Re: [Cooker] kernel and gcc3.2, a good match?

[Luca Olivetti]

Gwenole Beauchesne wrote:

On Wed, 20 Nov 2002, Luca Olivetti wrote:

Just tried with 2.4.20-0.1mdk, also compiled with gcc3.2, and it freezes.
It doesn't even print an Oops, just hangs, and alt+sysrq doesn't work.

What? where? That could simply be a broken module. When compiling with gcc

= 3.0 and something is failing at runtime, try to rebuild the culprit

with -fno-strict-aliasing. If this works with that option, then review code that may break aliasing rules.

Hmm, I'm rebuilding the 2.4.19-16mkd source rpm now using gcc3.2, and I see that it's already compiling with -fno-strict-aliasing.
Isn't this the case with 2.4.20-0.1mdk? Should I fetch the 2.4.20-0.1mdk.src and try with or can someone confirm that it's using or not -fno-strict-aliasing?

Fortunately, gcc3.2 has a new option to warn about code that breaks aliasing rule. Needless to say, such code needs fixing.

Could you tell me the warning message, so I check if I see it in the compilation messages? (and, will it appear with -fno-strict-aliasing?).

Bye
--
Luca Olivetti
Note.- This message reached you today, it may not tomorrow if you
are using MAPS services. They arbitrarily include in their lists
IP addresses not related in any way to spam, and in so doing are
disrupting Internet connectivity. Please stop supporting them.
See http://slashdot.org/article.pl?sid=01/05/21/1944247