[PATCH] gmux: Fix backlight set function

[PATCH] gmux: Fix backlight set function

[Grant Likely]

In commit 7e30ed6b, "gmux: Add generic write32 function", the new
gmux_write32() function didn't include the io offset when writing
registers. This means that backlight update writes never go to the right
place. This patch fixes it.

Tested on Apple Macbook 8,3

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Matthew Garrett <mjg@redhat.com>
---
 drivers/platform/x86/apple-gmux.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/platform/x86/apple-gmux.c b/drivers/platform/x86/apple-gmux.c
index dfb1a92..ee21056 100644
--- a/drivers/platform/x86/apple-gmux.c
+++ b/drivers/platform/x86/apple-gmux.c
@@ -101,7 +101,7 @@ static void gmux_pio_write32(struct apple_gmux_data *gmux_data, int port,
 
 	for (i = 0; i < 4; i++) {
 		tmpval = (val >> (i * 8)) & 0xff;
-		outb(tmpval, port + i);
+		outb(tmpval, gmux_data->iostart + port + i);
 	}
 }
 
-- 
1.7.9.5

Re: [PATCH] gmux: Fix backlight set function

[Seth Forshee]

On Thu, Sep 13, 2012 at 09:02:35PM +0100, Grant Likely wrote:
> In commit 7e30ed6b, "gmux: Add generic write32 function", the new
> gmux_write32() function didn't include the io offset when writing
> registers. This means that backlight update writes never go to the right
> place. This patch fixes it.
> 
> Tested on Apple Macbook 8,3
> 
> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
> Cc: Matthew Garrett <mjg@redhat.com>

I sent a patch for this several weeks ago [1], just waiting for Matthew
to pick it up.

Seth

[1] https://lkml.org/lkml/2012/8/21/721

Re: [PATCH] gmux: Fix backlight set function

[Matthew Garrett]

Sorry, I am literally catching up on these right now.

-- 
Matthew Garrett | mjg59@srcf.ucam.org

Re: [PATCH] gmux: Fix backlight set function

[David Woodhouse]

[-- Attachment #1: Type: text/plain, Size: 737 bytes --]

On Thu, 2012-09-13 at 21:02 +0100, Grant Likely wrote:
> In commit 7e30ed6b, "gmux: Add generic write32 function", the new
> gmux_write32() function didn't include the io offset when writing
> registers. This means that backlight update writes never go to the
> right
> place. This patch fixes it.
> 
> Tested on Apple Macbook 8,3 

Aha, thanks. I was going to have to get to that in the next few days.

I also noticed that since updating to the latest kernel and the latest
version of all the patches, I never get Intel graphics output working
after switching. And even the Radeon doesn't work right, if I switch
away and back. X just locks up after a few seconds and the graphics
engine isn't responding.

-- 
dwmw2

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 6171 bytes --]

Re: [PATCH] gmux: Fix backlight set function

[Grant Likely]

On Thu, Sep 13, 2012 at 9:35 PM, David Woodhouse <dwmw2@infradead.org> wrote:
> On Thu, 2012-09-13 at 21:02 +0100, Grant Likely wrote:
>> In commit 7e30ed6b, "gmux: Add generic write32 function", the new
>> gmux_write32() function didn't include the io offset when writing
>> registers. This means that backlight update writes never go to the
>> right
>> place. This patch fixes it.
>>
>> Tested on Apple Macbook 8,3
>
> Aha, thanks. I was going to have to get to that in the next few days.
>
> I also noticed that since updating to the latest kernel and the latest
> version of all the patches, I never get Intel graphics output working
> after switching. And even the Radeon doesn't work right, if I switch
> away and back. X just locks up after a few seconds and the graphics
> engine isn't responding.

What patches are you running on top of mainline for the mbp8,3? I'm
using the radeon blos loading patch and I've got the nasty direct io
writes in grub.cfg to disable the radeon before booting into the
kernel so that X picks up the i915 instead. I'm also passing
"i915.lvds_channel_mode=2 i915.modeset=1 i915.lvds_use_ssc=0
i915.i915_enable_rc6=1 vt.handoff=7" on the command line, but I
haven't experimented with it recently to see if that is still
required.

I've not had time to play with switcharoo on this system though.  :-(

g.

Re: [PATCH] gmux: Fix backlight set function

[David Woodhouse]

[-- Attachment #1: Type: text/plain, Size: 614 bytes --]

On Fri, 2012-09-14 at 14:06 +0100, Grant Likely wrote:
> What patches are you running on top of mainline for the mbp8,3? I'm
> using the radeon blos loading patch and I've got the nasty direct io
> writes in grub.cfg to disable the radeon before booting into the
> kernel so that X picks up the i915 instead. 

http://david.woodhou.se/mbp/ on top of the Fedora 3.5.3 kernel.

It was working with a similar set of patches on top of 3.5.0. Although
there was a patch from Matthew to make the i915 driver reprobe the LVDS
on switch, which I can't see in that list right now. Maybe that's it.

-- 
dwmw2

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 6171 bytes --]

Re: [PATCH] gmux: Fix backlight set function

[David Woodhouse]

[-- Attachment #1: Type: text/plain, Size: 1081 bytes --]

On Fri, 2012-09-14 at 14:06 +0100, Grant Likely wrote:
> What patches are you running on top of mainline for the mbp8,3?

A slightly simpler answer: I'm running the Fedora 18 3.6 kernel, with
these two patches backported:

commit f462ed939de67c20528bc08f11d2fc4f2d59c0d5
Author: Matthew Garrett <mjg@redhat.com>
Date:   Fri Jul 27 12:58:53 2012 -0400

    efifb: Skip DMI checks if the bootloader knows what it's doing

commit a7d92035a1d64242cbc1f79577e66f230fe59732
Author: Matthew Garrett <mjg@redhat.com>
Date:   Fri Jul 27 17:16:17 2012 -0400

    x86, EFI: Calculate the EFI framebuffer size instead of trusting the firmware

The F18 kernel has Matthew's stuff to handle PCI ROMs already
backported, although it's broken; it still prioritises the
IORESOURCE_ROM_SHADOW bit and returns 0xC0000 when there is a ROM from
EFI. You'll have to fix that too, so it matches commit 463edf567.

It still fails the same way — the Intel driver never manages to display.
So I'm back on my 3.5.0-based build until I have some time to work it
out.

-- 
dwmw2

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 6171 bytes --]