failed to send pre message kernel output delays booting/suspending/resuming

Submitted by MichaelLong

Assigned to Default DRI bug account

Link to original bug (#100891)

Description

Created attachment 131160
dmesg output (kernel 4.9)

Starting with kernel version 4.9 the boot process, s3 suspend and s3 resume is delayed by kernel messages like the following

[ 17.322912]
failed to send pre message 148 ret is 0
[ 17.543482]
failed to send message 148 ret is 0

for a very long time, e.g. 10 min (see attachment for a full dmesg log). When the system is fully booted everything else seems to be fine.


I'm not sure if it is related at all, but I did a quick bisection: between 4.9-rc2 and 4.9-rc3, with the introduction of commit 3b496626ee8f07919256a4e99cddf42ecd4ba891 I had to supply a new firmware file (topaz_k_smc.bin) otherwise my screen remains black. After supplying the missing firmware-file from from linux-firmware.git I noticed the long boot delay and the messages.

Kernel series 4.11-rcX reduces the delays drastically, down to around 1-2 minutes but after a resume from s3 the GPUs gets hot very quickly without fans kicking in until it reaches some sort of emergency mode with short bursts of jet engine like fans.

This happens on the kernels 4.9-rc3 up to the latest 4.11-rc, always tested with the latest set of firmware files.

My GPU-hardware is a Sapphire R9 380X:

05:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Amethyst XT [Radeon R9 M295X Mac Edition] (rev f1)
05:00.1 Audio device: Advanced Micro Devices, Inc. [AMD/ATI] Device aad8

Attachment 131160, "dmesg output (kernel 4.9)":
dmesg