Date: Sat, 15 Jan 2000 07:29:38 -0800 (PST) From: Mark Newton <newton@FreeBSD.org> To: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: cvs commit: src/sys/i386/svr4 svr4_machdep.c Message-ID: <200001151529.HAA32087@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
newton 2000/01/15 07:29:38 PST
Modified files:
sys/i386/svr4 svr4_machdep.c
Log:
Pre-freeze cleanups: Conditionalize debugging writes on DEBUG_SVR4,
eliminate warnings, etc.
Note that svr4_setcontext() and svr4_getcontext() currently don't work
correctly, which makes returning from signal handlers somewhat problematic
(for reference: the SysVR4 setcontext() and getcontext() syscalls behave
like a low-level version of setjmp() and longjmp(), in that they save and
restore process context. SysVR4 uses this to implement its signal handler
trampoline: The context which is saved before a signal handler is called
is restored by an implicit call to the setcontext() syscall when the signal
handler returns. That currently doesn't work right in this emulator,
although it used to).
I'm committing this anyway, with a caveat that it's buggy, so that I can
(a) note the bug for anyone who is wondering about it, and (b) get the
stuff I've done to improve (but not fix) the situation in the tree before
4.0 is released. If I have time to fix it totally RSN I'll talk to Jordan
to see if I can bend him into letting me commit a bugfix :-)
Note that the situation now is somewhat better than it was yesterday
anyway, because I've fixed the handling of svr4_sigsets which previously
was causing signal handlers to not get called at all. Stay tuned for
an upcoming commit to svr4_signal.c...
Revision Changes Path
1.13 +33 -12 src/sys/i386/svr4/svr4_machdep.c
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200001151529.HAA32087>