On Thu, Aug 29, 2024 at 11:09:44AM +0530, Prasad Pandit wrote: > On Wed, 28 Aug 2024 at 16:45, Michael S. Tsirkin <m...@redhat.com> wrote: > > > - Place QEMU_LOCK_GUARD near the vhost_user_write() calls, holding > > > the lock for longer fails some tests during rpmbuild(8). > > > > what do you mean fails rpmbuild? that qemu with this patch can not be > > compiled? > > * In V1 of this patch, QEMU_LOCK_GUARD was placed near beginning of > the function. But that caused some unit tests to fail reporting > TIMEOUT errors. In this V2, QEMU_LOCK_GUARD is placed near > vhost_user_write() calls, to reduce the time that lock is held. > > * Both (V1 & V2) compile well, but fail at '%check' stage while > running unit tests (on some machines), ie. rpm package is not built. > rpmbuild(8) on F40 machine failed, but koji scratch build with the > same SRPM worked fine. Those scratch builds are shared above. RHEL-9 > SRPM built well on RHEL-9 host, but failed to build on F40 machine > reporting failure at '%check' stage of rpmbuild(8). > > Thank you. > --- > - Prasad
Weird. Seems to indicate some kind of deadlock? -- MST
