Re: BUG: Bad page state (5)

From: Eric Biggers
Date: Thu Feb 28 2019 - 02:59:49 EST


On Thu, Feb 28, 2019 at 07:53:09AM +0100, 'Dmitry Vyukov' via syzkaller-bugs wrote:
> On Wed, Feb 27, 2019 at 9:53 PM Eric Biggers <ebiggers@xxxxxxxxxx> wrote:
> >
> > On Tue, Feb 26, 2019 at 10:21:30AM -0800, Eric Biggers wrote:
> > > On Wed, Feb 13, 2019 at 12:23:31PM -0800, Andrew Morton wrote:
> > > > On Wed, 13 Feb 2019 09:56:04 -0800 syzbot <syzbot+2cd2887ea471ed6e6995@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
> > > >
> > > > > Hello,
> > > > >
> > > > > syzbot found the following crash on:
> > > > >
> > > > > HEAD commit: c4f3ef3eb53f Add linux-next specific files for 20190213
> > > > > git tree: linux-next
> > > > > console output: https://syzkaller.appspot.com/x/log.txt?x=1130a124c00000
> > > > > kernel config: https://syzkaller.appspot.com/x/.config?x=9ec67976eb2df882
> > > > > dashboard link: https://syzkaller.appspot.com/bug?extid=2cd2887ea471ed6e6995
> > > > > compiler: gcc (GCC) 9.0.0 20181231 (experimental)
> > > > > syz repro: https://syzkaller.appspot.com/x/repro.syz?x=14ecdaa8c00000
> > > > > C reproducer: https://syzkaller.appspot.com/x/repro.c?x=12ebe178c00000
> > > > >
> > > > > IMPORTANT: if you fix the bug, please add the following tag to the commit:
> > > > > Reported-by: syzbot+2cd2887ea471ed6e6995@xxxxxxxxxxxxxxxxxxxxxxxxx
> > > >
> > > > It looks like a a memfd page was freed with a non-NULL ->mapping.
> > > >
> > > > Joel touched the memfd code with "mm/memfd: add an F_SEAL_FUTURE_WRITE
> > > > seal to memfd" but it would be surprising if syzbot tickled that code?
> > > >
> > > >
> > > > > BUG: Bad page state in process udevd pfn:472f0
> > > > > name:"memfd:"
> > > > > page:ffffea00011cbc00 count:0 mapcount:0 mapping:ffff88800df2ad40 index:0xf
> > > > > shmem_aops
> > > > > flags: 0x1fffc000008000c(uptodate|dirty|swapbacked)
> > > > > raw: 01fffc000008000c ffffea0000ac4f08 ffff8880a85af890 ffff88800df2ad40
> > > > > raw: 000000000000000f 0000000000000000 00000000ffffffff 0000000000000000
> > > > > page dumped because: non-NULL mapping
> > > > > Modules linked in:
> > > > > CPU: 1 PID: 7586 Comm: udevd Not tainted 5.0.0-rc6-next-20190213 #34
> > > > > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
> > > > > Google 01/01/2011
> > > > > Call Trace:
> > > > > __dump_stack lib/dump_stack.c:77 [inline]
> > > > > dump_stack+0x172/0x1f0 lib/dump_stack.c:113
> > > > > bad_page.cold+0xda/0xff mm/page_alloc.c:586
> > > > > free_pages_check_bad+0x142/0x1a0 mm/page_alloc.c:1014
> > > > > free_pages_check mm/page_alloc.c:1023 [inline]
> > > > > free_pages_prepare mm/page_alloc.c:1113 [inline]
> > > > > free_pcp_prepare mm/page_alloc.c:1138 [inline]
> > > > > free_unref_page_prepare mm/page_alloc.c:2991 [inline]
> > > > > free_unref_page_list+0x31d/0xc40 mm/page_alloc.c:3060
> > > > > name:"memfd:"
> > > > > release_pages+0x60d/0x1940 mm/swap.c:791
> > > > > pagevec_lru_move_fn+0x218/0x2a0 mm/swap.c:213
> > > > > __pagevec_lru_add mm/swap.c:917 [inline]
> > > > > lru_add_drain_cpu+0x2f7/0x520 mm/swap.c:581
> > > > > lru_add_drain+0x20/0x60 mm/swap.c:652
> > > > > exit_mmap+0x290/0x530 mm/mmap.c:3134
> > > > > __mmput kernel/fork.c:1047 [inline]
> > > > > mmput+0x15f/0x4c0 kernel/fork.c:1068
> > > > > exec_mmap fs/exec.c:1046 [inline]
> > > > > flush_old_exec+0x8d9/0x1c20 fs/exec.c:1279
> > > > > load_elf_binary+0x9bc/0x53f0 fs/binfmt_elf.c:864
> > > > > search_binary_handler fs/exec.c:1656 [inline]
> > > > > search_binary_handler+0x17f/0x570 fs/exec.c:1634
> > > > > exec_binprm fs/exec.c:1698 [inline]
> > > > > __do_execve_file.isra.0+0x1394/0x23f0 fs/exec.c:1818
> > > > > do_execveat_common fs/exec.c:1865 [inline]
> > > > > do_execve fs/exec.c:1882 [inline]
> > > > > __do_sys_execve fs/exec.c:1958 [inline]
> > > > > __se_sys_execve fs/exec.c:1953 [inline]
> > > > > __x64_sys_execve+0x8f/0xc0 fs/exec.c:1953
> > > > > do_syscall_64+0x103/0x610 arch/x86/entry/common.c:290
> > > > > entry_SYSCALL_64_after_hwframe+0x49/0xbe
> > > > > RIP: 0033:0x7fc7001ba207
> > > > > Code: Bad RIP value.
> > > > > RSP: 002b:00007ffe06aa13b8 EFLAGS: 00000206 ORIG_RAX: 000000000000003b
> > > > > RAX: ffffffffffffffda RBX: 00000000ffffffff RCX: 00007fc7001ba207
> > > > > RDX: 0000000001fd5fd0 RSI: 00007ffe06aa14b0 RDI: 00007ffe06aa24c0
> > > > > RBP: 0000000000625500 R08: 0000000000001c49 R09: 0000000000001c49
> > > > > R10: 0000000000000000 R11: 0000000000000206 R12: 0000000001fd5fd0
> > > > > R13: 0000000000000007 R14: 0000000001fc6250 R15: 0000000000000005
> > > > > BUG: Bad page state in process udevd pfn:2b13c
> > > > > page:ffffea0000ac4f00 count:0 mapcount:0 mapping:ffff88800df2ad40 index:0xe
> > > > > shmem_aops
> > > > > flags: 0x1fffc000008000c(uptodate|dirty|swapbacked)
> > > > > raw: 01fffc000008000c ffff8880a85af890 ffff8880a85af890 ffff88800df2ad40
> > > > > raw: 000000000000000e 0000000000000000 00000000ffffffff 0000000000000000
> > > > > page dumped because: non-NULL mapping
> > > > > Modules linked in:
> > > > > CPU: 1 PID: 7586 Comm: udevd Tainted: G B
> > > > > 5.0.0-rc6-next-20190213 #34
> > > > > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
> > > > > Google 01/01/2011
> > > > > Call Trace:
> > > > > __dump_stack lib/dump_stack.c:77 [inline]
> > > > > dump_stack+0x172/0x1f0 lib/dump_stack.c:113
> > > > > bad_page.cold+0xda/0xff mm/page_alloc.c:586
> > > > > name:"memfd:"
> > > > > free_pages_check_bad+0x142/0x1a0 mm/page_alloc.c:1014
> > > > > free_pages_check mm/page_alloc.c:1023 [inline]
> > > > > free_pages_prepare mm/page_alloc.c:1113 [inline]
> > > > > free_pcp_prepare mm/page_alloc.c:1138 [inline]
> > > > > free_unref_page_prepare mm/page_alloc.c:2991 [inline]
> > > > > free_unref_page_list+0x31d/0xc40 mm/page_alloc.c:3060
> > > > > release_pages+0x60d/0x1940 mm/swap.c:791
> > > > > pagevec_lru_move_fn+0x218/0x2a0 mm/swap.c:213
> > > > > __pagevec_lru_add mm/swap.c:917 [inline]
> > > > > lru_add_drain_cpu+0x2f7/0x520 mm/swap.c:581
> > > > > lru_add_drain+0x20/0x60 mm/swap.c:652
> > > > > exit_mmap+0x290/0x530 mm/mmap.c:3134
> > > > > __mmput kernel/fork.c:1047 [inline]
> > > > > mmput+0x15f/0x4c0 kernel/fork.c:1068
> > > > > exec_mmap fs/exec.c:1046 [inline]
> > > > > flush_old_exec+0x8d9/0x1c20 fs/exec.c:1279
> > > > > load_elf_binary+0x9bc/0x53f0 fs/binfmt_elf.c:864
> > > > > search_binary_handler fs/exec.c:1656 [inline]
> > > > > search_binary_handler+0x17f/0x570 fs/exec.c:1634
> > > > > exec_binprm fs/exec.c:1698 [inline]
> > > > > __do_execve_file.isra.0+0x1394/0x23f0 fs/exec.c:1818
> > > > > do_execveat_common fs/exec.c:1865 [inline]
> > > > > do_execve fs/exec.c:1882 [inline]
> > > > > __do_sys_execve fs/exec.c:1958 [inline]
> > > > > __se_sys_execve fs/exec.c:1953 [inline]
> > > > > __x64_sys_execve+0x8f/0xc0 fs/exec.c:1953
> > > > > do_syscall_64+0x103/0x610 arch/x86/entry/common.c:290
> > > > > entry_SYSCALL_64_after_hwframe+0x49/0xbe
> > > > > RIP: 0033:0x7fc7001ba207
> > > > > Code: Bad RIP value.
> > > > > RSP: 002b:00007ffe06aa13b8 EFLAGS: 00000206 ORIG_RAX: 000000000000003b
> > > > > RAX: ffffffffffffffda RBX: 00000000ffffffff RCX: 00007fc7001ba207
> > > > > RDX: 0000000001fd5fd0 RSI: 00007ffe06aa14b0 RDI: 00007ffe06aa24c0
> > > > > RBP: 0000000000625500 R08: 0000000000001c49 R09: 0000000000001c49
> > > > > R10: 0000000000000000 R11: 0000000000000206 R12: 0000000001fd5fd0
> > > > > R13: 0000000000000007 R14: 0000000001fc6250 R15: 0000000000000005
> > > > >
> > > > >
> > > > > ---
> > > > > This bug 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 syzkaller@xxxxxxxxxxxxxxxxx
> > > > >
> > > > > syzbot will keep track of this bug report. See:
> > > > > https://goo.gl/tpsmEJ#bug-status-tracking for how to communicate with
> > > > > syzbot.
> > > > > syzbot can test patches for this bug, for details see:
> > > > > https://goo.gl/tpsmEJ#testing-patches
> > > >
> > >
> > > It's apparently the bug in the io_uring patchset I reported yesterday (well, I
> > > stole it from another open syzbot bug...) and Jens is already planning to fix:
> > > https://marc.info/?l=linux-api&m=155115288114046&w=2. Reproducer is similar,
> > > and the crash bisects down to the same commit from the io_uring patchset:
> > > "block: implement bio helper to add iter bvec pages to bio".
> > >
> >
> > Fixed in next-20190227. The fix was folded into "block: implement bio helper to
> > add iter bvec pages to bio". Telling syzbot to invalidate this bug report:
> >
> > #syz invalid
>
> Was this discovered separately? We could also add Reported-by (or
> Tested-by) tag to the commit.
>

My report was based on a crash from the syzbot dashboard. However, there's no
fixing commit, as the fix was folded into the original patch. I.e. the mainline
git history (if/when the io_uring stuff is actually merged) won't show the bug
ever being introduced. Thus Reported-by isn't appropriate, and I used '#syz
invalid' instead of '#syz fix'. Nor did syzbot specifically test the new
version of the patch beyond fuzzing the next day's linux-next... So while I
personally might have added an informal note in the commit message, I don't
think those formal tags make sense for folded-in linux-next fixes like this.

- Eric