In aggregate kprobe case, when arm_kprobe failed,
we need set the kp->flags with KPROBE_FLAG_DISABLED again.
If not, the 'kp' kprobe will been considered as enabled
but it actually not enabled.
Signed-off-by: Li Qiang <[email protected]>
---
kernel/kprobes.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 08350e35a..333454df5 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -2424,8 +2424,11 @@ int enable_kprobe(struct kprobe *kp)
if (!kprobes_all_disarmed && kprobe_disabled(p)) {
p->flags &= ~KPROBE_FLAG_DISABLED;
ret = arm_kprobe(p);
- if (ret)
+ if (ret) {
p->flags |= KPROBE_FLAG_DISABLED;
+ if (p != kp)
+ kp->flags |= KPROBE_FLAG_DISABLED;
+ }
}
out:
mutex_unlock(&kprobe_mutex);
--
2.25.1
Kindly ping.
Li Qiang <[email protected]> 于2022年9月3日周六 00:00写道:
>
> In aggregate kprobe case, when arm_kprobe failed,
> we need set the kp->flags with KPROBE_FLAG_DISABLED again.
> If not, the 'kp' kprobe will been considered as enabled
> but it actually not enabled.
>
> Signed-off-by: Li Qiang <[email protected]>
> ---
> kernel/kprobes.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index 08350e35a..333454df5 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -2424,8 +2424,11 @@ int enable_kprobe(struct kprobe *kp)
> if (!kprobes_all_disarmed && kprobe_disabled(p)) {
> p->flags &= ~KPROBE_FLAG_DISABLED;
> ret = arm_kprobe(p);
> - if (ret)
> + if (ret) {
> p->flags |= KPROBE_FLAG_DISABLED;
> + if (p != kp)
> + kp->flags |= KPROBE_FLAG_DISABLED;
> + }
> }
> out:
> mutex_unlock(&kprobe_mutex);
> --
> 2.25.1
>
On Fri, 2 Sep 2022 08:58:20 -0700
Li Qiang <[email protected]> wrote:
> In aggregate kprobe case, when arm_kprobe failed,
> we need set the kp->flags with KPROBE_FLAG_DISABLED again.
> If not, the 'kp' kprobe will been considered as enabled
> but it actually not enabled.
>
> Signed-off-by: Li Qiang <[email protected]>
Looks good to me.
Acked-by: Masami Hiramatsu (Google) <[email protected]>
Thanks!
> ---
> kernel/kprobes.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index 08350e35a..333454df5 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -2424,8 +2424,11 @@ int enable_kprobe(struct kprobe *kp)
> if (!kprobes_all_disarmed && kprobe_disabled(p)) {
> p->flags &= ~KPROBE_FLAG_DISABLED;
> ret = arm_kprobe(p);
> - if (ret)
> + if (ret) {
> p->flags |= KPROBE_FLAG_DISABLED;
> + if (p != kp)
> + kp->flags |= KPROBE_FLAG_DISABLED;
> + }
> }
> out:
> mutex_unlock(&kprobe_mutex);
> --
> 2.25.1
>
--
Masami Hiramatsu (Google) <[email protected]>
Masami Hiramatsu <[email protected]> 于2022年9月9日周五 21:07写道:
>
> On Fri, 2 Sep 2022 08:58:20 -0700
> Li Qiang <[email protected]> wrote:
>
> > In aggregate kprobe case, when arm_kprobe failed,
> > we need set the kp->flags with KPROBE_FLAG_DISABLED again.
> > If not, the 'kp' kprobe will been considered as enabled
> > but it actually not enabled.
> >
> > Signed-off-by: Li Qiang <[email protected]>
>
> Looks good to me.
>
> Acked-by: Masami Hiramatsu (Google) <[email protected]>
Hello Masami,
Who should merge this, you? This issue will crash the kernel before:
0bc11ed5ab60c135aa764a62c02cd5ea68289de4 commit ("kprobes: Allow
kprobes coexist with
livepatch") and
8094029330a2f03fb406ecff80671cf27ce28d42 commit("libbpf: Cleanup the
legacy kprobe_event on failed add/attach_event()")
Thanks,
Li Qiang
>
> Thanks!
>
> > ---
> > kernel/kprobes.c | 5 ++++-
> > 1 file changed, 4 insertions(+), 1 deletion(-)
> >
> > diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> > index 08350e35a..333454df5 100644
> > --- a/kernel/kprobes.c
> > +++ b/kernel/kprobes.c
> > @@ -2424,8 +2424,11 @@ int enable_kprobe(struct kprobe *kp)
> > if (!kprobes_all_disarmed && kprobe_disabled(p)) {
> > p->flags &= ~KPROBE_FLAG_DISABLED;
> > ret = arm_kprobe(p);
> > - if (ret)
> > + if (ret) {
> > p->flags |= KPROBE_FLAG_DISABLED;
> > + if (p != kp)
> > + kp->flags |= KPROBE_FLAG_DISABLED;
> > + }
> > }
> > out:
> > mutex_unlock(&kprobe_mutex);
> > --
> > 2.25.1
> >
>
>
> --
> Masami Hiramatsu (Google) <[email protected]>
Hello all,
What's the status of this patch?
Li Qiang <[email protected]> 于2022年9月3日周六 00:00写道:
>
> In aggregate kprobe case, when arm_kprobe failed,
> we need set the kp->flags with KPROBE_FLAG_DISABLED again.
> If not, the 'kp' kprobe will been considered as enabled
> but it actually not enabled.
>
> Signed-off-by: Li Qiang <[email protected]>
> ---
> kernel/kprobes.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index 08350e35a..333454df5 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -2424,8 +2424,11 @@ int enable_kprobe(struct kprobe *kp)
> if (!kprobes_all_disarmed && kprobe_disabled(p)) {
> p->flags &= ~KPROBE_FLAG_DISABLED;
> ret = arm_kprobe(p);
> - if (ret)
> + if (ret) {
> p->flags |= KPROBE_FLAG_DISABLED;
> + if (p != kp)
> + kp->flags |= KPROBE_FLAG_DISABLED;
> + }
> }
> out:
> mutex_unlock(&kprobe_mutex);
> --
> 2.25.1
>