| Message ID | 1549207053-24190-6-git-send-email-philipp.tomsich@theobroma-systems.com |
|---|---|
| State | Accepted |
| Commit | 7e0e5c552f2afcb9b6eeb2836232dfc4bbe29865 |
| Delegated to: | Philipp Tomsich |
| Headers | show |
| Series | rockchip: Improve SPI-NOR read performance for the RK3399-Q7 | expand |
> The logic in the main transmit loop took a bit of reading the TRM to > fully understand (due to silent assumptions based in internal logic): > the "wait until idle" at the end of each iteration through the loop is > required for the transmit-path as each clearing of the ENA register > (to update run-length in the CTRLR1 register) will implicitly flush > the FIFOs... transmisson can therefore not overlap loop iterations. > > This change adds a comment to clarify the reason/need for waiting > until the controller becomes idle and wraps the entire check into an > 'if (out)' to make it clear that this is required for transfers with a > transmit-component only (for transfers having a receive-component, > completion of the transmit-side is trivially ensured by having > received the correct number of bytes). > > The change does not increase execution time measurably in any of my > tests. > > Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> > --- > > drivers/spi/rk_spi.c | 15 ++++++++++++--- > 1 file changed, 12 insertions(+), 3 deletions(-) > Applied to u-boot-rockchip, thanks!
diff --git a/drivers/spi/rk_spi.c b/drivers/spi/rk_spi.c index aaf244d..c807d78 100644 --- a/drivers/spi/rk_spi.c +++ b/drivers/spi/rk_spi.c @@ -379,9 +379,18 @@ static int rockchip_spi_xfer(struct udevice *dev, unsigned int bitlen, toread--; } } - ret = rkspi_wait_till_not_busy(regs); - if (ret) - break; + + /* + * In case that there's a transmit-component, we need to wait + * until the control goes idle before we can disable the SPI + * control logic (as this will implictly flush the FIFOs). + */ + if (out) { + ret = rkspi_wait_till_not_busy(regs); + if (ret) + break; + } + len -= todo; }
The logic in the main transmit loop took a bit of reading the TRM to fully understand (due to silent assumptions based in internal logic): the "wait until idle" at the end of each iteration through the loop is required for the transmit-path as each clearing of the ENA register (to update run-length in the CTRLR1 register) will implicitly flush the FIFOs... transmisson can therefore not overlap loop iterations. This change adds a comment to clarify the reason/need for waiting until the controller becomes idle and wraps the entire check into an 'if (out)' to make it clear that this is required for transfers with a transmit-component only (for transfers having a receive-component, completion of the transmit-side is trivially ensured by having received the correct number of bytes). The change does not increase execution time measurably in any of my tests. Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> --- drivers/spi/rk_spi.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-)