Re: [PATCH 17/25] spi: atcspi200: add optional interrupt support

[sashiko-bot@kernel.org]

> For PIO transfers, the driver busy-polls the FIFO empty/full flags and
> the SPI active bit for completion. When the controller's interrupt is
> available, use it instead of polling to keep the FIFO serviced and to
> detect the end of transfer.
> 
> Interrupts are optional. Controllers without an interrupt keep using the
> polling path, as do command-only transfers; the driver also falls back to
> polling when the interrupt cannot be requested.
> 
> The command register is now written when a transfer is started instead of
> in atcspi_prepare_trans(), so the interrupt path can enable interrupts
> before starting the transfer and the first FIFO or end-of-transfer
> event is not missed. As a result, it is also set at the beginning of
> atcspi_dma_trans() and atcspi_xfer_data_poll(), and in the command-only
> path of atcspi_exec_mem_op().
> 
> Introduce atcspi_reset_engine() as a helper to reset the SPI controller
> and the FIFOs, and use on transfer error and in atcspi_setup().
> 
> The interrupt path configures the FIFO thresholds for the transfer and
> restores them afterwards. It acknowledges the interrupt status before
> servicing the FIFOs so events arriving during servicing are not lost,
> and resets the controller if a transfer times out or completes with
> transmit/receive words left unhandled.
> 
> Originally-by: Jun Chang <jun.chang@metanoia-comm.com>
> Signed-off-by: Benoît Monin <benoit.monin@bootlin.com>

Sashiko has reviewed this patch and found no issues.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261106-mt2824-v1-0-86087eaf5af4@bootlin.com?part=17