diff options
author | Julius Werner <jwerner@chromium.org> | 2014-12-09 12:18:00 -0800 |
---|---|---|
committer | Patrick Georgi <pgeorgi@google.com> | 2015-04-20 10:19:56 +0200 |
commit | cdf92eacd5d2f85de473778cc1ebf35c2468a7f2 (patch) | |
tree | 6123dc35c53e3122b7433730b18f4169d02535c2 /src/soc/intel/baytrail/gfx.c | |
parent | 0b29a7b37c19200516dadc5c4bf0a1eb9d1290ec (diff) |
rk3288: Disable ramstage compression by default
The ramstage is loaded from romstage, so the LZMA scratchpad buffer used
to decompress it is part of the romstage BSS in SRAM. On RK3288, SRAM
cannot be cached which makes the decompression so slow that it's faster
to just load an uncompressed image from SPI. Disable ramstage
compression on this SoC to account for that.
[pg: implementation avoids restructuring all of Kconfig]
BRANCH=None
BUG=None
TEST=Built for Pinky and Falco, confirmed that the former didn't have
COMPRESS_RAMSTAGE in its .config and the latter still did. Measured a
speed-up of about 35ms on Pinky. (For some weird reason, the
decompression of the payload also takes way longer than on other
platforms, although not as long as the ramstage. I have no explanation
for that and can't really think of a good way to figure it out... maybe
the Cortex-A12 is just terrible at some operation that LZMA uses a lot?)
Change-Id: I9f67f7537696ec09496483b16b59a8b73f4cb11b
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/234192
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: http://review.coreboot.org/9792
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Diffstat (limited to 'src/soc/intel/baytrail/gfx.c')
0 files changed, 0 insertions, 0 deletions