| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <1df840eb-6d04-4023-8c48-ffdda45931d0@igalia.com>
Date: Tue, 27 May 2025 17:59:40 +0800
From: Gavin Guo <gavinguo@...lia.com>
To: Gavin Shan <gshan@...hat.com>, linux-mm@...ck.org
Cc: linux-kernel@...r.kernel.org, muchun.song@...ux.dev, osalvador@...e.de,
akpm@...ux-foundation.org, mike.kravetz@...cle.com, kernel-dev@...lia.com,
stable@...r.kernel.org, Hugh Dickins <hughd@...gle.com>,
Florent Revest <revest@...gle.com>
Subject: Re: [PATCH] mm/hugetlb: fix a deadlock with pagecache_folio and
hugetlb_fault_mutex_table
Hi Gavin,
On 5/26/25 12:41, Gavin Shan wrote:
> Hi Gavin,
>
> On 5/13/25 7:34 PM, Gavin Guo wrote:
>> The patch fixes a deadlock which can be triggered by an internal
>> syzkaller [1] reproducer and captured by bpftrace script [2] and its log
>> [3] in this scenario:
>>
>> Process 1 Process 2
>> --- ---
>> hugetlb_fault
>> mutex_lock(B) // take B
>> filemap_lock_hugetlb_folio
>> filemap_lock_folio
>> __filemap_get_folio
>> folio_lock(A) // take A
>> hugetlb_wp
>> mutex_unlock(B) // release B
>> ... hugetlb_fault
>> ... mutex_lock(B) // take B
>> filemap_lock_hugetlb_folio
>> filemap_lock_folio
>> __filemap_get_folio
>> folio_lock(A) // blocked
>> unmap_ref_private
>> ...
>> mutex_lock(B) // retake and blocked
>>
>> This is a ABBA deadlock involving two locks:
>> - Lock A: pagecache_folio lock
>> - Lock B: hugetlb_fault_mutex_table lock
>>
>> The deadlock occurs between two processes as follows:
>> 1. The first process (let’s call it Process 1) is handling a
>> copy-on-write (COW) operation on a hugepage via hugetlb_wp. Due to
>> insufficient reserved hugetlb pages, Process 1, owner of the reserved
>> hugetlb page, attempts to unmap a hugepage owned by another process
>> (non-owner) to satisfy the reservation. Before unmapping, Process 1
>> acquires lock B (hugetlb_fault_mutex_table lock) and then lock A
>> (pagecache_folio lock). To proceed with the unmap, it releases Lock B
>> but retains Lock A. After the unmap, Process 1 tries to reacquire Lock
>> B. However, at this point, Lock B has already been acquired by another
>> process.
>>
>> 2. The second process (Process 2) enters the hugetlb_fault handler
>> during the unmap operation. It successfully acquires Lock B
>> (hugetlb_fault_mutex_table lock) that was just released by Process 1,
>> but then attempts to acquire Lock A (pagecache_folio lock), which is
>> still held by Process 1.
>>
>> As a result, Process 1 (holding Lock A) is blocked waiting for Lock B
>> (held by Process 2), while Process 2 (holding Lock B) is blocked waiting
>> for Lock A (held by Process 1), constructing a ABBA deadlock scenario.
>>
>> The solution here is to unlock the pagecache_folio and provide the
>> pagecache_folio_unlocked variable to the caller to have the visibility
>> over the pagecache_folio status for subsequent handling.
>>
>> The error message:
>> INFO: task repro_20250402_:13229 blocked for more than 64 seconds.
>> Not tainted 6.15.0-rc3+ #24
>> "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
>> task:repro_20250402_ state:D stack:25856 pid:13229 tgid:13228
>> ppid:3513 task_flags:0x400040 flags:0x00004006
>> Call Trace:
>> <TASK>
>> __schedule+0x1755/0x4f50
>> schedule+0x158/0x330
>> schedule_preempt_disabled+0x15/0x30
>> __mutex_lock+0x75f/0xeb0
>> hugetlb_wp+0xf88/0x3440
>> hugetlb_fault+0x14c8/0x2c30
>> trace_clock_x86_tsc+0x20/0x20
>> do_user_addr_fault+0x61d/0x1490
>> exc_page_fault+0x64/0x100
>> asm_exc_page_fault+0x26/0x30
>> RIP: 0010:__put_user_4+0xd/0x20
>> copy_process+0x1f4a/0x3d60
>> kernel_clone+0x210/0x8f0
>> __x64_sys_clone+0x18d/0x1f0
>> do_syscall_64+0x6a/0x120
>> entry_SYSCALL_64_after_hwframe+0x76/0x7e
>> RIP: 0033:0x41b26d
>> </TASK>
>> INFO: task repro_20250402_:13229 is blocked on a mutex likely owned by
>> task repro_20250402_:13250.
>> task:repro_20250402_ state:D stack:28288 pid:13250 tgid:13228
>> ppid:3513 task_flags:0x400040 flags:0x00000006
>> Call Trace:
>> <TASK>
>> __schedule+0x1755/0x4f50
>> schedule+0x158/0x330
>> io_schedule+0x92/0x110
>> folio_wait_bit_common+0x69a/0xba0
>> __filemap_get_folio+0x154/0xb70
>> hugetlb_fault+0xa50/0x2c30
>> trace_clock_x86_tsc+0x20/0x20
>> do_user_addr_fault+0xace/0x1490
>> exc_page_fault+0x64/0x100
>> asm_exc_page_fault+0x26/0x30
>> RIP: 0033:0x402619
>> </TASK>
>> INFO: task repro_20250402_:13250 blocked for more than 65 seconds.
>> Not tainted 6.15.0-rc3+ #24
>> "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
>> task:repro_20250402_ state:D stack:28288 pid:13250 tgid:13228
>> ppid:3513 task_flags:0x400040 flags:0x00000006
>> Call Trace:
>> <TASK>
>> __schedule+0x1755/0x4f50
>> schedule+0x158/0x330
>> io_schedule+0x92/0x110
>> folio_wait_bit_common+0x69a/0xba0
>> __filemap_get_folio+0x154/0xb70
>> hugetlb_fault+0xa50/0x2c30
>> trace_clock_x86_tsc+0x20/0x20
>> do_user_addr_fault+0xace/0x1490
>> exc_page_fault+0x64/0x100
>> asm_exc_page_fault+0x26/0x30
>> RIP: 0033:0x402619
>> </TASK>
>>
>> Showing all locks held in the system:
>> 1 lock held by khungtaskd/35:
>> #0: ffffffff879a7440 (rcu_read_lock){....}-{1:3}, at:
>> debug_show_all_locks+0x30/0x180
>> 2 locks held by repro_20250402_/13229:
>> #0: ffff888017d801e0 (&mm->mmap_lock){++++}-{4:4}, at:
>> lock_mm_and_find_vma+0x37/0x300
>> #1: ffff888000fec848 (&hugetlb_fault_mutex_table[i]){+.+.}-{4:4},
>> at: hugetlb_wp+0xf88/0x3440
>> 3 locks held by repro_20250402_/13250:
>> #0: ffff8880177f3d08 (vm_lock){++++}-{0:0}, at:
>> do_user_addr_fault+0x41b/0x1490
>> #1: ffff888000fec848 (&hugetlb_fault_mutex_table[i]){+.+.}-{4:4},
>> at: hugetlb_fault+0x3b8/0x2c30
>> #2: ffff8880129500e8 (&resv_map->rw_sema){++++}-{4:4}, at:
>> hugetlb_fault+0x494/0x2c30
>>
>> Link: https://drive.google.com/file/d/1DVRnIW-
>> vSayU5J1re9Ct_br3jJQU6Vpb/view?usp=drive_link [1]
>> Link: https://github.com/bboymimi/bpftracer/blob/master/scripts/
>> hugetlb_lock_debug.bt [2]
>> Link: https://drive.google.com/file/d/1bWq2-8o-
>> BJAuhoHWX7zAhI6ggfhVzQUI/view?usp=sharing [3]
>> Fixes: 40549ba8f8e0 ("hugetlb: use new vma_lock for pmd sharing
>> synchronization")
>> Cc: <stable@...r.kernel.org>
>> Cc: Hugh Dickins <hughd@...gle.com>
>> Cc: Florent Revest <revest@...gle.com>
>> Cc: Gavin Shan <gshan@...hat.com>
>> Signed-off-by: Gavin Guo <gavinguo@...lia.com>
>> ---
>> mm/hugetlb.c | 33 ++++++++++++++++++++++++++++-----
>> 1 file changed, 28 insertions(+), 5 deletions(-)
>>
>
> I guess the change log can become concise after the kernel log is
> dropped. The summarized
> stack trace is sufficient to indicate how the dead locking scenario
> happens. Besides,
> it's no need to mention bpftrace and its output. So the changelog would
> be simplified
> to something like below. Please polish it a bit if you would to take it.
> The solution
> looks good except some nitpicks as below.
>
> ---
>
> There is ABBA dead locking scenario happening between hugetlb_fault()
> and hugetlb_wp() on
> the pagecache folio's lock and hugetlb global mutex, which is
> reproducible with syzkaller
> [1]. As below stack traces reveal, process-1 tries to take the hugetlb
> global mutex (A3),
> but with the pagecache folio's lock hold. Process-2 took the hugetlb
> global mutex but tries
> to take the pagecache folio's lock.
>
> [1] https://drive.google.com/file/d/1DVRnIW-vSayU5J1re9Ct_br3jJQU6Vpb/
> view?usp=drive_link
>
> Process-1 Process-2
> ========= =========
> hugetlb_fault
> mutex_lock (A1)
> filemap_lock_hugetlb_folio (B1)
> hugetlb_wp
> alloc_hugetlb_folio #error
> mutex_unlock (A2)
> hugetlb_fault
>
> mutex_lock (A4)
>
> filemap_lock_hugetlb_folio (B4)
> unmap_ref_private
> mutex_lock (A3)
>
> Fix it by releasing the pagecache folio's lock at (A2) of process-1 so
> that pagecache folio's
> lock is available to process-2 at (B4), to avoid the deadlock. In
> process-1, a new variable
> is added to track if the pagecache folio's lock has been released by its
> child function
> hugetlb_wp() to avoid double releases on the lock in hugetlb_fault().
> The similar changes
> are applied to hugetlb_no_page().
>
>> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
>> index e3e6ac991b9c..ad54a74aa563 100644
>> --- a/mm/hugetlb.c
>> +++ b/mm/hugetlb.c
>> @@ -6115,7 +6115,8 @@ static void unmap_ref_private(struct mm_struct
>> *mm, struct vm_area_struct *vma,
>> * Keep the pte_same checks anyway to make transition from the mutex
>> easier.
>> */
>> static vm_fault_t hugetlb_wp(struct folio *pagecache_folio,
>> - struct vm_fault *vmf)
>> + struct vm_fault *vmf,
>> + bool *pagecache_folio_unlocked)
>
> Nitpick: the variable may be renamed to 'pagecache_folio_locked' if
> you're happy
> with.
>
>> {
>> struct vm_area_struct *vma = vmf->vma;
>> struct mm_struct *mm = vma->vm_mm;
>> @@ -6212,6 +6213,22 @@ static vm_fault_t hugetlb_wp(struct folio
>> *pagecache_folio,
>> u32 hash;
>> folio_put(old_folio);
>> + /*
>> + * The pagecache_folio needs to be unlocked to avoid
> ^^^^^^^^
>
> has to be (?)
>
>> + * deadlock and we won't re-lock it in hugetlb_wp(). The
>> + * pagecache_folio could be truncated after being
>> + * unlocked. So its state should not be relied
> ^^^^^^
>
> reliable (?)
>> + * subsequently.
>> + *
>> + * Setting *pagecache_folio_unlocked to true allows the
>> + * caller to handle any necessary logic related to the
>> + * folio's unlocked state.
>> + */
>> + if (pagecache_folio) {
>> + folio_unlock(pagecache_folio);
>> + if (pagecache_folio_unlocked)
>> + *pagecache_folio_unlocked = true;
>> + }
>
> The second section of the comments looks a bit redundant since the code
> changes
> are self-explaining enough :-)
>
>> /*
>> * Drop hugetlb_fault_mutex and vma_lock before
>> * unmapping. unmapping needs to hold vma_lock
>> @@ -6566,7 +6583,7 @@ static vm_fault_t hugetlb_no_page(struct
>> address_space *mapping,
>> hugetlb_count_add(pages_per_huge_page(h), mm);
>> if ((vmf->flags & FAULT_FLAG_WRITE) && !(vma->vm_flags &
>> VM_SHARED)) {
>> /* Optimization, do the COW without a second fault */
>> - ret = hugetlb_wp(folio, vmf);
>> + ret = hugetlb_wp(folio, vmf, NULL);
>
> It's not certain if we have another deadlock between hugetlb_no_page()
> and hugetlb_wp(),
> similar to the existing one between hugetlb_fault() and hugetlb_wp(). So
> I think it's
> reasonable to pass '&pagecache_folio_locked' to hugetlb_wp() here and
> skip to unlock
> on pagecache_folio_locked == false in hugetlb_no_page(). It's not
> harmful at least.
Thank you very much for taking the time to review my patch! I appreciate
your feedback. :)
After carefully reviewing the hugetlb_no_page function, I've made an
observation regarding the pagecache_folio handling. Specifically, when
pagecache_folio is assigned to vmf->pte, the vmf->ptl lock is held. This
lock remains active when vmf->pte is later accessed in hugetlb_wp. The
ptl lock ensures that the pte value is the same as the pagecache_folio
assigned in hugetlb_no_page. As a result, the following comparison in
hugetlb_wp will always evaluate to false because old_folio and
pagecache_folio reference the same object:
if (is_vma_resv_set(vma, HPAGE_RESV_OWNER) &&
old_folio != pagecache_folio)
cow_from_owner = true;
Based on this analysis, passing pagecache_folio_locked in
hugetlb_no_page is unnecessary. Let me know if I missed anything. Other
comments look good to me. Thanks!
>
>> }
>> spin_unlock(vmf->ptl);
>> @@ -6638,6 +6655,7 @@ vm_fault_t hugetlb_fault(struct mm_struct *mm,
>> struct vm_area_struct *vma,
>> struct hstate *h = hstate_vma(vma);
>> struct address_space *mapping;
>> int need_wait_lock = 0;
>> + bool pagecache_folio_unlocked = false;
>> struct vm_fault vmf = {
>> .vma = vma,
>> .address = address & huge_page_mask(h),
>> @@ -6792,7 +6810,8 @@ vm_fault_t hugetlb_fault(struct mm_struct *mm,
>> struct vm_area_struct *vma,
>> if (flags & (FAULT_FLAG_WRITE|FAULT_FLAG_UNSHARE)) {
>> if (!huge_pte_write(vmf.orig_pte)) {
>> - ret = hugetlb_wp(pagecache_folio, &vmf);
>> + ret = hugetlb_wp(pagecache_folio, &vmf,
>> + &pagecache_folio_unlocked);
>> goto out_put_page;
>> } else if (likely(flags & FAULT_FLAG_WRITE)) {
>> vmf.orig_pte = huge_pte_mkdirty(vmf.orig_pte);
>> @@ -6809,10 +6828,14 @@ vm_fault_t hugetlb_fault(struct mm_struct *mm,
>> struct vm_area_struct *vma,
>> out_ptl:
>> spin_unlock(vmf.ptl);
>> - if (pagecache_folio) {
>> + /*
>> + * If the pagecache_folio is unlocked in hugetlb_wp(), we skip
>> + * folio_unlock() here.
>> + */
>> + if (pagecache_folio && !pagecache_folio_unlocked)
>> folio_unlock(pagecache_folio);
>> + if (pagecache_folio)
>> folio_put(pagecache_folio);
>> - }
>
> The comments seem redundant since the code changes are self-explaining.
> Besides, no need to validate 'pagecache_folio' for twice.
>
> if (pagecache_folio) {
> if (pagecache_folio_locked)
> folio_unlock(pagecache_folio);
>
> folio_put(pagecache_folio);
> }
>
>> out_mutex:
>> hugetlb_vma_unlock_read(vma);
>>
>> base-commit: d76bb1ebb5587f66b0f8b8099bfbb44722bc08b3
>
> Thanks,
> Gavin
>
Powered by blists - more mailing lists