diff options
author | Benjamin Herrenschmidt <benh@kernel.crashing.org> | 2014-01-13 09:49:17 +1100 |
---|---|---|
committer | Benjamin Herrenschmidt <benh@kernel.crashing.org> | 2014-01-13 09:49:17 +1100 |
commit | 10348f5976830e5d8f74e8abb04a9a057a5e8478 (patch) | |
tree | e1a0c5e5a9abe98f3799126c1b45333030024b33 /block | |
parent | f991db1cf1bdca43675b5d2df0af991719727029 (diff) | |
download | linux-10348f5976830e5d8f74e8abb04a9a057a5e8478.tar.gz linux-10348f5976830e5d8f74e8abb04a9a057a5e8478.tar.bz2 linux-10348f5976830e5d8f74e8abb04a9a057a5e8478.zip |
powerpc: Check return value of instance-to-package OF call
On PA-Semi firmware, the instance-to-package callback doesn't seem
to be implemented. We didn't check for error, however, thus
subsequently passed the -1 value returned into stdout_node to
thins like prom_getprop etc...
Thus caused the firmware to load values around 0 (physical) internally
as node structures. It somewhat "worked" as long as we had a NULL in the
right place (address 8) at the beginning of the kernel, we didn't "see"
the bug. But commit 5c0484e25ec03243d4c2f2d4416d4a13efc77f6a
"powerpc: Endian safe trampoline" changed the kernel entry point causing
that old bug to now cause a crash early during boot.
This fixes booting on PA-Semi board by properly checking the return
value from instance-to-package.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Tested-by: Olof Johansson <olof@lixom.net>
---
Diffstat (limited to 'block')
0 files changed, 0 insertions, 0 deletions