Bug 88501
| Summary: | AMD/ATI RS690M Console blanks to white | ||
|---|---|---|---|
| Product: | Drivers | Reporter: | business.kid |
| Component: | Video(DRI - non Intel) | Assignee: | drivers_video-dri |
| Status: | NEW --- | ||
| Severity: | normal | CC: | alexdeucher, fepaw95099, mlsemon35, pr0v1d3nc14 |
| Priority: | P1 | ||
| Hardware: | x86-64 | ||
| OS: | Linux | ||
| Kernel Version: | 3.6.0-rc5 to 3.18.0-rc4 | Subsystem: | |
| Regression: | Yes | Bisected commit-id: | |
| Attachments: |
Git Bisect log, & git visualise o/p
LSPCI -vnn Kernel Config 3.6.0-rc5+ possible fix Your patch slightly adjusted |
||
Created attachment 158191 [details]
LSPCI -vnn
Created attachment 158201 [details]
Kernel Config 3.6.0-rc5+
Does booting a recent kernel with radeon.backlight=0 fix the issue? Created attachment 158211 [details]
possible fix
Does this patch fix the issue (against 3.17 or newer kernel)?
<Does booting a recent kernel with radeon.backlight=0 fix the issue?> Yes, this fixes the issue on 3.18-rc4 <Does this patch fix the issue (against 3.17 or newer kernel)?> No, it does not. I get the same behaviour - first time it blanks to dark, every other time it blanks to white. I am a very rusty amateur at C, but I will adjust that patch very slightly here and have another go. Created attachment 158261 [details]
Your patch slightly adjusted
This seems to work. I know little of C, and less of graphics programming, but I got suspicious of the second 'if' construct inside a chain bracket, so I tweaked syntax. I built it here and tried it.
I now am sorted out, thank you. If you want something else tried, let me know. Otherwise, I assure you, I will *not* be building another kernel on that box for quite some time!
The code should be fine as is. Are you sure you tested the right kernel? Yes. I tested in order 1. 3.18.0-rc4 with no patch - white 2nd time. 2. 3.18.0-rc4 with the boot option - dark every time 3. 3.18.0-rc4 with your patch applied and no boot option - white 2nd time 4. 3.18.0-rc4 with a syntax error :-/ module did not load. 5. 3.18.0-rc4 with the attached suggested fix, and no boot option. - dark every time. I have built ~ 20 kernels in the last few days, and am not keeping any of them. The only reason I edited the patch was that I felt it unsafe to put an 'if' below an 'else' in the same chain bracket. I have met some compilers that don't respond to that as one would expect. I don't know what gcc does, and can't swear that the logic follows correctly. I just tried it. Have the same problem and no idea how to patch :( some may help? RS690M on AMD Thurion 64x2... The fix is already upstream. Alternatively, you can boot with radeon.backlight=0 on the kernel command line in grub. https://bugzilla.kernel.org/show_bug.cgi?id=202739 might be a fix. |
Created attachment 158181 [details] Git Bisect log, & git visualise o/p Without X running, linux consoles blank after a period adjustable by setterm. Mine does not. It goes white. If I invert the screen colours, it goes whiter. Backlight? I cloned the Linus-stable git and tried it. 3.18.0-rc4 twists this slightly - The FIRST time the screen blanks, it does so correctly; every other time, it blanks to white. I'm a hardware guy, not a software expert. Git bisect gives the output below. Details of the box also attached The box is a HP 6715S, twin core Turion, 3G ram. It's running Slackware-current as of Sept 30th, 2014. LSPCI output attached. Bios Version: ATI 02/13/08,1.