Am Dienstag, 15. August 2006 15:14 schrieb Lennart Sorensen: > On Tue, Aug 15, 2006 at 06:27:32AM +0200, Hans wrote: > > Sorry, I maybe did not ask correctly. It is not the problem, that the > > time is not shown correctly. The background is, that other timings are > > running in double speed, too (i.e. keyboard clock and some other). This > > is a known problem on AMD mobile processors and was discussed in earlier > > days. The solution of it, was to add "disable_timer_pin_1" in the boot > > vcommand on grub or lilo. Now I read about adding "noapictimer" should > > solve this, too. My question aimed to an technical answer, if the > > commands "disable_timer_pin_1" and "noapictimer" are doing the same, or > > if they both solve the mentioned problem in different ways. > > I believe the problem occoured with ATI chipsets on laptops. As far as > I have understood it, the problem is that the timer interrupts occour > both on the 8259 interrupt controller, and through the apic. I believe > 'disable_timer_pin_1' makes the kernel ignore the 8259 interrupt for the > timer, and that 'noapictimer' ignores the apic interrupt for the timer. > Since the problem seems to be getting two interrupts for every timer > event, one for each interrupt method, it makes sense that disabling > either one will solve the problem. It doesn't matter which one you > disable as long as you disable one of the two. > > It really looks like a bug in the design of the chipset, although it may > just be that the BIOS/ACPI tables are done wrong, which is rather common > it seems. Too often the ACPI tables for windows work, but are > incomplete or wrong for other operating systems. Apparently on such > systems, telling linux to lie to acpi and pretend to be windows xp often > solves such strange problems. > > -- > Len Sorensen
Hi Len ! It seems, that in last kernel 2.6.17 this problem is solved. I read the documentation of the kernel, and (as you wrote), these timers just misbehave if you have a chipset of ATI. This is the case on my notebook: ATI chipset ! Well, without this kernel-command it is running fine with the newest kernel. So there are no problems any more. I additionally did hope, that this could be the reason, why my 3D-accelertion with the "fglrx"-river is still slow. All 3D-functions are o.k., I get the box with the rotating wheels when starting "fgl_glxgears", but they are rather slow. About 50 FPS, that is speed as mesa-glx shows ! Should be 600 FPS, that would be o.k. "fglrxinfo" shows the correct driver-version, and "glxinfo |grep direct" shows "direct rendering: Yes" So verything seems t be o.k. with the software, and my idea was, something else, but not the driver is braking my system. You can believe me: I checked really everything, and do not know, where to look now. Additonally I tested every xorg.conf I found in the web and checked a whole bunch of settings. Sadly no success ! ATI really sux ! (This notebook I got as new of a guarantee as exchanged, I had Nvidia-card on the other noteook). And I have never heard of someone who got 3d-acceleration really fast running with a pure 64-bit-system with ATI...... Thats a little bit information about the background, maybe other people this could help, too Thanks for all your help ! best regards Hans -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]