diff options
author | imp <imp@FreeBSD.org> | 2008-12-11 06:27:18 +0000 |
---|---|---|
committer | imp <imp@FreeBSD.org> | 2008-12-11 06:27:18 +0000 |
commit | e6342fe8be30cb35eeaf40fa8ddb01e1afa2b8f6 (patch) | |
tree | 79a14abb4c843df22f3c08b85a2a403f45dc9089 /usr.bin/fetch/fetch.c | |
parent | 0d95748a8b7a6640e76e998a81187e8953d310e4 (diff) | |
download | FreeBSD-src-e6342fe8be30cb35eeaf40fa8ddb01e1afa2b8f6.zip FreeBSD-src-e6342fe8be30cb35eeaf40fa8ddb01e1afa2b8f6.tar.gz |
Update to the interrupt handling code:
o Try to be smarter about reading the ExCA CSC register. Now, we only
do it for 16-bit cards. Add some experimental code to treat it like
a power interrupt, but I'm not 100% sure that I like it. It may be
removed upon further testing. It seemed to help in one test case, but
the evidence may be inconclusive. This may be beneficial for cleaning up
exca_reset and exca_wait_ready.
o Check for CSTS events on the socket event register. We ask for it when
we're powering up a card, but I don't think we're otherwise using
it. Just ACK the interrupt for now. In theory, we can use it
instead of the busy wait we do in cbb_cardbus_reset. More research
is necessary to see if we can optimize things there when we're
waiting for the DEVVENDOR register to become valid.
o Rework the comments a bit. Minor tidying up. Etc.
Diffstat (limited to 'usr.bin/fetch/fetch.c')
0 files changed, 0 insertions, 0 deletions