https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113785
--- Comment #1 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot
Uni-Bielefeld.DE> ---
I've found what's going on: as described in Solaris makecontext(3C), the
function changed starting with Solaris 10:
NOTES
The semantics of the uc_stack member of the ucontext_t structure have
changed as they apply to inputs to makecontext(). Prior to Solaris 10,
the ss_sp member of the uc_stack structure represented the high memory
address of the area reserved for the stack. The ss_sp member now repre-
sents the base (low memory address), in keeping with other uses of
ss_sp.
This change in the meaning of ss_sp is now the default behavior. The
-D__MAKECONTEXT_V2_SOURCE compilation flag used in Solaris 9 update
releases to access this behavior is obsolete.
To preserve, binary compatiblity, the old behaviour of makecontext() has
been retained, and the following snippet in <ucontext.h> ensures that a
different function with the new behavious is called from user code
instead:
/* makecontext removed in SUSv4 (XPG7, UNIX V7) */
#if !defined(_XPG7) || defined(__EXTENSIONS__)
#ifdef __sparc
#ifdef __PRAGMA_REDEFINE_EXTNAME
#pragma redefine_extname makecontext __makecontext_v2
#else
#define makecontext __makecontext_v2
#endif
#endif /* __sparc */
#endif /* !defined(_XPG7) || ... */
However, libasan knows nothing about this and just intercepts unadorned
makecontext, but this interceptor is never called.
I've got a patch (attached for reference) that changes the makecontext
interceptor to call __makecontext_v2 instead on Solaris/SPARC. With
this, the c-c++-common/asan/swapcontext-test-1.c test now PASSes on
Solaris/SPARC.
I'm going to submit it upstream soon.
