[libvirt] Re: [PATCH 0/12] Standardized device addressing & SCSI controller/disk hotplug

Daniel P. Berrange wrote: > On Tue, Dec 15, 2009 at 01:06:51PM +0100, Wolfgang Mauerer wrote: >> Hi, >> >> Daniel P. Berrange wrote: >>> This patch series is a combination of series done by >>> Wolfgang Mauerer to support proper SCSI drive hotplug >>> and new work by myself to introduce generic addressing >>> for all devices. >>> >>> Wolfgang's most recent posting was >>> >>> http://www.redhat.com/archives/libvir-list/2009-November/msg00574.html >>> http://www.redhat.com/archives/libvir-list/2009-November/msg00701.html >>> >>> When testing that series I came across a few minor issues, >>> but more importantly it made me realize how important it is >>> that we introduce explicit device addressing in our XML format. >>> (...) >> essentially, this all looks good to me - thanks for the extensions. >> Unfortunately, I was not yet successful in fully testing the code >> because I have some issues with the underlying qemu that prevent my >> machines from booting correctly with recent qemu-kvms. However, I came >> across a small issue: When certain PCI devices are not present in the >> system, libvirt can crash at startup. The attached patch fixes this. > > Hmm, which particular device did you see it crash on? These > functions are only called with 'def' straight out of the > virDomainDefPtr struct, (eg, vm->def->disks[INDEX]) and those > are all supposed to be non-NULL at all times. So if one is NULL > then I've introduced a bug somewhere else ! I guess most likely > in the new controllers code. it crashed with a missing watchdog. I realise that the other checks are not strictly necessary when users stick to the calling conventions, but figured that one check too many in the startup phase doesn't do any harm and is better than a crash.