From: Minfei Huang <[email protected]>
It is more better that klp_disable_func returnes immediately, if
func->state and func->old_addr do not satisfy the condition.
We should robust the livepatch code, although the above situation never
happen in current code path.
Signed-off-by: Minfei Huang <[email protected]>
---
kernel/livepatch/core.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/kernel/livepatch/core.c b/kernel/livepatch/core.c
index c40ebcc..6e53441 100644
--- a/kernel/livepatch/core.c
+++ b/kernel/livepatch/core.c
@@ -348,8 +348,10 @@ static void klp_disable_func(struct klp_func *func)
{
struct klp_ops *ops;
- WARN_ON(func->state != KLP_ENABLED);
- WARN_ON(!func->old_addr);
+ if (WARN_ON(func->state != KLP_ENABLED))
+ return;
+ if (WARN_ON(!func->old_addr))
+ return;
ops = klp_find_ops(func->old_addr);
if (WARN_ON(!ops))
--
2.2.2
On Tue, Jul 14, 2015 at 11:15:37AM +0800, Minfei Huang wrote:
> From: Minfei Huang <[email protected]>
>
> It is more better that klp_disable_func returnes immediately, if
> func->state and func->old_addr do not satisfy the condition.
>
> We should robust the livepatch code, although the above situation never
> happen in current code path.
>
> Signed-off-by: Minfei Huang <[email protected]>
> ---
> kernel/livepatch/core.c | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/livepatch/core.c b/kernel/livepatch/core.c
> index c40ebcc..6e53441 100644
> --- a/kernel/livepatch/core.c
> +++ b/kernel/livepatch/core.c
> @@ -348,8 +348,10 @@ static void klp_disable_func(struct klp_func *func)
> {
> struct klp_ops *ops;
>
> - WARN_ON(func->state != KLP_ENABLED);
> - WARN_ON(!func->old_addr);
> + if (WARN_ON(func->state != KLP_ENABLED))
> + return;
> + if (WARN_ON(!func->old_addr))
> + return;
>
> ops = klp_find_ops(func->old_addr);
> if (WARN_ON(!ops))
Acked-by: Josh Poimboeuf <[email protected]>
On Tue, 14 Jul 2015, Minfei Huang wrote:
> From: Minfei Huang <[email protected]>
>
> It is more better that klp_disable_func returnes immediately, if
> func->state and func->old_addr do not satisfy the condition.
>
> We should robust the livepatch code, although the above situation never
> happen in current code path.
I have reworded the changelog a little bit and applied to
for-4.3/upstream.
Thanks,
--
Jiri Kosina
SUSE Labs