diff mbox series

[v2] cpufreq: qcom-hw: drop devm_xxx() calls from init/exit hooks

Message ID 20210119023925.22724-1-shawn.guo@linaro.org (mailing list archive)
State Accepted
Commit 67fc209b527d023db4d087c68e44e9790aa089ef
Headers show
Series [v2] cpufreq: qcom-hw: drop devm_xxx() calls from init/exit hooks | expand

Commit Message

Shawn Guo Jan. 19, 2021, 2:39 a.m. UTC
Commit f17b3e44320b ("cpufreq: qcom-hw: Use
devm_platform_ioremap_resource() to simplify code") introduces
a regression on platforms using the driver, by failing to initialise
a policy, when one is created post hotplug.

When all the CPUs of a policy are hoptplugged out, the call to .exit()
and later to devm_iounmap() does not release the memory region that was
requested during devm_platform_ioremap_resource().  Therefore,
a subsequent call to .init() will result in the following error, which
will prevent a new policy to be initialised:

[ 3395.915416] CPU4: shutdown
[ 3395.938185] psci: CPU4 killed (polled 0 ms)
[ 3399.071424] CPU5: shutdown
[ 3399.094316] psci: CPU5 killed (polled 0 ms)
[ 3402.139358] CPU6: shutdown
[ 3402.161705] psci: CPU6 killed (polled 0 ms)
[ 3404.742939] CPU7: shutdown
[ 3404.765592] psci: CPU7 killed (polled 0 ms)
[ 3411.492274] Detected VIPT I-cache on CPU4
[ 3411.492337] GICv3: CPU4: found redistributor 400 region 0:0x0000000017ae0000
[ 3411.492448] CPU4: Booted secondary processor 0x0000000400 [0x516f802d]
[ 3411.503654] qcom-cpufreq-hw 17d43000.cpufreq: can't request region for resource [mem 0x17d45800-0x17d46bff]

With that being said, the original code was tricky and skipping memory
region request intentionally to hide this issue.  The true cause is that
those devm_xxx() device managed functions shouldn't be used for cpufreq
init/exit hooks, because &pdev->dev is alive across the hooks and will
not trigger auto resource free-up.  Let's drop the use of device managed
functions and manually allocate/free resources, so that the issue can be
fixed properly.

Fixes: f17b3e44320b ("cpufreq: qcom-hw: Use devm_platform_ioremap_resource() to simplify code")
Suggested-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
---
Changes for v2:
- Move kzalloc() around to respect the original sequence which makes
  the local variables use reasonable.

 drivers/cpufreq/qcom-cpufreq-hw.c | 40 ++++++++++++++++++++++++-------
 1 file changed, 32 insertions(+), 8 deletions(-)

Comments

Shawn Guo Feb. 18, 2021, 3:19 a.m. UTC | #1
On Tue, Jan 19, 2021 at 10:39:25AM +0800, Shawn Guo wrote:
> Commit f17b3e44320b ("cpufreq: qcom-hw: Use
> devm_platform_ioremap_resource() to simplify code") introduces
> a regression on platforms using the driver, by failing to initialise
> a policy, when one is created post hotplug.
> 
> When all the CPUs of a policy are hoptplugged out, the call to .exit()
> and later to devm_iounmap() does not release the memory region that was
> requested during devm_platform_ioremap_resource().  Therefore,
> a subsequent call to .init() will result in the following error, which
> will prevent a new policy to be initialised:
> 
> [ 3395.915416] CPU4: shutdown
> [ 3395.938185] psci: CPU4 killed (polled 0 ms)
> [ 3399.071424] CPU5: shutdown
> [ 3399.094316] psci: CPU5 killed (polled 0 ms)
> [ 3402.139358] CPU6: shutdown
> [ 3402.161705] psci: CPU6 killed (polled 0 ms)
> [ 3404.742939] CPU7: shutdown
> [ 3404.765592] psci: CPU7 killed (polled 0 ms)
> [ 3411.492274] Detected VIPT I-cache on CPU4
> [ 3411.492337] GICv3: CPU4: found redistributor 400 region 0:0x0000000017ae0000
> [ 3411.492448] CPU4: Booted secondary processor 0x0000000400 [0x516f802d]
> [ 3411.503654] qcom-cpufreq-hw 17d43000.cpufreq: can't request region for resource [mem 0x17d45800-0x17d46bff]
> 
> With that being said, the original code was tricky and skipping memory
> region request intentionally to hide this issue.  The true cause is that
> those devm_xxx() device managed functions shouldn't be used for cpufreq
> init/exit hooks, because &pdev->dev is alive across the hooks and will
> not trigger auto resource free-up.  Let's drop the use of device managed
> functions and manually allocate/free resources, so that the issue can be
> fixed properly.
> 
> Fixes: f17b3e44320b ("cpufreq: qcom-hw: Use devm_platform_ioremap_resource() to simplify code")
> Suggested-by: Bjorn Andersson <bjorn.andersson@linaro.org>
> Signed-off-by: Shawn Guo <shawn.guo@linaro.org>

Hi Viresh,

What happened on this patch?  I thought we would have it in 5.11
release, but ...

Shawn
Viresh Kumar Feb. 18, 2021, 9:01 a.m. UTC | #2
On 18-02-21, 11:19, Shawn Guo wrote:
> What happened on this patch?  I thought we would have it in 5.11
> release, but ...

Just 2 days back I was about find the relevant thread and reply you
guys on the status of the patch that should fix the issue reported
earlier and it wasn't sent until now :(

Not sure how, but I missed this patch and yes it was in my inbox. This
doesn't normally happen as I tag such patches differently.

Any way, I will queue it now. Sorry about the delay.
patchwork-bot+linux-arm-msm@kernel.org March 1, 2021, 7:59 p.m. UTC | #3
Hello:

This patch was applied to qcom/linux.git (refs/heads/for-next):

On Tue, 19 Jan 2021 10:39:25 +0800 you wrote:
> Commit f17b3e44320b ("cpufreq: qcom-hw: Use
> devm_platform_ioremap_resource() to simplify code") introduces
> a regression on platforms using the driver, by failing to initialise
> a policy, when one is created post hotplug.
> 
> When all the CPUs of a policy are hoptplugged out, the call to .exit()
> and later to devm_iounmap() does not release the memory region that was
> requested during devm_platform_ioremap_resource().  Therefore,
> a subsequent call to .init() will result in the following error, which
> will prevent a new policy to be initialised:
> 
> [...]

Here is the summary with links:
  - [v2] cpufreq: qcom-hw: drop devm_xxx() calls from init/exit hooks
    https://git.kernel.org/qcom/c/67fc209b527d

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
diff mbox series

Patch

diff --git a/drivers/cpufreq/qcom-cpufreq-hw.c b/drivers/cpufreq/qcom-cpufreq-hw.c
index 9ed5341dc515..2726e77c9e5a 100644
--- a/drivers/cpufreq/qcom-cpufreq-hw.c
+++ b/drivers/cpufreq/qcom-cpufreq-hw.c
@@ -32,6 +32,7 @@  struct qcom_cpufreq_soc_data {
 
 struct qcom_cpufreq_data {
 	void __iomem *base;
+	struct resource *res;
 	const struct qcom_cpufreq_soc_data *soc_data;
 };
 
@@ -280,6 +281,7 @@  static int qcom_cpufreq_hw_cpu_init(struct cpufreq_policy *policy)
 	struct of_phandle_args args;
 	struct device_node *cpu_np;
 	struct device *cpu_dev;
+	struct resource *res;
 	void __iomem *base;
 	struct qcom_cpufreq_data *data;
 	int ret, index;
@@ -303,18 +305,33 @@  static int qcom_cpufreq_hw_cpu_init(struct cpufreq_policy *policy)
 
 	index = args.args[0];
 
-	base = devm_platform_ioremap_resource(pdev, index);
-	if (IS_ERR(base))
-		return PTR_ERR(base);
+	res = platform_get_resource(pdev, IORESOURCE_MEM, index);
+	if (!res) {
+		dev_err(dev, "failed to get mem resource %d\n", index);
+		return -ENODEV;
+	}
+
+	if (!request_mem_region(res->start, resource_size(res), res->name)) {
+		dev_err(dev, "failed to request resource %pR\n", res);
+		return -EBUSY;
+	}
 
-	data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
+	base = ioremap(res->start, resource_size(res));
+	if (IS_ERR(base)) {
+		dev_err(dev, "failed to map resource %pR\n", res);
+		ret = PTR_ERR(base);
+		goto release_region;
+	}
+
+	data = kzalloc(sizeof(*data), GFP_KERNEL);
 	if (!data) {
 		ret = -ENOMEM;
-		goto error;
+		goto unmap_base;
 	}
 
 	data->soc_data = of_device_get_match_data(&pdev->dev);
 	data->base = base;
+	data->res = res;
 
 	/* HW should be in enabled state to proceed */
 	if (!(readl_relaxed(base + data->soc_data->reg_enable) & 0x1)) {
@@ -349,7 +366,11 @@  static int qcom_cpufreq_hw_cpu_init(struct cpufreq_policy *policy)
 
 	return 0;
 error:
-	devm_iounmap(dev, base);
+	kfree(data);
+unmap_base:
+	iounmap(data->base);
+release_region:
+	release_mem_region(res->start, resource_size(res));
 	return ret;
 }
 
@@ -357,12 +378,15 @@  static int qcom_cpufreq_hw_cpu_exit(struct cpufreq_policy *policy)
 {
 	struct device *cpu_dev = get_cpu_device(policy->cpu);
 	struct qcom_cpufreq_data *data = policy->driver_data;
-	struct platform_device *pdev = cpufreq_get_driver_data();
+	struct resource *res = data->res;
+	void __iomem *base = data->base;
 
 	dev_pm_opp_remove_all_dynamic(cpu_dev);
 	dev_pm_opp_of_cpumask_remove_table(policy->related_cpus);
 	kfree(policy->freq_table);
-	devm_iounmap(&pdev->dev, data->base);
+	kfree(data);
+	iounmap(base);
+	release_mem_region(res->start, resource_size(res));
 
 	return 0;
 }