Unverified Commit a37fb516 authored by Kexin Sun's avatar Kexin Sun Committed by Mark Brown
Browse files

spi: pl022: update outdated references to pump_transfers() 



The pump_transfers tasklet was removed when the driver
switched to the SPI core message processing loop in commit
9b2ef250 ("spi: spl022: switch to use default
spi_transfer_one_message()").  The kdoc of
pl022_interrupt_handler() still describes the old flow: scheduling
the pump_transfers tasklet, calling giveback(), and flagging
STATE_ERROR — none of which exist in the current code.  Rewrite
the kdoc to reflect the current paths, which flag errors via
SPI_TRANS_FAIL_IO and call spi_finalize_current_transfer()
directly.

Assisted-by: unnamed:deepseek-v3.2 coccinelle
Signed-off-by: default avatarKexin Sun <kexinsun@smail.nju.edu.cn>
Reviewed-by: default avatarLinus Walleij <linusw@kernel.org>
Link: https://patch.msgid.link/20260321105940.8143-1-kexinsun@smail.nju.edu.cn


Signed-off-by: default avatarMark Brown <broonie@kernel.org>
parent 9861e15f

drivers/spi/spi-pl022.c

+5 −5
Original line number Diff line number Diff line
@@ -1127,11 +1127,11 @@ static inline void pl022_dma_remove(struct pl022 *pl022) 
 *

 * This function handles interrupts generated for an interrupt based transfer.

 * If a receive overrun (ROR) interrupt is there then we disable SSP, flag the

 * current message's state as STATE_ERROR and schedule the tasklet

 * pump_transfers which will do the postprocessing of the current message by

 * calling giveback(). Otherwise it reads data from RX FIFO till there is no

 * more data, and writes data in TX FIFO till it is not full. If we complete

 * the transfer we move to the next transfer and schedule the tasklet.

 * current transfer with SPI_TRANS_FAIL_IO and call

 * spi_finalize_current_transfer() to let the core finish the message.

 * Otherwise it reads data from RX FIFO till there is no more data, and writes

 * data in TX FIFO till it is not full. When the transfer is complete we call

 * spi_finalize_current_transfer() so the core can schedule the next one.

 */

static irqreturn_t pl022_interrupt_handler(int irq, void *dev_id)

{