Skip to contentSkip to conversation

v3.14-rc7: Backlight control broken on Dell XPS13

DISCUSSION Daniel Vetter 3 participants 7 messages Mar 17, 2014 1 message not in the archive

Subsystems (2)
lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

Please never just send a mail to maintainers, always cc mailing lists.

Quoted 7 lines · Romain Francoise

Moving from v3.13 to v3.14-rc7 breaks backlight control on my Dell XPS13, which I assume is because of this change:

http://git.kernel.org/linus/bc0bb9fd1c7810407ab810d204bbaecb255fddde

Is this right? Is backlight supposed to work correctly without the quirk now?

Yeah, it should work without the quirk. Have you tested with the patch reverted to confirm that it's indeed this commit and not something else (e.g. in the acpi code)?
-Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
#2 lore raw

Hi Daniel,

Quoted 1 line · Daniel Vetter

Please never just send a mail to maintainers, always cc mailing lists.

Yes, sorry, I sent this privately because it wasn't a fully-formed bug report; I just wanted Jani to confirm that it wasn't intentionally broken as of -rc7.

Quoted 2 lines

Have you tested with the patch reverted to confirm that it's indeed this commit and not something else (e.g. in the acpi code)?

I hadn't, but I've now reverted the revert and things work again as in v3.13. So as far as I can tell the quirk is still definitely needed.

Thanks!
-r

#3 lore raw
Quoted 10 lines · Romain Francoise

Hi Daniel,

Quoted 1 line · Daniel Vetter

Please never just send a mail to maintainers, always cc mailing lists.

Yes, sorry, I sent this privately because it wasn't a fully-formed bug report; I just wanted Jani to confirm that it wasn't intentionally broken as of -rc7.

Quoted 5 lines
Quoted 2 lines

Have you tested with the patch reverted to confirm that it's indeed this commit and not something else (e.g. in the acpi code)?

I hadn't, but I've now reverted the revert and things work again as in v3.13. So as far as I can tell the quirk is still definitely needed.

I still don't think having the quirk is the way to go. Let's fix this properly, shall we?

Please file a new bug on DRM/Intel component at [1]. Please attach intel_reg_dumper output (tool from intel-gpu-tools [2]) with and without the revert. Please also attach dmesg with drm.debug=0xe from early boot.

Do you have the snb or ivb model?

BR,
Jani.

[1] https://bugs.freedesktop.org/enter_bug.cgi?product=DRI
[2] http://cgit.freedesktop.org/xorg/app/intel-gpu-tools/

Quoted 3 lines

Thanks!
-r

Signature · 1 line
Jani Nikula, Intel Open Source Technology Center
#4 lore raw
Quoted 17 lines · Jani Nikula
Quoted 10 lines · Romain Francoise

Hi Daniel,

Quoted 1 line · Daniel Vetter

Please never just send a mail to maintainers, always cc mailing lists.

Yes, sorry, I sent this privately because it wasn't a fully-formed bug report; I just wanted Jani to confirm that it wasn't intentionally broken as of -rc7.

Quoted 5 lines
Quoted 2 lines

Have you tested with the patch reverted to confirm that it's indeed this commit and not something else (e.g. in the acpi code)?

I hadn't, but I've now reverted the revert and things work again as in v3.13. So as far as I can tell the quirk is still definitely needed.

Quoted 6 lines

I still don't think having the quirk is the way to go. Let's fix this properly, shall we?

Please file a new bug on DRM/Intel component at [1]. Please attach intel_reg_dumper output (tool from intel-gpu-tools [2]) with and without the revert. Please also attach dmesg with drm.debug=0xe from early boot.

The register dump from *before* i915 gets loaded is also interesting.

BR,
Jani.

Quoted 20 lines

Do you have the snb or ivb model?

BR,
Jani.

[1] https://bugs.freedesktop.org/enter_bug.cgi?product=DRI
[2] http://cgit.freedesktop.org/xorg/app/intel-gpu-tools/

Quoted 3 lines

Thanks!
-r

Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 1 line
Jani Nikula, Intel Open Source Technology Center
#5 lore raw
Quoted 4 lines · Jani Nikula

Please file a new bug on DRM/Intel component at [1]. Please attach intel_reg_dumper output (tool from intel-gpu-tools [2]) with and without the revert. Please also attach dmesg with drm.debug=0xe from early boot.

Done, https://bugs.freedesktop.org/show_bug.cgi?id=76276

Let me know if you need more info.

Quoted 1 line

Do you have the snb or ivb model?

IVB.

Quoted 1 line

The register dump from *before* i915 gets loaded is also interesting.

That's more difficult to procure, is it really necessary?

Thanks,
-r

#6 lore raw
In reply to Romain Francoise
Quoted 10 lines · Romain Francoise

Hi Daniel,

Quoted 1 line · Daniel Vetter

Please never just send a mail to maintainers, always cc mailing lists.

Yes, sorry, I sent this privately because it wasn't a fully-formed bug report; I just wanted Jani to confirm that it wasn't intentionally broken as of -rc7.

Quoted 5 lines
Quoted 2 lines

Have you tested with the patch reverted to confirm that it's indeed this commit and not something else (e.g. in the acpi code)?

I hadn't, but I've now reverted the revert and things work again as in v3.13. So as far as I can tell the quirk is still definitely needed.

Here's two patches to try, separately, *without* the quirk:

Patch #1:

drivers/gpu/drm/i915/intel_panel.c +1 -1
index cb058408c70e..ef26115dbe68 100644
@@ -659,7 +659,7 @@ static void pch_enable_backlight(struct intel_connector *connector)
 	/* This won't stick until the above enable. */
 	intel_panel_actually_set_backlight(connector, panel->backlight.level);
 
-	pch_ctl2 = panel->backlight.max << 16;
+	pch_ctl2 = panel->backlight.max << 16 | panel->backlight.level;
 	I915_WRITE(BLC_PWM_PCH_CTL2, pch_ctl2);
 
 	pch_ctl1 = 0;

Patch #2:

drivers/gpu/drm/i915/intel_panel.c +2 -0
index cb058408c70e..10849104fa1e 100644
@@ -669,6 +669,8 @@ static void pch_enable_backlight(struct intel_connector *connector)
 	I915_WRITE(BLC_PWM_PCH_CTL1, pch_ctl1);
 	POSTING_READ(BLC_PWM_PCH_CTL1);
 	I915_WRITE(BLC_PWM_PCH_CTL1, pch_ctl1 | BLM_PCH_PWM_ENABLE);
+
+	intel_panel_actually_set_backlight(connector, panel->backlight.level);
 }
 
 static void i9xx_enable_backlight(struct intel_connector *connector)
Signature · 1 line
Jani Nikula, Intel Open Source Technology Center
#7 lore raw
Quoted 1 line · Jani Nikula

Here's two patches to try, separately, *without* the quirk: [...]

Okay, I've tried both and neither works. However, I've noticed that while the backlight level never changes even though I'm changing it with Fn+F4/Fn+F5, it *does* get registered somewhere because if I bring it all the day down and reboot, the brightness is reduced in the next BIOS POST (loading i915 gets it back to 100%).

Thanks,
-r