There are a number of error paths in the probe function that do not call of_node_put() to decrement the np device node refcount, leading to memory leaks if those errors occur.
In order to ease backporting, the fix has been divided into two patches: the first one simply adds the missing calls to of_node_put(), and the second one adds the __free() macro to the existing device nodes to remove the need for of_node_put(), ensuring that the same bug will not arise in the future.
The issue was found by chance while analyzing the code, and I do not have the hardware to test it beyond compiling and static analysis tools. Although the issue is clear and the fix too, if someone wants to volunteer to test the series with real hardware, it would be great.
Signed-off-by: Javier Carrasco javier.carrasco.cruz@gmail.com --- Javier Carrasco (2): cpufreq: qcom-nvmem: fix memory leaks in probe error paths cpufreq: qcom-nvmem: eliminate uses of of_node_put()
drivers/cpufreq/qcom-cpufreq-nvmem.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) --- base-commit: 3689b0ef08b70e4e03b82ebd37730a03a672853a change-id: 20240523-qcom-cpufreq-nvmem_memleak-6b6821db52b1
Best regards,
The code refactoring added new error paths between the np device node allocation and the call to of_node_put(), which leads to memory leaks if any of those errors occur.
Add the missing of_node_put() in the error paths that require it.
Cc: stable@vger.kernel.org Fixes: 57f2f8b4aa0c ("cpufreq: qcom: Refactor the driver to make it easier to extend") Signed-off-by: Javier Carrasco javier.carrasco.cruz@gmail.com --- drivers/cpufreq/qcom-cpufreq-nvmem.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/drivers/cpufreq/qcom-cpufreq-nvmem.c b/drivers/cpufreq/qcom-cpufreq-nvmem.c index ea05d9d67490..5004e1dbc752 100644 --- a/drivers/cpufreq/qcom-cpufreq-nvmem.c +++ b/drivers/cpufreq/qcom-cpufreq-nvmem.c @@ -480,23 +480,30 @@ static int qcom_cpufreq_probe(struct platform_device *pdev)
drv = devm_kzalloc(&pdev->dev, struct_size(drv, cpus, num_possible_cpus()), GFP_KERNEL); - if (!drv) + if (!drv) { + of_node_put(np); return -ENOMEM; + }
match = pdev->dev.platform_data; drv->data = match->data; - if (!drv->data) + if (!drv->data) { + of_node_put(np); return -ENODEV; + }
if (drv->data->get_version) { speedbin_nvmem = of_nvmem_cell_get(np, NULL); - if (IS_ERR(speedbin_nvmem)) + if (IS_ERR(speedbin_nvmem)) { + of_node_put(np); return dev_err_probe(cpu_dev, PTR_ERR(speedbin_nvmem), "Could not get nvmem cell\n"); + }
ret = drv->data->get_version(cpu_dev, speedbin_nvmem, &pvs_name, drv); if (ret) { + of_node_put(np); nvmem_cell_put(speedbin_nvmem); return ret; }
On 23-05-24, 23:24, Javier Carrasco wrote:
There are a number of error paths in the probe function that do not call of_node_put() to decrement the np device node refcount, leading to memory leaks if those errors occur.
In order to ease backporting, the fix has been divided into two patches: the first one simply adds the missing calls to of_node_put(), and the second one adds the __free() macro to the existing device nodes to remove the need for of_node_put(), ensuring that the same bug will not arise in the future.
The issue was found by chance while analyzing the code, and I do not have the hardware to test it beyond compiling and static analysis tools. Although the issue is clear and the fix too, if someone wants to volunteer to test the series with real hardware, it would be great.
Signed-off-by: Javier Carrasco javier.carrasco.cruz@gmail.com
Javier Carrasco (2): cpufreq: qcom-nvmem: fix memory leaks in probe error paths cpufreq: qcom-nvmem: eliminate uses of of_node_put()
Applied. Thanks.
linux-stable-mirror@lists.linaro.org
-
Javier Carrasco -
Viresh Kumar