On 27/01/2014 20:17, Kevin Hilman wrote:
On Mon, Jan 27, 2014 at 9:59 AM, Kevin's boot bot khilman@linaro.org wrote:
Automated DT boot report for various ARM defconfigs.
Tree/Branch: mainline Git describe: v3.13-8631-gba635f8 Failed boot tests (console logs at the end) =========================================== armada-370-mirabox: FAIL: multi_v7_defconfig
This is a new(ish) boot failure that I finally bisected. Note that it only fails on multi_v7_defconfig, and not mvebu_defconfig.
When it fails, there's no useful console output, but it was bisected down to: first bad commit: [af8d1c63afcbf36eea06789c92e22d4af118d2fb] ARM: mvebu: Add support to get the ID and the revision of a SoC.
Based on an idea from Jason, I added CONFIG_PCI=y and CONFIG_MVEBU_PCI=y to multi_v7_defconfig and that made it start booting again, some something is probably hanging when accessing PCI registers?
This patch is needed: http://comments.gmane.org/gmane.linux.ports.arm.kernel/295642
It is part of the mvebu/fixes branch of Jason and it is also on linux-next but it has not been pushed yet to arm-soc and then to Linus.
(Actually we don't need the full PCI subsystem just the dt part which knows how to decode the assigned-address in of/address.c.)
Thanks for the report,
Gregory
Kevin