I have installed and regularly updated kernel 2.6.37 from early releases through 2.6.37-rc7-28.1. The kernel has run remarkably well on 11.3 install base, as well as 11.4 through MS5. By “remarkably well”, the performance is improved, boot time is quicker and the graphics and desktop effects seem smoother.
During the holidays, kernel upgrade 2.6.37-rc8-29.1 (and quickly thereafter, -29.2). After install, the display becomes (for lack of a better term) wonky. The display (actually on the physical monitor) is shifted right (and wrapped on the left), approximately 4 cm. The mouse/touchpad cannot move to the physical leftmost point, but can arrive there by moving (wrapping) to the right.
The actual “top” of the display exhibits jitter (vertical and horizontal) for approximately 2mm. Wondering at the possiblility of monitor problems, reboot to the default (2.6.34.7-05.1) works correctly, as does kernel 2.6.37-rc7-28.1.
This same problem occurs on a separate 11.3 Intel 855GM. As yet, this kernel update has not been applied to either of my ATI Radeon machines. Lastly, one of the more puzzling points is that a Gnome or KDE “screenshot” presents a perfect and normal display. (I can not even hazard a guess why !).
Kernel testing is in hiatus until this problem is corrected. As some may be aware, I am testing the advance kernel(s) for the Intel GMA HD (aka Arrandale/“Ironlake”). While also testing advance Xorg releases (for the Intel drivers), the above mentioned problem occurs on the supported Xorg release as well as the Factory versions.
SeanMc98, I am using the kernel 2.6.37-rc8 on openSUSE 11.3 without a problem. BUT, the very first download of this version produced a busted kernel for me. It would not run and my PC just locked up. I assume it was a download problem with the source files. I deleted ALL kernel 2.6.37-rc8 files, re-downloaded the kernel 2.6.37-rc8 source file again, recompiled it and all works well now. So, not the problem you saw, but might yours have been corrupted or something as well? How are you installing the new kernel by chance? Of course I am using SAKC myslef.
>
> SeanMc98, I am using the kernel 2.6.37-rc8 on openSUSE 11.3 without a
> problem. BUT, the very first download of this version produced a busted
> kernel for me. It would not run and my PC just locked up. I assume it
> was a download problem with the source files. I deleted ALL kernel
> 2.6.37-rc8 files, re-downloaded the kernel 2.6.37-rc8 source file again,
> recompiled it and all works well now. So, not the problem you saw, but
> might yours have been corrupted or something as well? How are you
> installing the new kernel by chance? Of course I am using SAKC myslef.
>
> Thank You,
>
>
here also kernel 2.6.37-rc8.29.2-x86_64 runs fine on openSUSE factory;
none of the disturbances the OP mentioned. it’s either a problem with his
local hardware or setup, or a corrupt download, as jdmcdaniel3 suggested.
i would definitely try to re-install all kernel related stuff & recompile
the video driver.
I installed the failing kernel (2.6.37-rc8-29.n) with YAST, including the Kernel\HEAD repositories, and installing from those repos. (I used the same procedure for all prior 2.6.37 kernels). Following your note, I removed all traces of the 2.6.37 kernels, and reinstalled from the Kernel\HEAD Factory repo (-29.2). Result is the same: right shifted/wrapped display, “Take Screenshot” shows correct.
On a somewhat positive note, this is the first readily apparent change in the handling of the Intel GMA HD (aka Arrandale/“Ironlake”). It is encouraging that this problem occurs before Xorg is initialized. I am looking at the boot.msg (and boot.omsg) logs.
The corresponding boot.msg for kernel 2.6.34.7-05.1 shows :
<6> 1.606797] vesafb: Truecolor: size=0:5:6:5, shift=0:11:5:0
<6> 1.606943] bootsplash 3.1.6-2004/03/31: looking for picture...
<6> 1.606945] bootsplash: silentjpeg size 157558 bytes
<4> 1.606946] bootsplash: does not fit into framebuffer.
<6> 1.617134] bootsplash: ...found (1024x768, 80639 bytes, v3).
<4> 1.685742] Console: switching to colour frame buffer device 124x44
<6> 1.743668] fb0: VESA VGA frame buffer device
<6> 1.744815] Non-volatile memory driver v1.3
<6> 1.744817] Linux agpgart interface v0.103
The highlighted lines in the failing boot.msg indicate a problem with the kernel intel interface (would this be considered a driver?). The line regarding “ERST” may be the telltale: ERST support is related to kernel support and firmware. (Reference: https://patchwork.kernel.org/patch/119192/ .
On a related note, my eradication of all 2.6.37 kernel(s), prior to a clean re-install, unfortunately removed the (functioning) 2.6.37-rc7-28.1 kernel. Other than compiling (SAKC’ing ?) this prior kernel, is their an EASY method to recreate this kernel ? (I may have just made a case for setting up multiple kernels! <sigh>).
On a related note, my eradication of all 2.6.37 kernel(s), prior to a clean re-install, unfortunately removed the (functioning) 2.6.37-rc7-28.1 kernel. Other than compiling (SAKC’ing ?) this prior kernel, is their an EASY method to recreate this kernel ? (I may have just made a case for setting up multiple kernels! <sigh>).
One reason for using SAKC was to be able to load different kernel versions besides what is at the head. For instance, this allows using the 2.6.36.2 while waiting for 2.6.37 to go final. If you have not used SAKC before, why not give it a try? It also had a companion file called KLIST, which simplifies the kernel selection, if you have more than one. Here is the pointer to SAKC, message #17 has the most recent version of SAKC and #23 for KLIST:
[quote="“jdmcdaniel3,post:6,topic:60407”]
One reason for using SAKC was to be able to load different kernel versions besides what is at the head. For instance, this allows using the 2.6.36.2 while waiting for 2.6.37 to go final. If you have not used SAKC before, why not give it a try? It also had a companion file called KLIST, which simplifies the kernel selection, if you have more than one. Here is the pointer to SAKC, message #17 has the most recent version of SAKC and #23 for KLIST:
I had put off using SAKC, and will attempt it presently. As to other kernels, as reported prior, 2.6.34.7-0.5 and -0.7 manifest correct display. Kernel 2.6.36-2.4.4 from Index of /repositories/Kernel:/stable/openSUSE_Factory presents similar correct behavior.
I will attempt SAKC to restore 2.6.37-rc7-28.1 (or any -rc7) presently. Should I report any problems on the referenced thread, or straight to you ?
Absolutely! Standard bugzilla, or should I add anything since this appears to be kernel-related, near to release ?
After a few delays, I finally got around to re-installing kernel 2.6.37-rc7 from archives. As reported earlier, this kernel correctly manages the Intel-driven display. This management is completely consistent with kernels 2.6.34.7-0.5.1, 2.6.34.7-0.7.1 and (stable repo) kernel 2.6.36.2-4.4. Further, kernels 2.6.37.rc8-29.1/29.2 continue to manifest the right-shift/wraparound of the display, along with image flicker and “jitter”.
As noted in post #9 (above, q.v.), this has been reported in bugzilla.
No activity seen on the above bug. Kernel 2.6.37 was released (Kernel 2.6.37 was release Today (1-5-2011) one day after the original report. I compiled/installed the release kernel, and the problem is either not present or corrected. I think that the bug report is now mooted.
After confirming that the OP was NOT a hardware/configuration problem, I compiled the problematic kernel (2.6.37-rc8), and the problem persisted. With the release, all is once again getting “closer”.