On Tue, 27 Oct 2015 13:51:20 +0200 "Michael S. Tsirkin" <[email protected]> wrote:
> On Tue, Oct 27, 2015 at 10:47:54AM +0200, Michael S. Tsirkin wrote: > > TL;DR: > > This fixes virtio in a way transparent to guest. > > We should now be able to revert commits aa8580cd and df0acded19ec which > > worked > > around it in a way that's not transparent. > > I didn't check dataplane BTW. Igor? Stefan? verified that series fixes virtio-[blk|scsi|net], all of them hit at least one descriptor(indirect) that crosses DIMM boundary and QEMU survived it. However as Stefan has said virtio-blk with dataplane enabled hangs guest instead of QEMU crashing and QEMU prints following error: "Failed to map descriptor addr 0x1045eb000 len 106496" I've used following CLI: qemu-system-x86_64 -enable-kvm -enable-kvm -m 128M,slots=250,maxmem=32G -drive if=none,id=hd,file=rhel72.img,cache=none,aio=native,format=raw -device virtio-blk,drive=hd,scsi=off,config-wce=off,x-data-plane=on `for i in $(seq 0 15); do echo -n "-object memory-backend-ram,id=m$i,size=10M -device pc-dimm,id=dimm$i,memdev=m$i "; done` it hangs at boot time or on executing 'dd if=/dev/vda of=/dev/null bs 32M'
