Re: [syzbot] [mm?] [io-uring?] WARNING in get_pte_pfn

[Yu Zhao <yuzhao@google.com>]

On Fri, Dec 22, 2023 at 7:55 AM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 12/22/23 1:11 AM, syzbot wrote:
> > Hello,
> >
> > syzbot found the following issue on:
> >
> > HEAD commit:    0e389834672c Merge tag 'for-6.7-rc5-tag' of git://git.kern..
> > git tree:       upstream
> > console output: https://syzkaller.appspot.com/x/log.txt?x=1454824ee80000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=f21aff374937e60e
> > dashboard link: https://syzkaller.appspot.com/bug?extid=03fd9b3f71641f0ebf2d
> > compiler:       gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40
> > syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=13b4ef49e80000
> > C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=118314d6e80000
> >
> > Downloadable assets:
> > disk image: https://storage.googleapis.com/syzbot-assets/e58cd74e152a/disk-0e389834.raw.xz
> > vmlinux: https://storage.googleapis.com/syzbot-assets/45d17ccb34bc/vmlinux-0e389834.xz
> > kernel image: https://storage.googleapis.com/syzbot-assets/b9b7105d4e08/bzImage-0e389834.xz
> >
> > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > Reported-by: syzbot+03fd9b3f71641f0ebf2d@syzkaller.appspotmail.com
> >
> > ------------[ cut here ]------------
> > WARNING: CPU: 1 PID: 5066 at mm/vmscan.c:3242 get_pte_pfn+0x1b5/0x3f0 mm/vmscan.c:3242
> > Modules linked in:
> > CPU: 1 PID: 5066 Comm: syz-executor668 Not tainted 6.7.0-rc5-syzkaller-00270-g0e389834672c #0
> > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 11/17/2023
> > RIP: 0010:get_pte_pfn+0x1b5/0x3f0 mm/vmscan.c:3242
> > Code: f3 74 2a e8 6d 78 cb ff 31 ff 48 b8 00 00 00 00 00 00 00 02 48 21 c5 48 89 ee e8 e6 73 cb ff 48 85 ed 74 4e e8 4c 78 cb ff 90 <0f> 0b 90 48 c7 c3 ff ff ff ff e8 3c 78 cb ff 48 b8 00 00 00 00 00
> > RSP: 0018:ffffc900041e6878 EFLAGS: 00010293
> > RAX: 0000000000000000 RBX: 000000000007891d RCX: ffffffff81bbf6e3
> > RDX: ffff88807d813b80 RSI: ffffffff81bbf684 RDI: 0000000000000005
> > RBP: 0000000000000000 R08: 0000000000000005 R09: 0000000000000000
> > R10: 0000000000000200 R11: 0000000000000003 R12: 0000000000000200
> > R13: 1ffff9200083cd0f R14: 0000000000010b21 R15: 0000000020ffc000
> > FS:  0000555555f4d480(0000) GS:ffff8880b9900000(0000) knlGS:0000000000000000
> > CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > CR2: 0000000000000000 CR3: 000000005fbfa000 CR4: 0000000000350ef0
> > Call Trace:
> >  <TASK>
> >  lru_gen_look_around+0x70d/0x11a0 mm/vmscan.c:4001
> >  folio_referenced_one+0x5a2/0xf70 mm/rmap.c:843
> >  rmap_walk_anon+0x225/0x570 mm/rmap.c:2485
> >  rmap_walk mm/rmap.c:2562 [inline]
> >  rmap_walk mm/rmap.c:2557 [inline]
> >  folio_referenced+0x28a/0x4b0 mm/rmap.c:960
> >  folio_check_references mm/vmscan.c:829 [inline]
> >  shrink_folio_list+0x1ace/0x3f00 mm/vmscan.c:1160
> >  evict_folios+0x6e7/0x1b90 mm/vmscan.c:4499
> >  try_to_shrink_lruvec+0x638/0xa10 mm/vmscan.c:4704
> >  lru_gen_shrink_lruvec mm/vmscan.c:4849 [inline]
> >  shrink_lruvec+0x314/0x2990 mm/vmscan.c:5622
> >  shrink_node_memcgs mm/vmscan.c:5842 [inline]
> >  shrink_node+0x811/0x3710 mm/vmscan.c:5877
> >  shrink_zones mm/vmscan.c:6116 [inline]
> >  do_try_to_free_pages+0x36c/0x1940 mm/vmscan.c:6178
> >  try_to_free_mem_cgroup_pages+0x31a/0x770 mm/vmscan.c:6493
> >  try_charge_memcg+0x3d3/0x11f0 mm/memcontrol.c:2742
> >  obj_cgroup_charge_pages mm/memcontrol.c:3255 [inline]
> >  __memcg_kmem_charge_page+0xdd/0x2a0 mm/memcontrol.c:3281
> >  __alloc_pages+0x263/0x2420 mm/page_alloc.c:4585
> >  alloc_pages_mpol+0x258/0x5f0 mm/mempolicy.c:2133
> >  __get_free_pages+0xc/0x40 mm/page_alloc.c:4615
> >  io_mem_alloc+0x33/0x60 io_uring/io_uring.c:2789
> >  io_allocate_scq_urings io_uring/io_uring.c:3842 [inline]
> >  io_uring_create io_uring/io_uring.c:4019 [inline]
> >  io_uring_setup+0x13ed/0x2430 io_uring/io_uring.c:4131
> >  __do_sys_io_uring_setup io_uring/io_uring.c:4158 [inline]
> >  __se_sys_io_uring_setup io_uring/io_uring.c:4152 [inline]
> >  __x64_sys_io_uring_setup+0x98/0x140 io_uring/io_uring.c:4152
> >  do_syscall_x64 arch/x86/entry/common.c:52 [inline]
> >  do_syscall_64+0x40/0x110 arch/x86/entry/common.c:83
> >  entry_SYSCALL_64_after_hwframe+0x63/0x6b
> > RIP: 0033:0x7f4b0e4778a9
> > Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 d1 1a 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
> > RSP: 002b:00007fff814fe868 EFLAGS: 00000202 ORIG_RAX: 00000000000001a9
> > RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 00007f4b0e4778a9
> > RDX: 0000000020000700 RSI: 0000000020000640 RDI: 0000000000005a19
> > RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000020000700
> > R10: 00007fff814fe8d0 R11: 0000000000000202 R12: 0000000020000640
> > R13: 0000000000000000 R14: 0000000000005a19 R15: 0000000020000700
> >  </TASK>
>
> Don't think this is io_uring related, test case looks like it's just
> setting up and tearing down big rings.

Can confirm it is an MM bug. Just posted the fix:
https://lore.kernel.org/20231223045647.1566043-1-yuzhao@google.com/