[PATCH] delay.h: add __must_check to msleep_interruptible

[PATCH] delay.h: add __must_check to msleep_interruptible

[Baruch Siach]

Code calling msleep_interruptible() must be aware that sleep time might be
shorter than intended as a result of a signal being caught. Code not checking
the return value of msleep_interruptible() is probably buggy, unless it's doing
the signal_pending() check itself, which is redundant.

Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
---
 include/linux/delay.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/include/linux/delay.h b/include/linux/delay.h
index a6ecb34..1be4994 100644
--- a/include/linux/delay.h
+++ b/include/linux/delay.h
@@ -44,7 +44,7 @@ static inline void ndelay(unsigned long x)
 extern unsigned long lpj_fine;
 void calibrate_delay(void);
 void msleep(unsigned int msecs);
-unsigned long msleep_interruptible(unsigned int msecs);
+unsigned long __must_check msleep_interruptible(unsigned int msecs);
 void usleep_range(unsigned long min, unsigned long max);
 
 static inline void ssleep(unsigned int seconds)
-- 
1.7.2.3

Re: [PATCH] delay.h: add __must_check to msleep_interruptible

[Baruch Siach]

Hi lkml,

On Thu, Nov 04, 2010 at 10:55:41AM +0200, Baruch Siach wrote:
> Code calling msleep_interruptible() must be aware that sleep time might be
> shorter than intended as a result of a signal being caught. Code not checking
> the return value of msleep_interruptible() is probably buggy, unless it's doing
> the signal_pending() check itself, which is redundant.

Ping?
Any reason not to apply this? This patch could have saved me from an incorrect 
handling of signaled userspace processes.

baruch

> Cc: Andrew Morton <akpm@linux-foundation.org>
> Signed-off-by: Baruch Siach <baruch@tkos.co.il>
> ---
>  include/linux/delay.h |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/include/linux/delay.h b/include/linux/delay.h
> index a6ecb34..1be4994 100644
> --- a/include/linux/delay.h
> +++ b/include/linux/delay.h
> @@ -44,7 +44,7 @@ static inline void ndelay(unsigned long x)
>  extern unsigned long lpj_fine;
>  void calibrate_delay(void);
>  void msleep(unsigned int msecs);
> -unsigned long msleep_interruptible(unsigned int msecs);
> +unsigned long __must_check msleep_interruptible(unsigned int msecs);
>  void usleep_range(unsigned long min, unsigned long max);
>  
>  static inline void ssleep(unsigned int seconds)
> -- 
> 1.7.2.3

-- 
                                                     ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch@tkos.co.il - tel: +972.2.679.5364, http://www.tkos.co.il -

Re: [PATCH] delay.h: add __must_check to msleep_interruptible

[Andrew Morton]

On Thu, 16 Dec 2010 08:17:11 +0200 Baruch Siach <baruch@tkos.co.il> wrote:

> On Thu, Nov 04, 2010 at 10:55:41AM +0200, Baruch Siach wrote:
> > Code calling msleep_interruptible() must be aware that sleep time might be
> > shorter than intended as a result of a signal being caught. Code not checking
> > the return value of msleep_interruptible() is probably buggy, unless it's doing
> > the signal_pending() check itself, which is redundant.
> 
> Ping?

Stuck in my backlog.  Delayed but not forgotten.

Re: [PATCH] delay.h: add __must_check to msleep_interruptible

[Andrew Morton]

On Thu,  4 Nov 2010 10:55:41 +0200
Baruch Siach <baruch@tkos.co.il> wrote:

> Code calling msleep_interruptible() must be aware that sleep time might be
> shorter than intended as a result of a signal being caught. Code not checking
> the return value of msleep_interruptible() is probably buggy, unless it's doing
> the signal_pending() check itself, which is redundant.
> 

True.  But there are around 250 callsites which don't check the
msleep_interruptible() return value.

I don't think I want to add 250 new warnings to the kernel build -
it'll take *years* to get them all weeded out and meanwhile it will
cause people to miss other warnings while they're ignoring the
msleep_interruptible() warnings.

So.  Some lucky duck needs to get down and start fixing all these
things first, please.  Meanwhile, a checkpatch rule which prevents new
occurrences would be good.

Except lots of developers and maintainers can't be assed running
checkpatch, so a volunteer who regularly runs linux-next.patch and
patch-2.6.Y-rcY through checkpatch and then hassles people would also be
good.

Re: [PATCH] delay.h: add __must_check to msleep_interruptible

[Andrew Morton]

On Fri, 17 Dec 2010 15:49:47 -0800
Andrew Morton <akpm@linux-foundation.org> wrote:

> On Thu,  4 Nov 2010 10:55:41 +0200
> Baruch Siach <baruch@tkos.co.il> wrote:
> 
> > Code calling msleep_interruptible() must be aware that sleep time might be
> > shorter than intended as a result of a signal being caught. Code not checking
> > the return value of msleep_interruptible() is probably buggy, unless it's doing
> > the signal_pending() check itself, which is redundant.
> > 
> 
> True.  But there are around 250 callsites which don't check the
> msleep_interruptible() return value.

One quick way of fixing this would be to add

/*
 * Used by callsites which are supposed to be calling msleep_interruptible(),
 * but which were failing to properly handle msleep_interruptible()'s return
 * value
 */
static inline void msleep_you_suck_fixme_please(unsigned int msecs)
{
	msleep(msecs);
}

and then patch all the offending msleep_interruptible() callsites to
use msleep_you_suck_fixme_please().  Then add the __must_check to
msleep_interruptible().

And note that 250 msleep_interruptible() -> msleep() changes will
probably fix lots of bugs.