[v2] kprobes: kretprobe events missing on 2-core KVM guest

  Default value of maxactive is set as num_possible_cpus() for nonpreemptable
systems. For a 2-core system, only 2 kretprobe instances would be allocated
in default, then these 2 instances for execve kretprobe are very likely to
be used up with a pipelined command.

Here's the testcase: a shell script was added to crontab, and the content
of the script is:

  #!/bin/sh
  do_something_magic `tr -dc a-z < /dev/urandom | head -c 10`

cron will trigger a series of program executions (4 times every hour). Then
events loss would be noticed normally after 3-4 hours of testings.

The issue is caused by a burst of series of execve requests. The best number
of kretprobe instances could be different case by case, and should be user's
duty to determine, but num_possible_cpus() as the default value is inadequate
especially for systems with small number of cpus.

This patch enables the logic for preemption as default, thus increases the
minimum of maxactive to 10 for nonpreemptable systems.

Signed-off-by: wuqiang <wuqiang.matt@bytedance.com>
---
 Documentation/trace/kprobes.rst |  3 +--
 kernel/kprobes.c                | 10 +++-------
 2 files changed, 4 insertions(+), 9 deletions(-)

--
2.34.1
  

On Thu, Nov 10, 2022 at 04:15:02PM +0800, wuqiang wrote:
> Default value of maxactive is set as num_possible_cpus() for nonpreemptable
> systems. For a 2-core system, only 2 kretprobe instances would be allocated
> in default, then these 2 instances for execve kretprobe are very likely to
> be used up with a pipelined command.
> 
> Here's the testcase: a shell script was added to crontab, and the content
> of the script is:
> 
>   #!/bin/sh
>   do_something_magic `tr -dc a-z < /dev/urandom | head -c 10`
> 
> cron will trigger a series of program executions (4 times every hour). Then
> events loss would be noticed normally after 3-4 hours of testings.
> 
> The issue is caused by a burst of series of execve requests. The best number
> of kretprobe instances could be different case by case, and should be user's
> duty to determine, but num_possible_cpus() as the default value is inadequate
> especially for systems with small number of cpus.
> 
> This patch enables the logic for preemption as default, thus increases the
> minimum of maxactive to 10 for nonpreemptable systems.
> 
> Signed-off-by: wuqiang <wuqiang.matt@bytedance.com>

Reviewed-by: Solar Designer <solar@openwall.com>

Thank you!

> ---
>  Documentation/trace/kprobes.rst |  3 +--
>  kernel/kprobes.c                | 10 +++-------
>  2 files changed, 4 insertions(+), 9 deletions(-)
> 
> diff --git a/Documentation/trace/kprobes.rst b/Documentation/trace/kprobes.rst
> index 48cf778a2468..fc7ce76eab65 100644
> --- a/Documentation/trace/kprobes.rst
> +++ b/Documentation/trace/kprobes.rst
> @@ -131,8 +131,7 @@ For example, if the function is non-recursive and is called with a
>  spinlock held, maxactive = 1 should be enough.  If the function is
>  non-recursive and can never relinquish the CPU (e.g., via a semaphore
>  or preemption), NR_CPUS should be enough.  If maxactive <= 0, it is
> -set to a default value.  If CONFIG_PREEMPT is enabled, the default
> -is max(10, 2*NR_CPUS).  Otherwise, the default is NR_CPUS.
> +set to a default value: max(10, 2*NR_CPUS).
>  
>  It's not a disaster if you set maxactive too low; you'll just miss
>  some probes.  In the kretprobe struct, the nmissed field is set to
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index a8b202f87e2d..1e80bddf2654 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -2212,11 +2212,7 @@ int register_kretprobe(struct kretprobe *rp)
>  	rp->kp.post_handler = NULL;
>  
>  	/* Pre-allocate memory for max kretprobe instances */
> -	if (rp->maxactive <= 0) {
> -#ifdef CONFIG_PREEMPTION
> +	if (rp->maxactive <= 0)
>  		rp->maxactive = max_t(unsigned int, 10, 2*num_possible_cpus());
> -#else
> -		rp->maxactive = num_possible_cpus();
> -#endif
> -	}
> +
>  #ifdef CONFIG_KRETPROBE_ON_RETHOOK
> --
> 2.34.1

Alexander
  

On Thu, 10 Nov 2022 15:52:23 +0100
Solar Designer <solar@openwall.com> wrote:

> On Thu, Nov 10, 2022 at 04:15:02PM +0800, wuqiang wrote:
> > Default value of maxactive is set as num_possible_cpus() for nonpreemptable
> > systems. For a 2-core system, only 2 kretprobe instances would be allocated
> > in default, then these 2 instances for execve kretprobe are very likely to
> > be used up with a pipelined command.
> > 
> > Here's the testcase: a shell script was added to crontab, and the content
> > of the script is:
> > 
> >   #!/bin/sh
> >   do_something_magic `tr -dc a-z < /dev/urandom | head -c 10`
> > 
> > cron will trigger a series of program executions (4 times every hour). Then
> > events loss would be noticed normally after 3-4 hours of testings.
> > 
> > The issue is caused by a burst of series of execve requests. The best number
> > of kretprobe instances could be different case by case, and should be user's
> > duty to determine, but num_possible_cpus() as the default value is inadequate
> > especially for systems with small number of cpus.
> > 
> > This patch enables the logic for preemption as default, thus increases the
> > minimum of maxactive to 10 for nonpreemptable systems.

Hm, OK. I think the enough value depends on where you put the probes on.
But maybe just making it NR_CPU -> 2 * NR_CPUS is reasonable.

Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>

Thanks,

> > 
> > Signed-off-by: wuqiang <wuqiang.matt@bytedance.com>
> 
> Reviewed-by: Solar Designer <solar@openwall.com>
> 
> Thank you!
> 
> > ---
> >  Documentation/trace/kprobes.rst |  3 +--
> >  kernel/kprobes.c                | 10 +++-------
> >  2 files changed, 4 insertions(+), 9 deletions(-)
> > 
> > diff --git a/Documentation/trace/kprobes.rst b/Documentation/trace/kprobes.rst
> > index 48cf778a2468..fc7ce76eab65 100644
> > --- a/Documentation/trace/kprobes.rst
> > +++ b/Documentation/trace/kprobes.rst
> > @@ -131,8 +131,7 @@ For example, if the function is non-recursive and is called with a
> >  spinlock held, maxactive = 1 should be enough.  If the function is
> >  non-recursive and can never relinquish the CPU (e.g., via a semaphore
> >  or preemption), NR_CPUS should be enough.  If maxactive <= 0, it is
> > -set to a default value.  If CONFIG_PREEMPT is enabled, the default
> > -is max(10, 2*NR_CPUS).  Otherwise, the default is NR_CPUS.
> > +set to a default value: max(10, 2*NR_CPUS).
> >  
> >  It's not a disaster if you set maxactive too low; you'll just miss
> >  some probes.  In the kretprobe struct, the nmissed field is set to
> > diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> > index a8b202f87e2d..1e80bddf2654 100644
> > --- a/kernel/kprobes.c
> > +++ b/kernel/kprobes.c
> > @@ -2212,11 +2212,7 @@ int register_kretprobe(struct kretprobe *rp)
> >  	rp->kp.post_handler = NULL;
> >  
> >  	/* Pre-allocate memory for max kretprobe instances */
> > -	if (rp->maxactive <= 0) {
> > -#ifdef CONFIG_PREEMPTION
> > +	if (rp->maxactive <= 0)
> >  		rp->maxactive = max_t(unsigned int, 10, 2*num_possible_cpus());
> > -#else
> > -		rp->maxactive = num_possible_cpus();
> > -#endif
> > -	}
> > +
> >  #ifdef CONFIG_KRETPROBE_ON_RETHOOK
> > --
> > 2.34.1
> 
> Alexander