From: Suraj Kandpal <suraj.kandpal@intel.com>
Date: Thu, 12 Sep 2024 01:25:45 +0000 (+0530)
Subject: drm/xe/display: Do not do intel_fbdev_set_suspend during runtime
X-Git-Tag: v6.13-rc1~122^2~19^2~162
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=5422d30957570b0f0283f8ad4d0dd45637c11db7;p=thirdparty%2Fkernel%2Flinux.git

drm/xe/display: Do not do intel_fbdev_set_suspend during runtime

Do not do intel_fbdev_set_suspend during runtime_suspend/resume
functions. This cause a big circular lock_dep splat.

kworker/0:4/198 is trying to acquire lock:
<4> [77.185594] ffffffff83398500 (console_lock){+.+.}-{0:0}, at:
intel_fbdev_set_suspend+0x169/0x1f0 [xe]
<4> [77.185947]
but task is already holding lock:
<4> [77.185949] ffffffffa09e9460
(xe_pm_runtime_lockdep_map){+.+.}-{0:0}, at:
xe_pm_runtime_suspend+0x51/0x3f0 [xe]
<4> [77.186262]
which lock already depends on the new lock.
<4> [77.186264]
the existing dependency chain (in reverse order) is:
<4> [77.186266]
-> #2 (xe_pm_runtime_lockdep_map){+.+.}-{0:0}:
<4> [77.186276]        pm_runtime_lockdep_prime+0x2f/0x50 [xe]
<4> [77.186572]        xe_pm_runtime_resume_and_get+0x29/0x90 [xe]
<4> [77.186867]        intelfb_create+0x150/0x390 [xe]
<4> [77.187197]
__drm_fb_helper_initial_config_and_unlock+0x31c/0x5e0 [drm_kms_helper]
<4> [77.187243]        drm_fb_helper_initial_config+0x3d/0x50
[drm_kms_helper]
<4> [77.187274]        intel_fbdev_client_hotplug+0xb1/0x140 [xe]
<4> [77.187603]        drm_client_register+0x87/0xd0 [drm]
<4> [77.187704]        intel_fbdev_setup+0x51c/0x640 [xe]
<4> [77.188033]        intel_display_driver_register+0xb7/0xf0 [xe]
<4> [77.188438]        xe_display_register+0x21/0x40 [xe]
<4> [77.188809]        xe_device_probe+0xa8d/0xbf0 [xe]
<4> [77.189035]        xe_pci_probe+0x333/0x5b0 [xe]
<4> [77.189330]        local_pci_probe+0x48/0xb0
<4> [77.189341]        pci_device_probe+0xc8/0x280
<4> [77.189351]        really_probe+0xf8/0x390
<4> [77.189362]        __driver_probe_device+0x8a/0x170
<4> [77.189373]        driver_probe_device+0x23/0xb0

Signed-off-by: Suraj Kandpal <suraj.kandpal@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240912012545.702032-3-suraj.kandpal@intel.com
---

diff --git a/drivers/gpu/drm/xe/display/xe_display.c b/drivers/gpu/drm/xe/display/xe_display.c
index cdfc87bd78bc2..b1730b5816561 100644
--- a/drivers/gpu/drm/xe/display/xe_display.c
+++ b/drivers/gpu/drm/xe/display/xe_display.c
@@ -333,7 +333,9 @@ void xe_display_pm_suspend(struct xe_device *xe, bool runtime)
 	 * properly.
 	 */
 	intel_power_domains_disable(xe);
-	intel_fbdev_set_suspend(&xe->drm, FBINFO_STATE_SUSPENDED, true);
+	if (!runtime)
+		intel_fbdev_set_suspend(&xe->drm, FBINFO_STATE_SUSPENDED, true);
+
 	if (!runtime && has_display(xe)) {
 		drm_kms_helper_poll_disable(&xe->drm);
 		intel_display_driver_disable_user_access(xe);
@@ -420,7 +422,8 @@ void xe_display_pm_resume(struct xe_device *xe, bool runtime)
 
 	intel_opregion_resume(display);
 
-	intel_fbdev_set_suspend(&xe->drm, FBINFO_STATE_RUNNING, false);
+	if (!runtime)
+		intel_fbdev_set_suspend(&xe->drm, FBINFO_STATE_RUNNING, false);
 
 	intel_power_domains_enable(xe);
 }