Public bug reported: SRU Justification:
Impact: qemu 1.6 adds support for virtio memory mapped devices (like net and block) on arm. While S will probably ship with qemu 1.5, we know it'll be provided in a ppa and we want to take advantage of the speedup offered by virtio, so 1) turn on virtio_mmio on armhf (generic and generic-lpae) 2) uniform VIRTIO* support among all the different archs, making it a compiled-in options like on i386 and amd64 Fix: apply the attached config patch and recompile a kernel with it Testcase: try booting a qemu kvm using a kernel without this patch: sudo qemu-system-arm -machine vexpress-a15 -cpu cortex-a15 -enable-kvm -m 512M -kernel /boot/vmlinuz-`uname -r` -append "console=ttyAMA0 earlyprintk=serial root=/dev/mmcblk0 ro rootfstype=ext4" -serial stdio -dtb /lib/firmware/`uname -r`/device-tree/vexpress-v2p-ca15-tc1.dtb -initrd /boot/initrd.img-`uname -r` -drive if=sd,cache=writeback,file =saucy-server-cloudimg-armhf.img -net nic -net user,hostfwd=tcp::2223-:22 -display none and then try to boot another instance using a kernel with the config patch applied: sudo qemu-system-arm -machine vexpress-a15 -cpu cortex-a15 -enable-kvm -m 512M -kernel /boot/vmlinuz-`uname -r` -append "console=ttyAMA0 earlyprintk=serial root=/dev/vda ro rootfstype=ext4" -serial stdio -dtb /lib/firmware/`uname -r`/device-tree/vexpress-v2p-ca15-tc1.dtb -initrd /boot/initrd.img-`uname -r` -drive if=none,cache=writeback,file=saucy- server-cloudimg-armhf.img,id=foo -device virtio-blk-device,drive=foo -display none -netdev tap,id=tap0 -device virtio-net-device,netdev=tap0 the difference is noticeable to the naked eye. ** Affects: linux (Ubuntu) Importance: Undecided Status: Incomplete -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1235277 Title: armhf: virtio_mmio support Status in “linux” package in Ubuntu: Incomplete Bug description: SRU Justification: Impact: qemu 1.6 adds support for virtio memory mapped devices (like net and block) on arm. While S will probably ship with qemu 1.5, we know it'll be provided in a ppa and we want to take advantage of the speedup offered by virtio, so 1) turn on virtio_mmio on armhf (generic and generic-lpae) 2) uniform VIRTIO* support among all the different archs, making it a compiled-in options like on i386 and amd64 Fix: apply the attached config patch and recompile a kernel with it Testcase: try booting a qemu kvm using a kernel without this patch: sudo qemu-system-arm -machine vexpress-a15 -cpu cortex-a15 -enable-kvm -m 512M -kernel /boot/vmlinuz-`uname -r` -append "console=ttyAMA0 earlyprintk=serial root=/dev/mmcblk0 ro rootfstype=ext4" -serial stdio -dtb /lib/firmware/`uname -r`/device-tree/vexpress-v2p-ca15-tc1.dtb -initrd /boot/initrd.img-`uname -r` -drive if=sd,cache=writeback,file =saucy-server-cloudimg-armhf.img -net nic -net user,hostfwd=tcp::2223-:22 -display none and then try to boot another instance using a kernel with the config patch applied: sudo qemu-system-arm -machine vexpress-a15 -cpu cortex-a15 -enable-kvm -m 512M -kernel /boot/vmlinuz-`uname -r` -append "console=ttyAMA0 earlyprintk=serial root=/dev/vda ro rootfstype=ext4" -serial stdio -dtb /lib/firmware/`uname -r`/device-tree/vexpress-v2p-ca15-tc1.dtb -initrd /boot/initrd.img-`uname -r` -drive if=none,cache=writeback,file=saucy-server-cloudimg-armhf.img,id=foo -device virtio-blk-device,drive=foo -display none -netdev tap,id=tap0 -device virtio-net-device,netdev=tap0 the difference is noticeable to the naked eye. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1235277/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp