Re: SpeedStep / Geyserville lockups?
>> I believe that Speedstep machines are much more stable if they are always
>> (re)booted while attached to wall power. Suspends and resumes should behave
>> much better, but *those* depend on whether your APM BIOS is crappy :(
>>
>> The reasoning is, boot time is when the kernel measures its bogomips, to
>> figure out what sorts of timings to use for hardware-driver matters.
>
> That's really interesting. I haven't had any lock-ups on this system at
> all (I finally got an IBM A21p). I have seen bogomips ratings of between
> 187 and 1690.
>
> Perhaps the BIOS manages to sort this out for me?
The BIOS wouldn't know dip about Bogomips - only the linux kernel would be
able to "sort those out" for you. Now, if the kernel has gotten smart enough
to recalc when you resume, or at any power event, I finally see a possible
compelling reason to make a jump to 2.4.x.
However, tricks it uses may not work on my beastie, which is strange. My
hubby's laptop, however, is definitely seeing better behavior in 2.4.4 - for
the first time, if he happens to get one of those dratted "hda lost interrupt"
when he resumes, he could actually shut down properly rather than flat-out
hang. And it's very hard to tell but he thinks he's getting them less.
> Suspend / resume have also worked flawlessly,
So you probably have a well-matched APM BIOS and motherboard combination...
With a *really* good APM BIOS, you can suspend and resume while using even a
non APM kernel at all and your processes won't lose the musical chairs
scheduling game. Not that I recommend staying that way, but it's so.
> apart from the (apparently)
> standard sound driver bugs, requiring ALSA to be restarted on resume.
timing bugs. Sigh. But none of this is really debian specific :(
> Regards,
> Andrew.
* Heather Stern * star@ many places...
Reply to: