Remove the extra conditions that cause some cases to fail prematurely like if the data number of bytes is odd. The controller can handle odd number of bytes data read in DTR Mode. Don't fail supports op for this condition.
Signed-off-by: Dhruva Gole d-gole@ti.com ---
For a deeper context, refer to a cover letter from an earlier patch series: https://lore.kernel.org/u-boot/20221019064759.493607-1-d-gole@ti.com/T/#m3fd... Here, I am trying to boot from OSPI Flash present on board the AM62-SK EVM. However, despite enabling the necessary configs and DT nodes, my boot flow seemed to be failing. I then came to know that the condition causing this failure was that my generated DTB was of odd number of bytes. So, while loading the dtb from the Octal SPI NOR Flash to the memory, the data.nbytes was odd which made the supports_op function return false. This check feels a little too strict at the spi-mem stage and we should let the controller decide what to do in case of odd bytes in DTR. After applying this patch, I was able to succesfully boot the AM62SK EVM without any issues.
I would also like to justify this patch by pointing the community to the equivalent code in the linux kernel, in drivers/spi/spi-mem.c, where this check is absent as well.
The controller does work with odd number of bytes and I have not seen any sort of bugs in the absence of this supports_op check. Hence, feel that it is safe enough to discard this check from here.
drivers/spi/spi-mem.c | 4 ---- 1 file changed, 4 deletions(-)
diff --git a/drivers/spi/spi-mem.c b/drivers/spi/spi-mem.c index 8e8995fc537f..eecc13bec90d 100644 --- a/drivers/spi/spi-mem.c +++ b/drivers/spi/spi-mem.c @@ -181,10 +181,6 @@ bool spi_mem_dtr_supports_op(struct spi_slave *slave, if (op->dummy.nbytes && op->dummy.buswidth == 8 && op->dummy.nbytes % 2) return false;
- if (op->data.dir != SPI_MEM_NO_DATA && - op->dummy.buswidth == 8 && op->data.nbytes % 2) - return false; - return spi_mem_check_buswidth(slave, op); } EXPORT_SYMBOL_GPL(spi_mem_dtr_supports_op);