diff options
Diffstat (limited to 'kernel')
| -rw-r--r-- | kernel/audit.c | 3 | ||||
| -rw-r--r-- | kernel/cpuset.c | 81 | ||||
| -rw-r--r-- | kernel/sched.c | 4 | ||||
| -rw-r--r-- | kernel/signal.c | 2 | ||||
| -rw-r--r-- | kernel/sysctl.c | 4 | ||||
| -rw-r--r-- | kernel/timer.c | 2 | ||||
| -rw-r--r-- | kernel/workqueue.c | 2 |
7 files changed, 68 insertions, 30 deletions
diff --git a/kernel/audit.c b/kernel/audit.c index ef35166..7f06997 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -514,7 +514,8 @@ static int __init audit_init(void) { printk(KERN_INFO "audit: initializing netlink socket (%s)\n", audit_default ? "enabled" : "disabled"); - audit_sock = netlink_kernel_create(NETLINK_AUDIT, audit_receive); + audit_sock = netlink_kernel_create(NETLINK_AUDIT, 0, audit_receive, + THIS_MODULE); if (!audit_sock) audit_panic("cannot initialize netlink socket"); diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 805fb90..8ab1b4e 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -398,21 +398,31 @@ static int cpuset_path(const struct cpuset *cs, char *buf, int buflen) * to continue to serve a useful existence. Next time it's released, * we will get notified again, if it still has 'notify_on_release' set. * - * Note final arg to call_usermodehelper() is 0 - that means - * don't wait. Since we are holding the global cpuset_sem here, - * and we are asking another thread (started from keventd) to rmdir a - * cpuset, we can't wait - or we'd deadlock with the removing thread - * on cpuset_sem. + * The final arg to call_usermodehelper() is 0, which means don't + * wait. The separate /sbin/cpuset_release_agent task is forked by + * call_usermodehelper(), then control in this thread returns here, + * without waiting for the release agent task. We don't bother to + * wait because the caller of this routine has no use for the exit + * status of the /sbin/cpuset_release_agent task, so no sense holding + * our caller up for that. + * + * The simple act of forking that task might require more memory, + * which might need cpuset_sem. So this routine must be called while + * cpuset_sem is not held, to avoid a possible deadlock. See also + * comments for check_for_release(), below. */ -static int cpuset_release_agent(char *cpuset_str) +static void cpuset_release_agent(const char *pathbuf) { char *argv[3], *envp[3]; int i; + if (!pathbuf) + return; + i = 0; argv[i++] = "/sbin/cpuset_release_agent"; - argv[i++] = cpuset_str; + argv[i++] = (char *)pathbuf; argv[i] = NULL; i = 0; @@ -421,17 +431,29 @@ static int cpuset_release_agent(char *cpuset_str) envp[i++] = "PATH=/sbin:/bin:/usr/sbin:/usr/bin"; envp[i] = NULL; - return call_usermodehelper(argv[0], argv, envp, 0); + call_usermodehelper(argv[0], argv, envp, 0); + kfree(pathbuf); } /* * Either cs->count of using tasks transitioned to zero, or the * cs->children list of child cpusets just became empty. If this * cs is notify_on_release() and now both the user count is zero and - * the list of children is empty, send notice to user land. + * the list of children is empty, prepare cpuset path in a kmalloc'd + * buffer, to be returned via ppathbuf, so that the caller can invoke + * cpuset_release_agent() with it later on, once cpuset_sem is dropped. + * Call here with cpuset_sem held. + * + * This check_for_release() routine is responsible for kmalloc'ing + * pathbuf. The above cpuset_release_agent() is responsible for + * kfree'ing pathbuf. The caller of these routines is responsible + * for providing a pathbuf pointer, initialized to NULL, then + * calling check_for_release() with cpuset_sem held and the address + * of the pathbuf pointer, then dropping cpuset_sem, then calling + * cpuset_release_agent() with pathbuf, as set by check_for_release(). */ -static void check_for_release(struct cpuset *cs) +static void check_for_release(struct cpuset *cs, char **ppathbuf) { if (notify_on_release(cs) && atomic_read(&cs->count) == 0 && list_empty(&cs->children)) { @@ -441,10 +463,9 @@ static void check_for_release(struct cpuset *cs) if (!buf) return; if (cpuset_path(cs, buf, PAGE_SIZE) < 0) - goto out; - cpuset_release_agent(buf); -out: - kfree(buf); + kfree(buf); + else + *ppathbuf = buf; } } @@ -606,6 +627,14 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) * Call with cpuset_sem held. May nest a call to the * lock_cpu_hotplug()/unlock_cpu_hotplug() pair. */ + +/* + * Hack to avoid 2.6.13 partial node dynamic sched domain bug. + * Disable letting 'cpu_exclusive' cpusets define dynamic sched + * domains, until the sched domain can handle partial nodes. + * Remove this #if hackery when sched domains fixed. + */ +#if 0 static void update_cpu_domains(struct cpuset *cur) { struct cpuset *c, *par = cur->parent; @@ -646,6 +675,11 @@ static void update_cpu_domains(struct cpuset *cur) partition_sched_domains(&pspan, &cspan); unlock_cpu_hotplug(); } +#else +static void update_cpu_domains(struct cpuset *cur) +{ +} +#endif static int update_cpumask(struct cpuset *cs, char *buf) { @@ -727,14 +761,14 @@ static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, char *buf) return 0; } -static int attach_task(struct cpuset *cs, char *buf) +static int attach_task(struct cpuset *cs, char *pidbuf, char **ppathbuf) { pid_t pid; struct task_struct *tsk; struct cpuset *oldcs; cpumask_t cpus; - if (sscanf(buf, "%d", &pid) != 1) + if (sscanf(pidbuf, "%d", &pid) != 1) return -EIO; if (cpus_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed)) return -ENOSPC; @@ -777,7 +811,7 @@ static int attach_task(struct cpuset *cs, char *buf) put_task_struct(tsk); if (atomic_dec_and_test(&oldcs->count)) - check_for_release(oldcs); + check_for_release(oldcs, ppathbuf); return 0; } @@ -801,6 +835,7 @@ static ssize_t cpuset_common_file_write(struct file *file, const char __user *us struct cftype *cft = __d_cft(file->f_dentry); cpuset_filetype_t type = cft->private; char *buffer; + char *pathbuf = NULL; int retval = 0; /* Crude upper limit on largest legitimate cpulist user might write. */ @@ -841,7 +876,7 @@ static ssize_t cpuset_common_file_write(struct file *file, const char __user *us retval = update_flag(CS_NOTIFY_ON_RELEASE, cs, buffer); break; case FILE_TASKLIST: - retval = attach_task(cs, buffer); + retval = attach_task(cs, buffer, &pathbuf); break; default: retval = -EINVAL; @@ -852,6 +887,7 @@ static ssize_t cpuset_common_file_write(struct file *file, const char __user *us retval = nbytes; out2: up(&cpuset_sem); + cpuset_release_agent(pathbuf); out1: kfree(buffer); return retval; @@ -1357,6 +1393,7 @@ static int cpuset_rmdir(struct inode *unused_dir, struct dentry *dentry) struct cpuset *cs = dentry->d_fsdata; struct dentry *d; struct cpuset *parent; + char *pathbuf = NULL; /* the vfs holds both inode->i_sem already */ @@ -1376,7 +1413,7 @@ static int cpuset_rmdir(struct inode *unused_dir, struct dentry *dentry) update_cpu_domains(cs); list_del(&cs->sibling); /* delete my sibling from parent->children */ if (list_empty(&parent->children)) - check_for_release(parent); + check_for_release(parent, &pathbuf); spin_lock(&cs->dentry->d_lock); d = dget(cs->dentry); cs->dentry = NULL; @@ -1384,6 +1421,7 @@ static int cpuset_rmdir(struct inode *unused_dir, struct dentry *dentry) cpuset_d_remove_dir(d); dput(d); up(&cpuset_sem); + cpuset_release_agent(pathbuf); return 0; } @@ -1483,10 +1521,13 @@ void cpuset_exit(struct task_struct *tsk) task_unlock(tsk); if (notify_on_release(cs)) { + char *pathbuf = NULL; + down(&cpuset_sem); if (atomic_dec_and_test(&cs->count)) - check_for_release(cs); + check_for_release(cs, &pathbuf); up(&cpuset_sem); + cpuset_release_agent(pathbuf); } else { atomic_dec(&cs->count); } diff --git a/kernel/sched.c b/kernel/sched.c index a646e4f..5f889d0 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -3378,8 +3378,8 @@ EXPORT_SYMBOL(set_user_nice); */ int can_nice(const task_t *p, const int nice) { - /* convert nice value [19,-20] to rlimit style value [0,39] */ - int nice_rlim = 19 - 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 || capable(CAP_SYS_NICE)); } diff --git a/kernel/signal.c b/kernel/signal.c index ca1186e..d282fea 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -692,7 +692,7 @@ static void handle_stop_signal(int sig, struct task_struct *p) { struct task_struct *t; - if (p->flags & SIGNAL_GROUP_EXIT) + if (p->signal->flags & SIGNAL_GROUP_EXIT) /* * The process is in the middle of dying already. */ diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 3e0bbee..8e56e24 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -31,6 +31,7 @@ #include <linux/smp_lock.h> #include <linux/init.h> #include <linux/kernel.h> +#include <linux/net.h> #include <linux/sysrq.h> #include <linux/highuid.h> #include <linux/writeback.h> @@ -136,9 +137,6 @@ static struct ctl_table_header root_table_header = static ctl_table kern_table[]; static ctl_table vm_table[]; -#ifdef CONFIG_NET -extern ctl_table net_table[]; -#endif static ctl_table proc_table[]; static ctl_table fs_table[]; static ctl_table debug_table[]; diff --git a/kernel/timer.c b/kernel/timer.c index f2a1188..5377f40 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -1023,7 +1023,7 @@ asmlinkage long sys_getppid(void) parent = me->group_leader->real_parent; for (;;) { pid = parent->tgid; -#ifdef CONFIG_SMP +#if defined(CONFIG_SMP) || defined(CONFIG_PREEMPT) { struct task_struct *old = parent; diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 259cf55..c7e36d4 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -308,8 +308,6 @@ struct workqueue_struct *__create_workqueue(const char *name, struct workqueue_struct *wq; struct task_struct *p; - BUG_ON(strlen(name) > 10); - wq = kmalloc(sizeof(*wq), GFP_KERNEL); if (!wq) return NULL; |
