Did a bisect on the drm i915 driver, came up with no root cause. Did a full bisect between 4.18 and 4.20, bisected down to a PTI commit:
7757d607c6b31867777de42e1fb0210b9c5d8b70 is the first bad commit commit 7757d607c6b31867777de42e1fb0210b9c5d8b70 Author: Joerg Roedel <email address hidden> Date: Wed Jul 18 11:41:14 2018 +0200
x86/pti: Allow CONFIG_PAGE_TABLE_ISOLATION for x86_32
Allow PTI to be compiled on x86_32.
This seems to make sense, it's x86_32 specific and explains the more random nature of the crashes.
Did a bisect on the drm i915 driver, came up with no root cause.
Did a full bisect between 4.18 and 4.20, bisected down to a PTI commit:
7757d607c6b3186 7777de42e1fb021 0b9c5d8b70 is the first bad commit 7777de42e1fb021 0b9c5d8b70
commit 7757d607c6b3186
Author: Joerg Roedel <email address hidden>
Date: Wed Jul 18 11:41:14 2018 +0200
x86/pti: Allow CONFIG_ PAGE_TABLE_ ISOLATION for x86_32
Allow PTI to be compiled on x86_32.
This seems to make sense, it's x86_32 specific and explains the more random nature of the crashes.