[igt-dev] ✗ Fi.CI.IGT: failure for lib/ioctl_wrappers: Handle PREAD/PWRITE ioctls not supported in gem_read/write (rev15)

[Dixit, Ashutosh]

On Fri, 22 Jan 2021 01:21:34 -0800, Chris Wilson wrote:
>
> Quoting Vudum, Lakshminarayana (2021-01-22 03:14:42)
> > Re-reported.
> >
> > -----Original Message-----
> > From: Dixit, Ashutosh <ashutosh.dixit at intel.com>
> > Sent: Thursday, January 21, 2021 12:32 AM
> > To: igt-dev at lists.freedesktop.org
> > Cc: Vudum, Lakshminarayana <lakshminarayana.vudum at intel.com>
> > Subject: Re: ✗ Fi.CI.IGT: failure for lib/ioctl_wrappers: Handle PREAD/PWRITE ioctls not supported in gem_read/write (rev15)
> >
> > On Wed, 20 Jan 2021 23:33:03 -0800, Patchwork wrote:
> > >
> > > Patch Details
> > >
> > >  Series:  lib/ioctl_wrappers: Handle PREAD/PWRITE ioctls not supported
> > > in gem_read/write (rev15)
> > >  URL:  https://patchwork.freedesktop.org/series/81152/
> > >  State:  failure
> > >  Details:
> > > https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_5411/index.html
> > >
> > > CI Bug Log - changes from IGT_5961_full -> IGTPW_5411_full
> > >
> > > Summary
> > >
> > > FAILURE
> > >
> > > Serious unknown changes coming with IGTPW_5411_full absolutely need to
> > > be verified manually.
> > >
> > > If you think the reported changes have nothing to do with the changes
> > > introduced in IGTPW_5411_full, please notify your bug team to allow
> > > them to document this new failure mode, which will reduce false positives in CI.
> > >
> > > External URL:
> > > https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_5411/index.html
> > >
> > > Possible new issues
> > >
> > > Here are the unknown changes that may have been introduced in IGTPW_5411_full:
> > >
> > > IGT changes
> > >
> > > Possible regressions
> > >
> > > * igt at gen7_exec_parse@basic-allocation:
> > >
> > >  * shard-hsw: PASS -> TIMEOUT
> >
> > Since this is an IGT patch it should not cause this failure which is
> > also accompanied by a kernel trace. Therefore this also looks like a
> > false positive. Thanks.
>
> Wrong.

Sorry, my bad misreading the dmesg trace where it's obvious the kernel
trace is the result of killing the hanging test.

Actually, on my old HSW the test actually passes but takes 85 seconds
instead of 22 seconds because of the patch. In the CI the execution time
exceeded 120 seconds.