Re: Race in mm/ksm.c

[Gabriel Ryan <gabe@cs.columbia.edu>]

[-- Attachment #1: Type: text/plain, Size: 5011 bytes --]

Yes, we are unable to trigger the reported race after applying your patch,
and by inspection the patch appears to have fixed the issue by extending
ksm_thread_mutex to cover the ksm_run flag check in __ksm_enter.

We have also contacted the kernel security team since the race may have
security implications.

Thanks,

Gabe

On Tue, Aug 2, 2022 at 8:18 AM Kefeng Wang <wangkefeng.wang@huawei.com>
wrote:

>
> On 2022/8/2 19:44, Gabriel Ryan wrote:
>
> Confirmed.
>
> It means that the following changes could fix the issue, right?
>
> If so, I will send a formal patch with your report and test, thanks.
>
>
> Thanks for the quick response and apologies for the delay!
>
> Best,
>
> Gabe
>
> On Thu, Jul 21, 2022 at 9:57 PM Kefeng Wang <wangkefeng.wang@huawei.com>
> wrote:
>
>>
>> On 2022/7/21 23:58, Abhishek Shah wrote:
>>
>> Dear Kernel Maintainers,
>>
>> We found a race in mm/ksm.c. During the execution of the function
>> *__ksm_run* which uses variable *ksm_run* to decide the list insertion
>> point, the variable *ksm_run* can be concurrently modified in the
>> function *run_store*, which we thought could be undesirable since “KSM
>> pages in newly forked mms can be missed” (See comment here:
>> https://elixir.bootlin.com/linux/v5.18-rc5/source/mm/ksm.c#L2498). We
>> would also like your thoughts on the security impact given it is a TOCTOU
>> bug.
>>
>> We provide more details below including the trace and reproducing
>> test cases.
>>
>>
>> Hello， could the following changes to avoid the data-race issue？
>>
>> diff --git a/mm/ksm.c b/mm/ksm.c
>> index 54f78c9eecae..f072753cbb3a 100644
>> --- a/mm/ksm.c
>> +++ b/mm/ksm.c
>> @@ -2497,6 +2497,7 @@ int __ksm_enter(struct mm_struct *mm)
>>  {
>>         struct mm_slot *mm_slot;
>>         int needs_wakeup;
>> +       bool ksm_run_merge;
>>
>>         mm_slot = alloc_mm_slot();
>>         if (!mm_slot)
>> @@ -2505,6 +2506,10 @@ int __ksm_enter(struct mm_struct *mm)
>>         /* Check ksm_run too?  Would need tighter locking */
>>         needs_wakeup = list_empty(&ksm_mm_head.mm_list);
>>
>> +       mutex_lock(&ksm_thread_mutex);
>> +       ksm_run_unmerge = !!(ksm_run & KSM_RUN_UNMERGE);
>> +       mutex_unlock(&ksm_thread_mutex);
>> +
>>         spin_lock(&ksm_mmlist_lock);
>>         insert_to_mm_slots_hash(mm, mm_slot);
>>         /*
>> @@ -2517,7 +2522,7 @@ int __ksm_enter(struct mm_struct *mm)
>>          * scanning cursor, otherwise KSM pages in newly forked mms will
>> be
>>          * missed: then we might as well insert at the end of the list.
>>          */
>> -       if (ksm_run & KSM_RUN_UNMERGE)
>> +       if (ksm_run_unmerge)
>>                 list_add_tail(&mm_slot->mm_list, &ksm_mm_head.mm_list);
>>         else
>>                 list_add_tail(&mm_slot->mm_list,
>> &ksm_scan.mm_slot->mm_list);
>>
>>
>>
>>
>> *Trace*
>> BUG: KCSAN: data-race in __ksm_enter / run_store
>> write to 0xffffffff881edae0 of 8 bytes by task 6542 on cpu 0:
>>  run_store+0x19a/0x2d0 mm/ksm.c:2897
>>  kobj_attr_store+0x44/0x60 lib/kobject.c:824
>>  sysfs_kf_write+0x16f/0x1a0 fs/sysfs/file.c:136
>>  kernfs_fop_write_iter+0x2ae/0x370 fs/kernfs/file.c:291
>>  call_write_iter include/linux/fs.h:2050 [inline]
>>  new_sync_write fs/read_write.c:504 [inline]
>>  vfs_write+0x779/0x900 fs/read_write.c:591
>>  ksys_write+0xde/0x190 fs/read_write.c:644
>>  __do_sys_write fs/read_write.c:656 [inline]
>>  __se_sys_write fs/read_write.c:653 [inline]
>>  __x64_sys_write+0x43/0x50 fs/read_write.c:653
>>  do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>>  do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80
>>  entry_SYSCALL_64_after_hwframe+0x44/0xae
>>
>> read to 0xffffffff881edae0 of 8 bytes by task 6541 on cpu 1:
>>  __ksm_enter+0x114/0x260 mm/ksm.c:2501
>>  ksm_madvise+0x291/0x350 mm/ksm.c:2451
>>  madvise_vma_behavior mm/madvise.c:1039 [inline]
>>  madvise_walk_vmas mm/madvise.c:1221 [inline]
>>  do_madvise+0x656/0xeb0 mm/madvise.c:1399
>>  __do_sys_madvise mm/madvise.c:1412 [inline]
>>  __se_sys_madvise mm/madvise.c:1410 [inline]
>>  __x64_sys_madvise+0x64/0x70 mm/madvise.c:1410
>>  do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>>  do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80
>>  entry_SYSCALL_64_after_hwframe+0x44/0xae
>>
>> Reported by Kernel Concurrency Sanitizer on:
>> CPU: 1 PID: 6541 Comm: syz-executor2-n Not tainted 5.18.0-rc5+ #107
>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1
>> 04/01/2014
>> ---------------------
>> *Inputs *
>> Input CPU 0:
>> r0 = openat$sysctl(0xffffff9c,
>> &(0x7f0000000100)='/sys/kernel/mm/ksm/run\x00', 0x1, 0x0)
>> write$sysctl(r0, &(0x7f0000000000)='2\x00', 0x2)
>>
>> Input CPU 1:
>> madvise(&(0x7f0000ffc000/0x4000)=nil, 0x4000, 0xc)
>> mlock2(&(0x7f0000ffe000/0x2000)=nil, 0x2000, 0x0)
>> madvise(&(0x7f0000ffd000/0x3000)=nil, 0x3000, 0x12)
>> clone(0x0, 0x0, 0x0, 0x0, 0x0)
>>
>>

[-- Attachment #2: Type: text/html, Size: 9359 bytes --]