Age | Commit message (Collapse) | Author |
|
<mark@edwards.homelinux.net>
- add a workaround for the bogus no-ack returvalue of the kernel-i2c bitbanging driver
- add sp887x.o to the Makefile
|
|
- use c99 initializers
- remove blank lines
- remove bt848.h -- it's already in driver/media/video
(this probably breaks 2.4 compilation)
|
|
- make DMA block pointer volatile
- use one local tasklet per card
- remove spinlock bt->s_lock, it protected an atomic variable only
- remove obsolete waitqueue code, the waitqueue was not used anymore
|
|
|
|
|
|
- remove flush_workqueue(wq) from dvb_net.c -- don't we have to ensure that the handler has finished? all other drivers don't do this as well...?
|
|
are deprecated -- to be removed.
register network device for bt8xx cards
|
|
"kernel shape":
I mainly changed all "typedef struct foo_s {...} foo_t"
to "struct foo {}". Please don't start a policy war, I know the pros
cons of typedefs and structures.
We should, however, stick to the coding rules of the kernel: typedefed
structs are only used for opaque data types, where the internal
representation is irrelevant and we have get()/set() functions for
accessing them (ie. spinlock_t)
|
|
nobody really needs them...
|
|
|
|
|
|
|
|
- Push sync handling down to DVB API
- Major fixes
|
|
|
|
|
|
|
|
|
|
|