[Bug 750805] Fails to build on ARM, needs to use default setjmp not ucontext

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.


https://bugzilla.redhat.com/show_bug.cgi?id=750805

--- Comment #9 from Henrik Nordström <henrik at henriknordstrom.net> 2011-11-05 04:52:26 EDT ---
It's a runtime issue detected by the test suite. I by accident did not have
FORTIFY_SOURCE enabled when testing the first change.

Coro (and pth) both abuses setjmp in the same manner for establishing new
stacks per context, but with FORTIFY_SOURCE enabled glibc verifies that setjmp
only sets the stack pointer to an earlier stack frame in the current stack and
aborts if the stack change looks odd.

*** longjmp causes uninitialized stack frame ***: /usr/bin/perl terminated

it's a valid error message in normal specified use of setjmp, but do not apply
to the special (outside specifications) use by Coro for establishing another
stack context entirely, outside the current stack.

-- 
Configure bugmail: https://bugzilla.redhat.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.