Bug 60161 - Fujitsu E753: Backlight dark onbootup
Summary: Fujitsu E753: Backlight dark onbootup
Status: CLOSED CODE_FIX
Alias: None
Product: ACPI
Classification: Unclassified
Component: Power-Video (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: Lan Tianyu
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-06-26 09:15 UTC by Jan Hinnerk Stosch
Modified: 2013-11-25 07:57 UTC (History)
2 users (show)

See Also:
Kernel Version: 3.9.7
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments
acpidump (256.50 KB, application/octet-stream)
2013-06-26 13:04 UTC, Jan Hinnerk Stosch
Details
errors (285 bytes, application/octet-stream)
2013-06-26 13:06 UTC, Jan Hinnerk Stosch
Details
acpidump without boot option (257.80 KB, application/octet-stream)
2013-06-26 13:13 UTC, Jan Hinnerk Stosch
Details
dmidecode (12.04 KB, application/octet-stream)
2013-06-28 13:23 UTC, Jan Hinnerk Stosch
Details

Description Jan Hinnerk Stosch 2013-06-26 09:15:42 UTC
On my Fujitsu Lifebook E753 I experienced some "regressions" with
current kernel versions (e.g. 3.9.7) compared to linux 3.5 (used in Ubuntu 12.04
for instance):
- on bootup backlight is extremely dark (I can change it though)
- the wireless button does not work any more
Recently I did a bisect and found out, that the "regression" was
introduced with commit "a57f7f9175b8ccbc9df83ac13860488913115de4".
Searching the internet I read about a bootloader option
"acpi_osi="!Windows 2012" which solved the issue for me.
I reported it on the linux-acpi mailing list and was asked to report it here and add a acpidump. Perhaps I have done something wrong but all I get is:

Linux version not implemented yet
Could not get ACPI tables, AE_SUPPORT

If you need some additional information tell me.

Thanks.
Comment 1 Lan Tianyu 2013-06-26 09:20:11 UTC
Please "git clone git://github.com/acpica/acpica.git" and "make && make install"
Try acpidump again.
Comment 2 Jan Hinnerk Stosch 2013-06-26 13:04:47 UTC
Created attachment 106081 [details]
acpidump

Ok, find attached the acpidump output.
Comment 3 Jan Hinnerk Stosch 2013-06-26 13:06:06 UTC
Created attachment 106091 [details]
errors

acpidump threw some errors which are attached.
Comment 4 Jan Hinnerk Stosch 2013-06-26 13:13:33 UTC
Created attachment 106101 [details]
acpidump without boot option

Booting without "acpi_osi="!Windows 2012" gives a different acpidump which is attached.
Comment 5 Lan Tianyu 2013-06-28 08:49:49 UTC
Please add kernel param "video.use_bios_initial_backlight=0" and check whether this resolves bootup backlight issue. For wireless key, it's in charge of fujitsu-laptop platform driver and it doesn't follow ACPI spec. So it is out of ACPI scope.
Comment 6 Jan Hinnerk Stosch 2013-06-28 09:09:41 UTC
Yes, using this kernel parameter also resolves the backlight issue. Thank you. However, if I remove the "acpi_osi="!Windows 2012" parameter, my wireless key is not working any more. So should I contact the fujitsu-laptop guys?
Comment 7 Lan Tianyu 2013-06-28 13:08:14 UTC
Please provide the output of dmidecode and I can do a quirk for this machine.
Your can apply a bug to drivers product and platform_x86 component for the wireless button bug.
Comment 8 Jan Hinnerk Stosch 2013-06-28 13:23:18 UTC
Created attachment 106221 [details]
dmidecode

Ok, thanks.
Find attached the dmidecode. It was created with boot parameter "acpi_osi="!Windows 2012"". If this has any influence and you need the dmidecode without or with another boot parameter - tell me.
Comment 9 Lan Tianyu 2013-06-28 13:30:25 UTC
Please try this patch without any kernel param and backlight should work on bootup.

diff --git a/drivers/acpi/video.c b/drivers/acpi/video.c
index 5d7075d..e441876 100644
--- a/drivers/acpi/video.c
+++ b/drivers/acpi/video.c
@@ -450,6 +450,14 @@ static struct dmi_system_id video_dmi_table[] __initdata = {
        },
        {
         .callback = video_ignore_initial_backlight,
+        .ident = "Fujitsu E753",
+        .matches = {
+               DMI_MATCH(DMI_BOARD_VENDOR, "FUJITSU"),
+               DMI_MATCH(DMI_PRODUCT_NAME, "LIFEBOOK E753"),
+               },
+       },
+       {
+        .callback = video_ignore_initial_backlight,
         .ident = "HP Pavilion dm4",
         .matches = {
                DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
Comment 10 Jan Hinnerk Stosch 2013-06-28 14:31:54 UTC
Yes, it works as intended now. Thank you.

I had to change one line to apply the patch, though:
.ident = "HP Pavilion dm4",
failed.
With:
.ident = "HP Pavilion m4",
it worked, so I guess it was a typo?
Comment 11 Lan Tianyu 2013-06-28 14:54:00 UTC
No, I think your tree lack the follow commit. I base on linux-pm bleeding-edge branch. Which kernel you apply the patch? Can you try v3.10-rc7?

commit 771d09b3c4c45d4d534a83a68e6331b97fd82e15
Author: Gustavo Maciel Dias Vieira <gustavo@sagui.org>
Date:   Mon Mar 4 15:23:37 2013 +0000

    ACPI video: ignore BIOS backlight value for HP dm4

    On a HP Pavilion dm4 laptop the BIOS sets minimum backlight on boot,
    completely dimming the screen. Ignore this initial value for this
    machine.

    Signed-off-by: Gustavo Maciel Dias Vieira <gustavo@sagui.org>
    Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
diff --git a/drivers/acpi/video.c b/drivers/acpi/video.c
index 313f959..f8a28b3 100644
--- a/drivers/acpi/video.c
+++ b/drivers/acpi/video.c
@@ -447,6 +447,14 @@ static struct dmi_system_id video_dmi_table[] __initdata = {
                DMI_MATCH(DMI_PRODUCT_NAME, "HP Folio 13 - 2000 Notebook PC"),
                },
        },
+       {
+        .callback = video_ignore_initial_backlight,
+        .ident = "HP Pavilion dm4",
+        .matches = {
+               DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
+               DMI_MATCH(DMI_PRODUCT_NAME, "HP Pavilion dm4 Notebook PC"),
+               },
+       },
        {}
 };
Comment 12 Jan Hinnerk Stosch 2013-06-28 15:10:20 UTC
You are right, sorry. I used the current stable tree.
With rc7 the patch applies.
Comment 13 Lan Tianyu 2013-06-28 16:05:55 UTC
Thanks for test. I will send this patch to ACPI maillist soon.
Comment 14 Jan Hinnerk Stosch 2013-06-28 17:28:04 UTC
You're welcome. Thank you for the quick response and help.
Comment 15 Aaron Lu 2013-07-22 01:15:18 UTC
Author: Lan Tianyu <tianyu.lan@intel.com>
Date:   Tue Jul 16 10:07:21 2013 +0800

    ACPI / video: ignore BIOS initial backlight value for Fujitsu E753

entered Linus's tree as of v3.11-rc2.
Comment 16 Aaron Lu 2013-11-13 01:41:25 UTC
We have a patch to solve all these initial black screen problem:

commit 2c62333a408f5badd2d2ffd7177f95deeccc5ca4
Author: Aaron Lu <aaron.lu@intel.com>
Date:   Wed Nov 6 09:07:10 2013 +0800

    ACPI / video: Quirk initial backlight level 0

It is now in Rafael's linux-next branch:
http://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git

Can someone please kindly give it a test? If confirmed, we can save an entry in the DMI table, thanks.

When using that branch, please manually remove the corresponding entry for your system in the DMI table, or the test is obviously useless :-)
Comment 17 Aaron Lu 2013-11-13 02:10:24 UTC
For easy testing, I've prepared a branch for you:
https://github.com/aaronlu/linux acpi_video_dmi

It is based on Rafael's linux-next branch(thus it already has the above mentioned fix patch), with only one patch on top:
Remove the affected system from the DMI table to test if the above mentioned fix patch can work for it.
Comment 18 Jan Hinnerk Stosch 2013-11-22 20:11:19 UTC
Sry for my late response. I saw you already applied the patch and I can confirm, that it works for my Fujitsu E753, too.
Thanks!

Just FYI:
There is one short stage in the boot process, when the screen gets darker. It is reset to normal after half a second or sth, but this is different to the behaviour beforehand.
Comment 19 Aaron Lu 2013-11-25 01:18:17 UTC
(In reply to Jan Hinnerk Stosch from comment #18)
> Sry for my late response. I saw you already applied the patch and I can
> confirm, that it works for my Fujitsu E753, too.
> Thanks!

Thanks for the confirm!

> 
> Just FYI:
> There is one short stage in the boot process, when the screen gets darker.
> It is reset to normal after half a second or sth, but this is different to
> the behaviour beforehand.

Not sure what caused this, that commit alone shouldn't affect it in my opinion.
Comment 20 Jan Hinnerk Stosch 2013-11-25 07:57:31 UTC
(In reply to Aaron Lu from comment #19)
> Not sure what caused this, that commit alone shouldn't affect it in my
> opinion.
Ok, I'll try to find out more details if I get some time. Thanks anyway..

Note You need to log in before you can comment on or make changes to this bug.