I've been trying to test the latest git under a "normal" workload for my
laptop and the only functionality change from my previous functioning kernel
(what I'm running right now, 2.6.24-rc7 updated a day or so before the 2.6.24
release) is the CPA stuff.
My previous testing kernel had been 25f4bdb4 - and besides problems with XChat
having a strange freeze and libata appearing to randomly decide to fail, it
worked well. With 36555adc there are strange latencies during boot (which
don't show in the dmesg, oddly - it's about 10 seconds at the console
waiting, but only a two second pause is recorded) and I have no functional
keyboard. The first pause is when the ATA subsystem initializes and the
second when the ACPI controls (lid switch, video, etc...) are loaded.
As said, the only differences in the configs for the kernels are the new
drivers and the CPA debugging code. Strangely the CPA debugging code reports
a "miss" just prior to a spate of APIC errors and then, after the APIC
errors, reports that the CPA system has passed all self tests.
Full dmesg from the boot and the config for the kernel in question is
attached. .git/refs/heads/master reports the git version as being:
24e1c13c93cbdd05e4b7ea921c0050b036555adc
DRH
--
Dialup is like pissing through a pipette. Slow and excruciatingly painful.