Description of problem: Running with upstream qemu I noticed errors when running avocado. It was caused by running: qemu-system-x86_64 -device scsi-hd,help which crashed. First bad commit is: commit ea9ce8934c5d2cc8925359a6d8d45eb829a9f27f (HEAD, refs/bisect/bad) Author: Marc-André Lureau <marcandre.lureau> Date: Mon Nov 26 22:04:32 2018 +0400 hw: apply accel compat properties without touching globals Instead of registering compat properties as globals, let's keep them in their own array, to avoid mixing with user globals. Introduce object_apply_global_props() function, to apply compatibility properties from a GPtrArray. Signed-off-by: Marc-André Lureau <marcandre.lureau> Reviewed-by: Igor Mammedov <imammedo> Reviewed-by: Cornelia Huck <cohuck> Acked-by: Eduardo Habkost <ehabkost> Version-Release number of selected component (if applicable): upstream qemu ea9ce8934c5d2cc8925359a6d8d45eb829a9f27f How reproducible: always (tested on s390 and x86_64) Steps to Reproduce: 1. clone qemu from git 2. git checkout ea9ce8934c5d2cc8925359a6d8d45eb829a9f27f 3. make 4. ./x86_64-softmmu/qemu-system-x86_64 -device scsi-hd,hel Actual results: Segmentation fault (core dumped [obraz paměti uložen]) Expected results: scsi-hd options: min_io_size=<uint16> secs=<uint32> removable=<bool> - on/off channel=<uint32> ...
Sent to Qemu ML: [PATCH] qdev: fix -device scsi-hd,help regression
(In reply to Marc-Andre Lureau from comment #3) > Sent to Qemu ML: > [PATCH] qdev: fix -device scsi-hd,help regression This is still a downstream BZ, so we can't move bugs to POST until it's either merged upstream (so we add "Fixed in version" here) or backported.