Re: [Qemu-devel] virtio block device is not working

[Daniel Henrique Barboza]

With a gdb attached I got the following backtrace:


Program received signal SIGABRT, Aborted.
0x00007ffff5f5eff0 in raise () from /lib64/libc.so.6
(gdb) bt
#0  0x00007ffff5f5eff0 in raise () at /lib64/libc.so.6
#1  0x00007ffff5f6136c in abort () at /lib64/libc.so.6
#2  0x000000001009e704 in kvm_mem_ioeventfd_del (listener=<optimized 
out>, section=0x7ffff2e6da00, match_data=<optimized out>, data=0, 
e=0x7ffff23d0080)
    at /home/danielhb/qemu/accel/kvm/kvm-all.c:813
#3  0x000000001008845c in address_space_add_del_ioeventfds 
(as=0x10c2df10 <address_space_memory>, fds_new=0x0, 
fds_new_nb=<optimized out>, fds_old=0x7fffe4078f20, 
fds_old_nb=<optimized out>) at /home/danielhb/qemu/memory.c:832
#4  0x0000000000000000 in  ()
(gdb)


At first glance it looks like kvm_mem_ioeventfd_del can't cleanup the 
EventNotifier
it received by itself. Reverting the patch that introduced it looks like 
a good
short term solution.



Daniel

On 01/23/2018 10:47 AM, Daniel Henrique Barboza wrote:
> Yeah, the problem looks like in virtio-pci-net. This is the output 
> when running in a Power 9
> host:
>
> sudo ./qemu-system-ppc64 -nographic -vga none -m 4G -M 
> pseries,accel=kvm -netdev type=user,id=net0 -device 
> virtio-net-pci,netdev=net0
>
>
> SLOF 
> **********************************************************************
> QEMU Starting
>  Build Date = Dec 18 2017 13:08:00
>  FW Version = git-fa981320a1e0968d
>  Press "s" to enter Open Firmware.
>
> Populating /vdevice methods
> Populating /vdevice/address@hidden
> Populating /vdevice/address@hidden
> Populating /vdevice/address@hidden
>        SCSI: Looking for devices
>           8200000000000000 CD-ROM   : "QEMU     QEMU CD-ROM 2.5+"
> Populating /address@hidden
>                      00 0000 (D) : 1af4 1000    virtio [ net ]
> Aborted
>
>
> As Ziviani mentioned, reverting 4fe6d78b2e fixes it:
>
>
> $ git diff
> diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
> index 071f4f5..f290f48 100644
> --- a/accel/kvm/kvm-all.c
> +++ b/accel/kvm/kvm-all.c
> @@ -812,10 +812,6 @@ static void kvm_mem_ioeventfd_del(MemoryListener 
> *listener,
>      if (r < 0) {
>          abort();
>      }
> -
> -    if (e->cleanup) {
> -        e->cleanup(e);
> -    }
>  }
>
> $ sudo ./qemu-system-ppc64 -nographic -vga none -m 4G -M 
> pseries,accel=kvm -netdev type=user,id=net0 -device 
> virtio-net-pci,netdev=net0
>
>
> SLOF 
> **********************************************************************
> QEMU Starting
>  Build Date = Dec 18 2017 13:08:00
>  FW Version = git-fa981320a1e0968d
>  Press "s" to enter Open Firmware.
>
> Populating /vdevice methods
> Populating /vdevice/address@hidden
> Populating /vdevice/address@hidden
> Populating /vdevice/address@hidden
>        SCSI: Looking for devices
>           8200000000000000 CD-ROM   : "QEMU     QEMU CD-ROM 2.5+"
> Populating /address@hidden
>                      00 0000 (D) : 1af4 1000    virtio [ net ]
> No NVRAM common partition, re-initializing...
> Scanning USB
> Using default console: /vdevice/address@hidden
>
>   Welcome to Open Firmware
>
>   Copyright (c) 2004, 2017 IBM Corporation All rights reserved.
>   This program and the accompanying materials are made available
>   under the terms of the BSD License available at
>   http://www.opensource.org/licenses/bsd-license.php
>
>
> Trying to load:  from: disk ...
> E3405: No such device
> Trying to load:  from: /vdevice/address@hidden/address@hidden 
> ... No medium !
>
> E3405: No such device
> Trying to load:  from: /address@hidden/address@hidden ...
>  Initializing NIC
>   Reading MAC address from device: 52:54:00:12:34:56
>   Requesting information via DHCP: done
>   Using IPv4 address: 10.0.2.15
>   Requesting file "" via TFTP from 10.0.2.2
>   Receiving data:  0 KBytes
> E3010 (net) TFTP access violation
>
> E3407: Load failed
>
>   Type 'boot' and press return to continue booting the system.
>   Type 'reset-all' and press return to reboot the system.
>
>
> Ready!
> 0 >
>
>
>
> On 01/23/2018 10:34 AM, address@hidden wrote:
> > On Tue, Jan 23, 2018 at 01:05:28AM -0200, address@hidden 
> > wrote:
> > > Hello people!
> > >
> > > I'm not able to boot any guest that sets a virtio block device like:
> > > (branch master)
> > >
> > > [PPC64]
> > > qemu-system-ppc64 -cpu POWER8 -nographic -vga none -m 4G -M 
> > > pseries,accel=kvm,kvm-type=PR -drive file=disk.qcow2,if=virtio
> > my bad, actually the command line is:
> >
> > qemu-system-ppc64 -cpu POWER8 -nographic -vga none -m 4G -M 
> > pseries,accel=kvm -netdev type=user,id=net0 -device 
> > virtio-net-pci,netdev=net0 -drive file=../disk.qcow2,if=virtio
> >
> > and the problem seem to be in virtio-net-pci, not in the block device
> >
> >
> > > QEMU Starting
> > >   Build Date = Dec 18 2017 13:08:00
> > >   FW Version = git-fa981320a1e0968d
> > >   Press "s" to enter Open Firmware.
> > >
> > > Populating /vdevice methods
> > > Populating /vdevice/address@hidden
> > > Populating /vdevice/address@hidden
> > > Populating /vdevice/address@hidden
> > >         SCSI: Looking for devices
> > >            8200000000000000 CD-ROM   : "QEMU     QEMU CD-ROM      2.5+"
> > > Populating /address@hidden
> > >                       00 0000 (D) : 1af4 1000    virtio [ net ]
> > > Aborted
> > >
> > > [x86]
> > >
> > > qemu-system-x86_64 -m 4G -enable-kvm -drive file=util.qcow2,if=virtio
> > > Running QEMU with GTK 2.x is deprecated, and will be removed
> > > in a future release. Please switch to GTK 3.x instead
> > > [1]    5282 abort
> > >
> > > [Cause]
> > >
> > > The commit 4fe6d78b2e introduces the
> > >
> > > ...
> > > kvm_mem_ioeventfd_del(...) {
> > >      ...
> > >      r = kvm_set_ioeventfd_mmio(fd, ...
> > >      if (r < 0) {
> > >          abort();
> > >      }
> > >
> > > +    if (e->cleanup) {
> > > +        e->cleanup(e);
> > > +    }
> > > }
> > >
> > > For some reason, not yet clear to me, cleanup() calls the same
> > > kvm_mem_ioeventfd_del again and again until kvm_set_ioeventfd_mmio
> > > returns < 0 and abort().
> > >
> > > I was going to send a patch to revert that 'if ()' but I think it could
> > > cause a regression. What do you guys think?
> > >
> > > Thanks,
> > >
> > > Jose Ricardo Ziviani