diff mbox series

[V2] cpufreq: dt: check the -EPROBE_DEFER error returned by dev_pm_opp_of_cpumask_add_table

Message ID 20210325064208.2291413-1-quanyang.wang@windriver.com (mailing list archive)
State New, archived
Delegated to: viresh kumar
Headers show
Series [V2] cpufreq: dt: check the -EPROBE_DEFER error returned by dev_pm_opp_of_cpumask_add_table | expand

Commit Message

Quanyang Wang March 25, 2021, 6:42 a.m. UTC 
From: Quanyang Wang <quanyang.wang@windriver.com>

The function dev_pm_opp_of_cpumask_add_table may return -EPROBE_DEFER
which comes from clk_get(dev, NULL) in _update_opp_table_clk. Ignoring
this error and call the next function  dev_pm_opp_of_cpumask_add_table
may cause dt_cpufreq_probe return -ENODEV instead of -EPROBE_DEFER.
And this will result cpufreq-dt driver probe failure.

Add check condition to fix this issue.

Signed-off-by: Quanyang Wang <quanyang.wang@windriver.com>
---
 drivers/cpufreq/cpufreq-dt.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

Comments

Viresh Kumar March 25, 2021, 7:12 a.m. UTC | #1 
On 25-03-21, 14:42, quanyang.wang@windriver.com wrote:
> From: Quanyang Wang <quanyang.wang@windriver.com>

Made some changes and applied this:

Author: Quanyang Wang <quanyang.wang@windriver.com>
Date:   Thu Mar 25 14:42:08 2021 +0800

    cpufreq: dt: dev_pm_opp_of_cpumask_add_table() may return -EPROBE_DEFER
    
    The function dev_pm_opp_of_cpumask_add_table() may return -EPROBE_DEFER,
    which needs to be propagated to the caller to try probing the driver
    later on.
    
    Signed-off-by: Quanyang Wang <quanyang.wang@windriver.com>
    [ Viresh: Massage changelog/subject, improve code. ]
    Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
 drivers/cpufreq/cpufreq-dt.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/drivers/cpufreq/cpufreq-dt.c b/drivers/cpufreq/cpufreq-dt.c
index b1e1bdc63b01..ece52863ba62 100644
--- a/drivers/cpufreq/cpufreq-dt.c
+++ b/drivers/cpufreq/cpufreq-dt.c
@@ -255,10 +255,15 @@ static int dt_cpufreq_early_init(struct device *dev, int cpu)
         * before updating priv->cpus. Otherwise, we will end up creating
         * duplicate OPPs for the CPUs.
         *
-        * OPPs might be populated at runtime, don't check for error here.
+        * OPPs might be populated at runtime, don't fail for error here unless
+        * it is -EPROBE_DEFER.
         */
-       if (!dev_pm_opp_of_cpumask_add_table(priv->cpus))
+       ret = dev_pm_opp_of_cpumask_add_table(priv->cpus);
+       if (!ret) {
                priv->have_static_opps = true;
+       } else if (ret == -EPROBE_DEFER) {
+               goto out;
+       }
 
        /*
         * The OPP table must be initialized, statically or dynamically, by this
diff mbox series

Patch

diff --git a/drivers/cpufreq/cpufreq-dt.c b/drivers/cpufreq/cpufreq-dt.c
index b1e1bdc63b01..25e46df92941 100644
--- a/drivers/cpufreq/cpufreq-dt.c
+++ b/drivers/cpufreq/cpufreq-dt.c
@@ -257,7 +257,12 @@  static int dt_cpufreq_early_init(struct device *dev, int cpu)
 	 *
 	 * OPPs might be populated at runtime, don't check for error here.
 	 */
-	if (!dev_pm_opp_of_cpumask_add_table(priv->cpus))
+	ret = dev_pm_opp_of_cpumask_add_table(priv->cpus);
+	if (ret) {
+		/* Return the -EPROBE_DEFER error to trigger the deferred probe. */
+		if (ret == -EPROBE_DEFER)
+			goto out;
+	} else
 		priv->have_static_opps = true;
 
 	/*