This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] New -fstack-check implementation
> If you split the patch into fixing static stack checking and the rest would
> it make the fixing static stack checking mostly middle-end independent?
Yes, static stack checking is basically entirely in back-ends whereas dynamic
stack checking is entirely in the middle-end (the latter has been rewritten
in the process so as to be immune to wrap-around). But some infrastructure
bits are needed in the middle-end for static stack checking though and the
bugs introduced since 3.x need to be fixed.
> I think fixing static statck checking should be uncontroversical and
> splitting it out for easier architecture maintainers review would be nice.
OK, I'll try and split it into as many pieces as possible.
> Is there a possibility to add C testcases (maybe architecture specific, for
> a popular architecture like x86) to incerase testing coverage of both
> features?
Yes. In fact I think that you can write C++ testcases throwing and catching a
stack overflow exception if compiled w/ -fstack-check -fnon-call-exceptions.
--
Eric Botcazou