[PATCH] prctl: return timerslack through pointer

[PATCH] prctl: return timerslack through pointer

[Smith, GeoffX]

This patch fixes the semantics of prctl() option PR_GET_TIMERSLACK
to pass the return value through *arg2.

With this change, the option now follows the same conventions as the
other "get" options added since 2.6.0, and also brings it into
conformance with the advice in chapter 16 of Documentation/CodingStyle.

In addition, it effectively doubles the range of time slack to 4.29
seconds (on a 32-bit architecture).

The timer slack feature was only added in Sep 2009 (new in 2.6.32.1), 
so there are not any production applications to break.  I have also 
repaired some unmatched signed/unsigned mismatches in timer slack
calculations.



diff --git a/kernel/sys.c b/kernel/sys.c
index 20ccfb5..e01da12 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -1531,7 +1531,8 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3,
 			error = perf_event_task_enable();
 			break;
 		case PR_GET_TIMERSLACK:
-			error = current->timer_slack_ns;
+			error = put_user(current->timer_slack_ns, 
+				(unsigned long __user *)arg2);
 			break;
 		case PR_SET_TIMERSLACK:
 			if (arg2 <= 0)
diff --git a/fs/select.c b/fs/select.c
index fd38ce2..5122364 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -44,10 +44,10 @@
 
 #define MAX_SLACK	(100 * NSEC_PER_MSEC)
 
-static long __estimate_accuracy(struct timespec *tv)
+static unsigned long __estimate_accuracy(struct timespec *tv)
 {
-	long slack;
-	int divfactor = 1000;
+	unsigned long slack;
+	unsigned int divfactor = 1000;
 
 	if (tv->tv_sec < 0)
 		return 0;
@@ -67,7 +67,7 @@ static long __estimate_accuracy(struct timespec *tv)
 	return slack;
 }
 
-static long estimate_accuracy(struct timespec *tv)
+static unsigned long estimate_accuracy(struct timespec *tv)
 {
 	unsigned long ret;
 	struct timespec now;

Re: [PATCH] prctl: return timerslack through pointer

[Andrew Morton]

On Thu, 17 Dec 2009 16:17:15 -0800
"Smith, GeoffX" <geoffx.smith@intel.com> wrote:

> This patch fixes the semantics of prctl() option PR_GET_TIMERSLACK
> to pass the return value through *arg2.
> 
> With this change, the option now follows the same conventions as the
> other "get" options added since 2.6.0, and also brings it into
> conformance with the advice in chapter 16 of Documentation/CodingStyle.
> 
> In addition, it effectively doubles the range of time slack to 4.29
> seconds (on a 32-bit architecture).
> 
> The timer slack feature was only added in Sep 2009 (new in 2.6.32.1), 
> so there are not any production applications to break.  I have also 
> repaired some unmatched signed/unsigned mismatches in timer slack
> calculations.
> 

If we're going to do this then we should backport the patch into
2.6.32.x as well.  And I agree that there is time to make this change.

> diff --git a/kernel/sys.c b/kernel/sys.c
> index 20ccfb5..e01da12 100644

You omitted the Signed-off-by: line (described in
Documentation/SubmittingPatches).  Please send one.

RE: [PATCH] prctl: return timerslack through pointer

[Smith, GeoffX]

Oops, my oversight.  In re: any patch I submitted 12/17/2009:

Signed-off-by: Geoff Smith <geoffx.smith@intel.com>

-----Original Message-----
From: Andrew Morton [mailto:akpm@linux-foundation.org] 
Sent: Tuesday, December 22, 2009 1:51 PM
To: Smith, GeoffX
Cc: linux-kernel@vger.kernel.org; Arjan van de Ven; stable@kernel.org
Subject: Re: [PATCH] prctl: return timerslack through pointer

On Thu, 17 Dec 2009 16:17:15 -0800
"Smith, GeoffX" <geoffx.smith@intel.com> wrote:

> This patch fixes the semantics of prctl() option PR_GET_TIMERSLACK
> to pass the return value through *arg2.
> 
> With this change, the option now follows the same conventions as the
> other "get" options added since 2.6.0, and also brings it into
> conformance with the advice in chapter 16 of Documentation/CodingStyle.
> 
> In addition, it effectively doubles the range of time slack to 4.29
> seconds (on a 32-bit architecture).
> 
> The timer slack feature was only added in Sep 2009 (new in 2.6.32.1), 
> so there are not any production applications to break.  I have also 
> repaired some unmatched signed/unsigned mismatches in timer slack
> calculations.
> 

If we're going to do this then we should backport the patch into
2.6.32.x as well.  And I agree that there is time to make this change.

> diff --git a/kernel/sys.c b/kernel/sys.c
> index 20ccfb5..e01da12 100644

You omitted the Signed-off-by: line (described in
Documentation/SubmittingPatches).  Please send one.

RE: [PATCH] prctl: return timerslack through pointer

[Smith, GeoffX]

On December 22, Andrew Morton <akpm@linux-foundation.org> wrote:

> To: Smith, GeoffX
> Cc: linux-kernel@vger.kernel.org; Arjan van de Ven; stable@kernel.org
> Subject: Re: [PATCH] prctl: return timerslack through pointer
>
> On Thu, 17 Dec 2009 16:17:15 -0800
> "Smith, GeoffX" <geoffx.smith@intel.com> wrote:
>
>> This patch fixes the semantics of prctl() option PR_GET_TIMERSLACK
>> to pass the return value through *arg2....
>
> If we're going to do this then we should backport the patch into
> 2.6.32.x as well.  And I agree that there is time to make this change.


The patch applies cleanly to 2.6.32.2 as well as to the mainline.