Re: [Xen-devel] [xen-unstable test] 142383: regressions - FAIL

[Jan Beulich <jbeulich@xxxxxxxx>]

On 08.10.2019 13:15, Roger Pau Monné  wrote:
> On Tue, Oct 08, 2019 at 12:42:25PM +0200, Jürgen Groß wrote:
>> On 07.10.19 22:56, osstest service owner wrote:
>>> flight 142383 xen-unstable real [real]
>>> http://logs.test-lab.xenproject.org/osstest/logs/142383/
>>>
>>> Regressions :-(
>>>
>>> Tests which did not succeed and are blocking,
>>> including tests which could not be run:
>>>   test-amd64-amd64-libvirt-pair 22 guest-migrate/src_host/dst_host fail 
>>> REGR. vs. 141822
>>>   test-amd64-i386-libvirt-pair 22 guest-migrate/src_host/dst_host fail 
>>> REGR. vs. 141822
>>
>> Hmm, test log says the guest didn't suspend.
>>
>> Could that be related to commit b183e180bce9303 ?
> 
> The libvirt libxl daemon spits:
> 
> 2019-10-07 12:31:09.953+0000: libxl-save-helper: starting save: Success
> 2019-10-07 12:31:09.953+0000: xc: fd 40, dom 1, flags 1, hvm 0
> 2019-10-07 12:31:09.954+0000: xc: Saving domain 1, type x86 PV
> 2019-10-07 12:31:09.954+0000: xc: 64 bits, 4 levels
> 2019-10-07 12:31:09.959+0000: xc: max_mfn 0x280000
> 2019-10-07 12:31:09.959+0000: xc: p2m list from 0xffffc90000000000 to 
> 0xffffc900000fffff, root at 0x27de0a
> 2019-10-07 12:31:09.959+0000: xc: max_pfn 0x1ffff, p2m_frames 256
> 2019-10-07 12:31:09.960+0000: xc: Failed to enable logdirty: 22,0,22 (22 = 
> Invalid argument): Internal error
> 2019-10-07 12:31:09.960+0000: xc: Save failed (22 = Invalid argument): 
> Internal error
> 2019-10-07 12:31:09.981+0000: libxl-save-helper: complete r=-1: Invalid 
> argument
> 2019-10-07 12:31:09.983+0000: libxl: libxl.c:752:libxl__fd_flags_restore: 
> fnctl F_SETFL of fd 40 to 0x2
> 2019-10-07 12:31:09.983+0000: libxl: libxl_event.c:1873:libxl__ao_complete: 
> ao 0x7f760c002110: complete, rc=-8
> 2019-10-07 12:31:09.983+0000: libxl: libxl_event.c:1842:libxl__ao__destroy: 
> ao 0x7f760c002110: destroy
> 
> Which seem to be related to the recent iommu changes, I would say it's
> likely hitting the is_iommu_enabled check in paging_log_dirty_enable
> and hence returning EINVAL.
> 
> Adding Paul and Jan who worked on the series.

Well, as mentioned in reply to Paul's patch earlier today, I intend
to commit the agreed upon half of it later today. Suggesting this was
actually a result of me noticing this log-dirty enabling failure when
I looked at the logs (another time) in the morning.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel