diff options
| author | Christian Krinitsin <mail@krinitsin.com> | 2025-06-03 12:04:13 +0000 |
|---|---|---|
| committer | Christian Krinitsin <mail@krinitsin.com> | 2025-06-03 12:04:13 +0000 |
| commit | 256709d2eb3fd80d768a99964be5caa61effa2a0 (patch) | |
| tree | 05b2352fba70923126836a64b6a0de43902e976a /results/classifier/105/assembly/1605 | |
| parent | 2ab14fa96a6c5484b5e4ba8337551bb8dcc79cc5 (diff) | |
| download | qemu-analysis-256709d2eb3fd80d768a99964be5caa61effa2a0.tar.gz qemu-analysis-256709d2eb3fd80d768a99964be5caa61effa2a0.zip | |
add new classifier result
Diffstat (limited to 'results/classifier/105/assembly/1605')
| -rw-r--r-- | results/classifier/105/assembly/1605 | 51 |
1 files changed, 51 insertions, 0 deletions
diff --git a/results/classifier/105/assembly/1605 b/results/classifier/105/assembly/1605 new file mode 100644 index 000000000..e18742c75 --- /dev/null +++ b/results/classifier/105/assembly/1605 @@ -0,0 +1,51 @@ +assembly: 0.734 +instruction: 0.734 +other: 0.725 +graphic: 0.703 +mistranslation: 0.695 +KVM: 0.688 +semantic: 0.681 +vnc: 0.676 +boot: 0.670 +socket: 0.634 +device: 0.634 +network: 0.554 + +On windows, 2nd kind vhdx-dyn bug, crash on Unexpected error in bdrv_check_qiov_request() in io.c +Description of problem: +On windows, 2nd kind vhdx-dyn bug, crash on Unexpected error in bdrv_check_qiov_request() in io.c +- qemu windows crashes during data copy + ```D:\tmpq\qemu\8.0.0-rc4\qemu\qemu-system-x86_64 -cpu qemu64 -m 4096 -machine "type=q35,kernel-irqchip=off" -accel whpx -smp "sockets=1,cores=8,threads=1" -bios D:\vstorage\win_m01_edk2-x8_64.fd -boot c -drive "index=0,if=virtio,media=disk,format=raw,file=D:\vstorage\m01_bootnoefi.raw.img" -drive "index=1,if=virtio,media=disk,format=raw,file=F:\m01_lnx.raw.img.vtoy" -drive "index=2,if=virtio,media=disk,format=vhdx,file=F:\gkpics01.vhdx" -drive "index=3,if=virtio,media=disk,format=vhdx,file=D:\test\sgdata.vhdx" -display sdl -vga virtio -rtc base=utc -netdev user,id=vmnic1,net=192.168.20.0/24,dns=192.168.20.3,dhcpstart=192.168.20.15,hostfwd=tcp::9551-:22 -device virtio-net,netdev=vmnic1 -chardev qemu-vdagent,id=ch1,name=vdagent,clipboard=on -device virtio-serial -device virtserialport,chardev=ch1,id=ch1,name=com.redhat.spice.0 -qmp "tcp:127.0.0.1:5955,server,nowait"``` + ``` ``` + ```Windows Hypervisor Platform accelerator is operational``` + ```Unexpected error in bdrv_check_qiov_request() at ../../../block/io.c:815:``` + ```D:\tmpq\qemu\8.0.0-rc4\qemu\qemu-system-x86_64.exe: offset is negative: -28983296``` + +. +- The **LINE NUMBER** : https://gitlab.com/qemu-project/qemu/-/blob/master/block/io.c#L815 +- qemu setup is ```qemu-w64-setup-20230414.exe ``` +Steps to reproduce: +1. have fresh vhdx ready create a vhdx in ```diskmgmt``` (also attached to [comment](https://gitlab.com/qemu-project/qemu/-/issues/727#note_1346341805)) +2. have vhdx with synthetic generated data ready (see process to generate sgdata in [comment](https://gitlab.com/qemu-project/qemu/-/issues/727#note_739930694) ) +3. start qemu, login, open terminal +4. Inside VM, start a terminal window, sudo root, +5. open```gdisk /dev/vdc``` create a ntfs partition +6. format as ntfs: ```mkfs.ntfs -Q -L fs_gkpics01 /dev/vdc1``` +7. mount the partition ```mount -t ntfs3 /dev/vdc1 /mnt/a -o uid=1000,gid=1000,defaults,umask=0002``` +8. mount the partition ```mount -t ntfs3 /dev/vdd2 /mnt/b -o uid=1000,gid=1000,defaults,umask=0002``` +9. In a user login, do rsync data-copy step + ```( fl="photos001" ; src="/mnt/b/sgdata" ; dst="/mnt/a" ; sdate=`date` ; echo "$sdate" ; cd "$src" ; rsync -avH "$fl" "$dst" ; echo "$sdate" ; date ; sudo -u gana DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/1000/bus DISPLAY=:0.0 -- notify-send "$src/$fl" "rsync $src/$fl" )``` + + +The bug is easily reproducible. +The moment of the crash may seems spurious, but is almost certainly bound to happen. +When it happens, it can be seen to be the same error message. +Sometimes the crash happens in ```gdisk``` step, sometimes during ```mkfs.ntfs``` sometimes partway through the ```rsync```-copy, not very long into it. +Additional information: +- This has been happening for some time. I haven't used/tested vhdx much in windows much since 7.0.0 on account of other corruption bugs/lack of dependability. +- This does not happen in Linux, as tested in #727 +- The fix of #727 is unrelated to this. It doesn't have the same feel/reproduction intuitive-signature. + - Happens before (on doing the same test) + - on 8.0.0-rc1 (line number of io.c there is L811) + - on 7.2.0 (line no of io.c there is [L971](https://gitlab.com/qemu-project/qemu/-/blob/ace5a161ea1c09d8eaa8b2a717528457dc924e83/block/io.c#L971)) +- It may be caused by other changes going into block code since 7.0 . |