linux/drivers
Rodrigo Siqueira ca08a1725d drm/amd/display: Remove wrong pipe control lock
When using a device based on DCN32/321,
we have an issue where a second
4k@60Hz display does not light up,
and the system becomes unresponsive
for a few minutes. In the debug process,
it was possible to see a hang
in the function dcn20_post_unlock_program_front_end
in this part:

for (j = 0; j < TIMEOUT_FOR_PIPE_ENABLE_MS*1000
	&& hubp->funcs->hubp_is_flip_pending(hubp); j++)
	mdelay(1);
}

The hubp_is_flip_pending always returns positive
for waiting pending flips which is a symptom of
pipe hang. Additionally, the dmesg log shows
this message after a few minutes:

  BUG: soft lockup - CPU#4 stuck for 26s!
  ...
  [  +0.000003]  dcn20_post_unlock_program_front_end+0x112/0x340 [amdgpu]
  [  +0.000171]  dc_commit_state_no_check+0x63d/0xbf0 [amdgpu]
  [  +0.000155]  ? dc_validate_global_state+0x358/0x3d0 [amdgpu]
  [  +0.000154]  dc_commit_state+0xe2/0xf0 [amdgpu]

This confirmed the hypothesis that we had a pipe
hanging somewhere. Next, after checking the
ftrace entries, we have the below weird
sequence:

 [..]
  2)               |        dcn10_lock_all_pipes [amdgpu]() {
  2)   0.120 us    |          optc1_is_tg_enabled [amdgpu]();
  2)               |          dcn20_pipe_control_lock [amdgpu]() {
  2)               |            dc_dmub_srv_clear_inbox0_ack [amdgpu]() {
  2)   0.121 us    |              amdgpu_dm_dmub_reg_write [amdgpu]();
  2)   0.551 us    |            }
  2)               |            dc_dmub_srv_send_inbox0_cmd [amdgpu]() {
  2)   0.110 us    |              amdgpu_dm_dmub_reg_write [amdgpu]();
  2)   0.511 us    |            }
  2)               |            dc_dmub_srv_wait_for_inbox0_ack [amdgpu]() {
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
 [..]

We are not expected to read from dmub register
so many times and for so long. From the trace log,
it was possible to identify that the function
dcn20_pipe_control_lock was triggering the dmub
operation when it was unnecessary and causing
the hang issue. This commit drops the unnecessary
dmub code and, consequently, fixes the second display not
lighting up the issue.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2022-10-24 14:44:03 -04:00
..
accessibility
acpi More ACPI updates for 6.1-rc1 2022-10-10 13:28:06 -07:00
amba
android Scheduler changes for v6.1: 2022-10-10 09:10:28 -07:00
ata ata changes for 6.1-rc1 2022-10-07 10:48:49 -07:00
atm
auxdisplay
base Interrupt subsystem updates: 2022-10-12 10:23:24 -07:00
bcma Interrupt subsystem updates: 2022-10-12 10:23:24 -07:00
block Random number generator fixes for Linux 6.1-rc1. 2022-10-16 15:27:07 -07:00
bluetooth
bus Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
cdrom
char prandom: remove unused functions 2022-10-11 17:42:58 -06:00
clk This is the final part of the clk patches for this merge window. 2022-10-16 11:08:19 -07:00
clocksource A boring time, timekeeping, timers update: 2022-10-10 10:16:00 -07:00
comedi
connector
counter gpio updates for v6.1-rc1 2022-10-08 09:46:29 -07:00
cpufreq Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
cpuidle RISC-V Patches for the 6.1 Merge Window, Part 1 2022-10-09 13:24:01 -07:00
crypto This update includes the following changes: 2022-10-10 13:04:25 -07:00
cxl
dax libnvdimm for 6.1 2022-10-14 18:41:41 -07:00
dca
devfreq
dio
dma treewide: use get_random_bytes() when possible 2022-10-11 17:42:58 -06:00
dma-buf whack-a-mole: cropped up open-coded file_inode() uses... 2022-10-06 17:22:11 -07:00
edac Merge patch series "Use composable cache instead of L2 cache" 2022-10-13 11:07:13 -07:00
eisa
extcon Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
firewire
firmware Merge branch 'dmi-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging 2022-10-11 10:44:20 -07:00
fpga Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
fsi
gnss
gpio Interrupt subsystem updates: 2022-10-12 10:23:24 -07:00
gpu drm/amd/display: Remove wrong pipe control lock 2022-10-24 14:44:03 -04:00
greybus
hid Input updates for 6.1 merge window: 2022-10-11 10:53:25 -07:00
hsi
hte
hv hyperv-next for 6.1 2022-10-10 13:59:01 -07:00
hwmon Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
hwspinlock
hwtracing Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
i2c - correct a variable type in the new pci1xxxx driver 2022-10-11 11:03:42 -07:00
i3c i3c: master: Remove the wrong place of reattach. 2022-10-12 23:45:29 +02:00
idle
iio Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
infiniband treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
input Input updates for 6.1 merge window: 2022-10-11 10:53:25 -07:00
interconnect
iommu IOMMU Updates for Linux v6.1: 2022-10-10 13:20:53 -07:00
ipack Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
irqchip Interrupt subsystem updates: 2022-10-12 10:23:24 -07:00
isdn mISDN: hfcpci: Fix use-after-free bug in hfcpci_softirq 2022-10-09 19:11:54 +01:00
leds This is very quiet release for LEDs, pca963 got blinking support and 2022-10-14 13:14:03 -07:00
macintosh powerpc updates for 6.1 2022-10-09 14:05:15 -07:00
mailbox mailbox: qcom-ipcc: flag IRQ NO_THREAD 2022-10-05 21:51:58 -05:00
mcb
md treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
media treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
memory
memstick
message
mfd - Core Frameworks 2022-10-07 11:24:20 -07:00
misc treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
mmc Random number generator fixes for Linux 6.1-rc1. 2022-10-16 15:27:07 -07:00
most
mtd Random number generator fixes for Linux 6.1-rc1. 2022-10-16 15:27:07 -07:00
mux
net Random number generator fixes for Linux 6.1-rc1. 2022-10-16 15:27:07 -07:00
nfc
ntb
nubus
nvdimm libnvdimm for 6.1 2022-10-14 18:41:41 -07:00
nvme Random number generator fixes for Linux 6.1-rc1. 2022-10-16 15:27:07 -07:00
nvmem
of Devicetree updates for v6.1: 2022-10-10 13:13:51 -07:00
opp
parisc parisc architecture fixes and updates for kernel v6.1-rc1: 2022-10-14 12:10:01 -07:00
parport
pci pci-v6.1-fixes-1 2022-10-15 16:36:38 -07:00
pcmcia
peci
perf arm64 fixes: 2022-10-14 12:38:03 -07:00
phy pci-v6.1-changes 2022-10-11 11:08:18 -07:00
pinctrl Interrupt subsystem updates: 2022-10-12 10:23:24 -07:00
platform LoongArch changes for v6.1 2022-10-12 10:35:20 -07:00
pnp Merge branches 'acpi-apei', 'acpi-wakeup', 'acpi-reboot' and 'acpi-thermal' 2022-10-10 18:11:11 +02:00
power power supply and reset changes for the v6.1 series 2022-10-07 11:48:30 -07:00
powercap Scheduler changes for v6.1: 2022-10-10 09:10:28 -07:00
pps
ps3
ptp ] ptp: ocp: remove symlink for second GNSS 2022-10-10 08:37:24 +01:00
pwm pwm: Changes for v6.1-rc1 2022-10-07 11:32:10 -07:00
rapidio
ras
regulator - Core Frameworks 2022-10-07 11:24:20 -07:00
remoteproc remoteproc: virtio: Fix warning on bindings by removing the of_match_table 2022-10-05 09:20:44 -06:00
reset Here's the main clk pull request for this merge window. We have some 2022-10-08 10:06:48 -07:00
rpmsg
rtc RTC for 6.1 2022-10-14 18:36:42 -07:00
s390 s390 updates for the 6.1 merge window #2 2022-10-14 11:36:05 -07:00
sbus
scsi treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
sh
siox
slimbus
soc Merge patch series "Use composable cache instead of L2 cache" 2022-10-13 11:07:13 -07:00
soundwire soundwire updates for 6.1-rc1 2022-10-07 16:13:55 -07:00
spi spi: Updates for v6.1 2022-10-04 19:36:53 -07:00
spmi spmi: pmic-arb: increase SPMI transaction timeout delay 2022-09-30 14:33:23 +02:00
ssb
staging Here's the main clk pull request for this merge window. We have some 2022-10-08 10:06:48 -07:00
target treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
tc
tee - Yu Zhao's Multi-Gen LRU patches are here. They've been under test in 2022-10-10 17:53:04 -07:00
thermal More thermal control updates for 6.1-rc1 2022-10-10 13:52:14 -07:00
thunderbolt treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
tty parisc architecture fixes and updates for kernel v6.1-rc1: 2022-10-14 12:10:01 -07:00
ufs SCSI misc on 20221007 2022-10-07 12:33:18 -07:00
uio
usb - Yu Zhao's Multi-Gen LRU patches are here. They've been under test in 2022-10-10 17:53:04 -07:00
vdpa virtio: fixes, features 2022-10-10 14:02:53 -07:00
vfio VFIO updates for v6.1-rc1 2022-10-12 14:46:48 -07:00
vhost virtio: fixes, features 2022-10-10 14:02:53 -07:00
video Random number generator fixes for Linux 6.1-rc1. 2022-10-16 15:27:07 -07:00
virt Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
virtio virtio_pci: use irq to detect interrupt support 2022-10-13 09:33:03 -04:00
vlynq
w1 Char/Misc and other driver changes for 6.1-rc1 2022-10-08 08:56:37 -07:00
watchdog linux-watchdog 6.1-rc1 tag 2022-10-13 10:31:13 -07:00
xen xen: branch for v6.1-rc1 2022-10-12 14:39:38 -07:00
zorro
Kconfig
Makefile