Regression? No, this hasn't worked since 2.6.11 (when S3 wake/sleep began to work for me). Distribution: Debian testing Hardware Environment: TP 600X, VGA controller: Neomagic Corporation NM2360 [MagicMedia 256ZX] Software Environment: XFree86 4.3.0 Problem Description: The VGA consoles are mostly illegible after S3 wake. During the wake, strange fine-mesh patterns start growing on the screen, sometimes with colored dots. Eventually it wakes up (and my wakeup script does 'chvt 1 ; chvt 7' just in case, to help reset the X console) and X is fine. But the vga console never recovers. If I switch to one with ctrl-alt-F1, say, the text jitters enough to be almost unreadable (though I can do basic sysadmin in a pinch). I can live with it, but it would be nice to figure out how to fix it. After a swsusp and resume (I don't use it right now until I fix the fan driver, see Bug #5000), the video card behaves properly. I tried the s3_bios, s3_mode tricks in Documentation/power/video.txt with no luck. Steps to reproduce:
Likely this is video card and video BIOS dependent. Unlikely that this can or will be addressed by the kernel. This is a ThinkPad 600X, yes? What video hardware has it got? Can you look into snagging the latest X and/or X video driver for it?
> This is a ThinkPad 600X, yes? What video hardware has it got? That's right. From lspci, the video hardware is: Neomagic Corporation NM2360 [MagicMedia 256ZX] > Can you look into snagging the latest X and/or X video driver for it? Sure, I can try the latest X.org. But would a different X video driver X help? I don't really understand the boundaries between kernel, bios, and X when it comes to the video card, but the illegible consoles happen even if I'm not running X (e.g. single user or runlevel 2). Is the idea that the X video driver can/will reset the vga adapter?
I just tried Xorg 6.8.2 with the same problem, plus a new one: Now the vga console does not go completely dark (I can see ghosts of its current state). Shutting the lid makes it go dark, but what is it doing before that?
Does the workarouds in Documentation/power/video.txt help in getting around the test console issue?
No -- with earlier kernels (2.6.11-13) I've tried the 4 combinations for acpi_sleep with s3_bios and s3_mode. I need s3_bios to get it to work at all, but s3_mode doesn't help the text console come back. -Sanjoy
Please try to use the s2ram tool (http://en.opensuse.org/S2ram) to solve the problem. If that helps, please close the bug.
Please reopen this bug if: - it is still present with kernel 2.6.19 and - you can provide the requested information.