| commit | f3ec33e587df02e25963755989cc473e3f07bf0d | |
| author | Al Viro <viro@ftp.linux.org.uk> | |
| Sun, 16 Dec 2007 23:30:08 +0000 (16 23:30 +0000) | ||
| committer | David S. Miller <davem@davemloft.net> | |
| Mon, 28 Jan 2008 23:07:09 +0000 (28 15:07 -0800) | ||
| tree | 5f176f664af9b9b8ad76ca5fdfd8fd68de6a1fe5 | treesnapshot (tar.gz zip) |
| parent | b710b43c306650261c01ad08100791afec78a7db | commitdiff |
sunhme endianness annotations
This one is interesting - SBUS and PCI variants have
opposite endianness in descriptors (SBUS is sparc-only, so there
host-endian == big-endian).
Solution: declare a bitwise type (hme32) and in accessor
helpers do typechecking and force-casts (once we know that the
type is right).
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
This one is interesting - SBUS and PCI variants have
opposite endianness in descriptors (SBUS is sparc-only, so there
host-endian == big-endian).
Solution: declare a bitwise type (hme32) and in accessor
helpers do typechecking and force-casts (once we know that the
type is right).
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
| drivers/net/sunhme.c | diffblobblamehistory | |
| drivers/net/sunhme.h | diffblobblamehistory |