From: zhong jiang <zhongjiang@huawei.com>
To: Michal Hocko <mhocko@suse.cz>
Cc: akpm@linux-foundation.org, vbabka@suse.cz, rientjes@google.com,
qiuxishi@huawei.com, linux-mm@kvack.org
Subject: Re: [PATCH] mm/hugetlb: fix race when migrate pages.
Date: Tue, 19 Jul 2016 19:31:26 +0800 [thread overview]
Message-ID: <578E0F8E.4030002@huawei.com> (raw)
In-Reply-To: <20160719111003.GG9486@dhcp22.suse.cz>
On 2016/7/19 19:10, Michal Hocko wrote:
> On Tue 19-07-16 17:52:50, zhong jiang wrote:
>> On 2016/7/19 17:17, Michal Hocko wrote:
> [...]
>>> OK, so this states the problem. Although it would be helpful to be
>>> specific about which BUG has triggered because the above line doesn't
>>> match any in the current code. I assume this is
>>>
>>> BUG_ON(pte && !pte_none(*pte) && !pte_huge(*pte))
>>>
>>> in huge_pte_alloc. Now the changelog is silent about what the actual
>>> problem is and what is the fix. Could you add this information please?
>> Yes, it hit this BUG_ON() you had mentioned. The pmd share function enable, when
>> I run online-offline memory , That will lead to pte_huge() return false. beacuse
>> it refer to the pmd may be ongoing migration.
> OK, I see. But is the proposed fix correct? AFAIU you are retrying the
> VMA walk and nothing really prevents huge_pte_offset returning the same
> spte, right?
Yes, I guess it may be exist other pmd share entry , The entry may be used . therefore,
I decide to access the others.
>> Thanks
>> zhong jiang
>>>> Signed-off-by: zhong jiang <zhongjiang@huawei.com>
>>>> ---
>>>> mm/hugetlb.c | 9 ++++++++-
>>>> 1 file changed, 8 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
>>>> index 6384dfd..1b54d7a 100644
>>>> --- a/mm/hugetlb.c
>>>> +++ b/mm/hugetlb.c
>>>> @@ -4213,13 +4213,14 @@ pte_t *huge_pmd_share(struct mm_struct *mm, unsigned long addr, pud_t *pud)
>>>> struct vm_area_struct *svma;
>>>> unsigned long saddr;
>>>> pte_t *spte = NULL;
>>>> - pte_t *pte;
>>>> + pte_t *pte, entry;
>>>> spinlock_t *ptl;
>>>>
>>>> if (!vma_shareable(vma, addr))
>>>> return (pte_t *)pmd_alloc(mm, pud, addr);
>>>>
>>>> i_mmap_lock_write(mapping);
>>>> +retry:
>>>> vma_interval_tree_foreach(svma, &mapping->i_mmap, idx, idx) {
>>>> if (svma == vma)
>>>> continue;
>>>> @@ -4240,6 +4241,12 @@ pte_t *huge_pmd_share(struct mm_struct *mm, unsigned long addr, pud_t *pud)
>>>>
>>>> ptl = huge_pte_lockptr(hstate_vma(vma), mm, spte);
>>>> spin_lock(ptl);
>>>> + entry = huge_ptep_get(spte);
>>>> + if (is_hugetlb_entry_migration(entry) ||
>>>> + is_hugetlb_entry_hwpoisoned(entry)) {
>>>> + spin_unlock(ptl);
>>>> + goto retry;
>>>> + }
>>>> if (pud_none(*pud)) {
>>>> pud_populate(mm, pud,
>>>> (pmd_t *)((unsigned long)spte & PAGE_MASK));
>>>> --
>>>> 1.8.3.1
>>>>
>>>> --
>>>> To unsubscribe, send a message with 'unsubscribe linux-mm' in
>>>> the body to majordomo@kvack.org. For more info on Linux MM,
>>>> see: http://www.linux-mm.org/ .
>>>> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2016-07-19 11:37 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-07-19 2:59 [PATCH] mm/hugetlb: fix race when migrate pages zhongjiang
2016-07-19 9:17 ` Michal Hocko
2016-07-19 9:52 ` zhong jiang
2016-07-19 11:10 ` Michal Hocko
2016-07-19 11:31 ` zhong jiang [this message]
2016-07-19 11:43 ` zhong jiang
2016-07-19 11:59 ` Michal Hocko
2016-07-19 12:04 ` zhong jiang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=578E0F8E.4030002@huawei.com \
--to=zhongjiang@huawei.com \
--cc=akpm@linux-foundation.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.cz \
--cc=qiuxishi@huawei.com \
--cc=rientjes@google.com \
--cc=vbabka@suse.cz \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).