From: Rongwei Wang <rongwei.wang@linux.alibaba.com>
To: Song Liu <song@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Matthew Wilcox <willy@infradead.org>,
Linux MM <linux-mm@kvack.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
William Kucharski <william.kucharski@oracle.com>,
Hugh Dickins <hughd@google.com>
Subject: Re: [PATCH v2 1/2] mm, thp: check page mapping when truncating page cache
Date: Wed, 29 Sep 2021 15:50:39 +0800 [thread overview]
Message-ID: <dde441c4-febe-cfa1-7729-b405fa331a4e@linux.alibaba.com> (raw)
In-Reply-To: <CAPhsuW4x2UzMLwZyioWH4dXqrYwNT-XKgzvrm+6YeWk9EgQmCQ@mail.gmail.com>
On 9/29/21 3:14 PM, Song Liu wrote:
> On Tue, Sep 28, 2021 at 9:20 AM Rongwei Wang
> <rongwei.wang@linux.alibaba.com> wrote:
>>
>>
>>
>> On 9/28/21 6:24 AM, Song Liu wrote:
>>> On Fri, Sep 24, 2021 at 12:12 AM Rongwei Wang
>>> <rongwei.wang@linux.alibaba.com> wrote:
>>>>
>>>>
>>>>
>>>> On 9/24/21 10:43 AM, Andrew Morton wrote:
>>>>> On Thu, 23 Sep 2021 01:04:54 +0800 Rongwei Wang <rongwei.wang@linux.alibaba.com> wrote:
>>>>>
>>>>>>
>>>>>>
>>>>>>> On Sep 22, 2021, at 7:37 PM, Matthew Wilcox <willy@infradead.org> wrote:
>>>>>>>
>>>>>>> On Wed, Sep 22, 2021 at 03:06:44PM +0800, Rongwei Wang wrote:
>>>>>>>> Transparent huge page has supported read-only non-shmem files. The file-
>>>>>>>> backed THP is collapsed by khugepaged and truncated when written (for
>>>>>>>> shared libraries).
>>>>>>>>
>>>>>>>> However, there is race in two possible places.
>>>>>>>>
>>>>>>>> 1) multiple writers truncate the same page cache concurrently;
>>>>>>>> 2) collapse_file rolls back when writer truncates the page cache;
>>>>>>>
>>>>>>> As I've said before, the bug here is that somehow there is a writable fd
>>>>>>> to a file with THPs. That's what we need to track down and fix.
>>>>>> Hi, Matthew
>>>>>> I am not sure get your means. We know “mm, thp: relax the VM_DENYWRITE constraint on file-backed THPs"
>>>>>> Introduced file-backed THPs for DSO. It is possible {very rarely} for DSO to be opened in writeable way.
>>>>>>
>>>>>> ...
>>>>>>
>>>>>>> https://lore.kernel.org/linux-mm/YUdL3lFLFHzC80Wt@casper.infradead.org/
>>>>>> All in all, what you mean is that we should solve this race at the source?
>>>>>
>>>>> Matthew is being pretty clear here: we shouldn't be permitting
>>>>> userspace to get a writeable fd for a thp-backed file.
>>>>>
>>>>> Why are we permitting the DSO to be opened writeably? If there's a
>>>>> legitimate case for doing this then presumably "mm, thp: relax the
>>>> There is a use case to stress file-backed THP within attachment.
>>>> I test this case in a system which has enabled CONFIG_READ_ONLY_THP_FOR_FS:
>>>>
>>>> $ gcc -Wall -g -o stress_madvise_dso stress_madvise_dso.c
>>>> $ ulimit -s unlimited
>>>> $ ./stress_madvise_dso 10000 <libtest.so>
>>>>
>>>> the meaning of above parameters:
>>>> 10000: the max test time;
>>>> <libtest.so>: the DSO that will been mapped into file-backed THP by
>>>> madvise. It recommended that the text segment of DSO to be tested is
>>>> greater than 2M.
>>>>
>>>> The crash will been triggered at once in the latest kernel. And this
>>>> case also can used to trigger the bug that mentioned in our another patch.
>>>
>>> Hmm.. I am not able to use the repro program to crash the system. Not
>>> sure what I did wrong.
>>>
>> Hi
>> I have tried to check my test case again. Can you make sure the DSO that
>> you test have THP mapping?
>>
>> If you are willing to try again, I can send my libtest.c which is used
>> to test by myself (actually, it shouldn't be target DSO problem).
>>
>> Thanks very much!
>>> OTOH, does it make sense to block writes within khugepaged, like:
>>>
>>> diff --git i/mm/khugepaged.c w/mm/khugepaged.c
>>> index 045cc579f724e..ad7c41ec15027 100644
>>> --- i/mm/khugepaged.c
>>> +++ w/mm/khugepaged.c
>>> @@ -51,6 +51,7 @@ enum scan_result {
>>> SCAN_CGROUP_CHARGE_FAIL,
>>> SCAN_TRUNCATED,
>>> SCAN_PAGE_HAS_PRIVATE,
>>> + SCAN_BUSY_WRITE,
>>> };
>>>
>>> #define CREATE_TRACE_POINTS
>>> @@ -1652,6 +1653,11 @@ static void collapse_file(struct mm_struct *mm,
>>> /* Only allocate from the target node */
>>> gfp = alloc_hugepage_khugepaged_gfpmask() | __GFP_THISNODE;
>>>
>>> + if (deny_write_access(file)) {
>>> + result = SCAN_BUSY_WRITE;
>>> + return;
>>> + }
>>> +
>> This can indeed avoid some possible races from source.
>>
>> But, I am thinking about whether this will lead to DDoS attack?
>> I remember the reason of DSO has ignored MAP_DENYWRITE in kernel
>> is that DDoS attack. In addition, 'deny_write_access' will change
>> the behavior, such as user will get 'Text file busy' during
>> collapse_file. I am not sure whether the behavior changing is acceptable
>> in user space.
>>
>> If it is acceptable, I am very willing to fix the races like your way.
>
> I guess we should not let the write get ETXTBUSY for khugepaged work.
>
> I am getting some segfault on stress_madvise_dso. And it doesn't really
> generate the bug stack in my vm (qemu-system-x86_64). Is there an newer
Hi, I can sure I am not update the stress_madvise_dso.c.
My test environment is vm (qemu-system-aarch64, 32 cores). And I can
think of the following possibilities:
(1) in thread_read()
printf("read %s\n", dso_path);
d = open(dso_path, O_RDONLY);
/* The start addr must be alignment with 2M */
void *p = mmap((void *)0x40000dc00000UL, 0x800000, PROT_READ |
PROT_EXEC,MAP_PRIVATE, fd, 0);
if (p == MAP_FAILED) {
perror("mmap");
goto out;
}
0x40000dc00000 is random setting by myself. I am not sure this address
is available in your vm.
(2) in thread_write()
int fd = open(dso_path, O_RDWR);
p = mmap(NULL, 0x800000, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0);
if (p == MAP_FAILED) {
perror("mmap");
goto out; /* fail */
}
because of I am sure the DSO is bigger than 0x800000, so directly map
the DSO using 0x800000. Maybe I had use '-z max-page-size=0x200000' to
compile the DSO? likes:
$ gcc -z max-page-size=0x200000 -o libtest.so -shared libtest.o
If you don't mind, you can send the segment fault log to me. And I will
find x86 environment to test.
Thanks!
> version of it?
>
> Thanks,
> Song
>
next prev parent reply other threads:[~2021-09-29 7:50 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-06 12:11 [PATCH 0/2] mm, thp: fix file-backed THP race in collapse_file Rongwei Wang
2021-09-06 12:11 ` [PATCH 1/2] mm, thp: check page mapping when truncating page cache Rongwei Wang
2021-09-07 2:49 ` Yu Xu
2021-09-07 18:08 ` Yang Shi
[not found] ` <38AF4DC8-5E6F-4568-B2E3-0434BD847BC9@linux.alibaba.com>
2021-09-08 21:48 ` Yang Shi
2021-09-13 14:49 ` [mm, thp] 20753096b6: BUG:unable_to_handle_page_fault_for_address kernel test robot
2021-09-06 12:12 ` [PATCH 2/2] mm, thp: bail out early in collapse_file for writeback page Rongwei Wang
2021-09-07 16:56 ` Yang Shi
[not found] ` <44BE85B4-692C-41E8-B5A0-C1E0B0272ACD@linux.alibaba.com>
2021-09-08 21:51 ` Yang Shi
2021-09-22 7:06 ` [PATCH v2 0/2] mm, thp: fix file-backed THP race in collapse_file and truncate pagecache Rongwei Wang
2021-09-22 7:06 ` [PATCH v2 1/2] mm, thp: check page mapping when truncating page cache Rongwei Wang
2021-09-22 11:37 ` Matthew Wilcox
2021-09-22 17:04 ` Rongwei Wang
2021-09-24 2:43 ` Andrew Morton
2021-09-24 3:08 ` Yang Shi
2021-09-24 3:35 ` Rongwei Wang
2021-09-24 7:12 ` Rongwei Wang
2021-09-27 22:24 ` Song Liu
2021-09-28 12:06 ` Matthew Wilcox
2021-09-28 16:59 ` Song Liu
2021-09-28 16:20 ` Rongwei Wang
2021-09-29 7:14 ` Song Liu
2021-09-29 7:50 ` Rongwei Wang [this message]
2021-09-29 16:59 ` Song Liu
2021-09-29 17:55 ` Matthew Wilcox
2021-09-29 23:41 ` Song Liu
2021-09-30 0:00 ` Matthew Wilcox
2021-09-30 0:41 ` Song Liu
2021-09-30 2:14 ` Rongwei Wang
2021-10-04 17:26 ` Rongwei Wang
2021-10-04 19:05 ` Matthew Wilcox
2021-10-05 1:58 ` Rongwei Wang
2021-10-04 20:26 ` Song Liu
2021-10-05 2:58 ` Hugh Dickins
2021-10-05 3:07 ` Matthew Wilcox
2021-10-05 9:03 ` Rongwei Wang
2021-09-30 1:54 ` Rongwei Wang
2021-09-30 3:26 ` Song Liu
2021-09-30 5:24 ` Hugh Dickins
2021-09-30 15:28 ` Matthew Wilcox
2021-09-30 16:49 ` Hugh Dickins
2021-09-30 17:39 ` Yang Shi
2021-10-02 17:08 ` Matthew Wilcox
2021-10-04 18:28 ` Yang Shi
2021-10-04 19:31 ` Matthew Wilcox
2021-10-05 2:26 ` Hugh Dickins
2021-10-02 2:22 ` Rongwei Wang
2021-09-22 7:06 ` [PATCH v2 2/2] mm, thp: bail out early in collapse_file for writeback page Rongwei Wang
2021-10-06 2:18 ` [PATCH v3 v3 0/2] mm, thp: fix file-backed THP race in collapse_file and truncate pagecache Rongwei Wang
2021-10-06 2:18 ` [PATCH v3 v3 1/2] mm, thp: lock filemap when truncating page cache Rongwei Wang
2021-10-06 2:18 ` [PATCH v3 v3 2/2] mm, thp: bail out early in collapse_file for writeback page Rongwei Wang
2021-10-06 2:41 ` Matthew Wilcox
2021-10-06 8:39 ` Rongwei Wang
2021-10-06 17:58 ` Yang Shi
2021-10-11 2:22 ` [PATCH v4 0/2] mm, thp: fix file-backed THP race in collapse_file and truncate pagecache Rongwei Wang
2021-10-11 2:22 ` [PATCH v4 1/2] mm, thp: lock filemap when truncating page cache Rongwei Wang
2021-10-13 7:55 ` Rongwei Wang
2021-10-11 2:22 ` [PATCH v4 2/2] mm, thp: bail out early in collapse_file for writeback page Rongwei Wang
2021-10-11 3:08 ` Matthew Wilcox
2021-10-11 3:22 ` Rongwei Wang
2021-10-11 5:08 ` [PATCH v4 RESEND " Rongwei Wang
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=dde441c4-febe-cfa1-7729-b405fa331a4e@linux.alibaba.com \
--to=rongwei.wang@linux.alibaba.com \
--cc=akpm@linux-foundation.org \
--cc=hughd@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=song@kernel.org \
--cc=william.kucharski@oracle.com \
--cc=willy@infradead.org \
/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).