[PATCH v2] ptrace: annotate siglock acquisition

[PATCH v2] ptrace: annotate siglock acquisition

[Namhyung Kim]

lock_task_sighand() conditionally acquires sighand->siglock in case of
returning non-NULL but unlock_task_sighand() releases it unconditionally.
This leads sparse to complain about the lock context imbalance. Annotate it
to make sparse happier.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
---

Roland, I sent the same patch a month ago and got your Acked-by but this
time I update it using __cond_lock() suggested by Al Viro. Please take a
look again, Thanks.

 kernel/ptrace.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/kernel/ptrace.c b/kernel/ptrace.c
index f34d798..8d9b7b2 100644
--- a/kernel/ptrace.c
+++ b/kernel/ptrace.c
@@ -435,7 +435,8 @@ static int ptrace_getsiginfo(struct task_struct *child, siginfo_t *info)
 	unsigned long flags;
 	int error = -ESRCH;
 
-	if (lock_task_sighand(child, &flags)) {
+	if (__cond_lock(&child->sighand->siglock,
+			lock_task_sighand(child, &flags))) {
 		error = -EINVAL;
 		if (likely(child->last_siginfo != NULL)) {
 			*info = *child->last_siginfo;
@@ -451,7 +452,8 @@ static int ptrace_setsiginfo(struct task_struct *child, const siginfo_t *info)
 	unsigned long flags;
 	int error = -ESRCH;
 
-	if (lock_task_sighand(child, &flags)) {
+	if (__cond_lock(&child->sighand->siglock,
+			lock_task_sighand(child, &flags))) {
 		error = -EINVAL;
 		if (likely(child->last_siginfo != NULL)) {
 			*child->last_siginfo = *info;
-- 
1.7.2.2

Re: [PATCH v2] ptrace: annotate siglock acquisition

[Oleg Nesterov]

On 09/25, Namhyung Kim wrote:
>
> --- a/kernel/ptrace.c
> +++ b/kernel/ptrace.c
> @@ -435,7 +435,8 @@ static int ptrace_getsiginfo(struct task_struct *child, siginfo_t *info)
>  	unsigned long flags;
>  	int error = -ESRCH;
>
> -	if (lock_task_sighand(child, &flags)) {
> +	if (__cond_lock(&child->sighand->siglock,
> +			lock_task_sighand(child, &flags))) {

Well, this looks fine, but if we are going to make sparse happy
then we have more callers to fix.

Perhaps we should just renam lock_task_sighand to __lock_task_sighand
and add

	#define lock_task_sighand(tsk, flagsp)	\
		__cond_lock((tsk)->sighand->siglock, __lock_task_sighand((tsk), (flagsp)))

?

Oleg.