Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 06 Dec 2024 16:38:38 +0000
From:      bugzilla-noreply@freebsd.org
To:        bugs@FreeBSD.org
Subject:   [Bug 283159] crash dump backtraces broken on arm64
Message-ID:  <bug-283159-227-8lTXsZiaE0@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-283159-227@https.bugs.freebsd.org/bugzilla/>
References:  <bug-283159-227@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D283159

--- Comment #2 from John Baldwin <jhb@FreeBSD.org> ---
Nothing has changed in unwinding for kgdb in a long while, and the only fra=
me
unwinding that is custom in kgdb is for exception frames.  Unwinding out of
doadump() is a "normal" frame just like a userspace frame and should be usi=
ng
DWARF unwind info.  The most likely regression here is an issue with the DW=
ARF.

Oh, is this with PAC?  The PC value looks like the upper N bits are "wrong":

#0  0xffff0000004b5aa8 in doadump (textdump=3D1) at
/usr/home/trasz/git/freebsd/sys/kern/kern_shutdown.c:404
#1  0x67fd0000004b5868 in ?? ()

The real PC for frame 1 probably starts with 'ffff' instead of '67fd'.  As a
workaround you can disable PAC for now (not sure if that's a thing you can =
do
currently).  I'll try to see if I can reproduce.

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-283159-227-8lTXsZiaE0>