Re: Egcs 1.0.3 & Linux

Khimenko Victor (
Tue, 19 May 1998 03:44:33 +0400 (MSD)

18-May-98 21:58 you wrote:
> The extremely lucid point that Khimenko has made here is that the
> response to "I compiled 2.1.102 with EGCS and now it doesn't boot
> at all" is not "use GCC" or the ever popular "EGCS is buggy", but
> should be "thanks for your bug report as it indicates there might
> be some ambiguous code in the kernel".
Unfortunatelly this is not good bug report :-(( This only means that
*somewhere* in kernel there are some such problem... Or may be egcs really is
buggy (but in fact stable versions of egcs is less buggy then gcc 2.8.x or may
be even then gcc 2.7.2.x :-) -- who knows ? If you could track down problem
and found -- what code caused this bug this will be good bug-report. gcc 2.7.2.x
are also "buggy" but since it's here for few YEARS all needed warkarounds in
kernel code are already there. egcs/pgcc/gcc 2.8.x is here only few MONTHS and
not all problems exposed by it's very agressive optimizing fixed in development
kernels. Saga of 2.0.x is near the end -- that's why 2.0.x probably will never
be reviewed for compatibility with egcs/pgcc/gcc 2.8.x ... 2.1.10x for now works
for most of us while compiled with egcs/pgcc/gcc 2.8.x but there are still few
places with such errors not discovered for sure. In you kernel hangs it's only
one more proof that such places are exists :-) But unfortunatelly this
information is already well-known :-(( The real problem is to find -- where
such places are...

> It is foolhardy to ignore such bug reports: they could be perhaps
> the most valuable bug reports possible. C is ambiguous, and it is
> full of hidden traps like Khimenko's example shows. Using varying
> C compilers is an excellent way of uncovering problems in Linux's
> source tree that would otherwise remain hidden.
Unfortunatelly you are wrong. This is cover only one side of problem: proof
that problem is really exists. Unfortunatelly this is easy part of problem :-(
In most cases other side -- finding place where in code problem placed -- is
far more troublesome...

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to