This is the mail archive of the cygwin-developers@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: -mpreferred-stack-boundary=2?


On Fri, Nov 02, 2001 at 11:28:16PM +0300, egor duda wrote:
>Hi!
>
>Friday, 02 November, 2001 Christopher Faylor cgf@redhat.com wrote:
>
>CF> I just added the -mpreferred-stack-boundary=2 option to Makefile.common.
>CF> Is there a good reason not to do this?  From the description in
>CF> gcc.info, it seems like the cases that they are talking about don't
>CF> affect the cygwin part of things, although it would conceivably be iffy
>CF> to use this switch in the newlib.
>
>CF> The code size reduction is noticeable.
>
>i have no problems with this. I'm getting a lot of "can't inline"
>warnings with gcc-2.95.2 due to to-be-inlined function is defined
>after it's called, however. gcc-3.0.2 doesn't have this problem but i
>wonder if it's feasible to reorganize function order to accommodate
>older compilers? patch attached.

Thank you, Egor!  I didn't realize that it was so easy to fix these.

>btw, what do you think about -Werror ?

I'm in favor of it, in general, but I think there's one bogus warning in
fork.cc currently -- at least under gcc 2.95.3.  It is the "setjmp/longjmp
may clobber local variable" warning (this is not the exact text but it is
something like that).

cgf


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]