Skip to contentSkip to conversation

hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list

PATCH Superseded Deepanshu Kartikey 3 participants 3 messages Sep 25, 2025

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list

hugetlb_vmdelete_list() uses trylock to acquire VMA locks during truncate operations. As per the original design in commit 40549ba8f8e0 ("hugetlb:
use new vma_lock for pmd sharing synchronization"), if the trylock fails or the VMA has no lock, it should skip that VMA. Any remaining mapped pages are handled by remove_inode_hugepages() which is called after hugetlb_vmdelete_list() and uses proper lock ordering to guarantee unmapping success.

Currently, when hugetlb_vma_trylock_write() returns success (1) for VMAs without shareable locks, the code proceeds to call unmap_hugepage_range(). This causes assertion failures in huge_pmd_unshare() → hugetlb_vma_assert_locked() because no lock is actually held:

  WARNING: CPU: 1 PID: 6594 Comm: syz.0.28 Not tainted
  Call Trace:
   hugetlb_vma_assert_locked+0x1dd/0x250
   huge_pmd_unshare+0x2c8/0x540
   __unmap_hugepage_range+0x6e3/0x1aa0
   unmap_hugepage_range+0x32e/0x410
   hugetlb_vmdelete_list+0x189/0x1f0

Fix by explicitly skipping VMAs without shareable locks after trylock succeeds, consistent with the original design where such VMAs are deferred to remove_inode_hugepages() for proper handling.

Reported-by [email protected]
Fixes 40549ba8f8e0 ("hugetlb: use new vma_lock for pmd sharing synchronization")
---
 fs/hugetlbfs/inode.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/hugetlbfs/inode.c +2 -1
index 9e0625167517..9ba98cab3388 100644
@@ -487,7 +487,8 @@ hugetlb_vmdelete_list(struct rb_root_cached *root, pgoff_t start, pgoff_t end,
 
 		if (!hugetlb_vma_trylock_write(vma))
 			continue;
-
+		if (!__vma_shareable_lock(vma))
+			continue;
 		v_start = vma_offset_start(vma, start);
 		v_end = vma_offset_end(vma, end);
 
Signature · 1 line
2.43.0
syzbot Tested-by [email protected] on v1 1/1
#1 lore raw
syzbot ci has tested the following series

syzbot ci has tested the following series

[v1] hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list
https://lore.kernel.org/all/[email protected]
* [PATCH] hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list

and found the following issue:
WARNING: lock held when returning to user space in hugetlb_vmdelete_list

Full report is available here:
https://ci.syzbot.org/series/7d3a090c-ec8f-4d0c-8f7b-8797e63f1fde

***

WARNING: lock held when returning to user space in hugetlb_vmdelete_list

tree:      torvalds
URL:       https://kernel.googlesource.com/pub/scm/linux/kernel/git/torvalds/linux
base:      07e27ad16399afcd693be20211b0dfae63e0615f
arch:      amd64
compiler:  Debian clang version 20.1.8 (++20250708063551+0c9f909b7976-1~exp1~20250708183702.136), Debian LLD 20.1.8
config:    https://ci.syzbot.org/builds/06671e7c-eebd-4d11-9f0d-d9dae5637c0a/config
C repro:   https://ci.syzbot.org/findings/75fcdb87-67fc-45ee-a70d-3064e8c1126c/c_repro
syz repro: https://ci.syzbot.org/findings/75fcdb87-67fc-45ee-a70d-3064e8c1126c/syz_repro

================================================
WARNING: lock held when returning to user space!
syzkaller #0 Not tainted
------------------------------------------------
syz.0.17/5991 is leaving the kernel with locks still held!
1 lock held by syz.0.17/5991:
#0: ffff88801f7dcce8 (&resv_map->rw_sema){+.+.}-{4:4}, at: hugetlb_vmdelete_list+0x67/0x230 fs/hugetlbfs/inode.c:490

***

If these findings have caused you to resend the series or submit a separate fix, please add the following tag to your commit message:

---
This report is generated by a bot. It may contain errors. syzbot ci engineers can be reached at [email protected].

maintainer
#2 lore raw
Quoted 25 lines · On Thu, 25 Sep 2025 14:06:49 -0700 syzbot ci…

syzbot ci has tested the following series

[v1] hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list
https://lore.kernel.org/all/[email protected]
* [PATCH] hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list

and found the following issue:
WARNING: lock held when returning to user space in hugetlb_vmdelete_list

Full report is available here:
https://ci.syzbot.org/series/7d3a090c-ec8f-4d0c-8f7b-8797e63f1fde

***

WARNING: lock held when returning to user space in hugetlb_vmdelete_list

tree:      torvalds
URL:       https://kernel.googlesource.com/pub/scm/linux/kernel/git/torvalds/linux
base:      07e27ad16399afcd693be20211b0dfae63e0615f
arch:      amd64
compiler:  Debian clang version 20.1.8 (++20250708063551+0c9f909b7976-1~exp1~20250708183702.136), Debian LLD 20.1.8
config:    https://ci.syzbot.org/builds/06671e7c-eebd-4d11-9f0d-d9dae5637c0a/config
C repro:   https://ci.syzbot.org/findings/75fcdb87-67fc-45ee-a70d-3064e8c1126c/c_repro
syz repro: https://ci.syzbot.org/findings/75fcdb87-67fc-45ee-a70d-3064e8c1126c/syz_repro
Quoted 3 lines

================================================
WARNING: lock held when returning to user space!
syzkaller #0 Not tainted

um yes. Presumably something like the below is needed.

I'm wondering how this passed runtime testing?

If resending, please let's add a comment telling readers why we're skipping !__vma_shareable_lock vmas.

I'll drop the patch, thanks.

fs/hugetlbfs/inode.c renamed from fs/hugetlbfs/inode.c~hugetlbfs-skip-vmas-without-shareable-locks-in-hugetlb_vmdelete_list-fix +2 -1
@@ -488,7 +488,7 @@ hugetlb_vmdelete_list(struct rb_root_cac
 		if (!hugetlb_vma_trylock_write(vma))
 			continue;
 		if (!__vma_shareable_lock(vma))
-			continue;
+			goto skip;
 		v_start = vma_offset_start(vma, start);
 		v_end = vma_offset_end(vma, end);
 
@@ -499,6 +499,7 @@ hugetlb_vmdelete_list(struct rb_root_cac
 		 * vmas.  Therefore, lock is not held when calling
 		 * unmap_hugepage_range for private vmas.
 		 */
+skip:
 		hugetlb_vma_unlock_write(vma);
 	}
 }

_