Re: [Qemu-devel] [PATCH] iothread: Stop threads before main() quits

2016-09-13 Thread Fam Zheng
On Tue, 09/13 09:39, Stefan Hajnoczi wrote: > On Thu, Sep 08, 2016 at 05:28:51PM +0800, Fam Zheng wrote: > > Right after main_loop ends, we release various things but keep iothread > > alive. The latter is not prepared to the sudden change of resources. > > > > Specifically, after

Re: [Qemu-devel] [PATCH] iothread: Stop threads before main() quits

2016-09-13 Thread Stefan Hajnoczi
On Thu, Sep 08, 2016 at 05:28:51PM +0800, Fam Zheng wrote: > Right after main_loop ends, we release various things but keep iothread > alive. The latter is not prepared to the sudden change of resources. > > Specifically, after bdrv_close_all(), virtio-scsi dataplane get a > surprise at the empty

Re: [Qemu-devel] [PATCH] iothread: Stop threads before main() quits

2016-09-13 Thread Stefan Hajnoczi
On Thu, Sep 08, 2016 at 05:28:51PM +0800, Fam Zheng wrote: > Right after main_loop ends, we release various things but keep iothread > alive. The latter is not prepared to the sudden change of resources. > > Specifically, after bdrv_close_all(), virtio-scsi dataplane get a > surprise at the empty

Re: [Qemu-devel] [PATCH] iothread: Stop threads before main() quits

2016-09-09 Thread Paolo Bonzini
> > Reviewed-by: Paolo Bonzini > > Thanks! Is there a coming PULL from you that this patch could sneak in? :) There is (next week), but iothread.c is one of those grey area of maintenance that I wouldn't mind passing to someone else. Such as the owner of block/io.c. :)

Re: [Qemu-devel] [PATCH] iothread: Stop threads before main() quits

2016-09-08 Thread Fam Zheng
On Thu, 09/08 11:34, Paolo Bonzini wrote: > > > On 08/09/2016 11:28, Fam Zheng wrote: > > Right after main_loop ends, we release various things but keep iothread > > alive. The latter is not prepared to the sudden change of resources. > > > > Specifically, after bdrv_close_all(), virtio-scsi

Re: [Qemu-devel] [PATCH] iothread: Stop threads before main() quits

2016-09-08 Thread Paolo Bonzini
On 08/09/2016 11:28, Fam Zheng wrote: > Right after main_loop ends, we release various things but keep iothread > alive. The latter is not prepared to the sudden change of resources. > > Specifically, after bdrv_close_all(), virtio-scsi dataplane get a > surprise at the empty BlockBackend: > >

[Qemu-devel] [PATCH] iothread: Stop threads before main() quits

2016-09-08 Thread Fam Zheng
Right after main_loop ends, we release various things but keep iothread alive. The latter is not prepared to the sudden change of resources. Specifically, after bdrv_close_all(), virtio-scsi dataplane get a surprise at the empty BlockBackend: (gdb) bt at