On Wed, Oct 30, 2002 at 11:57:58AM -0800, Martin J. Bligh wrote:
> >> > Can you also mention not using gcc 3.0.x (stack pointer handling bug)
> >>
> >> Any chance of putting this sort of thing as #error detection
> >> in the compile so it auto-breaks? I seem to recall that's done
> >> for some versions of GCC already ...
> >
> > And what arch is that for? Adding a nice facility for per-arch (and
> > maybe global) compiler / binutils testing would be nice, if we're going
> > to go down that road..
>
> Alan, would you consider something like (TOTALLY untested):
[snip]
So it's a global thing then?
> If you like it, I'll get it tested.
>
> Probably some things are per-arch and some are global, at a guess.
> Currently we have:
[snip]
And then ppc32 does some binutils checks, not-in-C and before we let the
kernel get too far. What I was proposing is we add something outside of
the direct kernel src (something in scripts? Generic make magic?) to
verify that a compiler / binutils / whatever tool is a good version.
This is rather easy for binutils + obvious instruction problem (missing
/ incorrect operands), and for compiler versions, something to generate
a test-file and then see if it compiles.
-- Tom Rini (TR1265) http://gate.crashing.org/~trini/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Oct 31 2002 - 22:00:49 EST