Re: [Qemu-devel] PCI regression in 2.9

[Michael S. Tsirkin]

On Tue, Apr 04, 2017 at 11:30:13AM -0400, Jeff Cody wrote:
> 
> I ran into this while running qemu iotests, and it looks like it is a
> regression from 2.8.
> 
> Here is the reproducer:
> 
> $ ./x86_64-softmmu/qemu-system-x86_64 -drive if=virtio
> qemu-system-x86_64: qemu-kvm/memory.c:2078: memory_region_del_subregion: 
> Assertion `subregion->container == mr' failed
> Aborted (core dumped)
> 
> The proper output is (with no abort):
> qemu-system-x86_64: -drive if=virtio: Device needs media, but drive is empty
> 
> I bisected it down to this commit:
> 
> commit c53598ed18e40a9609573b21f2a361221ca0f806
> Author: Alexey Kardashevskiy <address@hidden>
> Date:   Mon Mar 27 15:40:30 2017 +1100
> 
>     pci: Add missing drop of bus master AS reference
>     
>     The recent introduction of a bus master container added
>     memory_region_add_subregion() into the PCI device registering path but
>     missed memory_region_del_subregion() in the unregistering path leaving
>     a reference to the root memory region of the new container.
>     
>     This adds missing memory_region_del_subregion().
>     
>     Fixes: 3716d5902d743 ("pci: introduce a bus master container")
>     Signed-off-by: Alexey Kardashevskiy <address@hidden>
>     Reviewed-by: Michael S. Tsirkin <address@hidden>
>     Signed-off-by: Michael S. Tsirkin <address@hidden>
>     Reviewed-by: Paolo Bonzini <address@hidden>


Thanks for the report. I'm preparing a pull request with a fix,
will copy you so you can test.

Thanks!

-- 
MST