On Wed, Apr 07, 2021 at 09:00:33AM +0200, Geert Uytterhoeven wrote:
> Hi Uwe,
> 
> I'm not Mark, but I'd like to share my 2€c.
> 
> On Tue, Apr 6, 2021 at 3:43 PM Uwe Kleine-König
> <u.kleine-koenig@pengutronix.de> wrote:
> > On Wed, Mar 31, 2021 at 01:02:12PM +0100, Mark Brown wrote:
> > > On Tue, Mar 30, 2021 at 08:17:55PM +0200, Uwe Kleine-König wrote:
> > > > devm_clk_get_enabled() returns the clk already (prepared and) enabled
> > > > and the automatically called cleanup cares for disabling (and
> > > > unpreparing). So simplify .probe() and .remove() accordingly.
> > >
> > > Acked-by: Mark Brown <broonie@kernel.org>
> >
> > Thanks. I wonder what you think about this series. Is it more "Well, ok,
> > if you must, the change you did to this spi driver looks correct." or
> > "This is a good simplification and a similar change for nearly all other
> > spi drivers that make use of a clk is possible, too. Dear clk
> > maintainers, please go forward and apply this useful series."?
> 
> While this simplifies drivers, this makes it harder to add power
> management by controlling the clocks through Runtime PM later, as that
> will require reverting the s/devm_clk_get/devm_clk_get_enabled/ again.

Hmm, if you start with a driver that uses devm_clk_get_enabled() you
have to do:

diff --git a/drivers/spi/spi-davinci.c b/drivers/spi/spi-davinci.c
index 63ee918ecdb0..07855f89290e 100644
--- a/drivers/spi/spi-davinci.c
+++ b/drivers/spi/spi-davinci.c
@@ -936,7 +936,7 @@ static int davinci_spi_probe(struct platform_device *pdev)
 
 	dspi->bitbang.master = master;
 
-	dspi->clk = devm_clk_get_enabled(&pdev->dev, NULL);
+	dspi->clk = devm_clk_get(&pdev->dev, NULL);
 	if (IS_ERR(dspi->clk)) {
 		ret = -ENODEV;
 		goto free_master;

(+ adding runtime PM of course). When you start with the previous state
of the driver you have to do:

diff --git a/drivers/spi/spi-davinci.c b/drivers/spi/spi-davinci.c
index 7453a1dbbc06..07855f89290e 100644
--- a/drivers/spi/spi-davinci.c
+++ b/drivers/spi/spi-davinci.c
@@ -941,9 +941,6 @@ static int davinci_spi_probe(struct platform_device *pdev)
 		ret = -ENODEV;
 		goto free_master;
 	}
-	ret = clk_prepare_enable(dspi->clk);
-	if (ret)
-		goto free_master;
 
 	master->use_gpio_descriptors = true;
 	master->dev.of_node = pdev->dev.of_node;
@@ -968,7 +965,7 @@ static int davinci_spi_probe(struct platform_device *pdev)
 
 	ret = davinci_spi_request_dma(dspi);
 	if (ret == -EPROBE_DEFER) {
-		goto free_clk;
+		goto free_master;
 	} else if (ret) {
 		dev_info(&pdev->dev, "DMA is not supported (%d)\n", ret);
 		dspi->dma_rx = NULL;
@@ -1012,8 +1009,6 @@ static int davinci_spi_probe(struct platform_device *pdev)
 		dma_release_channel(dspi->dma_rx);
 		dma_release_channel(dspi->dma_tx);
 	}
-free_clk:
-	clk_disable_unprepare(dspi->clk);
 free_master:
 	spi_master_put(master);
 err:
@@ -1039,8 +1034,6 @@ static int davinci_spi_remove(struct platform_device *pdev)
 
 	spi_bitbang_stop(&dspi->bitbang);
 
-	clk_disable_unprepare(dspi->clk);
-
 	if (dspi->dma_rx) {
 		dma_release_channel(dspi->dma_rx);
 		dma_release_channel(dspi->dma_tx);

(+ again adding runtime PM of course). Do you really think the latter is
the easier approach? Or what am I missing?

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |