[PATCH] fix sigqueue accounting for posix-timers broken by new RLIMIT_SIGPENDING tracking code

From: Roland McGrath
Date: Fri Sep 10 2004 - 00:12:47 EST



The introduction of RLIMIT_SIGPENDING and the associated tracking code was
broken for the case of preallocated sigqueue elements, i.e. posix-timers.
It wrongly includes the timer's preallocated sigqueue structs in the count
towards the per-user when allocating them, but (rightly) does not decrement
the count when they are freed.

This patch fixes it by keeping preallocated sigqueue structs out of the
per-user accounting altogether.


Thanks,
Roland


Signed-off-by: Roland McGrath <roland@xxxxxxxxxx>

--- vanilla-linux-2.6/kernel/signal.c.orig 2004-09-09 22:07:56.492060454 -0700
+++ vanilla-linux-2.6/kernel/signal.c 2004-09-09 21:55:45.604292751 -0700
@@ -264,7 +264,7 @@ next_signal(struct sigpending *pending,
return sig;
}

-static struct sigqueue *__sigqueue_alloc(void)
+static struct sigqueue *__sigqueue_alloc(int flags)
{
struct sigqueue *q = NULL;

@@ -273,10 +273,13 @@ static struct sigqueue *__sigqueue_alloc
q = kmem_cache_alloc(sigqueue_cachep, GFP_ATOMIC);
if (q) {
INIT_LIST_HEAD(&q->list);
- q->flags = 0;
q->lock = NULL;
- q->user = get_uid(current->user);
- atomic_inc(&q->user->sigpending);
+ q->user = NULL;
+ q->flags = flags;
+ if (!(flags & SIGQUEUE_PREALLOC)) {
+ q->user = get_uid(current->user);
+ atomic_inc(&q->user->sigpending);
+ }
}
return(q);
}
@@ -1331,11 +1334,7 @@ kill_proc(pid_t pid, int sig, int priv)

struct sigqueue *sigqueue_alloc(void)
{
- struct sigqueue *q;
-
- if ((q = __sigqueue_alloc()))
- q->flags |= SIGQUEUE_PREALLOC;
- return(q);
+ return __sigqueue_alloc(SIGQUEUE_PREALLOC);
}

void sigqueue_free(struct sigqueue *q)
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/