On Wed, Jan 13, 2021 at 12:00:01PM +0530, Sumit Garg wrote:

On Wed, 13 Jan 2021 at 03:49, Will Deacon [off-list ref] wrote:

quoted

This reverts commit 367c820ef08082e68df8a3bc12e62393af21e4b5.

lockup_detector_init() makes heavy use of per-cpu variables and must be
called with preemption disabled. Usually, it's handled early during boot
in kernel_ionit_freeable(), before SMP has been initialised.

Since we do not know whether or not our PMU interrupt can be signalled
as an NMI until considerably later in the boot process, the Arm PMU
driver attempts to re-initialise the lockup detector off the back of a
device_initcall(). Unfortunately, this is called from preemptible
context and results in the following splat:

Can we consider the following fix (compile tested only) to bind the
call to lockup_detector_init() to a particular CPU instead of
reverting the hard lockup detection feature as a whole?

I think the revert is the right approach for 5.11, and you still have time
to post a new (preferably more than just compile-tested!) version for 5.12.

This is a new feature so we have time to get it right.

Thanks,

Will

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel