Wolfgang Grandegger wrote: > Kurt Van Dijck wrote: >> On Tue, Oct 06, 2009 at 02:04:16PM +0200, Wolfgang Grandegger wrote: >>> This patch makes the private functions alloc_can_skb() and >>> alloc_can_err_skb() of the at91_can driver public and adapts all >>> drivers to use these. While making the patch I realized, that >>> the skb's are *not* setup consistently. The skb's are now setup >>> as shown: >> [...] >>> Please check and comment. >>> >>> Marc, feel free to add your signed-off-by here. >>> Signed-off-by: Wolfgang Grandegger <[email protected]> >>> --- >>> kernel/2.6/drivers/net/can/at91_can.c | 32 ------------------ >>> kernel/2.6/drivers/net/can/cc770/cc770.c | 13 +------ >>> kernel/2.6/drivers/net/can/dev.c | 38 >>> ++++++++++++++++++++-- >>> kernel/2.6/drivers/net/can/esd_pci331.c | 14 +------- >>> kernel/2.6/drivers/net/can/mcp251x.c | 20 +---------- >>> kernel/2.6/drivers/net/can/mscan/mscan.c | 6 --- >>> kernel/2.6/drivers/net/can/sja1000/sja1000.c | 12 +----- >>> kernel/2.6/drivers/net/can/softing/softing_main.c | 8 +--- >>> kernel/2.6/drivers/net/can/usb/ems_usb.c | 16 +-------- >>> kernel/2.6/include/socketcan/can/dev.h | 4 ++ >>> 10 files changed, 54 insertions(+), 109 deletions(-) >>> >> Wolfgang, >> >> I'm not yet totally up with include/socketcan & include/linux >> seperation (which I don't like to question here). >> I noticed you do patcgh include/socketcan/can/dev.h >> but not include/linux/can/dev.h > > Oliver, what was the reason to maintain a redundant include/linux/can/dev.h?
The reason was to allow userspace applications to include that path. It would be very bad, if you would need to adapt the userspace apps to include the socketcan stuff, that is intended to separate the driver includes for older kernels. Maybe we should create symlinks for that ... Regards, Oliver _______________________________________________ Socketcan-core mailing list [email protected] https://lists.berlios.de/mailman/listinfo/socketcan-core
