There were actually two different problems:
1) the initial issue with 3.16: no interest from upstream, too old
=> use a newer kernel;
2) new kernel (and even 3.16 now with recent security updates)
hangs before the graphic context is restored (black screen).
This second issue is what I reported upstream in:
https://bugs.freedesktop.org/show_bug.cgi?id=91585
But it's actually not related to the Intel X driver nor graphic kernel
support. The patches in this kernel.org bug fix the issue for me:
https://bugzilla.kernel.org/show_bug.cgi?id=104771
The problem is related to page table handling on resume for amd64.
Without patch, on 15 hangs my laptop locked up on average after ~2
suspend/resume cycles, with a maximum of 4 successful resumes before
locking up. I tried the first patch 9172981 on an Intel DRM 4.7.0-rc4
kernel and did 43 successful resumes, no hang (and closed the upstream
FDO bug).
Then I applied the following newer patches to a Jessie Debian backport
4.6 kernel (applied cleanly):
https://patchwork.kernel.org/patch/9208541/ (update of the
previous one)
https://patchwork.kernel.org/patch/9202321/ (additional issue)
I'm at 11 successful resume in a row, no other issue seen. Plus the
previous test, it looks good.
The patch 9208541 is in state "Awaiting Upstream".
This resume issue seems common on Broadwell Thinkpads (X1 3rd gen, X250,
T450). So it seems useful to track this patch to make sure it's
backported to Jessie (at least on a 4.6+ backport, not sure it's
practical for 3.16). What is the recommended next step:
1) reassign this bugs to linux kernel, with updated title "hang on
resume from hibernate"?
2) close this bug, open a new one?
Thanks
