Issue Tracker

10/13/2020

MP 09:34 AM coreboot Bug #282 (New): lz4 romstage loading is not endian safe.

https://ticket.coreboot.org/issues/281 similar to this.
Without this commit:
https://github.com/hanetzer/coreboot/commit/272fc7e3f5d32c0323602682636b5494c1ab6ba0
I cannot boot a ppc64 be romstage if I enable lz4
romstage compress... Marty Plummer

10/09/2020

MP 12:11 AM coreboot Bug #281: cbfs stage loading||encoding is not endian safe.

Aaron Durbin wrote:
> I agree that this should be corrected, but there's not much we can do while still maintaining backwards compatibility. The endian calls, regardless of which one employed, would be required. The thing that always bu... Marty Plummer

10/08/2020

MP 03:34 AM coreboot Bug #281: cbfs stage loading||encoding is not endian safe.

This may be enough for now, but perhaps we should see about reworking
cbfs* to properly encode these things in big/little endian depending
on the target, and have the code just use the 'native' format instead
of doing these explicit r... Marty Plummer

MP 02:29 AM coreboot Bug #281: cbfs stage loading||encoding is not endian safe.

Patch seems to work. I get as far with a big endian coreboot as I do with
a little endian coreboot now (romstage->ramstage handoff fails for some reason). Marty Plummer

10/07/2020

MP 05:05 PM coreboot Bug #281: cbfs stage loading||encoding is not endian safe.

Aaron Durbin wrote:
> Original sighting:
> ...
I had to recompile; I've been building little-endian for the moment because otherwise I'd be stuck here.
> struct cbfs_stage starts at 0x202b8 for fallback/romstage file.
> ...
~~~
EL... Marty Plummer

MP 04:33 PM coreboot Bug #281: cbfs stage loading||encoding is not endian safe.

Aaron Durbin wrote:
> fsize is derived from the region_device size for the data in cbfs_for_each_file():
> ...
~~~
00020000 5f 5f 46 4d 41 50 5f 5f 01 01 00 00 00 00 00 00 |__FMAP__........|
00020010 00 00 00 00 00 04 46 4c 41 5... Marty Plummer

MP 02:59 AM coreboot Bug #281: cbfs stage loading||encoding is not endian safe.

This is also an issue if you build a big-endian coreboot
on a big-endian system (in my case a talos ii running be
gentoo), so something is definitely up.
Marty Plummer

MP 02:27 AM coreboot Bug #281 (New): cbfs stage loading||encoding is not endian safe.

https://github.com/hanetzer/coreboot/tree/power9-endian-demo
^ code in question.
Build a cross-pp64 toolchain, switch mainboards to qemu-powernv9.
Use https://github.com/hanetzer/qemu/tree/powernv-5.2 for qemu
(its the same as ht... Marty Plummer

10/24/2019

MP 06:47 PM coreboot Bug #238: coreboot.rom containing big-endian code has little-endian encoded fmap headers.

Ok, scratch that. I was under the impression that the four bytes at
the end of coreboot.rom(more properly the coreboot cbfs's end) was
from fmap, but that's from cbfstool add-master-header. Its encoded
as LE. Marty Plummer

10/20/2019

MP 07:51 PM coreboot Bug #238: coreboot.rom containing big-endian code has little-endian encoded fmap headers.

Marty Plummer wrote:
> So, as part of my work on fully porting coreboot for ppc64/power9/talos ii I've
> ...
Just checked against the emulation/qemu-power8 target (clean tree) and it also has
the fmap headers in LE while the cbfs head... Marty Plummer