This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug target/83302] i386 stack_probe has side effects
- From: "matz at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Thu, 07 Dec 2017 13:30:35 +0000
- Subject: [Bug target/83302] i386 stack_probe has side effects
- Auto-submitted: auto-generated
- References: <bug-83302-4@http.gcc.gnu.org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83302
--- Comment #2 from Michael Matz <matz at gcc dot gnu.org> ---
I guess the problem described in https://lkml.org/lkml/2017/11/10/188 is, that
the stack probe itself accesses a page which _doesn't_ belong to this threads
stack, but to something else. golang seems to use tiny stacks without guard
page, and the very idea of -fstack-check is that those probe instructions
either
access the own stack, or the guard page, but never anything else.
So, it's mismatching assumptions between golang and gcc's idea of stack layout,
they are simply incompatible. I think not using guard pages in golang is not a
very good idea.