Re: [PATCH v12 2/5] fs/proc/task_mmu: Implement IOCTL to get and optionally clear info about PTEs

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, 7 Apr 2023 at 11:35, Muhammad Usama Anjum
<usama.anjum@xxxxxxxxxxxxx> wrote:
> On 4/7/23 12:23 PM, Michał Mirosław wrote:
> > On Thu, 6 Apr 2023 at 23:12, Muhammad Usama Anjum
> > <usama.anjum@xxxxxxxxxxxxx> wrote:
> >> On 4/7/23 1:12 AM, Michał Mirosław wrote:
> >>> On Thu, 6 Apr 2023 at 09:40, Muhammad Usama Anjum
> >>> <usama.anjum@xxxxxxxxxxxxx> wrote:
> >>> [...]
> >>>> --- a/fs/proc/task_mmu.c
> >>>> +++ b/fs/proc/task_mmu.c
> >>> [...]
> >>>> +static int pagemap_scan_pmd_entry(pmd_t *pmd, unsigned long start,
> >>>> +                                 unsigned long end, struct mm_walk *walk)
> >>>> +{
> > [...]
> >>>> +#ifdef CONFIG_TRANSPARENT_HUGEPAGE
> >>>> +       ptl = pmd_trans_huge_lock(pmd, vma);
> >>>> +       if (ptl) {
> >>> [...]
> >>>> +               return ret;
> >>>> +       }
> >>>> +process_smaller_pages:
> >>>> +       if (pmd_trans_unstable(pmd))
> >>>> +               return 0;
> >>>
> >>> Why pmd_trans_unstable() is needed here and not only after split_huge_pmd()?
> >> I'm not entirely sure. But the idea is if THP is unstable, we should
> >> return. As it doesn't seem like after splitting THP can be unstable, we
> >> should not check it. Do you agree with the following?
> >
> > The description of pmd_trans_unstable() [1] seems to indicate that it
> > is needed only after split_huge_pmd().
> >
> > [1] https://elixir.bootlin.com/linux/v6.3-rc5/source/include/linux/pgtable.h#L1394
> Sorry, yeah pmd_trans_unstable() is need after split. But it is also needed
> in normal case when ptl is NULL to rule out the case if pmd is unstable
> before performing operation on normal pages:
>
> ptl = pmd_trans_huge_lock(pmd, vma);
> if (ptl) {
> ...
> }
> if (pmd_trans_unstable(pmd))
>         return 0;
>
> This file has usage examples of pmd_trans_unstable():
>
> https://elixir.bootlin.com/linux/v6.3-rc5/source/fs/proc/task_mmu.c#L634
> https://elixir.bootlin.com/linux/v6.3-rc5/source/fs/proc/task_mmu.c#L1195
> https://elixir.bootlin.com/linux/v6.3-rc5/source/fs/proc/task_mmu.c#L1543
> https://elixir.bootlin.com/linux/v6.3-rc5/source/fs/proc/task_mmu.c#L1887
>
> So we are good with what we have in this patch.

Shouldn't we signal ACTION_AGAIN then in order to call .pte_hole?

Best Regards
Michał Mirosław




[Index of Archives]     [Linux Ext4 Filesystem]     [Union Filesystem]     [Filesystem Testing]     [Ceph Users]     [Ecryptfs]     [NTFS 3]     [AutoFS]     [Kernel Newbies]     [Share Photos]     [Security]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux Cachefs]     [Reiser Filesystem]     [Linux RAID]     [NTFS 3]     [Samba]     [Device Mapper]     [CEPH Development]

  Powered by Linux