[PATCH][RFC] Check poll_select_set_timeout return value in sys_poll

[PATCH][RFC] Check poll_select_set_timeout return value in sys_poll

[Jesper Juhl]

poll_select_set_timeout() may return -EINVAL if it's given invalid 
arguments. In sys_poll() we don't check the return value. The way 
poll_select_set_timeout() is called here currently I don't think can ever 
actually fail, but it doesn't hurt to be defensive - code has a tendency 
to change over time. This patch checks the return value and returns 
-EINVAL just like we do elsewhere if it does fail.

What do you think? Is this just pointless overhead or actually worth it as 
an "insurance policy"?

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
---
 select.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/fs/select.c b/fs/select.c
index e56560d..333d5aa 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -922,8 +922,9 @@ SYSCALL_DEFINE3(poll, struct pollfd __user *, ufds, unsigned int, nfds,
 
 	if (timeout_msecs >= 0) {
 		to = &end_time;
-		poll_select_set_timeout(to, timeout_msecs / MSEC_PER_SEC,
-			NSEC_PER_MSEC * (timeout_msecs % MSEC_PER_SEC));
+		if (poll_select_set_timeout(to, timeout_msecs / MSEC_PER_SEC,
+			NSEC_PER_MSEC * (timeout_msecs % MSEC_PER_SEC)))
+			return -EINVAL;
 	}
 
 	ret = do_sys_poll(ufds, nfds, to);


-- 
Jesper Juhl <jj@chaosbits.net>            http://www.chaosbits.net/
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please.

Re: [PATCH][RFC] Check poll_select_set_timeout return value in sys_poll

[Jesper Juhl]

On Mon, 24 Jan 2011, Jesper Juhl wrote:

> poll_select_set_timeout() may return -EINVAL if it's given invalid 
> arguments. In sys_poll() we don't check the return value. The way 
> poll_select_set_timeout() is called here currently I don't think can ever 
> actually fail, but it doesn't hurt to be defensive - code has a tendency 
> to change over time. This patch checks the return value and returns 
> -EINVAL just like we do elsewhere if it does fail.
> 
> What do you think? Is this just pointless overhead or actually worth it as 
> an "insurance policy"?
> 

I guess that if we do this we should at least mark this branch as 
unlikely().

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
---
 select.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/fs/select.c b/fs/select.c
index e56560d..ebdca6e 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -922,8 +922,9 @@ SYSCALL_DEFINE3(poll, struct pollfd __user *, ufds, unsigned int, nfds,
 
 	if (timeout_msecs >= 0) {
 		to = &end_time;
-		poll_select_set_timeout(to, timeout_msecs / MSEC_PER_SEC,
-			NSEC_PER_MSEC * (timeout_msecs % MSEC_PER_SEC));
+		if (unlikely(poll_select_set_timeout(to, timeout_msecs / MSEC_PER_SEC,
+			NSEC_PER_MSEC * (timeout_msecs % MSEC_PER_SEC))))
+			return -EINVAL;
 	}
 
 	ret = do_sys_poll(ufds, nfds, to);


-- 
Jesper Juhl <jj@chaosbits.net>            http://www.chaosbits.net/
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please.