From: "Pandey, Radhey Shyam" <radheys@amd.com>
To: Srikanth Boyapally <srikanth.boyapally@amd.com>,
Mark Brown <broonie@kernel.org>
Cc: sxauwsk <sxauwsk@163.com>,
linux-spi@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] spi: spi-cadence: Move TX FIFO full busy-wait into FIFO
Date: Mon, 20 Jul 2026 21:25:12 +0530 [thread overview]
Message-ID: <81d818c5-c7f5-43b5-b8d2-202250ad213e@amd.com> (raw)
In-Reply-To: <20260720125510.60166-1-srikanth.boyapally@amd.com>
On 7/20/2026 6:25 PM, Srikanth Boyapally wrote:
> SPI host transfers could intermittently stall with spi_transfer timeouts.
> The TXFULL condition was checked only once in cdns_transfer_one() before
> cdns_spi_process_fifo(), so if the FIFO became full again during refill,
> writes could be dropped and the transfer would never complete.
>
> Move the TXFULL busy-wait into the TX path of cdns_spi_process_fifo() so
> the 10µs back-off is applied per FIFO entry during filling, ensuring
> forward progress and eliminating spurious timeouts.
>
> Restrict the delay to host mode using spi_controller_is_target(), the
> controller is passed into cdns_spi_process_fifo() so the check is made at
> the point of use. In target mode this delay must not run as it causes the
> target to miss its transfer window and corrupt data.
>
> Fixes: 49530e641178 ("spi: cadence: Add usleep_range() for cdns_spi_fill_tx_fifo()")
> Signed-off-by: Srikanth Boyapally <srikanth.boyapally@amd.com>
> ---
Reviewed-by: Radhey Shyam Pandey <radhey.shyam.pandey@amd.com>
Thanks!
> Note: checkpatch suggests usleep_range() over udelay(), but this code
> runs in interrupt context where sleeping is not allowed, so udelay() is
> intentional.
> ---
> drivers/spi/spi-cadence.c | 26 ++++++++++++++++----------
> 1 file changed, 16 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/spi/spi-cadence.c b/drivers/spi/spi-cadence.c
> index 9b4e5b7013ae..af1a05e78492 100644
> --- a/drivers/spi/spi-cadence.c
> +++ b/drivers/spi/spi-cadence.c
> @@ -388,11 +388,13 @@ static inline void cdns_spi_writer(struct cdns_spi *xspi)
>
> /**
> * cdns_spi_process_fifo - Fills the TX FIFO, and drain the RX FIFO
> + * @ctlr: Pointer to the spi_controller structure
> * @xspi: Pointer to the cdns_spi structure
> * @ntx: Number of bytes to pack into the TX FIFO
> * @nrx: Number of bytes to drain from the RX FIFO
> */
> -static void cdns_spi_process_fifo(struct cdns_spi *xspi, int ntx, int nrx)
> +static void cdns_spi_process_fifo(struct spi_controller *ctlr,
> + struct cdns_spi *xspi, int ntx, int nrx)
> {
> ntx = clamp(ntx, 0, xspi->tx_bytes);
> nrx = clamp(nrx, 0, xspi->rx_bytes);
> @@ -407,6 +409,16 @@ static void cdns_spi_process_fifo(struct cdns_spi *xspi, int ntx, int nrx)
> }
>
> if (ntx) {
> + /* When xspi in busy condition, bytes may send failed,
> + * then spi control didn't work thoroughly, add one byte
> + * delay. Only in host mode; in target mode this delay
> + * causes data corruption as the target fails to prepare
> + * data in time.
> + */
> + if (!spi_controller_is_target(ctlr) &&
> + (cdns_spi_read(xspi, CDNS_SPI_ISR) & CDNS_SPI_IXR_TXFULL))
> + udelay(10);
> +
> cdns_spi_writer(xspi);
> ntx--;
> }
> @@ -460,14 +472,14 @@ static irqreturn_t cdns_spi_irq(int irq, void *dev_id)
> cdns_spi_write(xspi, CDNS_SPI_THLD, 1);
>
> if (xspi->tx_bytes) {
> - cdns_spi_process_fifo(xspi, trans_cnt, trans_cnt);
> + cdns_spi_process_fifo(ctlr, xspi, trans_cnt, trans_cnt);
> } else {
> /* Fixed delay due to controller limitation with
> * RX_NEMPTY incorrect status
> * Xilinx AR:65885 contains more details
> */
> udelay(10);
> - cdns_spi_process_fifo(xspi, 0, trans_cnt);
> + cdns_spi_process_fifo(ctlr, xspi, 0, trans_cnt);
> cdns_spi_write(xspi, CDNS_SPI_IDR,
> CDNS_SPI_IXR_DEFAULT);
> spi_finalize_current_transfer(ctlr);
> @@ -520,17 +532,11 @@ static int cdns_transfer_one(struct spi_controller *ctlr,
> cdns_spi_write(xspi, CDNS_SPI_THLD, xspi->tx_fifo_depth >> 1);
> }
>
> - /* When xspi in busy condition, bytes may send failed,
> - * then spi control didn't work thoroughly, add one byte delay
> - */
> - if (cdns_spi_read(xspi, CDNS_SPI_ISR) & CDNS_SPI_IXR_TXFULL)
> - udelay(10);
> -
> xspi->n_bytes = cdns_spi_n_bytes(transfer);
> xspi->tx_bytes = DIV_ROUND_UP(xspi->tx_bytes, xspi->n_bytes);
> xspi->rx_bytes = DIV_ROUND_UP(xspi->rx_bytes, xspi->n_bytes);
>
> - cdns_spi_process_fifo(xspi, xspi->tx_fifo_depth, 0);
> + cdns_spi_process_fifo(ctlr, xspi, xspi->tx_fifo_depth, 0);
>
> cdns_spi_write(xspi, CDNS_SPI_IER, CDNS_SPI_IXR_DEFAULT);
> return transfer->len;
next prev parent reply other threads:[~2026-07-20 15:55 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-07-20 12:55 Srikanth Boyapally
2026-07-20 15:55 ` Pandey, Radhey Shyam [this message]
2026-07-27 17:52 ` Mark Brown
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=81d818c5-c7f5-43b5-b8d2-202250ad213e@amd.com \
--to=radheys@amd.com \
--cc=broonie@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-spi@vger.kernel.org \
--cc=srikanth.boyapally@amd.com \
--cc=sxauwsk@163.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®