Re: building kernel 2.0.x under potato
On Thu, 16 Sep 1999, Chris Rutter wrote:
chris>On Thu, 16 Sep 1999, John Lapeyre wrote:
chris>
chris>> The link to suse doesn't work at the moment, but I'll give it a try.
chris>> The blurb at cygnus does not look encouraging. I think it is claiming
chris>> that I have to "to change asm constructs" at various unspecified places
chris>> in the source.
chris>
chris>Nah, they're just trying to cover their backs, so that people can't
chris>whinge at them *if* things go wrong -- they're only *trying* to
chris>worry you. Everyone I know who uses the patches says they're
chris>fine.
Hmm. Well my two potato systems are slightly different. One just
compiled 2.0.36 with the patch. But the other one failed with the
message
fixed or forbidden register 2 (cx) was spilled for class CREG,
which seems to be the problem that cygnus was talking about. I
have not tracked down the difference. This is hard to believe.
Maybe I have still missed something. I tried reapply the patch
to a fresh source a couple of times.
chris>
chris>
chris>> What should work? gcc272 ? I have tried it on two current potato
chris>> machines. building with CC=gcc272 fails to build both 2.0.x and 2.2.x
chris>> kernels. Building with the default compiler (egcs 2.95) will only build
chris>> 2.2.x kernels. The kernel mailing list still claims that I should
chris>> build with 2.7.2 before sending a bug report about my corrupted fs.
chris>
chris>Yeah, 2.7.2.* is the canonical compiler for 2.0 kernels. Can you
chris>post what's actually going wrong?
I could. The system hangs when I tar and gzip a large directory. I
get no OOPS or any message in any log. It is hard to see what is
happening. But it happens everytime I make the tar file with 2.2.10 and
2.2.12, and never (I tried 10 times in a row ) with 2.0.36. There is
some noise on the kernel list of similar problems from people using an
AMD K6-2. I still have not tried swapping CPUs, and memory, etc.
chris>
chris>> I have an old 2.0.36 kernel, but I need to compile a module for a driver.
chris>> I think that given the number of instability reports regarding 2.2.x
chris>> kernels it might be nice to be able to compile 2.0.x somewhat easily.
chris>
chris>What module's that -- does it not work under 2.2? Yeah, it *should*
chris>be straightforward...
Yes the module works under 2.2, but 2.2 gives crashes for me. I was
running 2.0.34 happily for some time, but the ethernet card died, so I
got different model, and discovered that I can no longer compile 2.0.34
modules.
chris>
chris>> Am I being obtuse, or are things pretty fucked up regarding kernels and
chris>> compilers ?
chris>
chris>Er, a little, yeah. Unfortunately the Linux kernel is quite a
chris>stressful bit of code to compile (it needs to get good x86
chris>performance), and so things got really tested to the full, w.r.t.
chris>the compiler, but the compiler had bugs, and they had to be
chris>worked around, etc. It's not that pretty.
Well the patch worked on one machine, so I'll see if the new setup
is stable. I am just a bit frustrated because this is the first time
in four years that I can't just grab the source and type 'make zImage';
it has become much more complicated.
Thanks again.
John Lapeyre <lapeyre@physics.arizona.edu>
Tucson,AZ http://www.physics.arizona.edu/~lapeyre
Reply to: