Hi all,
just a little patch to the Documentation. I had some trouble
understanding the trailing "/1" on some lock class names of lock_stat
output, so I added something on this inside lockstat documentation.
Signed-off-by: Juri Lelli <[email protected]>
---
Documentation/lockstat.txt | 36 ++++++++++++++++++++++++++++++++++--
1 files changed, 34 insertions(+), 2 deletions(-)
diff --git a/Documentation/lockstat.txt b/Documentation/lockstat.txt
index 65f4c79..75eeb65 100644
--- a/Documentation/lockstat.txt
+++ b/Documentation/lockstat.txt
@@ -12,8 +12,9 @@ Because things like lock contention can severely
impact performance.
- HOW
Lockdep already has hooks in the lock functions and maps lock instances to
-lock classes. We build on that. The graph below shows the relation between
-the lock functions and the various hooks therein.
+lock classes. We build on that (see Documentation/lockdep-design.txt).
+The graph below shows the relation between the lock functions and the
various
+hooks therein.
__acquire
|
@@ -128,6 +129,37 @@ points are the points we're contending with.
The integer part of the time values is in us.
+Dealing with nested locks, subclasses may appear:
+
+32...............................................................................................................................................................................................
+33
+34 &rq->lock: 13128
13128 0.43 190.53 103881.26 97454
3453404 0.00 401.11 13224683.11
+35 ---------
+36 &rq->lock 645
[<ffffffff8103bfc4>] task_rq_lock+0x43/0x75
+37 &rq->lock 297
[<ffffffff8104ba65>] try_to_wake_up+0x127/0x25a
+38 &rq->lock 360
[<ffffffff8103c4c5>] select_task_rq_fair+0x1f0/0x74a
+39 &rq->lock 428
[<ffffffff81045f98>] scheduler_tick+0x46/0x1fb
+40 ---------
+41 &rq->lock 77
[<ffffffff8103bfc4>] task_rq_lock+0x43/0x75
+42 &rq->lock 174
[<ffffffff8104ba65>] try_to_wake_up+0x127/0x25a
+43 &rq->lock 4715
[<ffffffff8103ed4b>] double_rq_lock+0x42/0x54
+44 &rq->lock 893
[<ffffffff81340524>] schedule+0x157/0x7b8
+45
+46...............................................................................................................................................................................................
+47
+48 &rq->lock/1: 11526
11488 0.33 388.73 136294.31 21461
38404 0.00 37.93 109388.53
+49 -----------
+50 &rq->lock/1 11526
[<ffffffff8103ed58>] double_rq_lock+0x4f/0x54
+51 -----------
+52 &rq->lock/1 5645
[<ffffffff8103ed4b>] double_rq_lock+0x42/0x54
+53 &rq->lock/1 1224
[<ffffffff81340524>] schedule+0x157/0x7b8
+54 &rq->lock/1 4336
[<ffffffff8103ed58>] double_rq_lock+0x4f/0x54
+55 &rq->lock/1 181
[<ffffffff8104ba65>] try_to_wake_up+0x127/0x25a
+
+Line 48 shows statistics for the first subclass (/1) of &rq->lock
class, since
+in this case, as line 50 suggests, double_rq_lock actually acquires a
nested
+lock of two spinlocks.
+
View the top contending locks:
# grep : /proc/lock_stat | head
--
1.7.4.1
On Mon, 2011-05-16 at 12:01 +0200, Juri Lelli wrote:
> +The graph below shows the relation between the lock functions and the
> various
> +hooks therein.
That smells like the patch has whitespace damage..
* Peter Zijlstra <[email protected]> wrote:
> On Mon, 2011-05-16 at 12:01 +0200, Juri Lelli wrote:
> > +The graph below shows the relation between the lock functions and the
> > various
> > +hooks therein.
>
> That smells like the patch has whitespace damage..
Yeah, Documentation/email-clients.txt might help there.
Thanks,
Ingo
On Mon, May 16, 2011 at 6:01 PM, Juri Lelli <[email protected]> wrote:
> Hi all,
> just a little patch to the Documentation. I had some trouble understanding
> the trailing "/1" on some lock class names of lock_stat output, so I added
> something on this inside lockstat documentation.
>
> Signed-off-by: Juri Lelli <[email protected]>
>
> ---
> Documentation/lockstat.txt | 36 ++++++++++++++++++++++++++++++++++--
> 1 files changed, 34 insertions(+), 2 deletions(-)
>
> diff --git a/Documentation/lockstat.txt b/Documentation/lockstat.txt
> index 65f4c79..75eeb65 100644
> --- a/Documentation/lockstat.txt
> +++ b/Documentation/lockstat.txt
> @@ -12,8 +12,9 @@ Because things like lock contention can severely impact
> performance.
> - HOW
>
> Lockdep already has hooks in the lock functions and maps lock instances to
> -lock classes. We build on that. The graph below shows the relation between
> -the lock functions and the various hooks therein.
> +lock classes. We build on that (see Documentation/lockdep-design.txt).
> +The graph below shows the relation between the lock functions and the
> various
> +hooks therein.
>
> __acquire
> |
> @@ -128,6 +129,37 @@ points are the points we're contending with.
>
> The integer part of the time values is in us.
>
> +Dealing with nested locks, subclasses may appear:
> +
> +32...............................................................................................................................................................................................
> +33
> +34 &rq->lock: 13128 13128
> 0.43 190.53 103881.26 97454 3453404 0.00
> 401.11 13224683.11
> +35 ---------
> +36 &rq->lock 645
> [<ffffffff8103bfc4>] task_rq_lock+0x43/0x75
> +37 &rq->lock 297
> [<ffffffff8104ba65>] try_to_wake_up+0x127/0x25a
> +38 &rq->lock 360
> [<ffffffff8103c4c5>] select_task_rq_fair+0x1f0/0x74a
> +39 &rq->lock 428
> [<ffffffff81045f98>] scheduler_tick+0x46/0x1fb
> +40 ---------
> +41 &rq->lock 77
> [<ffffffff8103bfc4>] task_rq_lock+0x43/0x75
> +42 &rq->lock 174
> [<ffffffff8104ba65>] try_to_wake_up+0x127/0x25a
> +43 &rq->lock 4715
> [<ffffffff8103ed4b>] double_rq_lock+0x42/0x54
> +44 &rq->lock 893
> [<ffffffff81340524>] schedule+0x157/0x7b8
> +45
> +46...............................................................................................................................................................................................
> +47
> +48 &rq->lock/1: 11526 11488
> 0.33 388.73 136294.31 21461 38404 0.00
> 37.93 109388.53
> +49 -----------
> +50 &rq->lock/1 11526
> [<ffffffff8103ed58>] double_rq_lock+0x4f/0x54
> +51 -----------
> +52 &rq->lock/1 5645
> [<ffffffff8103ed4b>] double_rq_lock+0x42/0x54
> +53 &rq->lock/1 1224
> [<ffffffff81340524>] schedule+0x157/0x7b8
> +54 &rq->lock/1 4336
> [<ffffffff8103ed58>] double_rq_lock+0x4f/0x54
> +55 &rq->lock/1 181
> [<ffffffff8104ba65>] try_to_wake_up+0x127/0x25a
> +
> +Line 48 shows statistics for the first subclass (/1) of &rq->lock class,
Actually it's the second subclass because subclass starts from 0.
Thanks,
Yong
> since
> +in this case, as line 50 suggests, double_rq_lock actually acquires a
> nested
> +lock of two spinlocks.
> +
> View the top contending locks:
>
> # grep : /proc/lock_stat | head
> --
> 1.7.4.1
>
--
Only stand for myself
????{.n?+???????+%?????ݶ??w??{.n?+????{??G?????{ay? ʇڙ?,j??f???h?????????z_??(?階?ݢj"???m??????G????????????&???~???iO???z??v?^?m???? ????????I?
Sorry for my mistake with the email client and thanks for the hint!
May I resend a (I hope) corrected version of this small patch with the revision
made by Yong?
On 05/16/2011 12:35 PM, Ingo Molnar wrote:
>
> * Peter Zijlstra <[email protected]> wrote:
>
>> On Mon, 2011-05-16 at 12:01 +0200, Juri Lelli wrote:
>>> +The graph below shows the relation between the lock functions and the
>>> various
>>> +hooks therein.
>>
>> That smells like the patch has whitespace damage..
>
> Yeah, Documentation/email-clients.txt might help there.
>
> Thanks,
>
> Ingo
Thanks,
Juri
On Tue, May 17, 2011 at 4:39 PM, Juri Lelli <[email protected]> wrote:
> Sorry for my mistake with the email client and thanks for the hint!
> May I resend a (I hope) corrected version of this small patch with the revision
> made by Yong?
No problem on my side :)
Thanks,
Yong
>
> On 05/16/2011 12:35 PM, Ingo Molnar wrote:
>>
>> * Peter Zijlstra <[email protected]> wrote:
>>
>>> On Mon, 2011-05-16 at 12:01 +0200, Juri Lelli wrote:
>>>> +The graph below shows the relation between the lock functions and the
>>>> various
>>>> +hooks therein.
>>>
>>> That smells like the patch has whitespace damage..
>>
>> Yeah, Documentation/email-clients.txt might help there.
>>
>> Thanks,
>>
>> Ingo
>
> Thanks,
>
> Juri
>
--
Only stand for myself