lkml.org 
[lkml]   [2013]   [Mar]   [14]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH v2] mm/hugetlb: fix total hugetlbfs pages count when memory overcommit accouting
On Thu 14-03-13 19:24:11, Wanpeng Li wrote:
> On Thu, Mar 14, 2013 at 12:09:27PM +0100, Michal Hocko wrote:
> >On Thu 14-03-13 18:49:49, Wanpeng Li wrote:
> >> Changelog:
> >> v1 -> v2:
> >> * update patch description, spotted by Michal
> >>
> >> hugetlb_total_pages() does not account for all the supported hugepage
> >> sizes.
> >
> >> This can lead to incorrect calculation of the total number of
> >> page frames used by hugetlb. This patch corrects the issue.
> >
>
> Hi Michal,
>
> >Sorry to be so picky but this doesn't tell us much. Why do we need to
> >have the total number of hugetlb pages?
> >
> >What about the following:
> >"hugetlb_total_pages is used for overcommit calculations but the
> >current implementation considers only default hugetlb page size (which
> >is either the first defined hugepage size or the one specified by
> >default_hugepagesz kernel boot parameter).
> >
> >If the system is configured for more than one hugepage size (which is
> >possible since a137e1cc hugetlbfs: per mount huge page sizes) then
> >the overcommit estimation done by __vm_enough_memory (resp. shown by
> >meminfo_proc_show) is not precise - there is an impression of more
> >available/allowed memory. This can lead to an unexpected ENOMEM/EFAULT
> >resp. SIGSEGV when memory is accounted."
> >
>
> Fair enough, thanks. :-)
>
> >I think this is also worth pushing to the stable tree (it goes back to
> >2.6.27)
> >
>
> Yup, I will Cc Greg in next version.

Ccing Greg doesn't help. All that is required is:
Cc: stable@vger.kernel.org # 2.6.27+

> >> Testcase:
> >> boot: hugepagesz=1G hugepages=1
> >> before patch:
> >> egrep 'CommitLimit' /proc/meminfo
> >> CommitLimit: 55434168 kB
> >> after patch:
> >> egrep 'CommitLimit' /proc/meminfo
> >> CommitLimit: 54909880 kB
> >
> >This gives some more confusion to a reader because there is only
> >something like 500M difference here without any explanation.
> >
>
> the default overcommit ratio is 50.

And that part was missing in the description...
[...]
--
Michal Hocko
SUSE Labs


\
 
 \ /
  Last update: 2013-03-14 14:21    [W:0.297 / U:0.284 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site