The QSPI clocks are already enabled in probe() with clk_prepare_enable().
zynq_qspi_setup_op() enables them again with clk_enable(), but the extra
enable is never balanced. This leaves the clock enable count elevated and
can trigger warnings when unbinding the driver.

Remove the redundant clock enables from the setup callback.

Signed-off-by: Gaetan Florio <gaetansjo@gmail.com>
---
 drivers/spi/spi-zynq-qspi.c | 11 -----------
 1 file changed, 11 deletions(-)

diff --git a/drivers/spi/spi-zynq-qspi.c b/drivers/spi/spi-zynq-qspi.c
index 5f23e902b..2aa2c9621 100644
--- a/drivers/spi/spi-zynq-qspi.c
+++ b/drivers/spi/spi-zynq-qspi.c
@@ -437,21 +437,10 @@ static int zynq_qspi_setup_op(struct spi_device *spi)
 {
 	struct spi_controller *ctlr = spi->controller;
 	struct zynq_qspi *qspi = spi_controller_get_devdata(ctlr);
-	int ret;
 
 	if (ctlr->busy)
 		return -EBUSY;
 
-	ret = clk_enable(qspi->refclk);
-	if (ret)
-		return ret;
-
-	ret = clk_enable(qspi->pclk);
-	if (ret) {
-		clk_disable(qspi->refclk);
-		return ret;
-	}
-
 	zynq_qspi_write(qspi, ZYNQ_QSPI_ENABLE_OFFSET,
 			ZYNQ_QSPI_ENABLE_ENABLE_MASK);
 
-- 
2.43.0

On Mon, May 25, 2026 at 01:46:14PM -0400, Gaetan Florio wrote:
> The QSPI clocks are already enabled in probe() with clk_prepare_enable().
> zynq_qspi_setup_op() enables them again with clk_enable(), but the extra
> enable is never balanced. This leaves the clock enable count elevated and
> can trigger warnings when unbinding the driver.

This doesn't apply against current code, please check and resend - there
was a recent fix here so I think the issue is already fixed.