[Bug 225182] devel/powerpc64-xtoolchain-gcc: SLOF Data Storage Exception on boot with kernel built with devel/powerpc64-xtoolchain-gcc
bugzilla-noreply at freebsd.org
bugzilla-noreply at freebsd.org
Thu Jan 25 01:39:39 UTC 2018
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=225182
Mark Millard <marklmi26-fbsd at yahoo.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |marklmi26-fbsd at yahoo.com
--- Comment #6 from Mark Millard <marklmi26-fbsd at yahoo.com> ---
(In reply to Kenneth Salerno from comment #5)
Just as a point of reference for a
PowerMac G5 context:
# uname -apKU
FreeBSD FBSDG5L 12.0-CURRENT FreeBSD 12.0-CURRENT r327364M powerpc powerpc64
1200054 1200054
This was built via:
A) kernel: devel/powerpc64-xtoolchain-gcc based
B) world: system clang 5.0.1 based
[I experiment with clang targeting powerpc64
and powerpc.)
Both were using devel/powerpc64-binutils .
Both were cross builds from an amd64 context.
(As seen from the version number, its been a while since
I've updated. But it is not far back on the scale you report
exploring.)
I can list src.conf content used and the like if appropriate.
Note: I could also boot and run based on a clang 5.0.1 kernel
as long as I made sure no kernel module was dynamically
loaded. 5.0.1 is generating a new type of relocation
information that the kernel is not handling. With 5.0.0
I used to use a clang-based kernel.
It has been a while since I tried it, but I used to be able to
also build and install world based on
devel/powerpc64-xtoolchain-gcc and it worked.
Other notes:
If clang builds the system-library support for C++ exception
handling, then any thrown exception causes a process crash.
clang does not do everything it should for the built-ins
involved.
--
You are receiving this mail because:
You are on the CC list for the bug.
More information about the freebsd-ppc
mailing list