[SeaBIOS] [PATCH seabios 3/3] kvmtool: support larger virtio queues

Gerd Hoffmann kraxel at redhat.com
Fri Nov 3 16:34:26 CET 2017


On Fri, 2017-11-03 at 13:49 +0000, Jean-Philippe Brucker wrote:
> On 02/11/17 15:50, Gerd Hoffmann wrote:
> > Queues have 256 entries on kvmtool, support that.  Needs more
> > memory for
> > virtqueues now.  But with the move to 32bit drivers for virtio this
> > should not be much of an issue any more.
> > 
> > Known problems (probably kvmtool bugs):
> >  * Must bump to 260 entries to make things actually work,
> >    otherwise kvmtool segfaults.  Oops.
> 
> You mean setting MAX_QUEUE_NUM to 256 instead of 260 in seabios
> causes a
> kvmtool crash?

yes.

>  Do you have any more detail on the segfault?

Ok, lets have a look with gdb ...

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7f81caf3c700 (LWP 20234)]
virt_queue__get_head_iov (vq=vq at entry=0x7f82576be0a0, iov=iov at entry=0x7
f826770aae0, 
    out=out at entry=0x7f826770bae0, in=in at entry=0x7f826770bae2,
head=65104, kvm=kvm at entry=0x246eee0)
    at virtio/core.c:105
105             *out = *in = 0;
(gdb) bt
#0  0x000000000040c91b in virt_queue__get_head_iov (vq=vq at entry=0x7f825
76be0a0, iov=iov at entry=0x7f826770aae0, out=out at entry=0x7f826770bae0, in
=in at entry=0x7f826770bae2, head=65104, kvm=kvm at entry=0x246eee0) at
virtio/core.c:105
#1  0x000000000040bbf7 in virtio_blk_thread (bdev=0x7f82576be010,
vq=0x7f82576be0a0, kvm=0x246eee0)
    at virtio/blk.c:134
#2  0x000000000040bbf7 in virtio_blk_thread (dev=0x7f82576be010) at
virtio/blk.c:208
#3  0x00007f82571c6e25 in start_thread () at /lib64/libpthread.so.0
#4  0x00007f82543b134d in clone () at /lib64/libc.so.6
(gdb) print *vq
$1 = {vring = {num = 256, desc = 0x7f824cf3e000, avail =
0x7f824cf3f000, used = 0x7f824cf40000}, 
  pfn = 524285, last_avail_idx = 263, last_used_signalled = 1, endian =
1}

last_avail_idx looks bogus ...

> Since seabios doesn't use the event indexes, this would lead to
> missing signals, but not a segfault.

seabios polls anyway, so it doesn't need signals.

> >  * Linux kernel doesn't find virtio-blk devices after seabios
> >    initialized them.  virtio device reset not working properly?
> 
> No, reset isn't implemented at all... A lot of work is required to
> properly clear the state and threads of each device.

Hmm.  That is required for any kind of boot loader support though.

/me wonders what the kvmtool --firmware switch is good for then if a
direct kernel boot is apparently the only thing which actually works.

cheers,
  Gerd




More information about the SeaBIOS mailing list