Re: [xen-4.2-testing test] 31385: regressions - FAIL

[Ian Campbell <Ian.Campbell@citrix.com>]

On Mon, 2014-11-10 at 14:00 +0000, Ian Jackson wrote:
> Ian Campbell writes ("Re: [Xen-devel] [xen-4.2-testing test] 31385: regressions - FAIL"):
> > On Fri, 2014-11-07 at 11:31 +0000, Ian Jackson wrote:
> > > Hrm.  I have added the relevant host property to the moths too.  My
> > > osstest change got a push so results should soon start to include the
> > > `swiotlb=26422'.
> > 
> > Looks like things are still failing even with this? :-/
> 
> From serial-worm-moth.log:
> 
> Nov  5 14:10:40.550597 [    0.000000] Kernel command line: placeholder
> root=/dev/mapper/worm--moth-root ro BOOTIF=01-78-2b-cb-73-42-c2
> console=hvc0^M
> 
> Nov  5 14:10:40.578640 [    0.000000] software IO TLB [mem
> 0x1ba93000-0x1fa93000] (64MB) mapped at [dba93000-dfa92fff]^M
> 
> So the workaround still hasn't taken.  This is because we hadn't had a
> push of osstest at the time that report was made.  Looking at the
> database the first non-osstest-self-pushgate flight with the new
> osstest was 31432.

http://www.chiark.greenend.org.uk/~xensrcts/logs/31458/test-amd64-i386-pair/info.html seems to have the swiotlb= and has failed with
Nov 10 04:12:11.050621 [  985.934464] sd 4:1:0:0: pci_map_sg failed: request for 4096 bytes!
Nov 10 04:12:11.050678 [  985.952455] mpt2sas 0000:05:00.0: swiotlb buffer is full

on scape-moth.

Ian.