Commit 92d585ef067d ("numa: fix NULL pointer access and memory
leak in unregister_one_node()") added kfree() of node struct in
unregister_one_node(). But node struct is freed by node_device_release()
which is called in unregister_node(). So by adding the kfree(),
node struct is freed two times.
While hot removing memory, the commit leads the following BUG_ON():
kernel BUG at mm/slub.c:3346!
invalid opcode: 0000 [#1] SMP
[...]
Call Trace:
[...] unregister_one_node
[...] try_offline_node
[...] remove_memory
[...] acpi_memory_device_remove
[...] acpi_bus_trim
[...] acpi_bus_trim
[...] acpi_device_hotplug
[...] acpi_hotplug_work_fn
[...] process_one_work
[...] worker_thread
[...] ? rescuer_thread
[...] kthread
[...] ? kthread_create_on_node
[...] ret_from_fork
[...] ? kthread_create_on_node
This patch removes unnecessary kfree() from unregister_one_node().
Signed-off-by: Yasuaki Ishimatsu <[email protected]>
Cc: Xishi Qiu <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: [email protected] # v3.16+
Fixes: 92d585ef067d "numa: fix NULL pointer access and memory leak in unregister_one_node()"
---
drivers/base/node.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/base/node.c b/drivers/base/node.c
index c6d3ae0..d51c49c 100644
--- a/drivers/base/node.c
+++ b/drivers/base/node.c
@@ -603,7 +603,6 @@ void unregister_one_node(int nid)
return;
unregister_node(node_devices[nid]);
- kfree(node_devices[nid]);
node_devices[nid] = NULL;
}
--
1.8.3.1
On 2014/10/3 18:06, Yasuaki Ishimatsu wrote:
> Commit 92d585ef067d ("numa: fix NULL pointer access and memory
> leak in unregister_one_node()") added kfree() of node struct in
> unregister_one_node(). But node struct is freed by node_device_release()
> which is called in unregister_node(). So by adding the kfree(),
Hi,
Is this path?
unregister_node()
device_unregister()
device_del()
bus_remove_device()
device_release_driver()
__device_release_driver()
devres_release_all()
release_nodes()
dr->node.release(dev, dr->data);
then which function is be called?
Thanks,
Xishi Qiu
> node struct is freed two times.
>
> While hot removing memory, the commit leads the following BUG_ON():
>
> kernel BUG at mm/slub.c:3346!
> invalid opcode: 0000 [#1] SMP
> [...]
> Call Trace:
> [...] unregister_one_node
> [...] try_offline_node
> [...] remove_memory
> [...] acpi_memory_device_remove
> [...] acpi_bus_trim
> [...] acpi_bus_trim
> [...] acpi_device_hotplug
> [...] acpi_hotplug_work_fn
> [...] process_one_work
> [...] worker_thread
> [...] ? rescuer_thread
> [...] kthread
> [...] ? kthread_create_on_node
> [...] ret_from_fork
> [...] ? kthread_create_on_node
>
> This patch removes unnecessary kfree() from unregister_one_node().
>
> Signed-off-by: Yasuaki Ishimatsu <[email protected]>
> Cc: Xishi Qiu <[email protected]>
> Cc: Greg Kroah-Hartman <[email protected]>
> Cc: Andrew Morton <[email protected]>
> Cc: [email protected] # v3.16+
> Fixes: 92d585ef067d "numa: fix NULL pointer access and memory leak in unregister_one_node()"
> ---
> drivers/base/node.c | 1 -
> 1 file changed, 1 deletion(-)
>
> diff --git a/drivers/base/node.c b/drivers/base/node.c
> index c6d3ae0..d51c49c 100644
> --- a/drivers/base/node.c
> +++ b/drivers/base/node.c
> @@ -603,7 +603,6 @@ void unregister_one_node(int nid)
> return;
>
> unregister_node(node_devices[nid]);
> - kfree(node_devices[nid]);
> node_devices[nid] = NULL;
> }
>
(2014/10/09 13:10), Xishi Qiu wrote:
> On 2014/10/3 18:06, Yasuaki Ishimatsu wrote:
>
>> Commit 92d585ef067d ("numa: fix NULL pointer access and memory
>> leak in unregister_one_node()") added kfree() of node struct in
>> unregister_one_node(). But node struct is freed by node_device_release()
>> which is called in unregister_node(). So by adding the kfree(),
>
> Hi,
>
> Is this path?
> unregister_node()
> device_unregister()
> device_del()
> bus_remove_device()
> device_release_driver()
> __device_release_driver()
> devres_release_all()
> release_nodes()
> dr->node.release(dev, dr->data);
> then which function is be called?
node_device_release is called as follows:
unregister_one_node()
-> unregister_node()
-> device_unregister()
-> put_device()
-> kobject_put()
-> kref_put()
-> kref_sub()
-> kobject_release()
-> kobject_cleanup()
-> device_release()
-> node_device_release()
Thanks,
Yasuaki Ishimatsu
>
> Thanks,
> Xishi Qiu
>
>> node struct is freed two times.
>>
>> While hot removing memory, the commit leads the following BUG_ON():
>>
>> kernel BUG at mm/slub.c:3346!
>> invalid opcode: 0000 [#1] SMP
>> [...]
>> Call Trace:
>> [...] unregister_one_node
>> [...] try_offline_node
>> [...] remove_memory
>> [...] acpi_memory_device_remove
>> [...] acpi_bus_trim
>> [...] acpi_bus_trim
>> [...] acpi_device_hotplug
>> [...] acpi_hotplug_work_fn
>> [...] process_one_work
>> [...] worker_thread
>> [...] ? rescuer_thread
>> [...] kthread
>> [...] ? kthread_create_on_node
>> [...] ret_from_fork
>> [...] ? kthread_create_on_node
>>
>> This patch removes unnecessary kfree() from unregister_one_node().
>>
>> Signed-off-by: Yasuaki Ishimatsu <[email protected]>
>> Cc: Xishi Qiu <[email protected]>
>> Cc: Greg Kroah-Hartman <[email protected]>
>> Cc: Andrew Morton <[email protected]>
>> Cc: [email protected] # v3.16+
>> Fixes: 92d585ef067d "numa: fix NULL pointer access and memory leak in unregister_one_node()"
>> ---
>> drivers/base/node.c | 1 -
>> 1 file changed, 1 deletion(-)
>>
>> diff --git a/drivers/base/node.c b/drivers/base/node.c
>> index c6d3ae0..d51c49c 100644
>> --- a/drivers/base/node.c
>> +++ b/drivers/base/node.c
>> @@ -603,7 +603,6 @@ void unregister_one_node(int nid)
>> return;
>>
>> unregister_node(node_devices[nid]);
>> - kfree(node_devices[nid]);
>> node_devices[nid] = NULL;
>> }
>>
>
>
>