diff options
author | Stephen Warren <swarren@nvidia.com> | 2012-11-06 11:27:29 +0000 |
---|---|---|
committer | Andy Fleming <afleming@freescale.com> | 2012-11-27 17:26:48 -0600 |
commit | 84d35b2863455bedb9986c2b076241e8a441fc3e (patch) | |
tree | 9cb3af890b370091b0460d930b5829fe3d2abb62 /include | |
parent | 4ea7a09fafdf0592cb99428090946bf15128ea44 (diff) |
common: rework bouncebuf implementation
The current bouncebuf API requires all parameters to be passed to both
bounce_buffer_start() and bounce_buffer_stop(). Modify the bouncebuf
start function to accept a state structure as a parameter, and only
require that state struct to be passed to the stop function. This
simplifies usage of the bounce buffer by clients.
Don't modify the data pointer, but rather store the temporary buffer in
this state struct. The bouncebuf code ensures that client code can
always use a single buffer pointer in the state structure, irrespective
of whether a bounce buffer actually had to be allocated.
Move cache management logic into the bounce buffer code, so that each
client doesn't have to duplicate this. I believe there's no need to
invalidate the buffer before a DMA operation, since flushing the cache
should prevent any write-backs.
Update the MXS MMC driver for this change.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Diffstat (limited to 'include')
-rw-r--r-- | include/bouncebuf.h | 33 |
1 files changed, 24 insertions, 9 deletions
diff --git a/include/bouncebuf.h b/include/bouncebuf.h index aa2278c8f9..707b588c73 100644 --- a/include/bouncebuf.h +++ b/include/bouncebuf.h @@ -25,6 +25,8 @@ #ifndef __INCLUDE_BOUNCEBUF_H__ #define __INCLUDE_BOUNCEBUF_H__ +#include <linux/types.h> + /* * GEN_BB_READ -- Data are read from the buffer eg. by DMA hardware. * The source buffer is copied into the bounce buffer (if unaligned, otherwise @@ -51,23 +53,36 @@ */ #define GEN_BB_RW (GEN_BB_READ | GEN_BB_WRITE) +struct bounce_buffer { + /* Copy of data parameter passed to start() */ + void *user_buffer; + /* + * DMA-aligned buffer. This field is always set to the value that + * should be used for DMA; either equal to .user_buffer, or to a + * freshly allocated aligned buffer. + */ + void *bounce_buffer; + /* Copy of len parameter passed to start() */ + size_t len; + /* DMA-aligned buffer length */ + size_t len_aligned; + /* Copy of flags parameter passed to start() */ + unsigned int flags; +}; + /** * bounce_buffer_start() -- Start the bounce buffer session + * state: stores state passed between bounce_buffer_{start,stop} * data: pointer to buffer to be aligned * len: length of the buffer - * backup: pointer to backup buffer (the original value is stored here if - * needed * flags: flags describing the transaction, see above. */ -int bounce_buffer_start(void **data, size_t len, void **backup, uint8_t flags); +int bounce_buffer_start(struct bounce_buffer *state, void *data, + size_t len, unsigned int flags); /** * bounce_buffer_stop() -- Finish the bounce buffer session - * data: pointer to buffer that was aligned - * len: length of the buffer - * backup: pointer to backup buffer (the original value is stored here if - * needed - * flags: flags describing the transaction, see above. + * state: stores state passed between bounce_buffer_{start,stop} */ -int bounce_buffer_stop(void **data, size_t len, void **backup, uint8_t flags); +int bounce_buffer_stop(struct bounce_buffer *state); #endif |