Re: [syzbot] [mm] KMSAN: uninit-value in virtqueue_add (4)

[syzbot <syzbot+d7521c1e3841ed075a42@syzkaller.appspotmail.com>]

Hello,

syzbot has tested the proposed patch but the reproducer is still triggering an issue:
KMSAN: uninit-value in virtqueue_add

=====================================================
BUG: KMSAN: uninit-value in vring_map_one_sg drivers/virtio/virtio_ring.c:380 [inline]
BUG: KMSAN: uninit-value in virtqueue_add_split drivers/virtio/virtio_ring.c:614 [inline]
BUG: KMSAN: uninit-value in virtqueue_add+0x21c6/0x6530 drivers/virtio/virtio_ring.c:2210
 vring_map_one_sg drivers/virtio/virtio_ring.c:380 [inline]
 virtqueue_add_split drivers/virtio/virtio_ring.c:614 [inline]
 virtqueue_add+0x21c6/0x6530 drivers/virtio/virtio_ring.c:2210
 virtqueue_add_sgs+0x186/0x1a0 drivers/virtio/virtio_ring.c:2244
 __virtscsi_add_cmd drivers/scsi/virtio_scsi.c:467 [inline]
 virtscsi_add_cmd+0x817/0xa90 drivers/scsi/virtio_scsi.c:501
 virtscsi_queuecommand+0x896/0xa60 drivers/scsi/virtio_scsi.c:598
 scsi_dispatch_cmd drivers/scsi/scsi_lib.c:1516 [inline]
 scsi_queue_rq+0x4874/0x5790 drivers/scsi/scsi_lib.c:1758
 blk_mq_dispatch_rq_list+0x13f8/0x3600 block/blk-mq.c:2049
 __blk_mq_do_dispatch_sched block/blk-mq-sched.c:170 [inline]
 blk_mq_do_dispatch_sched block/blk-mq-sched.c:184 [inline]
 __blk_mq_sched_dispatch_requests+0x10af/0x2500 block/blk-mq-sched.c:309
 blk_mq_sched_dispatch_requests+0x160/0x2d0 block/blk-mq-sched.c:333
 blk_mq_run_work_fn+0xd0/0x280 block/blk-mq.c:2434
 process_one_work kernel/workqueue.c:2627 [inline]
 process_scheduled_works+0x104e/0x1e70 kernel/workqueue.c:2700
 worker_thread+0xf45/0x1490 kernel/workqueue.c:2781
 kthread+0x3ed/0x540 kernel/kthread.c:388
 ret_from_fork+0x66/0x80 arch/x86/kernel/process.c:147
 ret_from_fork_asm+0x11/0x20 arch/x86/entry/entry_64.S:242

Uninit was created at:
 __alloc_pages+0x9a4/0xe00 mm/page_alloc.c:4591
 alloc_pages_mpol+0x62b/0x9d0 mm/mempolicy.c:2133
 alloc_pages mm/mempolicy.c:2204 [inline]
 folio_alloc+0x1da/0x380 mm/mempolicy.c:2211
 filemap_alloc_folio+0xa5/0x430 mm/filemap.c:974
 __filemap_get_folio+0xa5a/0x1760 mm/filemap.c:1918
 ext4_da_write_begin+0x7f8/0xec0 fs/ext4/inode.c:2891
 generic_perform_write+0x3f5/0xc40 mm/filemap.c:3927
 ext4_buffered_write_iter+0x564/0xaa0 fs/ext4/file.c:299
 ext4_file_write_iter+0x20f/0x3460
 __kernel_write_iter+0x329/0x930 fs/read_write.c:517
 dump_emit_page fs/coredump.c:888 [inline]
 dump_user_range+0x593/0xcd0 fs/coredump.c:915
 elf_core_dump+0x528d/0x5a40 fs/binfmt_elf.c:2077
 do_coredump+0x32c9/0x4920 fs/coredump.c:764
 get_signal+0x2185/0x2d10 kernel/signal.c:2890
 arch_do_signal_or_restart+0x53/0xca0 arch/x86/kernel/signal.c:309
 exit_to_user_mode_loop+0xe8/0x320 kernel/entry/common.c:168
 exit_to_user_mode_prepare+0x163/0x220 kernel/entry/common.c:204
 irqentry_exit_to_user_mode+0xd/0x30 kernel/entry/common.c:309
 irqentry_exit+0x16/0x40 kernel/entry/common.c:412
 exc_page_fault+0x246/0x6f0 arch/x86/mm/fault.c:1566
 asm_exc_page_fault+0x2b/0x30 arch/x86/include/asm/idtentry.h:570

Bytes 0-1023 of 1024 are uninitialized
Memory access of size 1024 starts at ffff88801e7d9c00

CPU: 0 PID: 52 Comm: kworker/0:1H Not tainted 6.7.0-syzkaller-dirty #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/25/2024
Workqueue: kblockd blk_mq_run_work_fn
=====================================================


Tested on:

commit:         0dd3ee31 Linux 6.7
git tree:       https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git v6.7
console output: https://syzkaller.appspot.com/x/log.txt?x=15dee522180000
kernel config:  https://syzkaller.appspot.com/x/.config?x=373206b1ae2fe3d4
dashboard link: https://syzkaller.appspot.com/bug?extid=d7521c1e3841ed075a42
compiler:       Debian clang version 15.0.6, GNU ld (GNU Binutils for Debian) 2.40
patch:          https://syzkaller.appspot.com/x/patch.diff?x=1524ca02180000