Control: found -1 1:7.2+dfsg-7+deb12u14
On 25.08.2025 18:46, Moritz Poldrack wrote:
Package: qemu-system-x86
Version: 1:7.2+dfsg-7+deb12u15
Source: qemu
This occurs during live-migrating a guest onto a host with u15, migrating it
back fixes the softlocks. A reset is required to fix it but is only required
when the receiving host is on the latest version.
Last known good version: u13.
Ok, so this is a regression in 1:7.2+dfsg-7+deb12u14.
Does it work when migrating between two deb12u14 (or deb12u15, since
u14 and u15 is the same when qemu-system is concerned) ?
deb12u14 is 7.2.18, and it has a few migration- and virtio-net-related
changes:
https://salsa.debian.org/qemu-team/qemu/-/commits/v7.2.18?ref_type=tags
virtio: Call set_features during reset
https://salsa.debian.org/qemu-team/qemu/-/commit/ce1431615292dc735597db4062834bfb271381bc
virtio-net: Fix num_buffers for version 1
https://salsa.debian.org/qemu-team/qemu/-/commit/cefd67f2543045cfebe51f9e6feccfa56cda9797
migration: fix SEEK_CUR offset calculation in qio_channel_block_seek
https://salsa.debian.org/qemu-team/qemu/-/commit/d34c201195ecea4fa5c3328a8f5507ca6d2c0a96
The most interesting commit is the 2nd one above, "Fix num_buffers".
Are you able to (re)build qemu from the .dsc? If yes, can you please
create a reverse patch for that commit and verify it if will work?
I don't have an (easy) way to reproduce the prob sine I don't have
enough hosts to perform migration. Maybe I can do it between two
containers on the same machine, but it'll take some time to set up.
Thanks,
/mjt
