On 21/06/10 15:41, Paulraj, Sandeep wrote:
The following restructuring and optimisations increase the SPI read performance from 1.3MiB/s (on da850) to 2.87MiB/s (on da830):
Remove continual revaluation of driver state from the core of the copy loop. State can not change during the copy loop, so it is possible to move these evaluations to before the copy loop.
Cost is more code space as loop variants are required for each set of possible configurations. The loops are simpler however, so the extra is only 128bytes on da830 with CONFIG_SPI_HALF_DUPLEX defined.
Unrolling the first copy loop iteration allows the TX buffer to be pre-loaded reducing SPI clock starvation.
Unrolling the last copy loop iteration removes testing for the final loop iteration every time round the loop.
Using the RX buffer empty flag as a transfer throttle allows the assumption that it is always safe to write to the TX buffer, so polling of TX buffer full flag can be removed.
Signed-off-by: Nick Thompson nick.thompson@ge.com
da850 and da830 are similar devices. The SPI module is common to both, but da850 uses DDR and da830 uses SDRAM. The EVM's might not actually be comparable, but they appear to be at least similar.
The speed was tested with a 8MiB transfer from SPI FLASH using:
sf read 0xc0008000 0 0x800000
drivers/spi/davinci_spi.c | 195 +++++++++++++++++++++++++++++-----------
This patch does not apply against Wolfgang's next. The patch should be against u-boot/next.
It looks like Delio's patch ended up being pulled into main via Tom's tree and yours. On u-boot I see 9268236529161312c877e638a14c011fd3c883e1, but the same commit has id 23911740486c59851df57521c49bfd81ce1865ec on on u-boot-ti.
I guess this is the source of the confusion? My patch was against u-boot, not u-boot-ti.
Nick.