Message-ID: <9555d914-1d5a-e8aa-a67c-e84bd869e312@redhat.com>
Date:   Tue, 25 Oct 2022 10:54:01 +0100
From:   Andrew Price <anprice@...hat.com>
To:     agruenba@...hat.com, rpeterso@...hat.com
Cc:     syzbot <syzbot+f8bc4176e51e87e0928f@...kaller.appspotmail.com>,
        linux-kernel@...r.kernel.org, syzkaller-bugs@...glegroups.com,
        cluster-devel@...hat.com
Subject: Re: [Cluster-devel] [syzbot] WARNING in gfs2_ri_update

On 22/10/2022 12:11, syzbot wrote:
> syzbot has found a reproducer for the following issue on:
> 
> HEAD commit:    440b7895c990 Merge tag 'mm-hotfixes-stable-2022-10-20' of ..
> git tree:       upstream
> console+strace: https://syzkaller.appspot.com/x/log.txt?x=128087a4880000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=afc317c0f52ce670
> dashboard link: https://syzkaller.appspot.com/bug?extid=f8bc4176e51e87e0928f
> compiler:       Debian clang version 13.0.1-++20220126092033+75e33f71c2da-1~exp1~20220126212112.63, GNU ld (GNU Binutils for Debian) 2.35.2
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=13ed90f2880000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=16717fc2880000
> 
> Downloadable assets:
> disk image: https://storage.googleapis.com/syzbot-assets/105038975fc9/disk-440b7895.raw.xz
> vmlinux: https://storage.googleapis.com/syzbot-assets/edd7302c8fc8/vmlinux-440b7895.xz
> mounted in repro: https://storage.googleapis.com/syzbot-assets/95d6d27d2d50/mount_0.gz
> 
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+f8bc4176e51e87e0928f@...kaller.appspotmail.com
> 
> gfs2: fsid=syz:syz: Now mounting FS (format 1801)...
> gfs2: fsid=syz:syz.0: journal 0 mapped with 1 extents in 0ms
> gfs2: fsid=syz:syz.0: first mount done, others may mount
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 3611 at mm/page_alloc.c:5530 __alloc_pages+0x30a/0x560 mm/page_alloc.c:5530
> Modules linked in:
> CPU: 1 PID: 3611 Comm: syz-executor282 Not tainted 6.1.0-rc1-syzkaller-00158-g440b7895c990 #0
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/11/2022
> RIP: 0010:__alloc_pages+0x30a/0x560 mm/page_alloc.c:5530
> Code: 5c 24 04 0f 85 f3 00 00 00 44 89 e1 81 e1 7f ff ff ff a9 00 00 04 00 41 0f 44 cc 41 89 cc e9 e3 00 00 00 c6 05 a1 ab 29 0c 01 <0f> 0b 83 fb 0a 0f 86 c8 fd ff ff 31 db 48 c7 44 24 20 0e 36 e0 45
> RSP: 0018:ffffc90003c5f4a0 EFLAGS: 00010246
> RAX: ffffc90003c5f500 RBX: 0000000000000012 RCX: 0000000000000000
> RDX: 0000000000000028 RSI: 0000000000000000 RDI: ffffc90003c5f528
> RBP: ffffc90003c5f5b8 R08: dffffc0000000000 R09: ffffc90003c5f500
> R10: fffff5200078bea5 R11: 1ffff9200078bea0 R12: 0000000000040d40
> R13: 1ffff9200078be9c R14: dffffc0000000000 R15: 1ffff9200078be98
> FS:  0000555555781300(0000) GS:ffff8880b9b00000(0000) knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 00000000005d84c8 CR3: 0000000079874000 CR4: 00000000003506e0
> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> Call Trace:
>   <TASK>
>   __alloc_pages_node include/linux/gfp.h:223 [inline]
>   alloc_pages_node include/linux/gfp.h:246 [inline]
>   __kmalloc_large_node+0x8a/0x1a0 mm/slab_common.c:1098
>   __do_kmalloc_node mm/slab_common.c:943 [inline]
>   __kmalloc+0xfe/0x1a0 mm/slab_common.c:968
>   kmalloc_array include/linux/slab.h:628 [inline]
>   kcalloc include/linux/slab.h:659 [inline]
>   compute_bitstructs fs/gfs2/rgrp.c:766 [inline]
>   read_rindex_entry fs/gfs2/rgrp.c:931 [inline]

The fuzzer is writing 16777217 to an rindex entry's ri_length field so 
compute_bitstructs() is asking kmalloc for (16777217 * 40) bytes of 
memory, to which it says nope and dumps a warning.

Simpler version:
# mkfs.gfs2 -p lock_nolock -o align=0 /dev/vdc
# printf '\x01\x00\x00\x01' | dd of=/dev/vdc bs=1 seek=135676008 count=4
# mount /dev/vdc /mnt/test

I have some experimental code to add integrity checking to rindex 
entries[0] that would catch this but it would require an sb_fs_format 
bump before it can treat ri_crc == 0 usefully.

0. 
https://gitlab.com/andyprice/linux/-/commit/4e0557916733e59b2e6795775e94c9b733b51f6b

Andy

>   gfs2_ri_update+0x537/0x17f0 fs/gfs2/rgrp.c:1001
>   gfs2_rindex_update+0x313/0x3f0 fs/gfs2/rgrp.c:1051
>   init_inodes+0x242/0x340 fs/gfs2/ops_fstype.c:917
>   gfs2_fill_super+0x1bb2/0x2700 fs/gfs2/ops_fstype.c:1247
>   get_tree_bdev+0x400/0x620 fs/super.c:1323
>   gfs2_get_tree+0x50/0x210 fs/gfs2/ops_fstype.c:1330
>   vfs_get_tree+0x88/0x270 fs/super.c:1530
>   do_new_mount+0x289/0xad0 fs/namespace.c:3040
>   do_mount fs/namespace.c:3383 [inline]
>   __do_sys_mount fs/namespace.c:3591 [inline]
>   __se_sys_mount+0x2d3/0x3c0 fs/namespace.c:3568
>   do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>   do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
>   entry_SYSCALL_64_after_hwframe+0x63/0xcd
> RIP: 0033:0x7fb4bd4a797a
> Code: 83 c4 08 5b 5d c3 66 2e 0f 1f 84 00 00 00 00 00 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48
> RSP: 002b:00007ffc73f0a0b8 EFLAGS: 00000282 ORIG_RAX: 00000000000000a5
> RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 00007fb4bd4a797a
> RDX: 0000000020000000 RSI: 0000000020000100 RDI: 00007ffc73f0a0d0
> RBP: 00007ffc73f0a0d0 R08: 00007ffc73f0a110 R09: 00005555557812c0
> R10: 0000000000000000 R11: 0000000000000282 R12: 0000000000000004
> R13: 00007ffc73f0a110 R14: 000000000000091b R15: 000000002000dc88
>   </TASK>
> 

Powered by blists - more mailing lists