On Tue, Jan 12, 2010 at 11:06:33PM +0100, Christoph Hellwig wrote:
On Tue, Jan 12, 2010 at 09:50:55PM +0200, Michael S. Tsirkin wrote:
So the issue is that wrong block size (0x) was passed
to guest. Would it make sense to add some sanity checking in virtio-blk
to make it not crash
This patch causes 100% reproducible boot panics in a Linux guest using
virtio.
My qemu command line is:
/opt/qemu/bin/qemu-system-x86_64 \
-m 1500 \
-enable-kvm \
-drive
file=/dev/vg00/qemu-root,if=virtio,media=disk,cache=none,aio=threads \
-kernel
On Tue, Jan 12, 2010 at 07:09:27PM +0100, Christoph Hellwig wrote:
This patch causes 100% reproducible boot panics in a Linux guest using
virtio.
My qemu command line is:
/opt/qemu/bin/qemu-system-x86_64 \
-m 1500 \
-enable-kvm \
-drive
On Tue, Jan 12, 2010 at 07:09:27PM +0100, Christoph Hellwig wrote:
This patch causes 100% reproducible boot panics in a Linux guest using
virtio.
My qemu command line is:
/opt/qemu/bin/qemu-system-x86_64 \
-m 1500 \
-enable-kvm \
-drive
On Tue, Jan 12, 2010 at 09:50:55PM +0200, Michael S. Tsirkin wrote:
So the issue is that wrong block size (0x) was passed
to guest. Would it make sense to add some sanity checking in virtio-blk
to make it not crash but fail in probe? Which block size values
are sane?
Yes, I'll cook
Add feature bits as properties to virtio. This makes it possible to e.g. define
machine without indirect buffer support, which is required for 0.10
compatibility, or without hardware checksum support, which is required for 0.11
compatibility. Since default values for optional features are now set
