| Message ID | 20240907163630.55704-1-riyandhiman14@gmail.com (mailing list archive) |
|---|---|
| State | New |
| Delegated to: | viresh kumar |
| Headers | show |
| Series | [v2] cpufreq: mediatek: Fix resource leaks on mtk_cpu_create_freq_table() failure | expand |
On 07-09-24, 22:06, Riyan Dhiman wrote: > If mtk_cpu_create_freq_table() fails then there is a potential resource leak because > memory region is not released and IO memory is not unmapped. > Added error handling to ensure proper cleanup of all resources on failure, preventing potential leaks. > > Fixes: d776790a5536 (cpufreq: mediatek-hw: Fix double devm_remap in hotplug case) > > Signed-off-by: Riyan Dhiman <riyandhiman14@gmail.com> > --- > It is more of a extension of the above commit as this error handling was missing in that commit. > > v2: Fix commit message. > v1: Added error handling. > > drivers/cpufreq/mediatek-cpufreq-hw.c | 4 +++- > 1 file changed, 3 insertions(+), 1 deletion(-) > > diff --git a/drivers/cpufreq/mediatek-cpufreq-hw.c b/drivers/cpufreq/mediatek-cpufreq-hw.c > index 8925e096d5b9..3b1303f350ec 100644 > --- a/drivers/cpufreq/mediatek-cpufreq-hw.c > +++ b/drivers/cpufreq/mediatek-cpufreq-hw.c > @@ -207,13 +207,15 @@ static int mtk_cpu_resources_init(struct platform_device *pdev, > ret = mtk_cpu_create_freq_table(pdev, data); > if (ret) { > dev_info(dev, "Domain-%d failed to create freq table\n", index); > - return ret; > + goto unmap_region; > } > > policy->freq_table = data->table; > policy->driver_data = data; > > return 0; > +unmap_region: > + iounmap(base); Since this driver already uses devm_* APIs, what about using them instead ?
> Since this driver already uses devm_* APIs, what about using them > instead ? Since ioremap was called that's why I added a goto statement to handle the cleanup. In this file, devm_kzalloc and devm_regulator_get are called. Do I need to call this devm_platform_ioremap_resource instead of ioremap if devm_* api need to be used? Regards, Riyan Dhiman
On 01-10-24, 16:58, Riyan Dhiman wrote: > In this file, devm_kzalloc and devm_regulator_get are called. Do I need > to call this devm_platform_ioremap_resource instead of ioremap if devm_* > api need to be used? Yes, that would do.
diff --git a/drivers/cpufreq/mediatek-cpufreq-hw.c b/drivers/cpufreq/mediatek-cpufreq-hw.c index 8925e096d5b9..3b1303f350ec 100644 --- a/drivers/cpufreq/mediatek-cpufreq-hw.c +++ b/drivers/cpufreq/mediatek-cpufreq-hw.c @@ -207,13 +207,15 @@ static int mtk_cpu_resources_init(struct platform_device *pdev, ret = mtk_cpu_create_freq_table(pdev, data); if (ret) { dev_info(dev, "Domain-%d failed to create freq table\n", index); - return ret; + goto unmap_region; } policy->freq_table = data->table; policy->driver_data = data; return 0; +unmap_region: + iounmap(base); release_region: release_mem_region(res->start, resource_size(res)); return ret;
If mtk_cpu_create_freq_table() fails then there is a potential resource leak because memory region is not released and IO memory is not unmapped. Added error handling to ensure proper cleanup of all resources on failure, preventing potential leaks. Fixes: d776790a5536 (cpufreq: mediatek-hw: Fix double devm_remap in hotplug case) Signed-off-by: Riyan Dhiman <riyandhiman14@gmail.com> --- It is more of a extension of the above commit as this error handling was missing in that commit. v2: Fix commit message. v1: Added error handling. drivers/cpufreq/mediatek-cpufreq-hw.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)