Forward to upstream for more discussion.
於 五，2011-05-13 於 17:27 +0800，Michael Chang 提到：
Melchior,
I tried to puzzle out the possible cause of the problem, but sorry that no further progress to give you. I'd like to debug this issue, also did Joey. But there's more questions in my mind, made me feel not able to proceed any further.. :(
- is_backlight_combination_mode() really returns 0x4 ..? I understand
your words, but if this is true then Joey Lee provided patch should log out some information about changing the LPBC
set backlight lbpc = ...
but I didn't see any of the log output .. :(
- I dig the old mailing thread .. found a post
http://marc.info/?l=linux-kernel&m=130462815211684&w=2
If you rebase this patch to your currently git commit, will it fixes all problems for you now (only vanilla kernel with this patch)? (You replied that this patch works for you ..) Also another interesting finding...
- if (!lbpc)
combo_mode = 0;Looks like the purpose of it is to NOT touch the LPBC registers if lbpc is zero ...(since this forces is_backlight_combination_mode return 0). This would make contradiction about the assumption that you could change to back-light via LPBC .. (using setpci is actually access the LPBC registers .. IMHO).
Total a mess for me now .. Did I made anything wrong .. I know I may lose some details.
I'm thinking is it ok for us to discuss in IRC, the communication would be more efficient. Would this be convenient for you ? This is up to you.
Have nice weekend .. :)
Regards, Michael Chang
2011/5/10 Melchior FRANZ melchior.franz@gmail.com:
- Michael Chang -- Tuesday 10 May 2011:
Could you please try this patch and get the log ? We wonder why is_backlight_combination_mode () returns false.
This information was already buried in the bugzilla thread:
https://bugzilla.kernel.org/show_bug.cgi?id=31522 "It turned out that on this machine INTEL_INFO(dev)->gen equals 4, and is_backlight_combination_mode() returns 0x40000000."
But to say it again in your words: :-)
[drm:is_backlight_combination_mode], BLM_COMBINATION_MODE = 1073741824 (0x40000000)
6x during boot-up, and several times later when changing the backlight brightness.
This was with 8b061610dac3a3b89770c85ad63b481a47b0c38e. And now I have a little shocker for you (and me): because this was a vanilla kernel (apart from these debug messages), the screen went black again, like I knew it. But pressing the "brightness down" key turns the backlight on! I can't believe that I haven't tested that. I guess I've only tried "brightness up" and "display toggle". Those don't turn backlight on. Or maybe somethine else relevant meanwhile changed in the i915 drivers. (I've regularly been updating to HEAD.)
So, the problem was just the initial state all the time?
m.