[syzbot] [mm?] [cgroups?] WARNING in page_counter_uncharge (2)

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: 5597dd284ff8 net: ti: icssg-prueth: fix missing data copy ..
git tree: net
console output: https://syzkaller.appspot.com/x/log.txt?x=17f536da580000
kernel config: https://syzkaller.appspot.com/x/.config?x=6754c86e8d9e4c91
dashboard link: https://syzkaller.appspot.com/bug?extid=226c1f947186f8fef796
compiler: Debian clang version 21.1.8 (++20251221033036+2078da43e25a-1~exp1~20251221153213.50), Debian LLD 21.1.8
syz repro: https://syzkaller.appspot.com/x/repro.syz?x=131baeda580000
C reproducer: https://syzkaller.appspot.com/x/repro.c?x=167d6f72580000

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/b6c0ef6a1be9/disk-5597dd28.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/38b971059ff5/vmlinux-5597dd28.xz
kernel image: https://storage.googleapis.com/syzbot-assets/55dd4bd79e77/bzImage-5597dd28.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+226c1f...@syzkaller.appspotmail.com

------------[ cut here ]------------
page_counter underflow: -512 nr_pages=512
WARNING: mm/page_counter.c:61 at page_counter_cancel mm/page_counter.c:60 [inline], CPU#1: syz.0.3396/16434
WARNING: mm/page_counter.c:61 at page_counter_uncharge+0xd2/0x150 mm/page_counter.c:184, CPU#1: syz.0.3396/16434
Modules linked in:
CPU: 1 UID: 0 PID: 16434 Comm: syz.0.3396 Not tainted syzkaller #0 PREEMPT(full)
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 02/12/2026
RIP: 0010:page_counter_cancel mm/page_counter.c:60 [inline]
RIP: 0010:page_counter_uncharge+0xd8/0x150 mm/page_counter.c:184
Code: f7 e8 7c 88 f8 ff 4d 8b 36 4d 85 f6 74 6e e8 cf 3f 8e ff e9 6c ff ff ff e8 c5 3f 8e ff 48 8d 3d 2e ea df 0d 4c 89 fe 48 89 da <67> 48 0f b9 3a 4c 89 f7 be 08 00 00 00 e8 e6 8a f8 ff 4c 89 f0 48
RSP: 0018:ffffc9000da772b0 EFLAGS: 00010093
RAX: ffffffff82376eab RBX: 0000000000000200 RCX: ffff88807c631e80
RDX: 0000000000000200 RSI: fffffffffffffe00 RDI: ffffffff901758e0
RBP: fffffffffffffe00 R08: ffff888032fd5387 R09: 1ffff110065faa70
R10: dffffc0000000000 R11: ffffed10065faa71 R12: 0000000000000001
R13: dffffc0000000000 R14: ffff888032fd5380 R15: fffffffffffffe00
FS: 0000000000000000(0000) GS:ffff88812555a000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f14ec9d5ff8 CR3: 000000000e54c000 CR4: 00000000003526f0
Call Trace:
<TASK>
__hugetlb_cgroup_uncharge_folio+0x15e/0x510 mm/hugetlb_cgroup.c:354
free_huge_folio+0xaef/0x11e0 mm/hugetlb.c:1782
folios_put_refs+0x553/0x8d0 mm/swap.c:983
folio_batch_release include/linux/pagevec.h:101 [inline]
remove_inode_hugepages+0xf50/0x11a0 fs/hugetlbfs/inode.c:608
hugetlbfs_evict_inode+0xaf/0x260 fs/hugetlbfs/inode.c:623
evict+0x61e/0xb10 fs/inode.c:846
__dentry_kill+0x1a2/0x5e0 fs/dcache.c:670
finish_dput+0xc9/0x480 fs/dcache.c:879
__fput+0x691/0xa70 fs/file_table.c:477
task_work_run+0x1d9/0x270 kernel/task_work.c:233
exit_task_work include/linux/task_work.h:40 [inline]
do_exit+0x70f/0x23c0 kernel/exit.c:976
do_group_exit+0x21b/0x2d0 kernel/exit.c:1118
get_signal+0x1284/0x1330 kernel/signal.c:3034
arch_do_signal_or_restart+0xbc/0x830 arch/x86/kernel/signal.c:337
__exit_to_user_mode_loop kernel/entry/common.c:64 [inline]
exit_to_user_mode_loop+0x86/0x480 kernel/entry/common.c:98
__exit_to_user_mode_prepare include/linux/irq-entry-common.h:226 [inline]
syscall_exit_to_user_mode_prepare include/linux/irq-entry-common.h:256 [inline]
syscall_exit_to_user_mode include/linux/entry-common.h:325 [inline]
do_syscall_64+0x32d/0xf80 arch/x86/entry/syscall_64.c:100
entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x7f14ebb9c799
Code: Unable to access opcode bytes at 0x7f14ebb9c76f.
RSP: 002b:00007f14ec9d60e8 EFLAGS: 00000246 ORIG_RAX: 00000000000000ca
RAX: fffffffffffffe00 RBX: 00007f14ebe16098 RCX: 00007f14ebb9c799
RDX: 0000000000000000 RSI: 0000000000000080 RDI: 00007f14ebe16098
RBP: 00007f14ebe16090 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 00007f14ebe16128 R14: 00007ffc404c7910 R15: 00007ffc404c79f8
</TASK>
----------------
Code disassembly (best guess):
0: f7 e8 imul %eax
2: 7c 88 jl 0xffffff8c
4: f8 clc
5: ff 4d 8b decl -0x75(%rbp)
8: 36 4d 85 f6 ss test %r14,%r14
c: 74 6e je 0x7c
e: e8 cf 3f 8e ff call 0xff8e3fe2
13: e9 6c ff ff ff jmp 0xffffff84
18: e8 c5 3f 8e ff call 0xff8e3fe2
1d: 48 8d 3d 2e ea df 0d lea 0xddfea2e(%rip),%rdi # 0xddfea52
24: 4c 89 fe mov %r15,%rsi
27: 48 89 da mov %rbx,%rdx
* 2a: 67 48 0f b9 3a ud1 (%edx),%rdi <-- trapping instruction
2f: 4c 89 f7 mov %r14,%rdi
32: be 08 00 00 00 mov $0x8,%esi
37: e8 e6 8a f8 ff call 0xfff88b22
3c: 4c 89 f0 mov %r14,%rax
3f: 48 rex.W


---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

If the report is already addressed, let syzbot know by replying with:
#syz fix: exact-commit-title

If you want syzbot to run the reproducer, reply with:
#syz test: git://repo/address.git branch-or-commit-hash
If you attach or paste a git patch, syzbot will apply it before testing.

If you want to overwrite report's subsystems, reply with:
#syz set subsystems: new-subsystem
(See the list of subsystem names on the web dashboard)

If the report is a duplicate of another one, reply with:
#syz dup: exact-subject-of-another-report

If you want to undo deduplication, reply with:
#syz undup

[syzbot]

For archival purposes, forwarding an incoming command email to
linux-...@vger.kernel.org, syzkall...@googlegroups.com.

***

Subject: [PATCH] mm/hugetlb: fix hugetlb cgroup rsvd charge/uncharge mismatch
Author: karti...@gmail.com

#syz test: git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git main

In alloc_hugetlb_folio(), a single h_cg pointer is used for both
the rsvd and non-rsvd hugetlb cgroup charges. When map_chg is set,
hugetlb_cgroup_charge_cgroup_rsvd() stores the charged cgroup in
h_cg, but the immediately following hugetlb_cgroup_charge_cgroup()
overwrites h_cg with the non-rsvd cgroup pointer.

As a result, hugetlb_cgroup_commit_charge_rsvd() stores the wrong
(non-rsvd) cgroup pointer into the folio's rsvd slot.

When the folio is later freed, free_huge_folio() unconditionally
calls both hugetlb_cgroup_uncharge_folio() and
hugetlb_cgroup_uncharge_folio_rsvd(). The rsvd uncharge reads back
the wrong cgroup from the folio and decrements a counter that was
never charged for that cgroup, causing a page_counter underflow:

page_counter underflow: -512 nr_pages=512
WARNING: mm/page_counter.c:61 at page_counter_cancel

Fix this by introducing a separate h_cg_rsvd pointer exclusively
for the rsvd charge path, keeping the rsvd and non-rsvd charges
fully independent through their charge, commit, and error uncharge
paths.

Reported-by: syzbot+226c1f...@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=226c1f947186f8fef796
Signed-off-by: Deepanshu <yo...@email.com>
---
mm/hugetlb.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 327eaa4074d3..5be36a888e70 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -2915,6 +2915,7 @@ struct folio *alloc_hugetlb_folio(struct vm_area_struct *vma,
map_chg_state map_chg;
int ret, idx;
struct hugetlb_cgroup *h_cg = NULL;
+ struct hugetlb_cgroup *h_cg_rsvd = NULL;
gfp_t gfp = htlb_alloc_mask(h) | __GFP_RETRY_MAYFAIL;

idx = hstate_index(h);
@@ -2965,7 +2966,7 @@ struct folio *alloc_hugetlb_folio(struct vm_area_struct *vma,
*/
if (map_chg) {
ret = hugetlb_cgroup_charge_cgroup_rsvd(
- idx, pages_per_huge_page(h), &h_cg);
+ idx, pages_per_huge_page(h), &h_cg_rsvd);
if (ret)
goto out_subpool_put;
}
@@ -3007,7 +3008,7 @@ struct folio *alloc_hugetlb_folio(struct vm_area_struct *vma,
*/
if (map_chg) {
hugetlb_cgroup_commit_charge_rsvd(idx, pages_per_huge_page(h),
- h_cg, folio);
+ h_cg_rsvd, folio);
}

spin_unlock_irq(&hugetlb_lock);
@@ -3059,7 +3060,7 @@ struct folio *alloc_hugetlb_folio(struct vm_area_struct *vma,
out_uncharge_cgroup_reservation:
if (map_chg)
hugetlb_cgroup_uncharge_cgroup_rsvd(idx, pages_per_huge_page(h),
- h_cg);
+ h_cg_rsvd);
out_subpool_put:
/*
* put page to subpool iff the quota of subpool's rsv_hpages is used
--
2.43.0

[syzbot]

Hello,

syzbot has tested the proposed patch and the reproducer did not trigger any issue:

Reported-by: syzbot+226c1f...@syzkaller.appspotmail.com
Tested-by: syzbot+226c1f...@syzkaller.appspotmail.com

Tested on:

commit: dc9e9d61 Merge branch 'net-enetc-add-more-checks-to-en..
git tree: net
console output: https://syzkaller.appspot.com/x/log.txt?x=149f7f52580000

kernel config: https://syzkaller.appspot.com/x/.config?x=6754c86e8d9e4c91
dashboard link: https://syzkaller.appspot.com/bug?extid=226c1f947186f8fef796
compiler: Debian clang version 21.1.8 (++20251221033036+2078da43e25a-1~exp1~20251221153213.50), Debian LLD 21.1.8

patch: https://syzkaller.appspot.com/x/patch.diff?x=17fd4eda580000

Note: testing is done by a robot and is best-effort only.