lkml.org 
[lkml]   [2015]   [Dec]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] sched/core: Clear the root_domain cpumasks in init_rootdomain()

* Xunlei Pang <xlpang@redhat.com> wrote:

> Hi Ingo,
>
> On 12/04/2015 at 04:09 PM, Ingo Molnar wrote:
> > * Xunlei Pang <xlpang@redhat.com> wrote:
> >
> >>> Hm, is the alloc_cpumask_var() done in alloc_sched_domains() safe?
> >> Until now, I haven't found any other similar issues, but I will check further.
> >>
> >>> At least the usage pattern in init_sched_domains() looks unsafe:
> >>>
> >>> doms_cur = alloc_sched_domains(ndoms_cur);
> >>> if (!doms_cur)
> >>> doms_cur = &fallback_doms;
> >>> cpumask_andnot(doms_cur[0], cpu_map, cpu_isolated_map);
> > So is this pattern in init_sched_domains() correct, for OFFSTACK=y?
> >
> > It looks wrong to me, as alloc_sched_domains() allocates an uninitialized cpumask
> > via alloc_cpumask_var() and returns it:
> >
> > cpumask_var_t *alloc_sched_domains(unsigned int ndoms)
> > {
> > int i;
> > cpumask_var_t *doms;
> >
> > doms = kmalloc(sizeof(*doms) * ndoms, GFP_KERNEL);
> > if (!doms)
> > return NULL;
> > for (i = 0; i < ndoms; i++) {
> > if (!alloc_cpumask_var(&doms[i], GFP_KERNEL)) {
> > free_sched_domains(doms, i);
> > return NULL;
> > }
> > }
> > return doms;
> > }
> >
> > and then this code:
> >
> >>> cpumask_andnot(doms_cur[0], cpu_map, cpu_isolated_map);
> > uses it without first clearing it.
> >
> > So is this another such bug, or am I missing something?
>
> Yeah, I noticed that as well. But fortunately cpumask_andnot(),
> cpumask_and() and the like will clear doms_cur[] indirectly, also
> cpu_isolated_map, cpu_active_mask, etc doesn't contain any
> garbage bits. I also checked the use of it by cpuset, no extra such
> bug found by me so far.

Great, thanks for double checking!

Ingo


\
 
 \ /
  Last update: 2015-12-04 10:01    [W:0.245 / U:0.020 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site