qemu 7.1: assertion faillure with virtio-scsi in `blk_set_enable_write_cache`

Host environment

• Operating system: Debian 11
• OS/kernel version: 5.15.74
• Architecture: x86_64
• QEMU flavor: qemu-system-x86_64
• QEMU version: v7.1.0
• QEMU command line:

   ./qemu-system-x86_64 -enable-kvm -bios OVMF.fd -m 4G \
    -object iothread,id=io \
    -device virtio-scsi-pci,id=primary_stor_ctrl,iothread=io \
    -device scsi-hd,lun=1,drive=disk,bus=primary_stor_ctrl.0,id=disk \
    -drive file=/dev/loop0,if=none,format=raw,id=disk,cache=writethrough

Emulated/Virtualized environment

• Operating system: Windows 10
• OS/kernel version: 19041
• Architecture: x86_64
• virtio-win: 0.1.225

Description of problem

During the guest boot qemu crashes with the following error:

qemu-system-x86_64: ../src/block/block-backend.c:1949: blk_set_enable_write_cache: Assertion `qemu_in_main_thread()' failed.

Steps to reproduce

1. Start a windows guest

Additional information

Stacktrace:

#0  0x00007fd6c3515ce1 in raise () from /lib/x86_64-linux-gnu/libc.so.6
#1  0x00007fd6c34ff537 in abort () from /lib/x86_64-linux-gnu/libc.so.6
#2  0x00007fd6c34ff40f in ?? () from /lib/x86_64-linux-gnu/libc.so.6
#3  0x00007fd6c350e662 in __assert_fail () from /lib/x86_64-linux-gnu/libc.so.6
#4  0x000056149e2cea03 in blk_set_enable_write_cache (wce=true, blk=0x5614a01c27f0) at ../src/block/block-backend.c:1949
#5  0x000056149e2d0a67 in blk_set_enable_write_cache (blk=0x5614a01c27f0, wce=<optimized out>) at ../src/block/block-backend.c:1951
#6  0x000056149dfe9c59 in scsi_disk_apply_mode_select (p=0x7fd6b400c00e "\004", page=<optimized out>, s=<optimized out>) at ../src/hw/scsi/scsi-disk.c:1520
#7  mode_select_pages (change=true, len=18, p=0x7fd6b400c00e "\004", r=0x7fd6b4001ff0) at ../src/hw/scsi/scsi-disk.c:1570
#8  scsi_disk_emulate_mode_select (inbuf=<optimized out>, r=0x7fd6b4001ff0) at ../src/hw/scsi/scsi-disk.c:1640
#9  scsi_disk_emulate_write_data (req=0x7fd6b4001ff0) at ../src/hw/scsi/scsi-disk.c:1934
#10 0x000056149e18ff16 in virtio_scsi_handle_cmd_req_submit (req=<optimized out>, req=<optimized out>, s=0x5614a12f16b0) at ../src/hw/scsi/virtio-scsi.c:719
#11 virtio_scsi_handle_cmd_vq (vq=0x7fd6bab92140, s=0x5614a12f16b0) at ../src/hw/scsi/virtio-scsi.c:761
#12 virtio_scsi_handle_cmd (vq=<optimized out>, vdev=<optimized out>) at ../src/hw/scsi/virtio-scsi.c:775
#13 virtio_scsi_handle_cmd (vdev=0x5614a12f16b0, vq=0x7fd6bab92140) at ../src/hw/scsi/virtio-scsi.c:765
#14 0x000056149e1a8aa6 in virtio_queue_notify_vq (vq=0x7fd6bab92140) at ../src/hw/virtio/virtio.c:2365
#15 0x000056149e3ccea5 in aio_dispatch_handler (ctx=ctx@entry=0x5614a01babe0, node=<optimized out>) at ../src/util/aio-posix.c:369
#16 0x000056149e3cd868 in aio_dispatch_ready_handlers (ready_list=0x7fd6c09b2680, ctx=0x5614a01babe0) at ../src/util/aio-posix.c:399
#17 aio_poll (ctx=0x5614a01babe0, blocking=blocking@entry=true) at ../src/util/aio-posix.c:713
#18 0x000056149e2a7796 in iothread_run (opaque=opaque@entry=0x56149ffde500) at ../src/iothread.c:67
#19 0x000056149e3d0859 in qemu_thread_start (args=0x7fd6c09b26f0) at ../src/util/qemu-thread-posix.c:504
#20 0x00007fd6c36b9ea7 in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#21 0x00007fd6c35d9aef in clone () from /lib/x86_64-linux-gnu/libc.so.6

The crash was bisected to:

commit b1c073490553f80594b903ceedfc7c1aef6b1b19
Author: Hanna Reitz <hreitz@redhat.com>
Date:   Tue Mar 29 11:35:45 2022 +0200

    main-loop: Disable GLOBAL_STATE_CODE() assertions

I have not been able to reproduce the bug with a linux guest nor with a fresh windows installation.

The crashes appears with either writethrough and directsync cache modes but not with writeback none and unsafe.

Note: if needed I can extract (privately) provide a disk image demonstrating the behavior