2013-03-11 21:20:09

by Raphael S Carvalho

Subject: [PATCH 1/1] kernel/pid.c: Improve flow of a loop inside alloc_pidmap.

From: Raphael S.Carvalho <[email protected]>

Notes: find_next_offset searches for an available "cleaned bit"
in the respective pid bitmap (page), so returns the offset if found,
otherwise it returns a value equals to BITS_PER_PAGE.

For example, suppose find_next_offset didn't find any available
bit, so there's no purpose to call mk_pid (Wasteful Cpu Cycles).

Therefore, I found it could be better to call mk_pid after
the checking (offset < BITS_PER_PAGE) returned sucessfully!
Another point: If (offset < BITS_PER_PAGE) results in a "failure",
then mk_pid would be called again afterwards.

Signed-off-by: Raphael S. Carvalho <[email protected]>
---
kernel/pid.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/pid.c b/kernel/pid.c
index 047dc62..7ecb09a 100644
--- a/kernel/pid.c
+++ b/kernel/pid.c
@@ -190,8 +190,8 @@ static int alloc_pidmap(struct pid_namespace *pid_ns)
return pid;
}
offset = find_next_offset(map, offset);
- pid = mk_pid(pid_ns, map, offset);
- } while (offset < BITS_PER_PAGE && pid < pid_max);
+ } while (offset < BITS_PER_PAGE &&
+ (pid = mk_pid(pid_ns, map, offset)) < pid_max);
}
if (map < &pid_ns->pidmap[(pid_max-1)/BITS_PER_PAGE]) {
++map;
--
1.7.2.5

2013-03-11 21:57:37

by Andrew Morton

Subject: Re: [PATCH 1/1] kernel/pid.c: Improve flow of a loop inside alloc_pidmap.

On Mon, 11 Mar 2013 18:18:56 -0300 "Raphael S. Carvalho" <[email protected]> wrote:

> From: Raphael S.Carvalho <[email protected]>
>
> Notes: find_next_offset searches for an available "cleaned bit"
> in the respective pid bitmap (page), so returns the offset if found,
> otherwise it returns a value equals to BITS_PER_PAGE.
>
> For example, suppose find_next_offset didn't find any available
> bit, so there's no purpose to call mk_pid (Wasteful Cpu Cycles).
>
> Therefore, I found it could be better to call mk_pid after
> the checking (offset < BITS_PER_PAGE) returned sucessfully!
> Another point: If (offset < BITS_PER_PAGE) results in a "failure",
> then mk_pid would be called again afterwards.
>
> ...
>
> --- a/kernel/pid.c
> +++ b/kernel/pid.c
> @@ -190,8 +190,8 @@ static int alloc_pidmap(struct pid_namespace *pid_ns)
> return pid;
> }
> offset = find_next_offset(map, offset);
> - pid = mk_pid(pid_ns, map, offset);
> - } while (offset < BITS_PER_PAGE && pid < pid_max);
> + } while (offset < BITS_PER_PAGE &&
> + (pid = mk_pid(pid_ns, map, offset)) < pid_max);
> }
> if (map < &pid_ns->pidmap[(pid_max-1)/BITS_PER_PAGE]) {
> ++map;

Looks OK. But I think it's simpler and more straightforward to do it
this way?

for ( ; ; ) {
if (!test_and_set_bit(offset, map->page)) {
atomic_dec(&map->nr_free);
set_last_pid(pid_ns, last, pid);
return pid;
}
offset = find_next_offset(map, offset);
if (offset >= BITS_PER_PAGE)
break;
pid = mk_pid(pid_ns, map, offset);
if (pid >= pid_max)
break;
}

2013-03-11 22:01:28

by Raphael S Carvalho

Subject: Re: [PATCH 1/1] kernel/pid.c: Improve flow of a loop inside alloc_pidmap.

On Mon, Mar 11, 2013 at 6:57 PM, Andrew Morton
<[email protected]> wrote:
> On Mon, 11 Mar 2013 18:18:56 -0300 "Raphael S. Carvalho" <[email protected]> wrote:
>
>> From: Raphael S.Carvalho <[email protected]>
>>
>> Notes: find_next_offset searches for an available "cleaned bit"
>> in the respective pid bitmap (page), so returns the offset if found,
>> otherwise it returns a value equals to BITS_PER_PAGE.
>>
>> For example, suppose find_next_offset didn't find any available
>> bit, so there's no purpose to call mk_pid (Wasteful Cpu Cycles).
>>
>> Therefore, I found it could be better to call mk_pid after
>> the checking (offset < BITS_PER_PAGE) returned sucessfully!
>> Another point: If (offset < BITS_PER_PAGE) results in a "failure",
>> then mk_pid would be called again afterwards.
>>
>> ...
>>
>> --- a/kernel/pid.c
>> +++ b/kernel/pid.c
>> @@ -190,8 +190,8 @@ static int alloc_pidmap(struct pid_namespace *pid_ns)
>> return pid;
>> }
>> offset = find_next_offset(map, offset);
>> - pid = mk_pid(pid_ns, map, offset);
>> - } while (offset < BITS_PER_PAGE && pid < pid_max);
>> + } while (offset < BITS_PER_PAGE &&
>> + (pid = mk_pid(pid_ns, map, offset)) < pid_max);
>> }
>> if (map < &pid_ns->pidmap[(pid_max-1)/BITS_PER_PAGE]) {
>> ++map;
>
> Looks OK. But I think it's simpler and more straightforward to do it
> this way?

Yes, it looks much better.
>
> for ( ; ; ) {
> if (!test_and_set_bit(offset, map->page)) {
> atomic_dec(&map->nr_free);
> set_last_pid(pid_ns, last, pid);
> return pid;
> }
> offset = find_next_offset(map, offset);
> if (offset >= BITS_PER_PAGE)
> break;
> pid = mk_pid(pid_ns, map, offset);
> if (pid >= pid_max)
> break;
> }
>