Skip to contentSkip to conversation

clk: hisilicon: Fix of_iomap leak in hisi_clk_init

PATCH In review Zhen Ni 2 participants 2 messages Jul 29, 2025 – Aug 1, 2025

Subsystems (2)
lore
Share & tools
lore

Zhen Ni

author
#0 lore raw

PATCH 1/1 clk: hisilicon: Fix of_iomap leak in hisi_clk_init

Smatch reports:

drivers/clk/hisilicon/clk.c:87 hisi_clk_init() warn: 'base' from of_iomap() not released on lines: 87.

This is because in hisi_clk_init(), base is not released when handle error, which may cause a leak.
To fix this, iounmap is added when handle error.

---
 drivers/clk/hisilicon/clk.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/clk/hisilicon/clk.c +2 -0
index 09368fd32bef..c0d309060240 100644
@@ -84,6 +84,8 @@ struct hisi_clock_data *hisi_clk_init(struct device_node *np,
 err_data:
 	kfree(clk_data);
 err:
+	if (base)
+		iounmap(base);
 	return NULL;
 }
 EXPORT_SYMBOL_GPL(hisi_clk_init);
Signature · 1 line
2.20.1
#1 lore raw
Quoted 4 lines

Smatch reports:

drivers/clk/hisilicon/clk.c:87 hisi_clk_init() warn: 'base' from of_iomap() not released on lines: 87.

How do you think about to add any tags (like “Fixes” and “Cc”) accordingly?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.16#n145

…

Quoted 1 line

To fix this, iounmap is added when handle error.

See also once more:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.16#n94

…

drivers/clk/hisilicon/clk.c struct hisi_clock_data *hisi_clk_init(struct device_node *np, · line 90

Quoted 9 lines

+++ b/drivers/clk/hisilicon/clk.c

drivers/clk/hisilicon/clk.c +2 -0
@@ -84,6 +84,8 @@ struct hisi_clock_data *hisi_clk_init(struct device_node *np,
 err_data:
 	kfree(clk_data);
 err:
+	if (base)
+		iounmap(base);
 	return NULL;
 }

…

* Please return directly after a failed of_iomap() call.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/coding-style.rst?h=v6.16#n532

* Would it be nicer to use the label “unmap_io”?

* May the statement “clk_data->base = base;” be performed a bit later?

* How do you think about to replace an of_clk_add_provider() call by of_clk_add_hw_provider()?
https://elixir.bootlin.com/linux/v6.16/source/drivers/clk/clk.c#L4996-L5001

Regards,
Markus