> > > Don't build 2.0.34 with gcc 2.8.x/egcs.
> > ghm... Can we track that one down? Yes, I know that 2.8.x/egcs sometimes
> > generate wrong code, but we might want to help gcc/egcs team to discover
> > bugs, so that new generation of compilers will be usefull :)
>
> Its not the compiler thats the problem. Its the kernel code not stopping
> the new gcc from some additional optimising we dont want.
Few days ago, there was a patch on the list to set compiler
optimisations depending on the compiler / compiler version.
Is it possible to disable most/all optimisations if gcc 2.8.x
is detected (at least for 2.0.35) and slowly increase them
(at least for 2.1).
Then, we could know what to look for, for instance problems of
aliases in too_optimised_driver.c , when a bug report is submitted.
If there were bug report on a non-optimised compilation, then
it should be forwarded to gcc-bug.
My 0.02 Euro,
Etienne.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu