On 24/06/2016 10:12, Peter Maydell wrote: > On 24 June 2016 at 07:36, Paolo Bonzini <pbonz...@redhat.com> wrote: >> Mark, perhaps you can try to use migration to reduce the amount of >> logging? (Start QEMU with -snapshot, try to stop the vm before it >> fails. If you succeed, do a "migrate exec:cat>foo.sav" followed by >> "commit"; if you fail, try again). > > Why drag migration into it? I usually use 'savevm' and then > the -loadvm command line argument for this. (You need a > qcow2 disk image.)
Well, migration and savevm are the same. :) In this case IIUC the failure happens from a CDROM so migration lets you avoid the qcow2 image. In general I find it easier to manage migration files on disk than saved snapshots. Paolo >> It would be nice to have a mechanism to stop the VM after executing N >> basic blocks. Binary search on this value then can help with coming up >> with a more easily debuggable snapshot, possibly to a point where the >> difference between pre-patch and post-patch becomes deterministic. > > You can use the monitor and an expect script to say "take a > snapshot 0.7 seconds into boot", which I've found to be > a good enough approximation: > > https://translatedcode.wordpress.com/2015/07/06/tricks-for-debugging-qemu-savevm-snapshots/ > > thanks > -- PMM >