Hello guys, in 'update_domain_attr' function we can just return if
couldn't update domain attribute , rather then checking dattr and
again checking relax_domain_level of dattr , if not then return.
Signed-off-by: Md.Rakib H. Mullick ([email protected])
--- linux-2.6.26-rc9.orig/kernel/cpuset.c 2008-08-01 08:38:41.000000000 +0600
+++ linux-2.6.26-rc9/kernel/cpuset.c 2008-08-02 20:37:06.000000000 +0600
@@ -490,8 +490,6 @@ static int cpusets_overlap(struct cpuset
static void
update_domain_attr(struct sched_domain_attr *dattr, struct cpuset *c)
{
- if (!dattr)
- return;
if (dattr->relax_domain_level < c->relax_domain_level)
dattr->relax_domain_level = c->relax_domain_level;
return;
Rakib Mullick wrote:
> Hello guys, in 'update_domain_attr' function we can just return if
> couldn't update domain attribute , rather then checking dattr and
> again checking relax_domain_level of dattr , if not then return.
>
> Signed-off-by: Md.Rakib H. Mullick ([email protected])
>
> --- linux-2.6.26-rc9.orig/kernel/cpuset.c 2008-08-01 08:38:41.000000000 +0600
> +++ linux-2.6.26-rc9/kernel/cpuset.c 2008-08-02 20:37:06.000000000 +0600
> @@ -490,8 +490,6 @@ static int cpusets_overlap(struct cpuset
> static void
> update_domain_attr(struct sched_domain_attr *dattr, struct cpuset *c)
> {
> - if (!dattr)
> - return;
The above code no longer exists in latest-git.
> if (dattr->relax_domain_level < c->relax_domain_level)
> dattr->relax_domain_level = c->relax_domain_level;
> return;
> --
This PATCH has the exact same name 'Removes extra checking kernel/cpuset.c'
as another, different PATCH:
1) sent 7/31: kernel/cpuset.c guarantee_online_mems() and guarantee_online_cpus()
2) this one, sent 8/2: kernel/cpuset.c update_domain_attr().
Please use different names for different patches.
And this one, for update_domain_attr(), probably conflicts with other
work being done by Max K (I'll add him to the CC list, above) which relies
on this dattr check here, as commented in a recent patch of his:
cpuset: Rework sched domains and CPU hotplug handling (take 4)
-OR- it conflicts with the code version that Li Zefan was referring
to in his comment:
> The above code no longer exists in latest-git.
I'm not sure which conflict is still relevant here.
How about you wait for the dust to settle on Max K's above patch,
and then consider if this change is still useful? I suspect it
won't be useful.
--
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <[email protected]> 1.940.382.4214