[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] e1000e migration
From: |
Dmitry Fleytman |
Subject: |
Re: [Qemu-devel] e1000e migration |
Date: |
Mon, 15 May 2017 19:13:00 +0300 |
Hello Dave,
It looks like we identified the problem.
We are working on fix and will send it as soon as it is ready.
~Dmitry.
Sent from my iPhone
> On 15 May 2017, at 12:22, Dr. David Alan Gilbert <address@hidden> wrote:
>
> * Dmitry Fleytman (address@hidden) wrote:
>> Hello Dave,
>
> Hi Dmitry,
> Thanks for the reply.
>
>> We are trying to reproduce this issue on our systems but with no luck so far…
>
> Note our QE hit this with both a Win8.1 and a win2012r2 guest - although
> the 2012r2 is reported to have recoverd after a few minutes.
> 2016 apparently works OK.
>
>> From what you describe it looks like some bit in ICR is not being cleared by
>> the driver.
>> This usually means that this bit should never be set in that specific
>> interrupt mode.
>>
>> Could you please check which bit is not cleared and who sets it?
>
> The full set of e1000e_irq_pending_interrupts after migration is:
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x100000 (ICR:
> 0x80100082, IMS: 0x1f00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x80100082, IMS: 0x1e00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x80100082, IMS: 0x1e00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x200000 (ICR:
> 0x80300082, IMS: 0x1e00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x80100082, IMS: 0x1c00004)
> <repeated lots>
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x80300082, IMS: 0x1c00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x1000000 (ICR:
> 0x813000c2, IMS: 0x1c00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x1000000 (ICR:
> 0x813000c2, IMS: 0x1400004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x1000000 (ICR:
> 0x813000c2, IMS: 0x1000004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x813000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x813000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x813000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x813000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x813000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x200000 (ICR:
> 0x813000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0xa00004)
> <repeats>
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x200000 (ICR:
> 0x813000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0x800004)
> <repeats>
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x813000c2, IMS: 0x800004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x813000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x200000 (ICR:
> 0x813000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0x4)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x811000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x1000000 (ICR:
> 0x815000c2, IMS: 0x1a00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
> 0x815000c2, IMS: 0xa00004)
> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x1000000 (ICR:
> 0x815000c2, IMS: 0x1a00004)
>
> and then I think we get stuck in this cycle of this one always being the
> one that fires repeatedly. I think that's the 'other' firing, I think
> because of the receive-overrun. One thing I've not
> figured out is why the receive overrun happens - is that because we
> really have a very heavy packet rate or is it because something has
> stopped receiving them.
> The network I'm testing on does have a fair amount of broadcast traffic
> on.
>
> Dave
>
>> Regards,
>> Dmitry
>>
>>> On 11 May 2017, at 15:36 PM, Dr. David Alan Gilbert <address@hidden> wrote:
>>>
>>> Hi Dmitry,
>>> Have you seen any problems with e1000e migration under windows?
>>> I've got a repeatable case where after migration with e1000e windows
>>> hangs/almost hangs.
>>> I'm seeing the e1000e generate interrupts at a very very high
>>> rate (maybe ~1000 second ish?) after migration.
>>>
>>> Some versions of qemu do it and some dont, but my attempts
>>> at bisection lead me to code that should be irrelevant.
>>>
>>> Prior to migration I see:
>>>
>>> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x100000 (ICR:
>>> 0x80100082, IMS: 0x1f00004)
>>> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
>>> 0x80000082, IMS: 0x1a00004)
>>> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
>>> 0x80000082, IMS: 0x1f00004)
>>> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
>>> 0x80000082, IMS: 0x1a00004)
>>> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x0 (ICR:
>>> 0x80000082, IMS: 0x1f00004)
>>>
>>> which I think the ICR means:
>>> 31 - int asserted
>>> 20 - RxQ0 - receive queue 0 interrupt
>>> 7 - RXT0 - receiver timer interrupt
>>> 1 - TXQE - Transmit Queue empty
>>>
>>> after migration it varies more, I'm seeing mostly:
>>> address@hidden:e1000e_irq_pending_interrupts ICR PENDING: 0x1000000 (ICR:
>>> 0x815000c2, IMS: 0x1a00004)
>>> 31 - int asserted
>>> 24 - 'Other'
>>> 22 - TxQ0 interrupt
>>> 20 - RxQ0 interrupt
>>> 07 - RXT0 Receiver timer interrupt
>>> 06 - RX0 - Receiver overrun
>>> 01 - TXQE - Transmit queue empty
>>>
>>> For reference this is https://bugzilla.redhat.com/show_bug.cgi?id=1447935
>>>
>>> Dave
>>> --
>>> Dr. David Alan Gilbert / address@hidden / Manchester, UK
>>
> --
> Dr. David Alan Gilbert / address@hidden / Manchester, UK