Re: [libvirt PATCH 1/2] virQEMUCapsGetMachineTypesCaps: Use GPtrArray

On Fri, 2021-07-09 at 12:05 +0200, Michal Prívozník wrote: > On 7/8/21 4:28 PM, Tim Wiederhake wrote: >> This simplyfies the code a bit and removes one "goto", one >> "VIR_FREE", >> and one "VIR_INSERT_ELEMENT_COPY". >> >> Signed-off-by: Tim Wiederhake <twiederh(a)redhat.com&gt; >> --- >>  src/qemu/qemu_capabilities.c | 37 ++++++++++++++++++---------------- >> -- >>  1 file changed, 18 insertions(+), 19 deletions(-) >> > > I'm not exactly sure what is causing this, but with this patch libvirtd > crashes for me when I try to fetch capabilities: > > ==16567== Thread 3 rpc-worker: > ==16567== Invalid read of size 8 > ==16567==    at 0x49CB01A: virCapabilitiesFormatGuestXML > (capabilities.c:1259) > ==16567==    by 0x49CB6AB: virCapabilitiesFormatXML > (capabilities.c:1355) > ==16567==    by 0xAE898B1: qemuConnectGetCapabilities > (qemu_driver.c:1316) > ==16567==    by 0x4C47014: virConnectGetCapabilities (libvirt- > host.c:467) > ==16567==    by 0x1328FD: remoteDispatchConnectGetCapabilities > (remote_daemon_dispatch_stubs.h:766) > ==16567==    by 0x1328A5: remoteDispatchConnectGetCapabilitiesHelper > (remote_daemon_dispatch_stubs.h:748) > ==16567==    by 0x4AB4C0F: virNetServerProgramDispatchCall > (virnetserverprogram.c:428) > ==16567==    by 0x4AB478A: virNetServerProgramDispatch > (virnetserverprogram.c:302) > ==16567==    by 0x4ABBE71: virNetServerProcessMsg (virnetserver.c:135) > ==16567==    by 0x4ABBF31: virNetServerHandleJob (virnetserver.c:152) > ==16567==    by 0x49AC6D5: virThreadPoolWorker (virthreadpool.c:159) > ==16567==    by 0x49ABBEB: virThreadHelper (virthread.c:241) > ==16567==  Address 0x8 is not stack'd, malloc'd or (recently) free'd > > Michal > Weird. Pipeline passed for me: https://gitlab.com/twiederh/libvirt/-/pipelines/333827544 Do you maybe have a reproducer for me, so I can investigate what's going on here?