All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH bpf-next v6] selftests/bpf: trace_helpers.c: optimize kallsyms cache
@ 2023-08-21  6:45 Rong Tao
  2023-08-21  8:39 ` Jiri Olsa
  0 siblings, 1 reply; 5+ messages in thread
From: Rong Tao @ 2023-08-21  6:45 UTC (permalink / raw)
  To: andrii, daniel, sdf, yonghong.song
  Cc: Rong Tao, Alexei Starovoitov, Martin KaFai Lau, Song Liu,
	John Fastabend, KP Singh, Hao Luo, Jiri Olsa, Mykola Lysenko,
	Shuah Khan,
	open list:BPF [GENERAL] (Safe Dynamic Programs and Tools),
	open list, open list:KERNEL SELFTEST FRAMEWORK

From: Rong Tao <rongtao@cestc.cn>

Static ksyms often have problems because the number of symbols exceeds the
MAX_SYMS limit. Like changing the MAX_SYMS from 300000 to 400000 in
commit e76a014334a6("selftests/bpf: Bump and validate MAX_SYMS") solves
the problem somewhat, but it's not the perfect way.

This commit uses dynamic memory allocation, which completely solves the
problem caused by the limitation of the number of kallsyms.

Acked-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Rong Tao <rongtao@cestc.cn>
---
v6: Apply libbpf_ensure_mem()
v5: https://lore.kernel.org/lkml/tencent_0E9E1A1C0981678D5E7EA9E4BDBA8EE2200A@qq.com/
    Release the allocated memory once the load_kallsyms_refresh() upon error
    given it's dynamically allocated.
v4: https://lore.kernel.org/lkml/tencent_59C74613113F0C728524B2A82FE5540A5E09@qq.com/
    Make sure most cases we don't need the realloc() path to begin with,
    and check strdup() return value.
v3: https://lore.kernel.org/lkml/tencent_50B4B2622FE7546A5FF9464310650C008509@qq.com/
    Do not use structs and judge ksyms__add_symbol function return value.
v2: https://lore.kernel.org/lkml/tencent_B655EE5E5D463110D70CD2846AB3262EED09@qq.com/
    Do the usual len/capacity scheme here to amortize the cost of realloc, and
    don't free symbols.
v1: https://lore.kernel.org/lkml/tencent_AB461510B10CD484E0B2F62E3754165F2909@qq.com/
---
 samples/bpf/Makefile                        |  1 +
 tools/testing/selftests/bpf/trace_helpers.c | 59 +++++++++++++++++----
 2 files changed, 49 insertions(+), 11 deletions(-)

diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index 595b98d825ce..a49d0f759f5a 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -199,6 +199,7 @@ TPROGS_CFLAGS += -I$(srctree)/tools/testing/selftests/bpf/
 TPROGS_CFLAGS += -I$(LIBBPF_INCLUDE)
 TPROGS_CFLAGS += -I$(srctree)/tools/include
 TPROGS_CFLAGS += -I$(srctree)/tools/perf
+TPROGS_CFLAGS += -I$(srctree)/tools/lib
 TPROGS_CFLAGS += -DHAVE_ATTR_TEST=0
 
 ifdef SYSROOT
diff --git a/tools/testing/selftests/bpf/trace_helpers.c b/tools/testing/selftests/bpf/trace_helpers.c
index f83d9f65c65b..316a7874a12b 100644
--- a/tools/testing/selftests/bpf/trace_helpers.c
+++ b/tools/testing/selftests/bpf/trace_helpers.c
@@ -14,14 +14,48 @@
 #include <linux/limits.h>
 #include <libelf.h>
 #include <gelf.h>
+#ifndef __must_check
+#define __must_check
+#endif
+#include "bpf/libbpf_internal.h"
 
 #define TRACEFS_PIPE	"/sys/kernel/tracing/trace_pipe"
 #define DEBUGFS_PIPE	"/sys/kernel/debug/tracing/trace_pipe"
 
-#define MAX_SYMS 400000
-static struct ksym syms[MAX_SYMS];
+static struct ksym *syms;
+static size_t sym_cap;
 static int sym_cnt;
 
+static int ksyms__add_symbol(const char *name, unsigned long addr)
+{
+	void *tmp;
+
+	tmp = strdup(name);
+	if (!tmp)
+		return -ENOMEM;
+	syms[sym_cnt].addr = addr;
+	syms[sym_cnt].name = tmp;
+
+	sym_cnt++;
+
+	return 0;
+}
+
+static void ksyms__free(void)
+{
+	unsigned int i;
+
+	if (!syms)
+		return;
+
+	for (i = 0; i < sym_cnt; i++)
+		free(syms[i].name);
+	free(syms);
+	syms = NULL;
+	sym_cnt = 0;
+	sym_cap = 0;
+}
+
 static int ksym_cmp(const void *p1, const void *p2)
 {
 	return ((struct ksym *)p1)->addr - ((struct ksym *)p2)->addr;
@@ -33,9 +67,7 @@ int load_kallsyms_refresh(void)
 	char func[256], buf[256];
 	char symbol;
 	void *addr;
-	int i = 0;
-
-	sym_cnt = 0;
+	int ret;
 
 	f = fopen("/proc/kallsyms", "r");
 	if (!f)
@@ -46,17 +78,22 @@ int load_kallsyms_refresh(void)
 			break;
 		if (!addr)
 			continue;
-		if (i >= MAX_SYMS)
-			return -EFBIG;
 
-		syms[i].addr = (long) addr;
-		syms[i].name = strdup(func);
-		i++;
+		ret = libbpf_ensure_mem((void **) &syms, &sym_cap,
+					sizeof(struct ksym), sym_cnt + 1);
+		if (ret)
+			goto error;
+		ret = ksyms__add_symbol(func, (unsigned long)addr);
+		if (ret)
+			goto error;
 	}
 	fclose(f);
-	sym_cnt = i;
 	qsort(syms, sym_cnt, sizeof(struct ksym), ksym_cmp);
 	return 0;
+
+error:
+	ksyms__free();
+	return ret;
 }
 
 int load_kallsyms(void)
-- 
2.39.3


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH bpf-next v6] selftests/bpf: trace_helpers.c: optimize kallsyms cache
  2023-08-21  6:45 [PATCH bpf-next v6] selftests/bpf: trace_helpers.c: optimize kallsyms cache Rong Tao
@ 2023-08-21  8:39 ` Jiri Olsa
  2023-08-22  0:38   ` Rong Tao
  0 siblings, 1 reply; 5+ messages in thread
From: Jiri Olsa @ 2023-08-21  8:39 UTC (permalink / raw)
  To: Rong Tao
  Cc: andrii, daniel, sdf, yonghong.song, Rong Tao, Alexei Starovoitov,
	Martin KaFai Lau, Song Liu, John Fastabend, KP Singh, Hao Luo,
	Mykola Lysenko, Shuah Khan,
	open list:BPF [GENERAL] (Safe Dynamic Programs and Tools),
	open list, open list:KERNEL SELFTEST FRAMEWORK

On Mon, Aug 21, 2023 at 02:45:20PM +0800, Rong Tao wrote:
> From: Rong Tao <rongtao@cestc.cn>
> 
> Static ksyms often have problems because the number of symbols exceeds the
> MAX_SYMS limit. Like changing the MAX_SYMS from 300000 to 400000 in
> commit e76a014334a6("selftests/bpf: Bump and validate MAX_SYMS") solves
> the problem somewhat, but it's not the perfect way.
> 
> This commit uses dynamic memory allocation, which completely solves the
> problem caused by the limitation of the number of kallsyms.
> 
> Acked-by: Stanislav Fomichev <sdf@google.com>
> Signed-off-by: Rong Tao <rongtao@cestc.cn>
> ---
> v6: Apply libbpf_ensure_mem()
> v5: https://lore.kernel.org/lkml/tencent_0E9E1A1C0981678D5E7EA9E4BDBA8EE2200A@qq.com/
>     Release the allocated memory once the load_kallsyms_refresh() upon error
>     given it's dynamically allocated.
> v4: https://lore.kernel.org/lkml/tencent_59C74613113F0C728524B2A82FE5540A5E09@qq.com/
>     Make sure most cases we don't need the realloc() path to begin with,
>     and check strdup() return value.
> v3: https://lore.kernel.org/lkml/tencent_50B4B2622FE7546A5FF9464310650C008509@qq.com/
>     Do not use structs and judge ksyms__add_symbol function return value.
> v2: https://lore.kernel.org/lkml/tencent_B655EE5E5D463110D70CD2846AB3262EED09@qq.com/
>     Do the usual len/capacity scheme here to amortize the cost of realloc, and
>     don't free symbols.
> v1: https://lore.kernel.org/lkml/tencent_AB461510B10CD484E0B2F62E3754165F2909@qq.com/
> ---
>  samples/bpf/Makefile                        |  1 +
>  tools/testing/selftests/bpf/trace_helpers.c | 59 +++++++++++++++++----
>  2 files changed, 49 insertions(+), 11 deletions(-)
> 
> diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
> index 595b98d825ce..a49d0f759f5a 100644
> --- a/samples/bpf/Makefile
> +++ b/samples/bpf/Makefile
> @@ -199,6 +199,7 @@ TPROGS_CFLAGS += -I$(srctree)/tools/testing/selftests/bpf/
>  TPROGS_CFLAGS += -I$(LIBBPF_INCLUDE)
>  TPROGS_CFLAGS += -I$(srctree)/tools/include
>  TPROGS_CFLAGS += -I$(srctree)/tools/perf
> +TPROGS_CFLAGS += -I$(srctree)/tools/lib
>  TPROGS_CFLAGS += -DHAVE_ATTR_TEST=0
>  
>  ifdef SYSROOT
> diff --git a/tools/testing/selftests/bpf/trace_helpers.c b/tools/testing/selftests/bpf/trace_helpers.c
> index f83d9f65c65b..316a7874a12b 100644
> --- a/tools/testing/selftests/bpf/trace_helpers.c
> +++ b/tools/testing/selftests/bpf/trace_helpers.c
> @@ -14,14 +14,48 @@
>  #include <linux/limits.h>
>  #include <libelf.h>
>  #include <gelf.h>
> +#ifndef __must_check
> +#define __must_check
> +#endif

I think you need to fix this on samples/bpf side

I tried to play with the samples/bpf/ includes, but couldn't find a way to
make this work.. selftests base includes on tools/include, while samples
have $(objtree)/usr/include as first include and AFAICS the __must_check is
defined under __KERNEL__ ifdef

I guess the reason samples use $(objtree)/usr/include is to get some struct
definitions which are not in tools/include, but looks like some samples objects
already use vmlinux.h include, so that could be the way forward to fix that

or get rid of the trace_helpers.c or bpf/selftests dependency ;-)

> +#include "bpf/libbpf_internal.h"
>  
>  #define TRACEFS_PIPE	"/sys/kernel/tracing/trace_pipe"
>  #define DEBUGFS_PIPE	"/sys/kernel/debug/tracing/trace_pipe"
>  
> -#define MAX_SYMS 400000
> -static struct ksym syms[MAX_SYMS];
> +static struct ksym *syms;
> +static size_t sym_cap;
>  static int sym_cnt;

nit, sym_cnt should be size_t

jirka

>  
> +static int ksyms__add_symbol(const char *name, unsigned long addr)
> +{
> +	void *tmp;
> +
> +	tmp = strdup(name);
> +	if (!tmp)
> +		return -ENOMEM;
> +	syms[sym_cnt].addr = addr;
> +	syms[sym_cnt].name = tmp;
> +
> +	sym_cnt++;
> +
> +	return 0;
> +}
> +
> +static void ksyms__free(void)
> +{
> +	unsigned int i;
> +
> +	if (!syms)
> +		return;
> +
> +	for (i = 0; i < sym_cnt; i++)
> +		free(syms[i].name);
> +	free(syms);
> +	syms = NULL;
> +	sym_cnt = 0;
> +	sym_cap = 0;
> +}
> +
>  static int ksym_cmp(const void *p1, const void *p2)
>  {
>  	return ((struct ksym *)p1)->addr - ((struct ksym *)p2)->addr;
> @@ -33,9 +67,7 @@ int load_kallsyms_refresh(void)
>  	char func[256], buf[256];
>  	char symbol;
>  	void *addr;
> -	int i = 0;
> -
> -	sym_cnt = 0;
> +	int ret;
>  
>  	f = fopen("/proc/kallsyms", "r");
>  	if (!f)
> @@ -46,17 +78,22 @@ int load_kallsyms_refresh(void)
>  			break;
>  		if (!addr)
>  			continue;
> -		if (i >= MAX_SYMS)
> -			return -EFBIG;
>  
> -		syms[i].addr = (long) addr;
> -		syms[i].name = strdup(func);
> -		i++;
> +		ret = libbpf_ensure_mem((void **) &syms, &sym_cap,
> +					sizeof(struct ksym), sym_cnt + 1);
> +		if (ret)
> +			goto error;
> +		ret = ksyms__add_symbol(func, (unsigned long)addr);
> +		if (ret)
> +			goto error;
>  	}
>  	fclose(f);
> -	sym_cnt = i;
>  	qsort(syms, sym_cnt, sizeof(struct ksym), ksym_cmp);
>  	return 0;
> +
> +error:
> +	ksyms__free();
> +	return ret;
>  }
>  
>  int load_kallsyms(void)
> -- 
> 2.39.3
> 
> 

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH bpf-next v6] selftests/bpf: trace_helpers.c: optimize kallsyms cache
  2023-08-21  8:39 ` Jiri Olsa
@ 2023-08-22  0:38   ` Rong Tao
  2023-08-25  9:08     ` Jiri Olsa
  0 siblings, 1 reply; 5+ messages in thread
From: Rong Tao @ 2023-08-22  0:38 UTC (permalink / raw)
  To: olsajiri
  Cc: andrii, ast, bpf, daniel, haoluo, john.fastabend, kpsingh,
	linux-kernel, linux-kselftest, martin.lau, mykolal, rongtao,
	rtoax, sdf, shuah, song, yonghong.song

> I think you need to fix this on samples/bpf side
> 
> I tried to play with the samples/bpf/ includes, but couldn't find a way to
> make this work.. selftests base includes on tools/include, while samples
> have $(objtree)/usr/include as first include and AFAICS the __must_check is
> defined under __KERNEL__ ifdef
> 
> I guess the reason samples use $(objtree)/usr/include is to get some struct
> definitions which are not in tools/include, but looks like some samples objects
> already use vmlinux.h include, so that could be the way forward to fix that

I tried the method you suggested, and some unexpected problems occurred. Maybe,
we can apply v5 [0] first, and then solve this problem?

Best wishes,
Rong Tao

[0] v5: https://lore.kernel.org/lkml/tencent_0E9E1A1C0981678D5E7EA9E4BDBA8EE2200A@qq.com/


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH bpf-next v6] selftests/bpf: trace_helpers.c: optimize kallsyms cache
  2023-08-22  0:38   ` Rong Tao
@ 2023-08-25  9:08     ` Jiri Olsa
  2023-08-25 10:40       ` Rong Tao
  0 siblings, 1 reply; 5+ messages in thread
From: Jiri Olsa @ 2023-08-25  9:08 UTC (permalink / raw)
  To: Rong Tao
  Cc: olsajiri, andrii, ast, bpf, daniel, haoluo, john.fastabend,
	kpsingh, linux-kernel, linux-kselftest, martin.lau, mykolal,
	rongtao, sdf, shuah, song, yonghong.song

On Tue, Aug 22, 2023 at 08:38:58AM +0800, Rong Tao wrote:
> > I think you need to fix this on samples/bpf side
> > 
> > I tried to play with the samples/bpf/ includes, but couldn't find a way to
> > make this work.. selftests base includes on tools/include, while samples
> > have $(objtree)/usr/include as first include and AFAICS the __must_check is
> > defined under __KERNEL__ ifdef
> > 
> > I guess the reason samples use $(objtree)/usr/include is to get some struct
> > definitions which are not in tools/include, but looks like some samples objects
> > already use vmlinux.h include, so that could be the way forward to fix that
> 
> I tried the method you suggested, and some unexpected problems occurred. Maybe,
> we can apply v5 [0] first, and then solve this problem?

how about change below, seem to work for me

jirka


---
diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index 151ec320818b..3115f054dca7 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -316,6 +316,8 @@ XDP_SAMPLE_CFLAGS += -Wall -O2 \
 $(obj)/$(XDP_SAMPLE): TPROGS_CFLAGS = $(XDP_SAMPLE_CFLAGS)
 $(obj)/$(XDP_SAMPLE): $(src)/xdp_sample_user.h $(src)/xdp_sample_shared.h
 
+$(obj)/$(TRACE_HELPERS): TPROGS_CFLAGS := $(TPROGS_CFLAGS) -D__must_check=
+
 -include $(BPF_SAMPLES_PATH)/Makefile.target
 
 VMLINUX_BTF_PATHS ?= $(abspath $(if $(O),$(O)/vmlinux))				\
diff --git a/tools/testing/selftests/bpf/trace_helpers.c b/tools/testing/selftests/bpf/trace_helpers.c
index 316a7874a12b..551547ba6829 100644
--- a/tools/testing/selftests/bpf/trace_helpers.c
+++ b/tools/testing/selftests/bpf/trace_helpers.c
@@ -14,9 +14,6 @@
 #include <linux/limits.h>
 #include <libelf.h>
 #include <gelf.h>
-#ifndef __must_check
-#define __must_check
-#endif
 #include "bpf/libbpf_internal.h"
 
 #define TRACEFS_PIPE	"/sys/kernel/tracing/trace_pipe"

^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH bpf-next v6] selftests/bpf: trace_helpers.c: optimize kallsyms cache
  2023-08-25  9:08     ` Jiri Olsa
@ 2023-08-25 10:40       ` Rong Tao
  0 siblings, 0 replies; 5+ messages in thread
From: Rong Tao @ 2023-08-25 10:40 UTC (permalink / raw)
  To: olsajiri
  Cc: andrii, ast, bpf, daniel, haoluo, john.fastabend, kpsingh,
	linux-kernel, linux-kselftest, martin.lau, mykolal, rongtao,
	rtoax, sdf, shuah, song, yonghong.song

Thanks a lot, jirka. I just submit v7 [0], please review.

Rong Tao.

[0] https://lore.kernel.org/lkml/tencent_BD6E19C00BF565CD5C36A9A0BD828CFA210A@qq.com/


^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2023-08-25 10:47 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-08-21  6:45 [PATCH bpf-next v6] selftests/bpf: trace_helpers.c: optimize kallsyms cache Rong Tao
2023-08-21  8:39 ` Jiri Olsa
2023-08-22  0:38   ` Rong Tao
2023-08-25  9:08     ` Jiri Olsa
2023-08-25 10:40       ` Rong Tao

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.