From 43cf38eb5cea91245502df3fcee4dbfc1c74dd1c Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 2 Feb 2010 14:38:57 +0900 Subject: percpu: add __percpu sparse annotations to core kernel subsystems Add __percpu sparse annotations to core subsystems. These annotations are to make sparse consider percpu variables to be in a different address space and warn if accessed without going through percpu accessors. This patch doesn't affect normal builds. Signed-off-by: Tejun Heo Reviewed-by: Christoph Lameter Acked-by: Paul E. McKenney Cc: Jens Axboe Cc: linux-mm@kvack.org Cc: Rusty Russell Cc: Dipankar Sarma Cc: Peter Zijlstra Cc: Andrew Morton Cc: Eric Biederman --- kernel/sched.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 3a8fb30a91b1..978edfd35a96 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1566,7 +1566,7 @@ static unsigned long cpu_avg_load_per_task(int cpu) #ifdef CONFIG_FAIR_GROUP_SCHED -static __read_mostly unsigned long *update_shares_data; +static __read_mostly unsigned long __percpu *update_shares_data; static void __set_se_shares(struct sched_entity *se, unsigned long shares); @@ -10683,7 +10683,7 @@ struct cgroup_subsys cpu_cgroup_subsys = { struct cpuacct { struct cgroup_subsys_state css; /* cpuusage holds pointer to a u64-type object on every cpu */ - u64 *cpuusage; + u64 __percpu *cpuusage; struct percpu_counter cpustat[CPUACCT_STAT_NSTATS]; struct cpuacct *parent; }; -- cgit v1.2.3 From 78d7d407b62a021e6d2e8dc24c0b90e390ab58a1 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Fri, 5 Mar 2010 13:42:54 -0800 Subject: kernel core: use helpers for rlimits Make sure compiler won't do weird things with limits. E.g. fetching them twice may return 2 different values after writable limits are implemented. I.e. either use rlimit helpers added in commit 3e10e716abf3 ("resource: add helpers for fetching rlimits") or ACCESS_ONCE if not applicable. Signed-off-by: Jiri Slaby Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: john stultz Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 10 ++++++---- kernel/perf_event.c | 2 +- kernel/posix-cpu-timers.c | 10 ++++++---- kernel/sched.c | 4 ++-- kernel/sched_rt.c | 5 +++-- kernel/signal.c | 2 +- kernel/sys.c | 3 +-- 7 files changed, 20 insertions(+), 16 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/fork.c b/kernel/fork.c index bab7b254ad39..b0ec34abc0bb 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -828,6 +828,8 @@ void __cleanup_sighand(struct sighand_struct *sighand) */ static void posix_cpu_timers_init_group(struct signal_struct *sig) { + unsigned long cpu_limit; + /* Thread group counters. */ thread_group_cputime_init(sig); @@ -842,9 +844,9 @@ static void posix_cpu_timers_init_group(struct signal_struct *sig) sig->cputime_expires.virt_exp = cputime_zero; sig->cputime_expires.sched_exp = 0; - if (sig->rlim[RLIMIT_CPU].rlim_cur != RLIM_INFINITY) { - sig->cputime_expires.prof_exp = - secs_to_cputime(sig->rlim[RLIMIT_CPU].rlim_cur); + cpu_limit = ACCESS_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur); + if (cpu_limit != RLIM_INFINITY) { + sig->cputime_expires.prof_exp = secs_to_cputime(cpu_limit); sig->cputimer.running = 1; } @@ -1037,7 +1039,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, #endif retval = -EAGAIN; if (atomic_read(&p->real_cred->user->processes) >= - p->signal->rlim[RLIMIT_NPROC].rlim_cur) { + task_rlimit(p, RLIMIT_NPROC)) { if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) && p->real_cred->user != INIT_USER) goto bad_fork_free; diff --git a/kernel/perf_event.c b/kernel/perf_event.c index a661e7991865..8e352c756ba7 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -2610,7 +2610,7 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma) if (user_locked > user_lock_limit) extra = user_locked - user_lock_limit; - lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur; + lock_limit = rlimit(RLIMIT_MEMLOCK); lock_limit >>= PAGE_SHIFT; locked = vma->vm_mm->locked_vm + extra; diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index dbb16bf15c45..1a22dfd42df9 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -1031,9 +1031,10 @@ static void check_thread_timers(struct task_struct *tsk, /* * Check for the special case thread timers. */ - soft = sig->rlim[RLIMIT_RTTIME].rlim_cur; + soft = ACCESS_ONCE(sig->rlim[RLIMIT_RTTIME].rlim_cur); if (soft != RLIM_INFINITY) { - unsigned long hard = sig->rlim[RLIMIT_RTTIME].rlim_max; + unsigned long hard = + ACCESS_ONCE(sig->rlim[RLIMIT_RTTIME].rlim_max); if (hard != RLIM_INFINITY && tsk->rt.timeout > DIV_ROUND_UP(hard, USEC_PER_SEC/HZ)) { @@ -1194,10 +1195,11 @@ static void check_process_timers(struct task_struct *tsk, SIGPROF); check_cpu_itimer(tsk, &sig->it[CPUCLOCK_VIRT], &virt_expires, utime, SIGVTALRM); - soft = sig->rlim[RLIMIT_CPU].rlim_cur; + soft = ACCESS_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur); if (soft != RLIM_INFINITY) { unsigned long psecs = cputime_to_secs(ptime); - unsigned long hard = sig->rlim[RLIMIT_CPU].rlim_max; + unsigned long hard = + ACCESS_ONCE(sig->rlim[RLIMIT_CPU].rlim_max); cputime_t x; if (psecs >= hard) { /* diff --git a/kernel/sched.c b/kernel/sched.c index abb36b16b93b..b47ceeec1a91 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4353,7 +4353,7 @@ int can_nice(const struct task_struct *p, const int nice) /* convert nice value [19,-20] to rlimit style value [1,40] */ int nice_rlim = 20 - nice; - return (nice_rlim <= p->signal->rlim[RLIMIT_NICE].rlim_cur || + return (nice_rlim <= task_rlimit(p, RLIMIT_NICE) || capable(CAP_SYS_NICE)); } @@ -4530,7 +4530,7 @@ recheck: if (!lock_task_sighand(p, &flags)) return -ESRCH; - rlim_rtprio = p->signal->rlim[RLIMIT_RTPRIO].rlim_cur; + rlim_rtprio = task_rlimit(p, RLIMIT_RTPRIO); unlock_task_sighand(p, &flags); /* can't set/change the rt policy */ diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index bf3e38fdbe6d..5a6ed1f0990a 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -1662,8 +1662,9 @@ static void watchdog(struct rq *rq, struct task_struct *p) if (!p->signal) return; - soft = p->signal->rlim[RLIMIT_RTTIME].rlim_cur; - hard = p->signal->rlim[RLIMIT_RTTIME].rlim_max; + /* max may change after cur was read, this will be fixed next tick */ + soft = task_rlimit(p, RLIMIT_RTTIME); + hard = task_rlimit_max(p, RLIMIT_RTTIME); if (soft != RLIM_INFINITY) { unsigned long next; diff --git a/kernel/signal.c b/kernel/signal.c index 5bb9baffa4f1..dbd7fe073c55 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -245,7 +245,7 @@ __sigqueue_alloc(int sig, struct task_struct *t, gfp_t flags, int override_rlimi if (override_rlimit || atomic_read(&user->sigpending) <= - t->signal->rlim[RLIMIT_SIGPENDING].rlim_cur) { + task_rlimit(t, RLIMIT_SIGPENDING)) { q = kmem_cache_alloc(sigqueue_cachep, flags); } else { print_dropped_signal(sig); diff --git a/kernel/sys.c b/kernel/sys.c index 877fe4f8e05e..9814e43fb23b 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -571,8 +571,7 @@ static int set_user(struct cred *new) if (!new_user) return -EAGAIN; - if (atomic_read(&new_user->processes) >= - current->signal->rlim[RLIMIT_NPROC].rlim_cur && + if (atomic_read(&new_user->processes) >= rlimit(RLIMIT_NPROC) && new_user != INIT_USER) { free_uid(new_user); return -EAGAIN; -- cgit v1.2.3 From c9be0a36f9bf392a7984473124a67a12964df11f Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Tue, 5 Jan 2010 12:47:58 +0100 Subject: sysdev: Pass attribute in sysdev_class attributes show/store Passing the attribute to the low level IO functions allows all kinds of cleanups, by sharing low level IO code without requiring an own function for every piece of data. Also drivers can extend the attributes with own data fields and use that in the low level function. Similar to sysdev_attributes and normal attributes. This is a tree-wide sweep, converting everything in one go. No functional changes in this patch other than passing the new argument everywhere. Tested on x86, the non x86 parts are uncompiled. Signed-off-by: Andi Kleen Signed-off-by: Greg Kroah-Hartman --- arch/mips/txx9/generic/7segled.c | 5 +++- arch/s390/kernel/smp.c | 8 +++++-- arch/s390/kernel/time.c | 49 ++++++++++++++++++++++++++++++---------- drivers/base/cpu.c | 9 +++++--- drivers/base/node.c | 17 ++++++++++---- drivers/base/sys.c | 4 ++-- drivers/cpuidle/sysfs.c | 4 ++++ include/linux/sysdev.h | 6 +++-- kernel/perf_event.c | 13 ++++++++--- kernel/sched.c | 4 ++++ 10 files changed, 89 insertions(+), 30 deletions(-) (limited to 'kernel/sched.c') diff --git a/arch/mips/txx9/generic/7segled.c b/arch/mips/txx9/generic/7segled.c index 727ab21b6618..7f8416f86222 100644 --- a/arch/mips/txx9/generic/7segled.c +++ b/arch/mips/txx9/generic/7segled.c @@ -58,13 +58,16 @@ static ssize_t raw_store(struct sys_device *dev, static SYSDEV_ATTR(ascii, 0200, NULL, ascii_store); static SYSDEV_ATTR(raw, 0200, NULL, raw_store); -static ssize_t map_seg7_show(struct sysdev_class *class, char *buf) +static ssize_t map_seg7_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { memcpy(buf, &txx9_seg7map, sizeof(txx9_seg7map)); return sizeof(txx9_seg7map); } static ssize_t map_seg7_store(struct sysdev_class *class, + struct sysdev_class_attribute *attr, const char *buf, size_t size) { if (size != sizeof(txx9_seg7map)) diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c index 8b10127c00ad..e2121099f03b 100644 --- a/arch/s390/kernel/smp.c +++ b/arch/s390/kernel/smp.c @@ -1020,7 +1020,9 @@ out: return rc; } -static ssize_t __ref rescan_store(struct sysdev_class *class, const char *buf, +static ssize_t __ref rescan_store(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + const char *buf, size_t count) { int rc; @@ -1041,7 +1043,9 @@ static ssize_t dispatching_show(struct sysdev_class *class, char *buf) return count; } -static ssize_t dispatching_store(struct sysdev_class *dev, const char *buf, +static ssize_t dispatching_store(struct sysdev_class *dev, + struct sysdev_class_attribute *attr, + const char *buf, size_t count) { int val, rc; diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c index a8f93f1705ad..75894c281710 100644 --- a/arch/s390/kernel/time.c +++ b/arch/s390/kernel/time.c @@ -1116,14 +1116,18 @@ static struct sys_device etr_port1_dev = { /* * ETR class attributes */ -static ssize_t etr_stepping_port_show(struct sysdev_class *class, char *buf) +static ssize_t etr_stepping_port_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { return sprintf(buf, "%i\n", etr_port0.esw.p); } static SYSDEV_CLASS_ATTR(stepping_port, 0400, etr_stepping_port_show, NULL); -static ssize_t etr_stepping_mode_show(struct sysdev_class *class, char *buf) +static ssize_t etr_stepping_mode_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { char *mode_str; @@ -1584,7 +1588,9 @@ static struct sysdev_class stp_sysclass = { .name = "stp", }; -static ssize_t stp_ctn_id_show(struct sysdev_class *class, char *buf) +static ssize_t stp_ctn_id_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { if (!stp_online) return -ENODATA; @@ -1594,7 +1600,9 @@ static ssize_t stp_ctn_id_show(struct sysdev_class *class, char *buf) static SYSDEV_CLASS_ATTR(ctn_id, 0400, stp_ctn_id_show, NULL); -static ssize_t stp_ctn_type_show(struct sysdev_class *class, char *buf) +static ssize_t stp_ctn_type_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { if (!stp_online) return -ENODATA; @@ -1603,7 +1611,9 @@ static ssize_t stp_ctn_type_show(struct sysdev_class *class, char *buf) static SYSDEV_CLASS_ATTR(ctn_type, 0400, stp_ctn_type_show, NULL); -static ssize_t stp_dst_offset_show(struct sysdev_class *class, char *buf) +static ssize_t stp_dst_offset_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { if (!stp_online || !(stp_info.vbits & 0x2000)) return -ENODATA; @@ -1612,7 +1622,9 @@ static ssize_t stp_dst_offset_show(struct sysdev_class *class, char *buf) static SYSDEV_CLASS_ATTR(dst_offset, 0400, stp_dst_offset_show, NULL); -static ssize_t stp_leap_seconds_show(struct sysdev_class *class, char *buf) +static ssize_t stp_leap_seconds_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { if (!stp_online || !(stp_info.vbits & 0x8000)) return -ENODATA; @@ -1621,7 +1633,9 @@ static ssize_t stp_leap_seconds_show(struct sysdev_class *class, char *buf) static SYSDEV_CLASS_ATTR(leap_seconds, 0400, stp_leap_seconds_show, NULL); -static ssize_t stp_stratum_show(struct sysdev_class *class, char *buf) +static ssize_t stp_stratum_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { if (!stp_online) return -ENODATA; @@ -1630,7 +1644,9 @@ static ssize_t stp_stratum_show(struct sysdev_class *class, char *buf) static SYSDEV_CLASS_ATTR(stratum, 0400, stp_stratum_show, NULL); -static ssize_t stp_time_offset_show(struct sysdev_class *class, char *buf) +static ssize_t stp_time_offset_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { if (!stp_online || !(stp_info.vbits & 0x0800)) return -ENODATA; @@ -1639,7 +1655,9 @@ static ssize_t stp_time_offset_show(struct sysdev_class *class, char *buf) static SYSDEV_CLASS_ATTR(time_offset, 0400, stp_time_offset_show, NULL); -static ssize_t stp_time_zone_offset_show(struct sysdev_class *class, char *buf) +static ssize_t stp_time_zone_offset_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { if (!stp_online || !(stp_info.vbits & 0x4000)) return -ENODATA; @@ -1649,7 +1667,9 @@ static ssize_t stp_time_zone_offset_show(struct sysdev_class *class, char *buf) static SYSDEV_CLASS_ATTR(time_zone_offset, 0400, stp_time_zone_offset_show, NULL); -static ssize_t stp_timing_mode_show(struct sysdev_class *class, char *buf) +static ssize_t stp_timing_mode_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { if (!stp_online) return -ENODATA; @@ -1658,7 +1678,9 @@ static ssize_t stp_timing_mode_show(struct sysdev_class *class, char *buf) static SYSDEV_CLASS_ATTR(timing_mode, 0400, stp_timing_mode_show, NULL); -static ssize_t stp_timing_state_show(struct sysdev_class *class, char *buf) +static ssize_t stp_timing_state_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { if (!stp_online) return -ENODATA; @@ -1667,12 +1689,15 @@ static ssize_t stp_timing_state_show(struct sysdev_class *class, char *buf) static SYSDEV_CLASS_ATTR(timing_state, 0400, stp_timing_state_show, NULL); -static ssize_t stp_online_show(struct sysdev_class *class, char *buf) +static ssize_t stp_online_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { return sprintf(buf, "%i\n", stp_online); } static ssize_t stp_online_store(struct sysdev_class *class, + struct sysdev_class_attribute *attr, const char *buf, size_t count) { unsigned int value; diff --git a/drivers/base/cpu.c b/drivers/base/cpu.c index 958bd1540c30..fd1b2f9b7b8f 100644 --- a/drivers/base/cpu.c +++ b/drivers/base/cpu.c @@ -151,7 +151,8 @@ static ssize_t print_cpus_map(char *buf, const struct cpumask *map) } #define print_cpus_func(type) \ -static ssize_t print_cpus_##type(struct sysdev_class *class, char *buf) \ +static ssize_t print_cpus_##type(struct sysdev_class *class, \ + struct sysdev_class_attribute *attr, char *buf) \ { \ return print_cpus_map(buf, cpu_##type##_mask); \ } \ @@ -165,7 +166,8 @@ print_cpus_func(present); /* * Print values for NR_CPUS and offlined cpus */ -static ssize_t print_cpus_kernel_max(struct sysdev_class *class, char *buf) +static ssize_t print_cpus_kernel_max(struct sysdev_class *class, + struct sysdev_class_attribute *attr, char *buf) { int n = snprintf(buf, PAGE_SIZE-2, "%d\n", NR_CPUS - 1); return n; @@ -175,7 +177,8 @@ static SYSDEV_CLASS_ATTR(kernel_max, 0444, print_cpus_kernel_max, NULL); /* arch-optional setting to enable display of offline cpus >= nr_cpu_ids */ unsigned int total_cpus; -static ssize_t print_cpus_offline(struct sysdev_class *class, char *buf) +static ssize_t print_cpus_offline(struct sysdev_class *class, + struct sysdev_class_attribute *attr, char *buf) { int n = 0, len = PAGE_SIZE-2; cpumask_var_t offline; diff --git a/drivers/base/node.c b/drivers/base/node.c index 70122791683d..85c9d30d7004 100644 --- a/drivers/base/node.c +++ b/drivers/base/node.c @@ -544,23 +544,29 @@ static ssize_t print_nodes_state(enum node_states state, char *buf) return n; } -static ssize_t print_nodes_possible(struct sysdev_class *class, char *buf) +static ssize_t print_nodes_possible(struct sysdev_class *class, + struct sysdev_class_attribute *attr, char *buf) { return print_nodes_state(N_POSSIBLE, buf); } -static ssize_t print_nodes_online(struct sysdev_class *class, char *buf) +static ssize_t print_nodes_online(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { return print_nodes_state(N_ONLINE, buf); } static ssize_t print_nodes_has_normal_memory(struct sysdev_class *class, - char *buf) + struct sysdev_class_attribute *attr, + char *buf) { return print_nodes_state(N_NORMAL_MEMORY, buf); } -static ssize_t print_nodes_has_cpu(struct sysdev_class *class, char *buf) +static ssize_t print_nodes_has_cpu(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { return print_nodes_state(N_CPU, buf); } @@ -573,7 +579,8 @@ static SYSDEV_CLASS_ATTR(has_cpu, 0444, print_nodes_has_cpu, NULL); #ifdef CONFIG_HIGHMEM static ssize_t print_nodes_has_high_memory(struct sysdev_class *class, - char *buf) + struct sysdev_class_attribute *attr, + char *buf) { return print_nodes_state(N_HIGH_MEMORY, buf); } diff --git a/drivers/base/sys.c b/drivers/base/sys.c index 0d903909af7e..a38445c0f8c5 100644 --- a/drivers/base/sys.c +++ b/drivers/base/sys.c @@ -89,7 +89,7 @@ static ssize_t sysdev_class_show(struct kobject *kobj, struct attribute *attr, struct sysdev_class_attribute *class_attr = to_sysdev_class_attr(attr); if (class_attr->show) - return class_attr->show(class, buffer); + return class_attr->show(class, class_attr, buffer); return -EIO; } @@ -100,7 +100,7 @@ static ssize_t sysdev_class_store(struct kobject *kobj, struct attribute *attr, struct sysdev_class_attribute *class_attr = to_sysdev_class_attr(attr); if (class_attr->store) - return class_attr->store(class, buffer, count); + return class_attr->store(class, class_attr, buffer, count); return -EIO; } diff --git a/drivers/cpuidle/sysfs.c b/drivers/cpuidle/sysfs.c index 97b003839fb6..c9cefacabf37 100644 --- a/drivers/cpuidle/sysfs.c +++ b/drivers/cpuidle/sysfs.c @@ -22,6 +22,7 @@ static int __init cpuidle_sysfs_setup(char *unused) __setup("cpuidle_sysfs_switch", cpuidle_sysfs_setup); static ssize_t show_available_governors(struct sysdev_class *class, + struct sysdev_class_attribute *attr, char *buf) { ssize_t i = 0; @@ -41,6 +42,7 @@ out: } static ssize_t show_current_driver(struct sysdev_class *class, + struct sysdev_class_attribute *attr, char *buf) { ssize_t ret; @@ -56,6 +58,7 @@ static ssize_t show_current_driver(struct sysdev_class *class, } static ssize_t show_current_governor(struct sysdev_class *class, + struct sysdev_class_attribute *attr, char *buf) { ssize_t ret; @@ -71,6 +74,7 @@ static ssize_t show_current_governor(struct sysdev_class *class, } static ssize_t store_current_governor(struct sysdev_class *class, + struct sysdev_class_attribute *attr, const char *buf, size_t count) { char gov_name[CPUIDLE_NAME_LEN]; diff --git a/include/linux/sysdev.h b/include/linux/sysdev.h index f395bb3fa2f2..c2458fa8376c 100644 --- a/include/linux/sysdev.h +++ b/include/linux/sysdev.h @@ -41,8 +41,10 @@ struct sysdev_class { struct sysdev_class_attribute { struct attribute attr; - ssize_t (*show)(struct sysdev_class *, char *); - ssize_t (*store)(struct sysdev_class *, const char *, size_t); + ssize_t (*show)(struct sysdev_class *, struct sysdev_class_attribute *, + char *); + ssize_t (*store)(struct sysdev_class *, struct sysdev_class_attribute *, + const char *, size_t); }; #define _SYSDEV_CLASS_ATTR(_name,_mode,_show,_store) \ diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 8e352c756ba7..f40560b86544 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -5481,13 +5481,16 @@ void __init perf_event_init(void) register_cpu_notifier(&perf_cpu_nb); } -static ssize_t perf_show_reserve_percpu(struct sysdev_class *class, char *buf) +static ssize_t perf_show_reserve_percpu(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { return sprintf(buf, "%d\n", perf_reserved_percpu); } static ssize_t perf_set_reserve_percpu(struct sysdev_class *class, + struct sysdev_class_attribute *attr, const char *buf, size_t count) { @@ -5516,13 +5519,17 @@ perf_set_reserve_percpu(struct sysdev_class *class, return count; } -static ssize_t perf_show_overcommit(struct sysdev_class *class, char *buf) +static ssize_t perf_show_overcommit(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { return sprintf(buf, "%d\n", perf_overcommit); } static ssize_t -perf_set_overcommit(struct sysdev_class *class, const char *buf, size_t count) +perf_set_overcommit(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + const char *buf, size_t count) { unsigned long val; int err; diff --git a/kernel/sched.c b/kernel/sched.c index b47ceeec1a91..150b6988de49 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -7406,11 +7406,13 @@ static ssize_t sched_power_savings_store(const char *buf, size_t count, int smt) #ifdef CONFIG_SCHED_MC static ssize_t sched_mc_power_savings_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, char *page) { return sprintf(page, "%u\n", sched_mc_power_savings); } static ssize_t sched_mc_power_savings_store(struct sysdev_class *class, + struct sysdev_class_attribute *attr, const char *buf, size_t count) { return sched_power_savings_store(buf, count, 0); @@ -7422,11 +7424,13 @@ static SYSDEV_CLASS_ATTR(sched_mc_power_savings, 0644, #ifdef CONFIG_SCHED_SMT static ssize_t sched_smt_power_savings_show(struct sysdev_class *dev, + struct sysdev_class_attribute *attr, char *page) { return sprintf(page, "%u\n", sched_smt_power_savings); } static ssize_t sched_smt_power_savings_store(struct sysdev_class *dev, + struct sysdev_class_attribute *attr, const char *buf, size_t count) { return sched_power_savings_store(buf, count, 1); -- cgit v1.2.3 From cd3d8031eb4311e516329aee03c79a08333141f1 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 12 Mar 2010 16:15:36 +0900 Subject: sched: sched_getaffinity(): Allow less than NR_CPUS length [ Note, this commit changes the syscall ABI for > 1024 CPUs systems. ] Recently, some distro decided to use NR_CPUS=4096 for mysterious reasons. Unfortunately, glibc sched interface has the following definition: # define __CPU_SETSIZE 1024 # define __NCPUBITS (8 * sizeof (__cpu_mask)) typedef unsigned long int __cpu_mask; typedef struct { __cpu_mask __bits[__CPU_SETSIZE / __NCPUBITS]; } cpu_set_t; It mean, if NR_CPUS is bigger than 1024, cpu_set_t makes an ABI issue ... More recently, Sharyathi Nagesh reported following test program makes misterious syscall failure: ----------------------------------------------------------------------- #define _GNU_SOURCE #include #include #include int main() { cpu_set_t set; if (sched_getaffinity(0, sizeof(cpu_set_t), &set) < 0) printf("\n Call is failing with:%d", errno); } ----------------------------------------------------------------------- Because the kernel assumes len argument of sched_getaffinity() is bigger than NR_CPUS. But now it is not correct. Now we are faced with the following annoying dilemma, due to the limitations of the glibc interface built in years ago: (1) if we change glibc's __CPU_SETSIZE definition, we lost binary compatibility of _all_ application. (2) if we don't change it, we also lost binary compatibility of Sharyathi's use case. Then, I would propse to change the rule of the len argument of sched_getaffinity(). Old: len should be bigger than NR_CPUS New: len should be bigger than maximum possible cpu id This creates the following behavior: (A) In the real 4096 cpus machine, the above test program still return -EINVAL. (B) NR_CPUS=4096 but the machine have less than 1024 cpus (almost all machines in the world), the above can run successfully. Fortunatelly, BIG SGI machine is mainly used for HPC use case. It means they can rebuild their programs. IOW we hope they are not annoyed by this issue ... Reported-by: Sharyathi Nagesh Signed-off-by: KOSAKI Motohiro Acked-by: Ulrich Drepper Acked-by: Peter Zijlstra Cc: Linus Torvalds Cc: Andrew Morton Cc: Jack Steiner Cc: Russ Anderson Cc: Mike Travis LKML-Reference: <20100312161316.9520.A69D9226@jp.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 9ab3cd7858d3..6eaef3df2d19 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4902,7 +4902,9 @@ SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, int ret; cpumask_var_t mask; - if (len < cpumask_size()) + if (len < nr_cpu_ids) + return -EINVAL; + if (len & (sizeof(unsigned long)-1)) return -EINVAL; if (!alloc_cpumask_var(&mask, GFP_KERNEL)) @@ -4910,10 +4912,12 @@ SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, ret = sched_getaffinity(pid, mask); if (ret == 0) { - if (copy_to_user(user_mask_ptr, mask, cpumask_size())) + int retlen = min(len, cpumask_size()); + + if (copy_to_user(user_mask_ptr, mask, retlen)) ret = -EFAULT; else - ret = cpumask_size(); + ret = retlen; } free_cpumask_var(mask); -- cgit v1.2.3 From c890692bf37671b5b78a1870d55d6d87e1c8a509 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Thu, 11 Mar 2010 14:08:43 -0800 Subject: kernel/sched.c: Suppress unused var warning On UP: kernel/sched.c: In function 'wake_up_new_task': kernel/sched.c:2631: warning: unused variable 'cpu' Signed-off-by: Andrew Morton Cc: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 6eaef3df2d19..82975b5b42f7 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2650,7 +2650,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) { unsigned long flags; struct rq *rq; - int cpu = get_cpu(); + int cpu __maybe_unused = get_cpu(); #ifdef CONFIG_SMP /* -- cgit v1.2.3 From 8bc037fb89bb3104b9ae290d18c877624cd7d9cc Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Wed, 17 Mar 2010 09:36:58 +0900 Subject: sched: Use proper type in sched_getaffinity() Using the proper type fixes the following compiler warning: kernel/sched.c:4850: warning: comparison of distinct pointer types lacks a cast Signed-off-by: KOSAKI Motohiro Cc: torvalds@linux-foundation.org Cc: travis@sgi.com Cc: peterz@infradead.org Cc: drepper@redhat.com Cc: rja@sgi.com Cc: sharyath@in.ibm.com Cc: steiner@sgi.com LKML-Reference: <20100317090046.4C79.A69D9226@jp.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 82975b5b42f7..49d2fa7b687a 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4912,7 +4912,7 @@ SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, ret = sched_getaffinity(pid, mask); if (ret == 0) { - int retlen = min(len, cpumask_size()); + size_t retlen = min_t(size_t, len, cpumask_size()); if (copy_to_user(user_mask_ptr, mask, retlen)) ret = -EFAULT; -- cgit v1.2.3