Excellent (I didn't know about http://linux/dialnet.net/LMP before now)!!
Anyways, Don't get me wrong... I really appreciate the work that goes into
the kernel development and I realize that all the kernels will always have
bugs of some nature. I can live with the bugs present in 2.0.29+ISS but I
can't live with a locked machine hence my frustration. As long as I know
2.0.29 doesn't crash I can wait forever for 2.0.33 or other fixes to come
out.
I guess the only real thing I was posting about was the "apparent" lack of
2.0.x support. I know lots of people still work on it but it is hard to
determine whether your problem is real or what the progress on the issue
is like. The above website is excellent though (And the boxing equipped
penguin is a great mascot).
Sorry if some of this sounded like whining (certainly don't want to
provoke another David Miller like disaster circa 2.0.31 ;-) but I'm
glad to see that my post generated some discussion on the topic where
my first post went unanswered.
Off to get the neat patches from dialnet and see if they don't fix my
system... (Or maybe I'll just junk that 3c905 card I have in favor of
the tulip cards).
- Jeff
Doug Ledford wrote:
>
> On 02-Dec-97 Jeff Wiegley wrote:
> >My advice is: 2.0.3* are all broken in so many ways that the impression
> >I gather from this list is nobody has a real handle on exactly what is
> >broken and how to fix it. 2.0.32 is a good step except I'm seeing several
> >people like myself with solid lockups (SMP or not) and this type of
> >failure is completely unacceptable for my purposes.
>
> Actually, some of us are working on the 2.0.x kernel series still....we just
> don't make a lot of noise about it :)
>
> >But I still think some serious work needs to go into the 2.0 kernel
> >since it is suppose to be the stable kernel but isn't.
>
> Work is going into it :)
>
> >The amount of attention given 2.0 by the gurus is so small compared to the
> >effort going on in 2.1 that I would feel more comfortable running and
> >updating a 2.1 bleeding-edge kernel because I know problems will get
> >fixed quickly.
> >
> >My post on the 2.0.32 lockup and the others I've seen have pretty much
> >been ignored (or so it seems since there hasn't been further discussion
> >on whats wrong and how to fix it).
>
> Actually, not all of them have gone unanswered, and they haven't gone
> ignored. So far we've been able to develope the following two items in the
> patterns that I've seen anyway. Either the aic7xxx driver is involved, or
> the 3c59x driver is involved. There may be those out there that don't fit
> this pattern, but I don't recall your particular hardare listing. As to the
> aic7xxx driver, the patch found on the LMP web site has been reported to
> help with the lock-ups (and well it should, I fixed several infinite loop
> problems in the stock driver with that patch as well as added bounding to
> the majority of the loops in case anything goes haywire and creates a
> circumstance I haven't been able to create yet). As for the 3c59x driver,
> well, that's just a fine and dandy can of worms. I have yet to hear of a
> single version of that driver that works on *all* the cards it is suppossed
> to support.....
>
> If you want to see the problems we've identified and are working on, then
> check out the LMP web page (http://linux.dialnet.net/LMP).