On 17/06/10 04:52, [email protected] wrote: > > Hi All, > > Saw this post a few days ago and started poking around > and google'ing. Unfortunately i didn't write down what > i was doing and this post is from a cold addled memory. > > You'll definitely need to verify anything i say in this > post. > > IIRC it would seem that gcc-4.5.0 has changed its default > behavior for stack alignment. Instead of 16 byte boundaries > its using 32 byte boundaries. I'm not sure if this only > applies to x86_64 machines, all machines, or what. Not even > sure if the numbers "16" and "32" used above are correct or > that it was the alignment of the "stack". > > I think someone suggested using something like "-mstack-align" > on the gcc-4.5.0 command to cause the normally expected behavior. >
Thanks, I hadn't connected it with gcc-4.5.0. Googling on that led to a very interesting thread here: http://mail.madler.net/pipermail/zlib-devel_madler.net/2010-May/002277.html It seems zlib-1.2.5 needs the cflag -mstackrealign (not tested it yet) Andy -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
