[Skiboot] [PATCH V2 0/5] BT Improvements
stewart at linux.vnet.ibm.com
Tue Nov 17 17:56:03 AEDT 2015
Cyril Bur <cyril.bur at au1.ibm.com> writes:
> Tested this on current AMI firmware and boots the host past petitboot
> without issue.
> The first two patches are cleanups and improve debugging output which was
> difficult to work with. No functional change.
> The third patch simplifies by removing the state machine.
> The fourth patch puts all 0xff in the fifo when a response to a message
> skiboot doesn't know about arrives, this will hopefully save a double send
> of a valid request and help to catch this situation in debugging as all
> 0xff will be easy to spot.
> The fifth patch asks the BMC for its BT capabilities. It should be noted
> here that AMI firmware responds with an error to this message which means
> we fall back to default values, the same values we've always been using.
> Cleanups from Alistairs comments.
> Added clear the write buffer when clearing the fifo
> Cyril Bur (5):
> hw/bt: Improve debug printing
> hw/bt: Convert retry_count to send_count
> hw/bt: Remove state machine and use only B2H_ATN and H2B_ATN.
> hw/bt: Clear the fifo when an unknown response is received
> hw/bt: Ask the BMC for its BT interface capabilities
> hw/bt.c | 209 ++++++++++++++++++++++++++++++++++++++++++++-------------
> include/ipmi.h | 1 +
> 2 files changed, 164 insertions(+), 46 deletions(-)
Thanks, merged to master as of fb457c957258
If this horribly breaks all of BT when my palmetto finishes flashing a
non-hacky hostboot... I reserve the right to be all whiny (not
winey... although I reserve that right too).
More information about the Skiboot