2023-04-28 03:13:02

by Randy Dunlap

[permalink] [raw]
Subject: [PATCH -next] sched: fix cid_lock kernel-doc warnings

Fix kernel-doc warnings for cid_lock and use_cid_lock.
These comments are not in kernel-doc format.

kernel/sched/core.c:11496: warning: Cannot understand * @cid_lock: Guarantee forward-progress of cid allocation.
on line 11496 - I thought it was a doc line
kernel/sched/core.c:11505: warning: Cannot understand * @use_cid_lock: Select cid allocation behavior: lock-free vs spinlock.
on line 11505 - I thought it was a doc line

Fixes: 223baf9d17f2 ("sched: Fix performance regression introduced by mm_cid")
Signed-off-by: Randy Dunlap <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
Cc: "Peter Zijlstra (Intel)" <[email protected]>
Cc: Ingo Molnar <[email protected]>
---
kernel/sched/core.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff -- a/kernel/sched/core.c b/kernel/sched/core.c
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -11492,7 +11492,7 @@ void call_trace_sched_update_nr_running(

#ifdef CONFIG_SCHED_MM_CID

-/**
+/*
* @cid_lock: Guarantee forward-progress of cid allocation.
*
* Concurrency ID allocation within a bitmap is mostly lock-free. The cid_lock
@@ -11501,7 +11501,7 @@ void call_trace_sched_update_nr_running(
*/
DEFINE_RAW_SPINLOCK(cid_lock);

-/**
+/*
* @use_cid_lock: Select cid allocation behavior: lock-free vs spinlock.
*
* When @use_cid_lock is 0, the cid allocation is lock-free. When contention is


2023-05-01 13:12:41

by Mathieu Desnoyers

[permalink] [raw]
Subject: Re: [PATCH -next] sched: fix cid_lock kernel-doc warnings

On 2023-04-27 23:11, Randy Dunlap wrote:
> Fix kernel-doc warnings for cid_lock and use_cid_lock.
> These comments are not in kernel-doc format.
>
> kernel/sched/core.c:11496: warning: Cannot understand * @cid_lock: Guarantee forward-progress of cid allocation.
> on line 11496 - I thought it was a doc line
> kernel/sched/core.c:11505: warning: Cannot understand * @use_cid_lock: Select cid allocation behavior: lock-free vs spinlock.
> on line 11505 - I thought it was a doc line
>

Right. It made sense to have those as kernel-doc when those were
mm_struct fields in previous versions of the patch, but not anymore now
that those are stand-alone static variables.

Acked-by: Mathieu Desnoyers <[email protected]>

Thanks,

Mathieu

> Fixes: 223baf9d17f2 ("sched: Fix performance regression introduced by mm_cid")
> Signed-off-by: Randy Dunlap <[email protected]>
> Cc: Mathieu Desnoyers <[email protected]>
> Cc: "Peter Zijlstra (Intel)" <[email protected]>
> Cc: Ingo Molnar <[email protected]>
> ---
> kernel/sched/core.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff -- a/kernel/sched/core.c b/kernel/sched/core.c
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -11492,7 +11492,7 @@ void call_trace_sched_update_nr_running(
>
> #ifdef CONFIG_SCHED_MM_CID
>
> -/**
> +/*
> * @cid_lock: Guarantee forward-progress of cid allocation.
> *
> * Concurrency ID allocation within a bitmap is mostly lock-free. The cid_lock
> @@ -11501,7 +11501,7 @@ void call_trace_sched_update_nr_running(
> */
> DEFINE_RAW_SPINLOCK(cid_lock);
>
> -/**
> +/*
> * @use_cid_lock: Select cid allocation behavior: lock-free vs spinlock.
> *
> * When @use_cid_lock is 0, the cid allocation is lock-free. When contention is

--
Mathieu Desnoyers
EfficiOS Inc.
https://www.efficios.com

2023-05-01 13:24:30

by Peter Zijlstra

[permalink] [raw]
Subject: Re: [PATCH -next] sched: fix cid_lock kernel-doc warnings

On Thu, Apr 27, 2023 at 08:11:11PM -0700, Randy Dunlap wrote:
> Fix kernel-doc warnings for cid_lock and use_cid_lock.
> These comments are not in kernel-doc format.
>
> kernel/sched/core.c:11496: warning: Cannot understand * @cid_lock: Guarantee forward-progress of cid allocation.
> on line 11496 - I thought it was a doc line
> kernel/sched/core.c:11505: warning: Cannot understand * @use_cid_lock: Select cid allocation behavior: lock-free vs spinlock.
> on line 11505 - I thought it was a doc line
>
> Fixes: 223baf9d17f2 ("sched: Fix performance regression introduced by mm_cid")
> Signed-off-by: Randy Dunlap <[email protected]>

Blergh, so mostly I avoid kerneldoc like the plague; because then you
get random people doing 'fixups' that end up making the actual comment
unreadable garbage.

Mostly my answer to any such patch is to simply remove the extra * and
call it fixed.

But now the thing presumes to know better? :-(

Anyway, in this case I suppose we can add it on and see, there's still a
fair number of actual kerneldoc comments in that file.

> kernel/sched/core.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff -- a/kernel/sched/core.c b/kernel/sched/core.c
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -11492,7 +11492,7 @@ void call_trace_sched_update_nr_running(
>
> #ifdef CONFIG_SCHED_MM_CID
>
> -/**
> +/*
> * @cid_lock: Guarantee forward-progress of cid allocation.
> *
> * Concurrency ID allocation within a bitmap is mostly lock-free. The cid_lock
> @@ -11501,7 +11501,7 @@ void call_trace_sched_update_nr_running(
> */
> DEFINE_RAW_SPINLOCK(cid_lock);
>
> -/**
> +/*
> * @use_cid_lock: Select cid allocation behavior: lock-free vs spinlock.
> *
> * When @use_cid_lock is 0, the cid allocation is lock-free. When contention is

2023-05-01 13:43:54

by Peter Zijlstra

[permalink] [raw]
Subject: Re: [PATCH -next] sched: fix cid_lock kernel-doc warnings

On Mon, May 01, 2023 at 03:20:20PM +0200, Peter Zijlstra wrote:
> On Thu, Apr 27, 2023 at 08:11:11PM -0700, Randy Dunlap wrote:
> > Fix kernel-doc warnings for cid_lock and use_cid_lock.
> > These comments are not in kernel-doc format.
> >
> > kernel/sched/core.c:11496: warning: Cannot understand * @cid_lock: Guarantee forward-progress of cid allocation.
> > on line 11496 - I thought it was a doc line
> > kernel/sched/core.c:11505: warning: Cannot understand * @use_cid_lock: Select cid allocation behavior: lock-free vs spinlock.
> > on line 11505 - I thought it was a doc line
> >
> > Fixes: 223baf9d17f2 ("sched: Fix performance regression introduced by mm_cid")
> > Signed-off-by: Randy Dunlap <[email protected]>
>
> Blergh, so mostly I avoid kerneldoc like the plague; because then you
> get random people doing 'fixups' that end up making the actual comment
> unreadable garbage.
>
> Mostly my answer to any such patch is to simply remove the extra * and
> call it fixed.
>
> But now the thing presumes to know better? :-(

Bah, I should go sleep more.. this removes them. All good.

Subject: [tip: sched/urgent] sched: fix cid_lock kernel-doc warnings

The following commit has been merged into the sched/urgent branch of tip:

Commit-ID: 0019a2d4b7e37a983d133d42b707b8a3018ae6f4
Gitweb: https://git.kernel.org/tip/0019a2d4b7e37a983d133d42b707b8a3018ae6f4
Author: Randy Dunlap <[email protected]>
AuthorDate: Thu, 27 Apr 2023 20:11:11 -07:00
Committer: Peter Zijlstra <[email protected]>
CommitterDate: Mon, 08 May 2023 10:58:28 +02:00

sched: fix cid_lock kernel-doc warnings

Fix kernel-doc warnings for cid_lock and use_cid_lock.
These comments are not in kernel-doc format.

kernel/sched/core.c:11496: warning: Cannot understand * @cid_lock: Guarantee forward-progress of cid allocation.
on line 11496 - I thought it was a doc line
kernel/sched/core.c:11505: warning: Cannot understand * @use_cid_lock: Select cid allocation behavior: lock-free vs spinlock.
on line 11505 - I thought it was a doc line

Fixes: 223baf9d17f2 ("sched: Fix performance regression introduced by mm_cid")
Signed-off-by: Randy Dunlap <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Link: https://lkml.kernel.org/r/[email protected]
---
kernel/sched/core.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 944c3ae..a68d127 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -11492,7 +11492,7 @@ void call_trace_sched_update_nr_running(struct rq *rq, int count)

#ifdef CONFIG_SCHED_MM_CID

-/**
+/*
* @cid_lock: Guarantee forward-progress of cid allocation.
*
* Concurrency ID allocation within a bitmap is mostly lock-free. The cid_lock
@@ -11501,7 +11501,7 @@ void call_trace_sched_update_nr_running(struct rq *rq, int count)
*/
DEFINE_RAW_SPINLOCK(cid_lock);

-/**
+/*
* @use_cid_lock: Select cid allocation behavior: lock-free vs spinlock.
*
* When @use_cid_lock is 0, the cid allocation is lock-free. When contention is