On my MSI rs480m2 mobo using all kernels up to and including 2.6.15-gentoo-r5 I get a number of errors in dmesg that look like time.c: Lost 13 timer tick(s)! rip 0x95f51e) time.c: Lost 15 timer tick(s)! rip default_idle+0x36/0x80) APIC error on CPU0: 04(40) time.c: Lost 13 timer tick(s)! rip 0x823d49b) time.c: Lost 15 timer tick(s)! rip default_idle+0x36/0x80) time.c: Lost 13 timer tick(s)! rip 0x5b9bf2) time.c: Lost 15 timer tick(s)! rip default_idle+0x36/0x80) APIC error on CPU0: 40(40) time.c: Lost 1 timer tick(s)! rip default_idle+0x36/0x80) time.c: Lost 14 timer tick(s)! rip clear_page+0x7/0x10) time.c: Lost 15 timer tick(s)! rip default_idle+0x36/0x80) for example. I will attach hardware/config info below. Raphael
Created attachment 80157 [details] output of lspci -v
Created attachment 80158 [details] /var/log/messages for current boot
Created attachment 80159 [details] .config for current compile
Created attachment 80160 [details] /proc/cpuinfo
Created attachment 80161 [details] /proc/interrupts
Created attachment 80163 [details] dmesg with acpi=off and nolapic
Created attachment 80165 [details] dmesg with FREQ disabled and nolapic acpi=off
Created attachment 80169 [details] dmesg enable_timer_pin_1 clock=pmtmr
Created attachment 80171 [details] enable_timer_pin_1 clock=pmtmr no_timer_check
Created attachment 80172 [details] disable_timer_pin_1 clock=pmtmr no_timer_check
Created attachment 80361 [details] dmesg 2.6.16-rc4
Created attachment 80441 [details] Picture of 2.6.16-rc4-mm1 failing to boot 2.6.16-rc4-mm1 does not boot at all. Are there any more tests that I can do?
Created attachment 80442 [details] dmesg acpi=off nolapic of 2.6.16-rc4-mm1 I could get mm1 to boot with acpi off and nolapic
Created attachment 81749 [details] dmesg of 2.6.16-rc5 Status at 2.6.16-rc5: Bug still present.
I also have the MSI RS480M2 motherboard and have had similar APIC and timing issues for a while. I'm currently using the sys-kernel/gentoo-sources-2.6.15-r7 kernel and booting with just the "disable_timer_pin_1" option. I haven't done extensive testing but booting without this option means that the clock runs very fast (double speed) and so everything is very unstable. With this option enabled, however, the clock runs normally and system performance seems ok. I'm not convinced that the problem is totally solved for me though because I still get the following log messages (dmesg): APIC error on CPU0: 40(40) Losing some ticks... checking if CPU frequency changed. The first of these messages is logged every few minutes, i.e. very frequent, but the second is only seen maybe once every few days. I have this motherboard running in a headless system with the AMD Athlon64 3000+ processor, 512MB of RAM, one 80Gb IDE hard drive and one 200Gb SATA hard drive. The fact that it is headless (and isolated) probably limits what debugging I can do; i.e. it will take a lot of effort to get the system back up if any particular boot option prevents the system booting and network coming up. I'd be happy to post any logs/etc. if it is of any help.
Raphael, Can we get a status on your situation with this error? Do you still utilize the hardware and have the same errors with more recent kernels?
Please reopen if this is still an issue on the latest kernels.