Hi Paul,

On 12 May 2015, at 5:00 am, Paul Walmsley pwalmsley@nvidia.com wrote:

Hi folks

Tegra124 Jetson TK1 hangs during boot in next-20150508 and beyond:

http://nvt.pwsan.com/experimental/linux-next/testlogs/test_next-20150508/201... http://nvt.pwsan.com/experimental/linux-next/testlogs/test_next-20150511/201...

Weirdly enough, it seems to work in kernelci: http://kernelci.org/boot/tegra124-jetson-tk1_rootfs:nfs/job/next/kernel/next...

Though maybe it's just good fortune that Kevin doesn't have a display plugged into his TK1.

Reverting commit 672cb1d6aec7da2799afd1b529d5136d84ed2561 ("drm: Add reference counting to blob properties") gets it booting again:

http://nvt.pwsan.com/pub/pwalmsley-tester/testlogs/test_20150511204703_38382...

I'm not a DRM expert but maybe one of you knows what's going on here?

Thanks for the catch. I have a Nyan-Big Chromebook here, so will try to work it out locally and let you know.

Sorry for the inconvenience!

Cheers, Daniel

Hi Paul,

On 12 May 2015 at 05:00, Paul Walmsley pwalmsley@nvidia.com wrote:

Tegra124 Jetson TK1 hangs during boot in next-20150508 and beyond:

http://nvt.pwsan.com/experimental/linux-next/testlogs/test_next-20150508/201... http://nvt.pwsan.com/experimental/linux-next/testlogs/test_next-20150511/201...

Reverting commit 672cb1d6aec7da2799afd1b529d5136d84ed2561 ("drm: Add reference counting to blob properties") gets it booting again:

http://nvt.pwsan.com/pub/pwalmsley-tester/testlogs/test_20150511204703_38382...

I'm not a DRM expert but maybe one of you knows what's going on here?

Thierry found and fixed the rather stupid bug: not using the cached 'dev' variable in the one place it would've been useful!

http://paste.debian.net/175522/ is the fix, which has been squashed in to the branch and should appear in -next from today.

Thanks for the report.

Cheers, Daniel