lkml.org 
[lkml]   [2012]   [Jun]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    From
    SubjectRe: [PATCH -V9 04/15] hugetlb: use mmu_gather instead of a temporary linked list for accumulating pages
    Date
    Michal Hocko <mhocko@suse.cz> writes:

    > On Wed 13-06-12 16:59:23, Michal Hocko wrote:
    >> On Wed 13-06-12 15:57:23, Aneesh Kumar K.V wrote:
    >> > From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
    >> >
    >> > Use a mmu_gather instead of a temporary linked list for accumulating
    >> > pages when we unmap a hugepage range
    >>
    >> Sorry for coming up with the comment that late but you owe us an
    >> explanation _why_ you are doing this.
    >>
    >> I assume that this fixes a real problem when we take i_mmap_mutex
    >> already up in
    >> unmap_mapping_range
    >> mutex_lock(&mapping->i_mmap_mutex);
    >> unmap_mapping_range_tree | unmap_mapping_range_list
    >> unmap_mapping_range_vma
    >> zap_page_range_single
    >> unmap_single_vma
    >> unmap_hugepage_range
    >> mutex_lock(&vma->vm_file->f_mapping->i_mmap_mutex);
    >>
    >> And that this should have been marked for stable as well (I haven't
    >> checked when this has been introduced).
    >>
    >> But then I do not see how this help when you still do this:
    >> [...]
    >> > diff --git a/mm/memory.c b/mm/memory.c
    >> > index 1b7dc66..545e18a 100644
    >> > --- a/mm/memory.c
    >> > +++ b/mm/memory.c
    >> > @@ -1326,8 +1326,11 @@ static void unmap_single_vma(struct mmu_gather *tlb,
    >> > * Since no pte has actually been setup, it is
    >> > * safe to do nothing in this case.
    >> > */
    >> > - if (vma->vm_file)
    >> > - unmap_hugepage_range(vma, start, end, NULL);
    >> > + if (vma->vm_file) {
    >> > + mutex_lock(&vma->vm_file->f_mapping->i_mmap_mutex);
    >> > + __unmap_hugepage_range(tlb, vma, start, end, NULL);
    >> > + mutex_unlock(&vma->vm_file->f_mapping->i_mmap_mutex);
    >> > + }
    >> > } else
    >> > unmap_page_range(tlb, vma, start, end, details);
    >> > }
    >
    > Ahhh, you are removing the lock in the next patch. Really confusing and
    > not nice for the stable backport.
    > Could you merge those two patches and add Cc: stable?
    > Then you can add my
    > Reviewed-by: Michal Hocko <mhocko@suse.cz>
    >

    In the last review cycle I was asked to see if we can get a lockdep
    report for the above and what I found was we don't really cause the
    above deadlock with the current codebase because for hugetlb we don't
    directly call unmap_mapping_range. But still it is good to remove the
    i_mmap_mutex, because we don't need that protection now. I didn't
    mark it for stable because of the above reason.

    -aneesh



    \
     
     \ /
      Last update: 2012-06-13 19:21    [W:3.270 / U:0.040 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site