(parent); /* First get the #interrupt-cells property of the current cursor @@ -670,6 +673,9 @@ int of_irq_map_raw(struct device_node *parent, u32 *intspec, u32 *addr, DBG("of_irq_map_raw: ipar=%s, size=%d\n", ipar->full_name, intsize); + if (ointsize != intsize) + return -EINVAL; + /* Look for this #address-cells. We have to implement the old linux * trick of looking for the parent here as some device-trees rely on it */ @@ -875,12 +881,15 @@ int of_irq_map_one(struct device_node *device, int index, struct of_irq *out_irq } intsize = *tmp; + DBG(" intsize=%d intlen=%d\n", intsize, intlen); + /* Check index */ if ((index + 1) * intsize > intlen) return -EINVAL; /* Get new specifier and map it */ - res = of_irq_map_raw(p, intspec + index * intsize, addr, out_irq); + res = of_irq_map_raw(p, intspec + index * intsize, intsize, + addr, out_irq); of_node_put(p); return res; } @@ -965,7 +974,7 @@ int of_irq_map_pci(struct pci_dev *pdev, struct of_irq *out_irq) laddr[0] = (pdev->bus->number << 16) | (pdev->devfn << 8); laddr[1] = laddr[2] = 0; - return of_irq_map_raw(ppnode, &lspec, laddr, out_irq); + return of_irq_map_raw(ppnode, &lspec, 1, laddr, out_irq); } EXPORT_SYMBOL_GPL(of_irq_map_pci); #endif /* CONFIG_PCI */ -- cgit v1.2.1 From 11e9ed43ca8a741c2858c33d12120cf8817d3bff Mon Sep 17 00:00:00 2001 From: Will Schmidt Date: Fri, 25 Aug 2006 15:46:59 -0500 Subject: [POWERPC] Fix up ibm_architecture_vec definition This problem was noticed by one of the Phyp firmware folks. Our ibm,client-architecture-support call was failing. This corrects the vector length parameters being passed in. Signed-off-by: Will Schmidt Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/prom_init.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c index 462bced40c12..4394e545f9f7 100644 --- a/arch/powerpc/kernel/prom_init.c +++ b/arch/powerpc/kernel/prom_init.c @@ -646,13 +646,13 @@ static unsigned char ibm_architecture_vec[] = { 5 - 1, /* 5 option vectors */ /* option vector 1: processor architectures supported */ - 3 - 1, /* length */ + 3 - 2, /* length */ 0, /* don't ignore, don't halt */ OV1_PPC_2_00 | OV1_PPC_2_01 | OV1_PPC_2_02 | OV1_PPC_2_03 | OV1_PPC_2_04 | OV1_PPC_2_05, /* option vector 2: Open Firmware options supported */ - 34 - 1, /* length */ + 34 - 2, /* length */ OV2_REAL_MODE, 0, 0, W(0xffffffff), /* real_base */ @@ -666,16 +666,16 @@ static unsigned char ibm_architecture_vec[] = { 48, /* max log_2(hash table size) */ /* option vector 3: processor options supported */ - 3 - 1, /* length */ + 3 - 2, /* length */ 0, /* don't ignore, don't halt */ OV3_FP | OV3_VMX, /* option vector 4: IBM PAPR implementation */ - 2 - 1, /* length */ + 2 - 2, /* length */ 0, /* don't halt */ /* option vector 5: PAPR/OF options */ - 3 - 1, /* length */ + 3 - 2, /* length */ 0, /* don't ignore, don't halt */ OV5_LPAR | OV5_SPLPAR | OV5_LARGE_PAGES, }; -- cgit v1.2.1 From fea23bfefb4e98efd3c36f00ccc0b60281dc99da Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Wed, 30 Aug 2006 14:45:35 +1000 Subject: [POWERPC] Restore copyright notice in arch/powerpc/kernel/fpu.S This code got moved from head.S but the copyright notice on head.S didn't get transferred with it. Noticed by Cort Dougan . Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/fpu.S | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/fpu.S b/arch/powerpc/kernel/fpu.S index 7e2c9fe44ac1..821e152e093c 100644 --- a/arch/powerpc/kernel/fpu.S +++ b/arch/powerpc/kernel/fpu.S @@ -2,6 +2,11 @@ * FPU support code, moved here from head.S so that it can be used * by chips which use other head-whatever.S files. * + * Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org) + * Copyright (C) 1996 Cort Dougan + * Copyright (C) 1996 Paul Mackerras. + * Copyright (C) 1997 Dan Malek (dmalek@jlc.net). + * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version -- cgit v1.2.1 From e0d872d536bb93335d5905b09fe374a163486d43 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Wed, 30 Aug 2006 15:55:32 +1000 Subject: [POWERPC] Fix problem with time not advancing on 32-bit platforms This fixes a problem introduced in 5db9fa9593e2ff69f2b95f9d59229dc4faaa564d. The last_jiffy per-cpu variable is only 32 bits on 32-bit machines, but it was being compared with a 64-bit quantity (tb_next_jiffy), which resulted in time not advancing. This fixes it by changing last_jiffy to be 64 bits on all platforms. With this, we no longer need tb_last_stamp as a 32-bit version of tb_last_jiffy, so this gets rid of tb_last_stamp and we just use tb_last_jiffy instead. This also fixes a bug when the boot cpu is not online, because using tb_last_stamp could have caused the wrong timebase origin value to be used when calculating the time of day. Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/time.c | 25 ++++++++----------------- 1 file changed, 8 insertions(+), 17 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 18e59e43d2b3..a124499e65d9 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -125,15 +125,8 @@ static long timezone_offset; unsigned long ppc_proc_freq; unsigned long ppc_tb_freq; -u64 tb_last_jiffy __cacheline_aligned_in_smp; -unsigned long tb_last_stamp; - -/* - * Note that on ppc32 this only stores the bottom 32 bits of - * the timebase value, but that's enough to tell when a jiffy - * has passed. - */ -DEFINE_PER_CPU(unsigned long, last_jiffy); +static u64 tb_last_jiffy __cacheline_aligned_in_smp; +static DEFINE_PER_CPU(u64, last_jiffy); #ifdef CONFIG_VIRT_CPU_ACCOUNTING /* @@ -458,7 +451,7 @@ void do_gettimeofday(struct timeval *tv) do { seq = read_seqbegin_irqsave(&xtime_lock, flags); sec = xtime.tv_sec; - nsec = xtime.tv_nsec + tb_ticks_since(tb_last_stamp); + nsec = xtime.tv_nsec + tb_ticks_since(tb_last_jiffy); } while (read_seqretry_irqrestore(&xtime_lock, seq, flags)); usec = nsec / 1000; while (usec >= 1000000) { @@ -700,7 +693,6 @@ void timer_interrupt(struct pt_regs * regs) tb_next_jiffy = tb_last_jiffy + tb_ticks_per_jiffy; if (per_cpu(last_jiffy, cpu) >= tb_next_jiffy) { tb_last_jiffy = tb_next_jiffy; - tb_last_stamp = per_cpu(last_jiffy, cpu); do_timer(regs); timer_recalc_offset(tb_last_jiffy); timer_check_rtc(); @@ -749,7 +741,7 @@ void __init smp_space_timers(unsigned int max_cpus) int i; unsigned long half = tb_ticks_per_jiffy / 2; unsigned long offset = tb_ticks_per_jiffy / max_cpus; - unsigned long previous_tb = per_cpu(last_jiffy, boot_cpuid); + u64 previous_tb = per_cpu(last_jiffy, boot_cpuid); /* make sure tb > per_cpu(last_jiffy, cpu) for all cpus always */ previous_tb -= tb_ticks_per_jiffy; @@ -830,7 +822,7 @@ int do_settimeofday(struct timespec *tv) * and therefore the (jiffies - wall_jiffies) computation * has been removed. */ - tb_delta = tb_ticks_since(tb_last_stamp); + tb_delta = tb_ticks_since(tb_last_jiffy); tb_delta = mulhdu(tb_delta, do_gtod.varp->tb_to_xs); /* in xsec */ new_nsec -= SCALE_XSEC(tb_delta, 1000000000); @@ -950,8 +942,7 @@ void __init time_init(void) if (__USE_RTC()) { /* 601 processor: dec counts down by 128 every 128ns */ ppc_tb_freq = 1000000000; - tb_last_stamp = get_rtcl(); - tb_last_jiffy = tb_last_stamp; + tb_last_jiffy = get_rtcl(); } else { /* Normal PowerPC with timebase register */ ppc_md.calibrate_decr(); @@ -959,7 +950,7 @@ void __init time_init(void) ppc_tb_freq / 1000000, ppc_tb_freq % 1000000); printk(KERN_DEBUG "time_init: processor frequency = %lu.%.6lu MHz\n", ppc_proc_freq / 1000000, ppc_proc_freq % 1000000); - tb_last_stamp = tb_last_jiffy = get_tb(); + tb_last_jiffy = get_tb(); } tb_ticks_per_jiffy = ppc_tb_freq / HZ; @@ -1036,7 +1027,7 @@ void __init time_init(void) do_gtod.varp = &do_gtod.vars[0]; do_gtod.var_idx = 0; do_gtod.varp->tb_orig_stamp = tb_last_jiffy; - __get_cpu_var(last_jiffy) = tb_last_stamp; + __get_cpu_var(last_jiffy) = tb_last_jiffy; do_gtod.varp->stamp_xsec = (u64) xtime.tv_sec * XSEC_PER_SEC; do_gtod.tb_ticks_per_sec = tb_ticks_per_sec; do_gtod.varp->tb_to_xs = tb_to_xs; -- cgit v1.2.1 From 467c37801c453849a2fe243c3226476ee3985868 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Wed, 30 Aug 2006 16:10:47 +1000 Subject: [POWERPC] Fix irq enable/disable in smp_generic_take_timebase Eran Ben-Avi pointed out that the arch/ppc version of smp_generic_take_timebase disables interrupts on entry but exits without restoring them. However, both it and the arch/powerpc version have another problem, which is that they use local_irq_disable/enable rather than local_irq_save/restore, and they are called with interrupts disabled. This fixes both problems; it changes a return to a break in the arch/ppc version, and changes both versions to use local_irq_save/restore. Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/smp-tbsync.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/smp-tbsync.c b/arch/powerpc/kernel/smp-tbsync.c index f19e2e0e61e7..de59c6c31a5b 100644 --- a/arch/powerpc/kernel/smp-tbsync.c +++ b/arch/powerpc/kernel/smp-tbsync.c @@ -45,8 +45,9 @@ void __devinit smp_generic_take_timebase(void) { int cmd; u64 tb; + unsigned long flags; - local_irq_disable(); + local_irq_save(flags); while (!running) barrier(); rmb(); @@ -70,7 +71,7 @@ void __devinit smp_generic_take_timebase(void) set_tb(tb >> 32, tb & 0xfffffffful); enter_contest(tbsync->mark, -1); } - local_irq_enable(); + local_irq_restore(flags); } static int __devinit start_contest(int cmd, long offset, int num) -- cgit v1.2.1 From eb36c2884a1a2190791afe65fd833b2d3cd4b999 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Wed, 30 Aug 2006 16:13:16 +1000 Subject: [PATCH] ppc32: fix last_jiffy time comparison This fixes a hang on ppc32. The problem was that I was comparing a 32-bit quantity with a 64-bit quantity, and consequently time wasn't advancing. This makes us use a 64-bit quantity on all platforms, which ends up simplifying the code since we can now get rid of the tb_last_stamp variable (which actually fixes another bug that Ben H and I noticed while going carefully through the code). This works fine on my G4 tibook. Let me know how it goes on your machines. Acked-by: Olaf Hering Acked-by: Mikael Pettersson Signed-off-by: Linus Torvalds --- arch/powerpc/kernel/time.c | 25 ++++++++----------------- 1 file changed, 8 insertions(+), 17 deletions(-) (limited to 'arch/powerpc/kernel') diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 18e59e43d2b3..a124499e65d9 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -125,15 +125,8 @@ static long timezone_offset; unsigned long ppc_proc_freq; unsigned long ppc_tb_freq; -u64 tb_last_jiffy __cacheline_aligned_in_smp; -unsigned long tb_last_stamp; - -/* - * Note that on ppc32 this only stores the bottom 32 bits of - * the timebase value, but that's enough to tell when a jiffy - * has passed. - */ -DEFINE_PER_CPU(unsigned long, last_jiffy); +static u64 tb_last_jiffy __cacheline_aligned_in_smp; +static DEFINE_PER_CPU(u64, last_jiffy); #ifdef CONFIG_VIRT_CPU_ACCOUNTING /* @@ -458,7 +451,7 @@ void do_gettimeofday(struct timeval *tv) do { seq = read_seqbegin_irqsave(&xtime_lock, flags); sec = xtime.tv_sec; - nsec = xtime.tv_nsec + tb_ticks_since(tb_last_stamp); + nsec = xtime.tv_nsec + tb_ticks_since(tb_last_jiffy); } while (read_seqretry_irqrestore(&xtime_lock, seq, flags)); usec = nsec / 1000; while (usec >= 1000000) { @@ -700,7 +693,6 @@ void timer_interrupt(struct pt_regs * regs) tb_next_jiffy = tb_last_jiffy + tb_ticks_per_jiffy; if (per_cpu(last_jiffy, cpu) >= tb_next_jiffy) { tb_last_jiffy = tb_next_jiffy; - tb_last_stamp = per_cpu(last_jiffy, cpu); do_timer(regs); timer_recalc_offset(tb_last_jiffy); timer_check_rtc(); @@ -749,7 +741,7 @@ void __init smp_space_timers(unsigned int max_cpus) int i; unsigned long half = tb_ticks_per_jiffy / 2; unsigned long offset = tb_ticks_per_jiffy / max_cpus; - unsigned long previous_tb = per_cpu(last_jiffy, boot_cpuid); + u64 previous_tb = per_cpu(last_jiffy, boot_cpuid); /* make sure tb > per_cpu(last_jiffy, cpu) for all cpus always */ previous_tb -= tb_ticks_per_jiffy; @@ -830,7 +822,7 @@ int do_settimeofday(struct timespec *tv) * and therefore the (jiffies - wall_jiffies) computation * has been removed. */ - tb_delta = tb_ticks_since(tb_last_stamp); + tb_delta = tb_ticks_since(tb_last_jiffy); tb_delta = mulhdu(tb_delta, do_gtod.varp->tb_to_xs); /* in xsec */ new_nsec -= SCALE_XSEC(tb_delta, 1000000000); @@ -950,8 +942,7 @@ void __init time_init(void) if (__USE_RTC()) { /* 601 processor: dec counts down by 128 every 128ns */ ppc_tb_freq = 1000000000; - tb_last_stamp = get_rtcl(); - tb_last_jiffy = tb_last_stamp; + tb_last_jiffy = get_rtcl(); } else { /* Normal PowerPC with timebase register */ ppc_md.calibrate_decr(); @@ -959,7 +950,7 @@ void __init time_init(void) ppc_tb_freq / 1000000, ppc_tb_freq % 1000000); printk(KERN_DEBUG "time_init: processor frequency = %lu.%.6lu MHz\n", ppc_proc_freq / 1000000, ppc_proc_freq % 1000000); - tb_last_stamp = tb_last_jiffy = get_tb(); + tb_last_jiffy = get_tb(); } tb_ticks_per_jiffy = ppc_tb_freq / HZ; @@ -1036,7 +1027,7 @@ void __init time_init(void) do_gtod.varp = &do_gtod.vars[0]; do_gtod.var_idx = 0; do_gtod.varp->tb_orig_stamp = tb_last_jiffy; - __get_cpu_var(last_jiffy) = tb_last_stamp; + __get_cpu_var(last_jiffy) = tb_last_jiffy; do_gtod.varp->stamp_xsec = (u64) xtime.tv_sec * XSEC_PER_SEC; do_gtod.tb_ticks_per_sec = tb_ticks_per_sec; do_gtod.varp->tb_to_xs = tb_to_xs; -- cgit v1.2.1