Le 24/05/2021 à 15:59, Ulf Hansson a écrit :

On Fri, 21 May 2021 at 14:36, Christophe JAILLET
[off-list ref] wrote:

quoted

The 'reg_vmmc' regulator is never disabled. Neither in the error handling
of the probe, nor in the remove function.

Add a managed action to do the required clean-up before a 'regulator_put()'
call.

Fixes: 4dc5a79f1350 ("mmc: mxs-mmc: enable regulator for mmc slot")
Signed-off-by: Christophe JAILLET <redacted>
---
  drivers/mmc/host/mxs-mmc.c | 11 +++++++++++
  1 file changed, 11 insertions(+)

diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c
index 947581de7860..b043d53dd728 100644
--- a/drivers/mmc/host/mxs-mmc.c
+++ b/drivers/mmc/host/mxs-mmc.c

@@ -552,6 +552,13 @@ static const struct of_device_id mxs_mmc_dt_ids[] = {
  };
  MODULE_DEVICE_TABLE(of, mxs_mmc_dt_ids);

+static void regulator_disable_action(void *_data)
+{
+       struct regulator *regulator = _data;
+
+       regulator_disable(regulator);
+}
+
  static int mxs_mmc_probe(struct platform_device *pdev)
  {
         struct device_node *np = pdev->dev.of_node;

@@ -591,6 +598,10 @@ static int mxs_mmc_probe(struct platform_device *pdev)
                                 "Failed to enable vmmc regulator: %d\n", ret);
                         goto out_mmc_free;
                 }
+               ret = devm_add_action_or_reset(&pdev->dev,
+                                       regulator_disable_action, reg_vmmc);
+               if (ret)
+                       goto out_mmc_free;

Even if this improves the behaviour, there is a standardized way for
how we deal with regulators for mmc.

1. Call mmc_regulator_get_supply() during probe to fetch the optional
regulator. If a regulator is found a corresponding OCR mask, in
host->ocr_avail is assigned.

2. In the ->set_ios() callback, invoke mmc_regulator_set_ocr(). This
will also set the correct voltage-level and turn on/off the regulator,
depending on the requested OCR/voltage-level.

Hi,
I don't know this API.
I've tried to look at a few drivers to see how it was used, but it 
didn't help me either.

So, I won't be able to provide any other proposal on this. It would only 
be trial/error.
It is yours if something needs to be fixed here.

Anyway, thanks for your time for answering and trying to show the right 
direction.

CJ

quoted

         }

         ssp->clk = devm_clk_get(&pdev->dev, NULL);
--
2.30.2

Kind regards
Uffe

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel