[PATCH bpf-next] bpf: Prevent bpf program recursion for raw tracepoint probes

[PATCH bpf-next] bpf: Prevent bpf program recursion for raw tracepoint probes

[Jiri Olsa]

We got report from sysbot [1] about warnings that were caused by
bpf program attached to contention_begin raw tracepoint triggering
the same tracepoint by using bpf_trace_printk helper that takes
trace_printk_lock lock.

 Call Trace:
  <TASK>
  ? trace_event_raw_event_bpf_trace_printk+0x5f/0x90
  bpf_trace_printk+0x2b/0xe0
  bpf_prog_a9aec6167c091eef_prog+0x1f/0x24
  bpf_trace_run2+0x26/0x90
  native_queued_spin_lock_slowpath+0x1c6/0x2b0
  _raw_spin_lock_irqsave+0x44/0x50
  bpf_trace_printk+0x3f/0xe0
  bpf_prog_a9aec6167c091eef_prog+0x1f/0x24
  bpf_trace_run2+0x26/0x90
  native_queued_spin_lock_slowpath+0x1c6/0x2b0
  _raw_spin_lock_irqsave+0x44/0x50
  bpf_trace_printk+0x3f/0xe0
  bpf_prog_a9aec6167c091eef_prog+0x1f/0x24
  bpf_trace_run2+0x26/0x90
  native_queued_spin_lock_slowpath+0x1c6/0x2b0
  _raw_spin_lock_irqsave+0x44/0x50
  bpf_trace_printk+0x3f/0xe0
  bpf_prog_a9aec6167c091eef_prog+0x1f/0x24
  bpf_trace_run2+0x26/0x90
  native_queued_spin_lock_slowpath+0x1c6/0x2b0
  _raw_spin_lock_irqsave+0x44/0x50
  __unfreeze_partials+0x5b/0x160
  ...

The can be reproduced by attaching bpf program as raw tracepoint on
contention_begin tracepoint. The bpf prog calls bpf_trace_printk
helper. Then by running perf bench the spin lock code is forced to
take slowpath and call contention_begin tracepoint.

Fixing this by skipping execution of the bpf program if it's
already running, Using bpf prog 'active' field, which is being
currently used by trampoline programs for the same reason.

Reported-by: syzbot+2251879aa068ad9c960d@syzkaller.appspotmail.com
[1] https://lore.kernel.org/bpf/YxhFe3EwqchC%2FfYf@krava/T/#t
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
 include/linux/bpf.h      | 1 +
 kernel/bpf/trampoline.c  | 6 +++---
 kernel/trace/bpf_trace.c | 6 ++++++
 3 files changed, 10 insertions(+), 3 deletions(-)

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 48ae05099f36..4737bd0fcbb8 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -2640,4 +2640,5 @@ static inline void bpf_cgroup_atype_get(u32 attach_btf_id, int cgroup_atype) {}
 static inline void bpf_cgroup_atype_put(int cgroup_atype) {}
 #endif /* CONFIG_BPF_LSM */
 
+void notrace bpf_prog_inc_misses_counter(struct bpf_prog *prog);
 #endif /* _LINUX_BPF_H */
diff --git a/kernel/bpf/trampoline.c b/kernel/bpf/trampoline.c
index ad76940b02cc..a098bdc33209 100644
--- a/kernel/bpf/trampoline.c
+++ b/kernel/bpf/trampoline.c
@@ -863,7 +863,7 @@ static __always_inline u64 notrace bpf_prog_start_time(void)
 	return start;
 }
 
-static void notrace inc_misses_counter(struct bpf_prog *prog)
+void notrace bpf_prog_inc_misses_counter(struct bpf_prog *prog)
 {
 	struct bpf_prog_stats *stats;
 	unsigned int flags;
@@ -896,7 +896,7 @@ u64 notrace __bpf_prog_enter(struct bpf_prog *prog, struct bpf_tramp_run_ctx *ru
 	run_ctx->saved_run_ctx = bpf_set_run_ctx(&run_ctx->run_ctx);
 
 	if (unlikely(this_cpu_inc_return(*(prog->active)) != 1)) {
-		inc_misses_counter(prog);
+		bpf_prog_inc_misses_counter(prog);
 		return 0;
 	}
 	return bpf_prog_start_time();
@@ -967,7 +967,7 @@ u64 notrace __bpf_prog_enter_sleepable(struct bpf_prog *prog, struct bpf_tramp_r
 	might_fault();
 
 	if (unlikely(this_cpu_inc_return(*(prog->active)) != 1)) {
-		inc_misses_counter(prog);
+		bpf_prog_inc_misses_counter(prog);
 		return 0;
 	}
 
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index 68e5cdd24cef..c8cd1aa7b112 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -2042,9 +2042,15 @@ static __always_inline
 void __bpf_trace_run(struct bpf_prog *prog, u64 *args)
 {
 	cant_sleep();
+	if (unlikely(this_cpu_inc_return(*(prog->active)) != 1)) {
+		bpf_prog_inc_misses_counter(prog);
+		goto out;
+	}
 	rcu_read_lock();
 	(void) bpf_prog_run(prog, args);
 	rcu_read_unlock();
+out:
+	this_cpu_dec(*(prog->active));
 }
 
 #define UNPACK(...)			__VA_ARGS__
-- 
2.37.3

Re: [PATCH bpf-next] bpf: Prevent bpf program recursion for raw tracepoint probes

[sdf]

On 09/08, Jiri Olsa wrote:
> We got report from sysbot [1] about warnings that were caused by
> bpf program attached to contention_begin raw tracepoint triggering
> the same tracepoint by using bpf_trace_printk helper that takes
> trace_printk_lock lock.

>   Call Trace:
>    <TASK>
>    ? trace_event_raw_event_bpf_trace_printk+0x5f/0x90
>    bpf_trace_printk+0x2b/0xe0
>    bpf_prog_a9aec6167c091eef_prog+0x1f/0x24
>    bpf_trace_run2+0x26/0x90
>    native_queued_spin_lock_slowpath+0x1c6/0x2b0
>    _raw_spin_lock_irqsave+0x44/0x50
>    bpf_trace_printk+0x3f/0xe0
>    bpf_prog_a9aec6167c091eef_prog+0x1f/0x24
>    bpf_trace_run2+0x26/0x90
>    native_queued_spin_lock_slowpath+0x1c6/0x2b0
>    _raw_spin_lock_irqsave+0x44/0x50
>    bpf_trace_printk+0x3f/0xe0
>    bpf_prog_a9aec6167c091eef_prog+0x1f/0x24
>    bpf_trace_run2+0x26/0x90
>    native_queued_spin_lock_slowpath+0x1c6/0x2b0
>    _raw_spin_lock_irqsave+0x44/0x50
>    bpf_trace_printk+0x3f/0xe0
>    bpf_prog_a9aec6167c091eef_prog+0x1f/0x24
>    bpf_trace_run2+0x26/0x90
>    native_queued_spin_lock_slowpath+0x1c6/0x2b0
>    _raw_spin_lock_irqsave+0x44/0x50
>    __unfreeze_partials+0x5b/0x160
>    ...

> The can be reproduced by attaching bpf program as raw tracepoint on
> contention_begin tracepoint. The bpf prog calls bpf_trace_printk
> helper. Then by running perf bench the spin lock code is forced to
> take slowpath and call contention_begin tracepoint.

> Fixing this by skipping execution of the bpf program if it's
> already running, Using bpf prog 'active' field, which is being
> currently used by trampoline programs for the same reason.

Makes sense to me and seems to address Alexei's earlier point
about bpf_prog_active.

Reviewed-by: Stanislav Fomichev <sdf@google.com>

> Reported-by: syzbot+2251879aa068ad9c960d@syzkaller.appspotmail.com
> [1] https://lore.kernel.org/bpf/YxhFe3EwqchC%2FfYf@krava/T/#t
> Signed-off-by: Jiri Olsa <jolsa@kernel.org>
> ---
>   include/linux/bpf.h      | 1 +
>   kernel/bpf/trampoline.c  | 6 +++---
>   kernel/trace/bpf_trace.c | 6 ++++++
>   3 files changed, 10 insertions(+), 3 deletions(-)

> diff --git a/include/linux/bpf.h b/include/linux/bpf.h
> index 48ae05099f36..4737bd0fcbb8 100644
> --- a/include/linux/bpf.h
> +++ b/include/linux/bpf.h
> @@ -2640,4 +2640,5 @@ static inline void bpf_cgroup_atype_get(u32  
> attach_btf_id, int cgroup_atype) {}
>   static inline void bpf_cgroup_atype_put(int cgroup_atype) {}
>   #endif /* CONFIG_BPF_LSM */

> +void notrace bpf_prog_inc_misses_counter(struct bpf_prog *prog);
>   #endif /* _LINUX_BPF_H */
> diff --git a/kernel/bpf/trampoline.c b/kernel/bpf/trampoline.c
> index ad76940b02cc..a098bdc33209 100644
> --- a/kernel/bpf/trampoline.c
> +++ b/kernel/bpf/trampoline.c
> @@ -863,7 +863,7 @@ static __always_inline u64 notrace  
> bpf_prog_start_time(void)
>   	return start;
>   }

> -static void notrace inc_misses_counter(struct bpf_prog *prog)
> +void notrace bpf_prog_inc_misses_counter(struct bpf_prog *prog)
>   {
>   	struct bpf_prog_stats *stats;
>   	unsigned int flags;
> @@ -896,7 +896,7 @@ u64 notrace __bpf_prog_enter(struct bpf_prog *prog,  
> struct bpf_tramp_run_ctx *ru
>   	run_ctx->saved_run_ctx = bpf_set_run_ctx(&run_ctx->run_ctx);

>   	if (unlikely(this_cpu_inc_return(*(prog->active)) != 1)) {
> -		inc_misses_counter(prog);
> +		bpf_prog_inc_misses_counter(prog);
>   		return 0;
>   	}
>   	return bpf_prog_start_time();
> @@ -967,7 +967,7 @@ u64 notrace __bpf_prog_enter_sleepable(struct  
> bpf_prog *prog, struct bpf_tramp_r
>   	might_fault();

>   	if (unlikely(this_cpu_inc_return(*(prog->active)) != 1)) {
> -		inc_misses_counter(prog);
> +		bpf_prog_inc_misses_counter(prog);
>   		return 0;
>   	}

> diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
> index 68e5cdd24cef..c8cd1aa7b112 100644
> --- a/kernel/trace/bpf_trace.c
> +++ b/kernel/trace/bpf_trace.c
> @@ -2042,9 +2042,15 @@ static __always_inline
>   void __bpf_trace_run(struct bpf_prog *prog, u64 *args)
>   {
>   	cant_sleep();
> +	if (unlikely(this_cpu_inc_return(*(prog->active)) != 1)) {
> +		bpf_prog_inc_misses_counter(prog);
> +		goto out;
> +	}
>   	rcu_read_lock();
>   	(void) bpf_prog_run(prog, args);
>   	rcu_read_unlock();
> +out:
> +	this_cpu_dec(*(prog->active));
>   }

>   #define UNPACK(...)			__VA_ARGS__
> --
> 2.37.3

Re: [PATCH bpf-next] bpf: Prevent bpf program recursion for raw tracepoint probes

[kernel test robot]

Hi Jiri,

I love your patch! Yet something to improve:

[auto build test ERROR on bpf-next/master]

url:    https://github.com/intel-lab-lkp/linux/commits/Jiri-Olsa/bpf-Prevent-bpf-program-recursion-for-raw-tracepoint-probes/20220908-194832
base:   https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git master
config: arm64-buildonly-randconfig-r002-20220907 (https://download.01.org/0day-ci/archive/20220909/202209091236.avgRKOSj-lkp@intel.com/config)
compiler: aarch64-linux-gcc (GCC) 12.1.0
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # https://github.com/intel-lab-lkp/linux/commit/f68b567cfb6572c20e431242a440cc5f01452485
        git remote add linux-review https://github.com/intel-lab-lkp/linux
        git fetch --no-tags linux-review Jiri-Olsa/bpf-Prevent-bpf-program-recursion-for-raw-tracepoint-probes/20220908-194832
        git checkout f68b567cfb6572c20e431242a440cc5f01452485
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=arm64 SHELL=/bin/bash

If you fix the issue, kindly add following tag where applicable
Reported-by: kernel test robot <lkp@intel.com>

All errors (new ones prefixed by >>):

   aarch64-linux-ld: Unexpected GOT/PLT entries detected!
   aarch64-linux-ld: Unexpected run-time procedure linkages detected!
   aarch64-linux-ld: kernel/trace/bpf_trace.o: in function `__bpf_trace_run':
>> kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
>> aarch64-linux-ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
>> aarch64-linux-ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
>> aarch64-linux-ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
>> aarch64-linux-ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
   aarch64-linux-ld: kernel/trace/bpf_trace.o:kernel/trace/bpf_trace.c:2046: more undefined references to `bpf_prog_inc_misses_counter' follow


vim +2046 kernel/trace/bpf_trace.c

  2040	
  2041	static __always_inline
  2042	void __bpf_trace_run(struct bpf_prog *prog, u64 *args)
  2043	{
  2044		cant_sleep();
  2045		if (unlikely(this_cpu_inc_return(*(prog->active)) != 1)) {
> 2046			bpf_prog_inc_misses_counter(prog);
  2047			goto out;
  2048		}
  2049		rcu_read_lock();
  2050		(void) bpf_prog_run(prog, args);
  2051		rcu_read_unlock();
  2052	out:
  2053		this_cpu_dec(*(prog->active));
  2054	}
  2055	

-- 
0-DAY CI Kernel Test Service
https://01.org/lkp

Re: [PATCH bpf-next] bpf: Prevent bpf program recursion for raw tracepoint probes

[kernel test robot]

Hi Jiri,

I love your patch! Yet something to improve:

[auto build test ERROR on bpf-next/master]

url:    https://github.com/intel-lab-lkp/linux/commits/Jiri-Olsa/bpf-Prevent-bpf-program-recursion-for-raw-tracepoint-probes/20220908-194832
base:   https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git master
config: x86_64-randconfig-c022 (https://download.01.org/0day-ci/archive/20220909/202209091544.TU8KWEUM-lkp@intel.com/config)
compiler: gcc-11 (Debian 11.3.0-5) 11.3.0
reproduce (this is a W=1 build):
        # https://github.com/intel-lab-lkp/linux/commit/f68b567cfb6572c20e431242a440cc5f01452485
        git remote add linux-review https://github.com/intel-lab-lkp/linux
        git fetch --no-tags linux-review Jiri-Olsa/bpf-Prevent-bpf-program-recursion-for-raw-tracepoint-probes/20220908-194832
        git checkout f68b567cfb6572c20e431242a440cc5f01452485
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash

If you fix the issue, kindly add following tag where applicable
Reported-by: kernel test robot <lkp@intel.com>

All errors (new ones prefixed by >>):

   ld: kernel/trace/bpf_trace.o: in function `__bpf_trace_run':
   kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
>> ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
>> ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
>> ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
>> ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
   ld: kernel/trace/bpf_trace.o:kernel/trace/bpf_trace.c:2046: more undefined references to `bpf_prog_inc_misses_counter' follow

-- 
0-DAY CI Kernel Test Service
https://01.org/lkp

Re: [PATCH bpf-next] bpf: Prevent bpf program recursion for raw tracepoint probes

[Jiri Olsa]

On Fri, Sep 09, 2022 at 03:27:57PM +0800, kernel test robot wrote:
> Hi Jiri,
> 
> I love your patch! Yet something to improve:
> 
> [auto build test ERROR on bpf-next/master]
> 
> url:    https://github.com/intel-lab-lkp/linux/commits/Jiri-Olsa/bpf-Prevent-bpf-program-recursion-for-raw-tracepoint-probes/20220908-194832
> base:   https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git master
> config: x86_64-randconfig-c022 (https://download.01.org/0day-ci/archive/20220909/202209091544.TU8KWEUM-lkp@intel.com/config)
> compiler: gcc-11 (Debian 11.3.0-5) 11.3.0
> reproduce (this is a W=1 build):
>         # https://github.com/intel-lab-lkp/linux/commit/f68b567cfb6572c20e431242a440cc5f01452485
>         git remote add linux-review https://github.com/intel-lab-lkp/linux
>         git fetch --no-tags linux-review Jiri-Olsa/bpf-Prevent-bpf-program-recursion-for-raw-tracepoint-probes/20220908-194832
>         git checkout f68b567cfb6572c20e431242a440cc5f01452485
>         # save the config file
>         mkdir build_dir && cp config build_dir/.config
>         make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash
> 
> If you fix the issue, kindly add following tag where applicable
> Reported-by: kernel test robot <lkp@intel.com>
> 
> All errors (new ones prefixed by >>):
> 
>    ld: kernel/trace/bpf_trace.o: in function `__bpf_trace_run':
>    kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
> >> ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
> >> ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
> >> ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
> >> ld: kernel/trace/bpf_trace.c:2046: undefined reference to `bpf_prog_inc_misses_counter'
>    ld: kernel/trace/bpf_trace.o:kernel/trace/bpf_trace.c:2046: more undefined references to `bpf_prog_inc_misses_counter' follow

ah right, trampoline.o is for JIT config only, will move
bpf_prog_inc_misses_counter to some common place

jirka

> 
> -- 
> 0-DAY CI Kernel Test Service
> https://01.org/lkp