On 01/28/14 17:12, Dr. David Alan Gilbert wrote: > * Michael S. Tsirkin (m...@redhat.com) wrote: >> On Tue, Jan 28, 2014 at 03:20:39PM +0000, Dr. David Alan Gilbert (git) wrote: >>> From: "Dr. David Alan Gilbert" <dgilb...@redhat.com> >>> >>> If enabled, set the thread name at creation (on GNU systems with >>> pthread_set_np) >>> Fix up all the callers with a thread name >>> >>> Signed-off-by: Dr. David Alan Gilbert <dgilb...@redhat.com> >> >> Thanks for the patch. >> >> It worries me that tool might start assuming specific >> thread names - this effectively becomes part of >> management interface. >> >> We avoided this in the past except for VCPU threads - >> in particular we only expose thread id for VCPU threads. >> How about some generic name for non-VCPU threads >> to avoid this issue? > > Since I'm doing migration development, restriction to VCPU > threads doesn't help me much.
I'm not doing migration development, but I agree that the feature is only really useful if *all* threads have names. (IOW when it completely saves the developer the work to figure out which thread is which.) > Putting big scary warnings somewhere (where?) to say that > the names aren't guaranteed is all I can think of. > (I did put that warning in the cover letter). > > I guess I could change the option name to debug-threads > to make it clear it's for debug. > >> Also - should we put VCPU # in the thread name? > > Yeh that's something I could add. Would be very useful. Thanks Laszlo
