On 01/18/2013 10:19 AM, Eric Blake wrote:
On 01/18/2013 08:45 AM, Viktor Mihajlovski wrote:
This should have been commit 56fd513 already, but was missed by initially. The driver unlock call in the cleanup section of
s/by //
DomainManagedSave does actually belong to DomainSendKey.
Please also call out commit 8c5d2ba as the source of the original problem.
@@ -3174,7 +3175,6 @@ qemuDomainManagedSave(virDomainPtr dom, unsigned int flags) cleanup: if (vm) virObjectUnlock(vm); - qemuDriverUnlock(driver); VIR_FREE(name);
This is in qemuDomainManagedSave, which was not mentioned in the commit message of 8c5d2ba, was not touched in 56fd513, but does fix a bug (unlocking without owning the lock). Now I have to go dig which commit introduced _this_ problem...
Found it - commit 2745177 was the culprit here.
Scratch that; I just reread the function. qemuDomainManagedSave used qemuDomObjFromDomainDriver(), which returns driver locked; so commit 2745177 is correct, and there is no locking bug here, and _your_ hunk would make it wrong.
Yuck. You _still_ haven't fixed DomainHasManagedSaveImage (which 8c5d2ba claimed to fix), and we've found yet another bogus commit. We need a v2 of this patch that scrubs ALL of the bugs at once.
This is still true - we need a v2 of the patch that fixes all of the bugs, it's just that there are fewer bugs than I was worried about, because right now it looks like there is only one bogus commit, not two. -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org