Subject: Re: trouble compiling arm32 kernels
To: Izumi Tsutsui <>
From: David Brownlee <>
List: tech-toolchain
Date: 04/13/2000 14:06:48
	We could import a new egcs into a different place in the tree and
	switch ports across to it as they check out OK.

	At the very least we should have a version available in pkgsrc so
	people can be testing and feeding results back to the egcs (gcc :)


On Thu, 13 Apr 2000, Izumi Tsutsui wrote:

> In <037001bfa4ac$84c1b240$>
> wrote:
> > Okay, I was able to revert to version 1.26 of vfscanf.c in
> > src/lib/libc/stdio, which doesn't use strtoull(), and this seems to be an
> > acceptable workaround for at least kernel builds.
> > 
> > Is anyone actively working on resolving it the right way? I think I have to
> > agree with Earnshaw that the problem appears deeper than originally thought.
> > I'm willing to help test fixes and grok some of the code, but I suspect
> > others are better than me at this, so far as dealing with gcc is concerned.
> Of cource, the "right way" is importing newer egcs sources
> to our source tree but it is too hard to do right now.
> I think a simple "workaround" is just to disable schedule-insns
> optimization, which would cause this bug.
> (i.e. add "-fno-schedule-insns" to COPTS in mk.conf or elsewhere)
> ---
> Izumi Tsutsui