From 7835e98b2e3c66dba79cb0ff8ebb90a2fe030c29 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Wed, 22 Mar 2006 00:08:40 -0800 Subject: [PATCH] remove set_page_count() outside mm/ set_page_count usage outside mm/ is limited to setting the refcount to 1. Remove set_page_count from outside mm/, and replace those users with init_page_count() and set_page_refcounted(). This allows more debug checking, and tighter control on how code is allowed to play around with page->_count. Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/sparc/kernel/sun4m_smp.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/sparc/kernel/sun4m_smp.c') diff --git a/arch/sparc/kernel/sun4m_smp.c b/arch/sparc/kernel/sun4m_smp.c index a21f27d10e5..fbbd8a474c4 100644 --- a/arch/sparc/kernel/sun4m_smp.c +++ b/arch/sparc/kernel/sun4m_smp.c @@ -233,21 +233,21 @@ void __init smp4m_boot_cpus(void) /* Free unneeded trap tables */ if (!cpu_isset(i, cpu_present_map)) { ClearPageReserved(virt_to_page(trapbase_cpu1)); - set_page_count(virt_to_page(trapbase_cpu1), 1); + init_page_count(virt_to_page(trapbase_cpu1)); free_page((unsigned long)trapbase_cpu1); totalram_pages++; num_physpages++; } if (!cpu_isset(2, cpu_present_map)) { ClearPageReserved(virt_to_page(trapbase_cpu2)); - set_page_count(virt_to_page(trapbase_cpu2), 1); + init_page_count(virt_to_page(trapbase_cpu2)); free_page((unsigned long)trapbase_cpu2); totalram_pages++; num_physpages++; } if (!cpu_isset(3, cpu_present_map)) { ClearPageReserved(virt_to_page(trapbase_cpu3)); - set_page_count(virt_to_page(trapbase_cpu3), 1); + init_page_count(virt_to_page(trapbase_cpu3)); free_page((unsigned long)trapbase_cpu3); totalram_pages++; num_physpages++; -- cgit v1.2.3 From 394e3902c55e667945f6f1c2bdbc59842cce70f7 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Thu, 23 Mar 2006 03:01:05 -0800 Subject: [PATCH] more for_each_cpu() conversions When we stop allocating percpu memory for not-possible CPUs we must not touch the percpu data for not-possible CPUs at all. The correct way of doing this is to test cpu_possible() or to use for_each_cpu(). This patch is a kernel-wide sweep of all instances of NR_CPUS. I found very few instances of this bug, if any. But the patch converts lots of open-coded test to use the preferred helper macros. Cc: Mikael Starvik Cc: David Howells Acked-by: Kyle McMartin Cc: Anton Blanchard Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Martin Schwidefsky Cc: Heiko Carstens Cc: Paul Mundt Cc: "David S. Miller" Cc: William Lee Irwin III Cc: Andi Kleen Cc: Christian Zankel Cc: Philippe Elie Cc: Nathan Scott Cc: Jens Axboe Cc: Eric Dumazet Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/sparc/kernel/sun4m_smp.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'arch/sparc/kernel/sun4m_smp.c') diff --git a/arch/sparc/kernel/sun4m_smp.c b/arch/sparc/kernel/sun4m_smp.c index fbbd8a474c4..1dde312eebd 100644 --- a/arch/sparc/kernel/sun4m_smp.c +++ b/arch/sparc/kernel/sun4m_smp.c @@ -218,10 +218,8 @@ void __init smp4m_boot_cpus(void) cpu_present_map = cpumask_of_cpu(smp_processor_id()); } else { unsigned long bogosum = 0; - for(i = 0; i < NR_CPUS; i++) { - if (cpu_isset(i, cpu_present_map)) - bogosum += cpu_data(i).udelay_val; - } + for_each_present_cpu(i) + bogosum += cpu_data(i).udelay_val; printk("Total of %d Processors activated (%lu.%02lu BogoMIPS).\n", cpucount + 1, bogosum/(500000/HZ), -- cgit v1.2.3 From a54123e27779049d27d21e6c8adfee73aa2c0734 Mon Sep 17 00:00:00 2001 From: Bob Breuer Date: Thu, 23 Mar 2006 22:36:19 -0800 Subject: [SPARC]: Try to start getting SMP back into shape. Todo items: - IRQ_INPROGRESS flag - use sparc64 irq buckets, or generic irq_desc? - sun4d - re-indent large chunks of sun4m_smp.c - some places assume sequential cpu numbering (i.e. 0,1 instead of 0,2) Last I checked (with 2.6.14), random programs segfault with dual HyperSPARC. And with SuperSPARC II's, it seems stable but will eventually die from a write lock error (wrong lock owner or something). I haven't tried the HyperSPARC + highmem combination recently, so that may still be a problem. Signed-off-by: David S. Miller --- arch/sparc/kernel/sun4m_smp.c | 181 ++++++++++++++++++------------------------ 1 file changed, 76 insertions(+), 105 deletions(-) (limited to 'arch/sparc/kernel/sun4m_smp.c') diff --git a/arch/sparc/kernel/sun4m_smp.c b/arch/sparc/kernel/sun4m_smp.c index 1dde312eebd..70b375a4c2c 100644 --- a/arch/sparc/kernel/sun4m_smp.c +++ b/arch/sparc/kernel/sun4m_smp.c @@ -40,15 +40,11 @@ extern ctxd_t *srmmu_ctx_table_phys; extern void calibrate_delay(void); extern volatile int smp_processors_ready; -extern int smp_num_cpus; extern volatile unsigned long cpu_callin_map[NR_CPUS]; extern unsigned char boot_cpu_id; -extern int smp_activated; -extern volatile int __cpu_number_map[NR_CPUS]; -extern volatile int __cpu_logical_map[NR_CPUS]; -extern volatile unsigned long ipi_count; -extern volatile int smp_process_available; -extern volatile int smp_commenced; + +extern cpumask_t smp_commenced_mask; + extern int __smp4m_processor_id(void); /*#define SMP_DEBUG*/ @@ -77,8 +73,6 @@ void __init smp4m_callin(void) local_flush_cache_all(); local_flush_tlb_all(); - set_irq_udt(boot_cpu_id); - /* Get our local ticker going. */ smp_setup_percpu_timer(); @@ -95,8 +89,9 @@ void __init smp4m_callin(void) * to call the scheduler code. */ /* Allow master to continue. */ - swap((unsigned long *)&cpu_callin_map[cpuid], 1); + swap(&cpu_callin_map[cpuid], 1); + /* XXX: What's up with all the flushes? */ local_flush_cache_all(); local_flush_tlb_all(); @@ -111,13 +106,14 @@ void __init smp4m_callin(void) atomic_inc(&init_mm.mm_count); current->active_mm = &init_mm; - while(!smp_commenced) - barrier(); - - local_flush_cache_all(); - local_flush_tlb_all(); + while (!cpu_isset(cpuid, smp_commenced_mask)) + mb(); local_irq_enable(); + + cpu_set(cpuid, cpu_online_map); + /* last one in gets all the interrupts (for testing) */ + set_irq_udt(boot_cpu_id); } extern void init_IRQ(void); @@ -134,102 +130,76 @@ extern unsigned long trapbase_cpu3[]; void __init smp4m_boot_cpus(void) { - int cpucount = 0; - int i, mid; + smp_setup_percpu_timer(); + local_flush_cache_all(); +} - printk("Entering SMP Mode...\n"); +int smp4m_boot_one_cpu(int i) +{ + extern unsigned long sun4m_cpu_startup; + unsigned long *entry = &sun4m_cpu_startup; + struct task_struct *p; + int timeout; + int cpu_node; - local_irq_enable(); - cpus_clear(cpu_present_map); + cpu_find_by_mid(i, &cpu_node); + + /* Cook up an idler for this guy. */ + p = fork_idle(i); + current_set[i] = task_thread_info(p); + /* See trampoline.S for details... */ + entry += ((i-1) * 3); - for (i = 0; !cpu_find_by_instance(i, NULL, &mid); i++) - cpu_set(mid, cpu_present_map); + /* + * Initialize the contexts table + * Since the call to prom_startcpu() trashes the structure, + * we need to re-initialize it for each cpu + */ + smp_penguin_ctable.which_io = 0; + smp_penguin_ctable.phys_addr = (unsigned int) srmmu_ctx_table_phys; + smp_penguin_ctable.reg_size = 0; - for(i=0; i < NR_CPUS; i++) { - __cpu_number_map[i] = -1; - __cpu_logical_map[i] = -1; + /* whirrr, whirrr, whirrrrrrrrr... */ + printk("Starting CPU %d at %p\n", i, entry); + local_flush_cache_all(); + prom_startcpu(cpu_node, + &smp_penguin_ctable, 0, (char *)entry); + + /* wheee... it's going... */ + for(timeout = 0; timeout < 10000; timeout++) { + if(cpu_callin_map[i]) + break; + udelay(200); } - __cpu_number_map[boot_cpu_id] = 0; - __cpu_logical_map[0] = boot_cpu_id; - current_thread_info()->cpu = boot_cpu_id; + if (!(cpu_callin_map[i])) { + printk("Processor %d is stuck.\n", i); + return -ENODEV; + } - smp_store_cpu_info(boot_cpu_id); - set_irq_udt(boot_cpu_id); - smp_setup_percpu_timer(); local_flush_cache_all(); - if(cpu_find_by_instance(1, NULL, NULL)) - return; /* Not an MP box. */ - for(i = 0; i < NR_CPUS; i++) { - if(i == boot_cpu_id) - continue; - - if (cpu_isset(i, cpu_present_map)) { - extern unsigned long sun4m_cpu_startup; - unsigned long *entry = &sun4m_cpu_startup; - struct task_struct *p; - int timeout; - - /* Cook up an idler for this guy. */ - p = fork_idle(i); - cpucount++; - current_set[i] = task_thread_info(p); - /* See trampoline.S for details... */ - entry += ((i-1) * 3); - - /* - * Initialize the contexts table - * Since the call to prom_startcpu() trashes the structure, - * we need to re-initialize it for each cpu - */ - smp_penguin_ctable.which_io = 0; - smp_penguin_ctable.phys_addr = (unsigned int) srmmu_ctx_table_phys; - smp_penguin_ctable.reg_size = 0; - - /* whirrr, whirrr, whirrrrrrrrr... */ - printk("Starting CPU %d at %p\n", i, entry); - local_flush_cache_all(); - prom_startcpu(cpu_data(i).prom_node, - &smp_penguin_ctable, 0, (char *)entry); - - /* wheee... it's going... */ - for(timeout = 0; timeout < 10000; timeout++) { - if(cpu_callin_map[i]) - break; - udelay(200); - } - if(cpu_callin_map[i]) { - /* Another "Red Snapper". */ - __cpu_number_map[i] = i; - __cpu_logical_map[i] = i; - } else { - cpucount--; - printk("Processor %d is stuck.\n", i); - } - } - if(!(cpu_callin_map[i])) { - cpu_clear(i, cpu_present_map); - __cpu_number_map[i] = -1; + return 0; +} + +void __init smp4m_smp_done(void) +{ + int i, first; + int *prev; + + /* setup cpu list for irq rotation */ + first = 0; + prev = &first; + for (i = 0; i < NR_CPUS; i++) { + if (cpu_online(i)) { + *prev = i; + prev = &cpu_data(i).next; } } + *prev = first; local_flush_cache_all(); - if(cpucount == 0) { - printk("Error: only one Processor found.\n"); - cpu_present_map = cpumask_of_cpu(smp_processor_id()); - } else { - unsigned long bogosum = 0; - for_each_present_cpu(i) - bogosum += cpu_data(i).udelay_val; - printk("Total of %d Processors activated (%lu.%02lu BogoMIPS).\n", - cpucount + 1, - bogosum/(500000/HZ), - (bogosum/(5000/HZ))%100); - smp_activated = 1; - smp_num_cpus = cpucount + 1; - } /* Free unneeded trap tables */ - if (!cpu_isset(i, cpu_present_map)) { + if (!cpu_isset(1, cpu_present_map)) { ClearPageReserved(virt_to_page(trapbase_cpu1)); init_page_count(virt_to_page(trapbase_cpu1)); free_page((unsigned long)trapbase_cpu1); @@ -263,6 +233,9 @@ void __init smp4m_boot_cpus(void) */ void smp4m_irq_rotate(int cpu) { + int next = cpu_data(cpu).next; + if (next != cpu) + set_irq_udt(next); } /* Cross calls, in order to work efficiently and atomically do all @@ -289,7 +262,7 @@ void smp4m_message_pass(int target, int msg, unsigned long data, int wait) smp_cpu_in_msg[me]++; if(target == MSG_ALL_BUT_SELF || target == MSG_ALL) { - mask = cpu_present_map; + mask = cpu_online_map; if(target == MSG_ALL_BUT_SELF) cpu_clear(me, mask); for(i = 0; i < 4; i++) { @@ -314,8 +287,8 @@ static struct smp_funcall { unsigned long arg3; unsigned long arg4; unsigned long arg5; - unsigned long processors_in[NR_CPUS]; /* Set when ipi entered. */ - unsigned long processors_out[NR_CPUS]; /* Set when ipi exited. */ + unsigned long processors_in[SUN4M_NCPUS]; /* Set when ipi entered. */ + unsigned long processors_out[SUN4M_NCPUS]; /* Set when ipi exited. */ } ccall_info; static DEFINE_SPINLOCK(cross_call_lock); @@ -324,8 +297,7 @@ static DEFINE_SPINLOCK(cross_call_lock); void smp4m_cross_call(smpfunc_t func, unsigned long arg1, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5) { - if(smp_processors_ready) { - register int ncpus = smp_num_cpus; + register int ncpus = SUN4M_NCPUS; unsigned long flags; spin_lock_irqsave(&cross_call_lock, flags); @@ -340,7 +312,7 @@ void smp4m_cross_call(smpfunc_t func, unsigned long arg1, unsigned long arg2, /* Init receive/complete mapping, plus fire the IPI's off. */ { - cpumask_t mask = cpu_present_map; + cpumask_t mask = cpu_online_map; register int i; cpu_clear(smp_processor_id(), mask); @@ -373,7 +345,6 @@ void smp4m_cross_call(smpfunc_t func, unsigned long arg1, unsigned long arg2, } spin_unlock_irqrestore(&cross_call_lock, flags); - } } /* Running cross calls. */ -- cgit v1.2.3