| Message ID | 20230323174026.950622-1-krzysztof.kozlowski@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Delegated to: | viresh kumar |
| Headers | show |
| Series | cpufreq: qcom-cpufreq-hw: fix double IO unmap and resource release on exit | expand |
On Thu, Mar 23, 2023 at 06:40:26PM +0100, Krzysztof Kozlowski wrote: > Commit 054a3ef683a1 ("cpufreq: qcom-hw: Allocate qcom_cpufreq_data > during probe") moved getting memory resource and iomap from > qcom_cpufreq_hw_cpu_init() to the probe function, however it left > untouched cleanup in qcom_cpufreq_hw_cpu_exit(). > > During device unbind this will lead to doule release of resource and > double iounmap(), first by qcom_cpufreq_hw_cpu_exit() and second via > managed resources: > > resource: Trying to free nonexistent resource <0x0000000018593000-0x0000000018593fff> > Trying to vunmap() nonexistent vm area (0000000088a7d4dc) > ... > vunmap (mm/vmalloc.c:2771 (discriminator 1)) > iounmap (mm/ioremap.c:60) > devm_ioremap_release (lib/devres.c:19) > devres_release_all (drivers/base/devres.c:506 drivers/base/devres.c:535) > device_unbind_cleanup (drivers/base/dd.c:523) > device_release_driver_internal (drivers/base/dd.c:1248 drivers/base/dd.c:1263) > device_driver_detach (drivers/base/dd.c:1300) > unbind_store (drivers/base/bus.c:243) > drv_attr_store (drivers/base/bus.c:127) > sysfs_kf_write (fs/sysfs/file.c:137) > kernfs_fop_write_iter (fs/kernfs/file.c:334) > vfs_write (include/linux/fs.h:1851 fs/read_write.c:491 fs/read_write.c:584) > ksys_write (fs/read_write.c:637) > __arm64_sys_write (fs/read_write.c:646) > invoke_syscall (arch/arm64/include/asm/current.h:19 arch/arm64/kernel/syscall.c:57) > el0_svc_common.constprop.0 (arch/arm64/include/asm/daifflags.h:28 arch/arm64/kernel/syscall.c:150) > do_el0_svc (arch/arm64/kernel/syscall.c:194) > el0_svc (arch/arm64/include/asm/daifflags.h:28 arch/arm64/kernel/entry-common.c:133 arch/arm64/kernel/entry-common.c:142 arch/arm64/kernel/entry-common.c:638) > el0t_64_sync_handler (arch/arm64/kernel/entry-common.c:656) > el0t_64_sync (arch/arm64/kernel/entry.S:591) > > Fixes: 054a3ef683a1 ("cpufreq: qcom-hw: Allocate qcom_cpufreq_data during probe") > Cc: <stable@vger.kernel.org> > Cc: Manivannan Sadhasivam <mani@kernel.org> > Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Manivannan Sadhasivam <mani@kernel.org> Thanks, Mani > --- > drivers/cpufreq/qcom-cpufreq-hw.c | 11 ++--------- > 1 file changed, 2 insertions(+), 9 deletions(-) > > diff --git a/drivers/cpufreq/qcom-cpufreq-hw.c b/drivers/cpufreq/qcom-cpufreq-hw.c > index 2f581d2d617d..b2d2907200a9 100644 > --- a/drivers/cpufreq/qcom-cpufreq-hw.c > +++ b/drivers/cpufreq/qcom-cpufreq-hw.c > @@ -43,7 +43,6 @@ struct qcom_cpufreq_soc_data { > > struct qcom_cpufreq_data { > void __iomem *base; > - struct resource *res; > > /* > * Mutex to synchronize between de-init sequence and re-starting LMh > @@ -590,16 +589,12 @@ 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 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); > qcom_cpufreq_hw_lmh_exit(data); > kfree(policy->freq_table); > kfree(data); > - iounmap(base); > - release_mem_region(res->start, resource_size(res)); > > return 0; > } > @@ -718,17 +713,15 @@ static int qcom_cpufreq_hw_driver_probe(struct platform_device *pdev) > for (i = 0; i < num_domains; i++) { > struct qcom_cpufreq_data *data = &qcom_cpufreq.data[i]; > struct clk_init_data clk_init = {}; > - struct resource *res; > void __iomem *base; > > - base = devm_platform_get_and_ioremap_resource(pdev, i, &res); > + base = devm_platform_ioremap_resource(pdev, i); > if (IS_ERR(base)) { > - dev_err(dev, "Failed to map resource %pR\n", res); > + dev_err(dev, "Failed to map resource index %d\n", i); > return PTR_ERR(base); > } > > data->base = base; > - data->res = res; > > /* Register CPU clock for each frequency domain */ > clk_init.name = kasprintf(GFP_KERNEL, "qcom_cpufreq%d", i); > -- > 2.34.1 >
On Thu, Mar 23, 2023 at 06:40:26PM +0100, Krzysztof Kozlowski wrote: > Commit 054a3ef683a1 ("cpufreq: qcom-hw: Allocate qcom_cpufreq_data > during probe") moved getting memory resource and iomap from > qcom_cpufreq_hw_cpu_init() to the probe function, however it left > untouched cleanup in qcom_cpufreq_hw_cpu_exit(). > > During device unbind this will lead to doule release of resource and > double iounmap(), first by qcom_cpufreq_hw_cpu_exit() and second via > managed resources: > > resource: Trying to free nonexistent resource <0x0000000018593000-0x0000000018593fff> > Trying to vunmap() nonexistent vm area (0000000088a7d4dc) > ... > vunmap (mm/vmalloc.c:2771 (discriminator 1)) > iounmap (mm/ioremap.c:60) > devm_ioremap_release (lib/devres.c:19) > devres_release_all (drivers/base/devres.c:506 drivers/base/devres.c:535) > device_unbind_cleanup (drivers/base/dd.c:523) > device_release_driver_internal (drivers/base/dd.c:1248 drivers/base/dd.c:1263) > device_driver_detach (drivers/base/dd.c:1300) > unbind_store (drivers/base/bus.c:243) > drv_attr_store (drivers/base/bus.c:127) > sysfs_kf_write (fs/sysfs/file.c:137) > kernfs_fop_write_iter (fs/kernfs/file.c:334) > vfs_write (include/linux/fs.h:1851 fs/read_write.c:491 fs/read_write.c:584) > ksys_write (fs/read_write.c:637) > __arm64_sys_write (fs/read_write.c:646) > invoke_syscall (arch/arm64/include/asm/current.h:19 arch/arm64/kernel/syscall.c:57) > el0_svc_common.constprop.0 (arch/arm64/include/asm/daifflags.h:28 arch/arm64/kernel/syscall.c:150) > do_el0_svc (arch/arm64/kernel/syscall.c:194) > el0_svc (arch/arm64/include/asm/daifflags.h:28 arch/arm64/kernel/entry-common.c:133 arch/arm64/kernel/entry-common.c:142 arch/arm64/kernel/entry-common.c:638) > el0t_64_sync_handler (arch/arm64/kernel/entry-common.c:656) > el0t_64_sync (arch/arm64/kernel/entry.S:591) > > Fixes: 054a3ef683a1 ("cpufreq: qcom-hw: Allocate qcom_cpufreq_data during probe") > Cc: <stable@vger.kernel.org> > Cc: Manivannan Sadhasivam <mani@kernel.org> > Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Bjorn Andersson <andersson@kernel.org> Regards, Bjorn > --- > drivers/cpufreq/qcom-cpufreq-hw.c | 11 ++--------- > 1 file changed, 2 insertions(+), 9 deletions(-) > > diff --git a/drivers/cpufreq/qcom-cpufreq-hw.c b/drivers/cpufreq/qcom-cpufreq-hw.c > index 2f581d2d617d..b2d2907200a9 100644 > --- a/drivers/cpufreq/qcom-cpufreq-hw.c > +++ b/drivers/cpufreq/qcom-cpufreq-hw.c > @@ -43,7 +43,6 @@ struct qcom_cpufreq_soc_data { > > struct qcom_cpufreq_data { > void __iomem *base; > - struct resource *res; > > /* > * Mutex to synchronize between de-init sequence and re-starting LMh > @@ -590,16 +589,12 @@ 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 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); > qcom_cpufreq_hw_lmh_exit(data); > kfree(policy->freq_table); > kfree(data); > - iounmap(base); > - release_mem_region(res->start, resource_size(res)); > > return 0; > } > @@ -718,17 +713,15 @@ static int qcom_cpufreq_hw_driver_probe(struct platform_device *pdev) > for (i = 0; i < num_domains; i++) { > struct qcom_cpufreq_data *data = &qcom_cpufreq.data[i]; > struct clk_init_data clk_init = {}; > - struct resource *res; > void __iomem *base; > > - base = devm_platform_get_and_ioremap_resource(pdev, i, &res); > + base = devm_platform_ioremap_resource(pdev, i); > if (IS_ERR(base)) { > - dev_err(dev, "Failed to map resource %pR\n", res); > + dev_err(dev, "Failed to map resource index %d\n", i); > return PTR_ERR(base); > } > > data->base = base; > - data->res = res; > > /* Register CPU clock for each frequency domain */ > clk_init.name = kasprintf(GFP_KERNEL, "qcom_cpufreq%d", i); > -- > 2.34.1 >
On 23-03-23, 18:40, Krzysztof Kozlowski wrote: > Commit 054a3ef683a1 ("cpufreq: qcom-hw: Allocate qcom_cpufreq_data > during probe") moved getting memory resource and iomap from > qcom_cpufreq_hw_cpu_init() to the probe function, however it left > untouched cleanup in qcom_cpufreq_hw_cpu_exit(). Applied. Thanks.
diff --git a/drivers/cpufreq/qcom-cpufreq-hw.c b/drivers/cpufreq/qcom-cpufreq-hw.c index 2f581d2d617d..b2d2907200a9 100644 --- a/drivers/cpufreq/qcom-cpufreq-hw.c +++ b/drivers/cpufreq/qcom-cpufreq-hw.c @@ -43,7 +43,6 @@ struct qcom_cpufreq_soc_data { struct qcom_cpufreq_data { void __iomem *base; - struct resource *res; /* * Mutex to synchronize between de-init sequence and re-starting LMh @@ -590,16 +589,12 @@ 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 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); qcom_cpufreq_hw_lmh_exit(data); kfree(policy->freq_table); kfree(data); - iounmap(base); - release_mem_region(res->start, resource_size(res)); return 0; } @@ -718,17 +713,15 @@ static int qcom_cpufreq_hw_driver_probe(struct platform_device *pdev) for (i = 0; i < num_domains; i++) { struct qcom_cpufreq_data *data = &qcom_cpufreq.data[i]; struct clk_init_data clk_init = {}; - struct resource *res; void __iomem *base; - base = devm_platform_get_and_ioremap_resource(pdev, i, &res); + base = devm_platform_ioremap_resource(pdev, i); if (IS_ERR(base)) { - dev_err(dev, "Failed to map resource %pR\n", res); + dev_err(dev, "Failed to map resource index %d\n", i); return PTR_ERR(base); } data->base = base; - data->res = res; /* Register CPU clock for each frequency domain */ clk_init.name = kasprintf(GFP_KERNEL, "qcom_cpufreq%d", i);
Commit 054a3ef683a1 ("cpufreq: qcom-hw: Allocate qcom_cpufreq_data during probe") moved getting memory resource and iomap from qcom_cpufreq_hw_cpu_init() to the probe function, however it left untouched cleanup in qcom_cpufreq_hw_cpu_exit(). During device unbind this will lead to doule release of resource and double iounmap(), first by qcom_cpufreq_hw_cpu_exit() and second via managed resources: resource: Trying to free nonexistent resource <0x0000000018593000-0x0000000018593fff> Trying to vunmap() nonexistent vm area (0000000088a7d4dc) ... vunmap (mm/vmalloc.c:2771 (discriminator 1)) iounmap (mm/ioremap.c:60) devm_ioremap_release (lib/devres.c:19) devres_release_all (drivers/base/devres.c:506 drivers/base/devres.c:535) device_unbind_cleanup (drivers/base/dd.c:523) device_release_driver_internal (drivers/base/dd.c:1248 drivers/base/dd.c:1263) device_driver_detach (drivers/base/dd.c:1300) unbind_store (drivers/base/bus.c:243) drv_attr_store (drivers/base/bus.c:127) sysfs_kf_write (fs/sysfs/file.c:137) kernfs_fop_write_iter (fs/kernfs/file.c:334) vfs_write (include/linux/fs.h:1851 fs/read_write.c:491 fs/read_write.c:584) ksys_write (fs/read_write.c:637) __arm64_sys_write (fs/read_write.c:646) invoke_syscall (arch/arm64/include/asm/current.h:19 arch/arm64/kernel/syscall.c:57) el0_svc_common.constprop.0 (arch/arm64/include/asm/daifflags.h:28 arch/arm64/kernel/syscall.c:150) do_el0_svc (arch/arm64/kernel/syscall.c:194) el0_svc (arch/arm64/include/asm/daifflags.h:28 arch/arm64/kernel/entry-common.c:133 arch/arm64/kernel/entry-common.c:142 arch/arm64/kernel/entry-common.c:638) el0t_64_sync_handler (arch/arm64/kernel/entry-common.c:656) el0t_64_sync (arch/arm64/kernel/entry.S:591) Fixes: 054a3ef683a1 ("cpufreq: qcom-hw: Allocate qcom_cpufreq_data during probe") Cc: <stable@vger.kernel.org> Cc: Manivannan Sadhasivam <mani@kernel.org> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> --- drivers/cpufreq/qcom-cpufreq-hw.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-)