Linux Virtualization - Aug 2014 - [virtio-dev] Re: [PATCH v2 0/4] virtio: Clean up scatterlists and use the DMA API

Il 28/08/2014 20:29, Christian Borntraeger ha scritto:
>>> >> block seems to work, with net a simple ping works, iperf
causes this:
>> > 
>> > I neither see the bug, nor can I reproduce it on x86_64 on KVM.  I
>> > doubt that dma vs. non-dma is relevant.  I tried -net user a -net
tap
>> > with iperf running in both directions.  I also tried switching
>> > virtio_pci into non-DMA-API mode.  No errors.
>> > 
>> > Is there any chance that you could instrument that BUG_ON to print
n,
>> > i, in_sgs, out_sgs, total_sg, total_in, and total_out?  I assume
this
>> > is some oddity with patch 1, but I'm mystified.
> Yes, its triggered by patch 1.
What version of QEMU?  Can you two cat the "features" file in the
virtio-net device's sysfs directory?

Paolo

[removed virtio-dev -- I'm sick of bounces]

On Thu, Aug 28, 2014 at 11:55 AM, Paolo Bonzini <pbonzini at redhat.com>
wrote:
> Il 28/08/2014 20:29, Christian Borntraeger ha scritto:
>>>> >> block seems to work, with net a simple ping works,
iperf causes this:
>>> >
>>> > I neither see the bug, nor can I reproduce it on x86_64 on
KVM.  I
>>> > doubt that dma vs. non-dma is relevant.  I tried -net user a
-net tap
>>> > with iperf running in both directions.  I also tried switching
>>> > virtio_pci into non-DMA-API mode.  No errors.
>>> >
>>> > Is there any chance that you could instrument that BUG_ON to
print n,
>>> > i, in_sgs, out_sgs, total_sg, total_in, and total_out?  I
assume this
>>> > is some oddity with patch 1, but I'm mystified.
>> Yes, its triggered by patch 1.
>
> What version of QEMU?  Can you two cat the "features" file in the
> virtio-net device's sysfs directory?
With vhost=on:

# cat /sys/class/net/eth0/device/features
0000011000000001111101010001110000000000000000000000000000000000

Without vhost=on:

# cat /sys/class/net/eth0/device/features
0000011000000001111101010001110000000000000000000000000000000000

Instrumenting the code says that indirect=1.

--Andy

On 28/08/14 21:03, Andy Lutomirski wrote:
> [removed virtio-dev -- I'm sick of bounces]
> 
> On Thu, Aug 28, 2014 at 11:55 AM, Paolo Bonzini <pbonzini at
redhat.com> wrote:
>> Il 28/08/2014 20:29, Christian Borntraeger ha scritto:
>>>>>>> block seems to work, with net a simple ping works,
iperf causes this:
>>>>>
>>>>> I neither see the bug, nor can I reproduce it on x86_64 on
KVM.  I
>>>>> doubt that dma vs. non-dma is relevant.  I tried -net user
a -net tap
>>>>> with iperf running in both directions.  I also tried
switching
>>>>> virtio_pci into non-DMA-API mode.  No errors.
>>>>>
>>>>> Is there any chance that you could instrument that BUG_ON
to print n,
>>>>> i, in_sgs, out_sgs, total_sg, total_in, and total_out?  I
assume this
>>>>> is some oddity with patch 1, but I'm mystified.
>>> Yes, its triggered by patch 1.
>>
>> What version of QEMU?  Can you two cat the "features" file in
the
>> virtio-net device's sysfs directory?
> 
> With vhost=on:
> 
> # cat /sys/class/net/eth0/device/features
> 0000011000000001111101010001110000000000000000000000000000000000
  > 
> Without vhost=on:
> 
> # cat /sys/class/net/eth0/device/features
> 0000011000000001111101010001110000000000000000000000000000000000
> 
> Instrumenting the code says that indirect=1.
> 
> --Andy
> 
I have 

1100011111111111111101010001110000000000000000000000000000000000