diff options
author | Stefan Reinauer <reinauer@chromium.org> | 2012-04-30 16:42:07 -0700 |
---|---|---|
committer | Stefan Reinauer <stefan.reinauer@coreboot.org> | 2012-05-01 20:08:00 +0200 |
commit | 816d081760d4cccf0db0c952728f5ec6926d6c85 (patch) | |
tree | 12542d99189f5b412ad39fb25a78427dcfb4aaa5 /src/mainboard/tyan/s2891/chip.h | |
parent | 8508cff0359dcd99d319407f15e5034ff896a716 (diff) |
Fix SATA port map to only enable port 0
The sata controller comes up in legacy/normal mode and
is currently put into AHCI mode in romstage.
If that is removed and the controller is left alone until the
ramstage driver (like we do on Stumpy/Lumpy) then the resource
allocator will have configured the device for IDE mode with an
IO address in BAR5. Then when the ramstage driver puts the
controller into AHCI mode it will not have the correct resources
to do the rest of the AHCI setup.
So the controller mode needs to be changed in the enable stage
rather than in the init phase. This same register contains
the port map and it is a R/WO (write once) field so the configured
port map must be written at the same time. For non-AHCI mode
the devicetree map was ignored before but it is used now.
Since the port map register is now written at enable step it
does not need to be written again during init.
With this change the sata port map can be reduced to just port 0
and then U-boot does not have to probe all available ports.
Change-Id: I977952cd88797ab4cea79202e832ecbb5c37e0bd
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/977
Tested-by: build bot (Jenkins)
Diffstat (limited to 'src/mainboard/tyan/s2891/chip.h')
0 files changed, 0 insertions, 0 deletions