Skip to contentSkip to conversation

libbpf: Fix is_pow_of_2

PATCH Applied Ian Rogers 4 participants 10 messages Jun 3, 2022 – Jun 17, 2022

Subsystems (4)
lore
Share & tools
lore

Ian Rogers

author
#0 lore raw

PATCH v2 1/1 libbpf: Fix is_pow_of_2

From: Yuze Chi <[email protected]>

Move the correct definition from linker.c into libbpf_internal.h.

---
 tools/lib/bpf/libbpf.c          | 5 -----
 tools/lib/bpf/libbpf_internal.h | 5 +++++
 tools/lib/bpf/linker.c          | 5 -----
 3 files changed, 5 insertions(+), 10 deletions(-)
tools/lib/bpf/libbpf.c +0 -5
index 3f4f18684bd3..346f941bb995 100644
@@ -4954,11 +4954,6 @@ bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 
 static void bpf_map__destroy(struct bpf_map *map);
 
-static bool is_pow_of_2(size_t x)
-{
-	return x && (x & (x - 1));
-}
-
 static size_t adjust_ringbuf_sz(size_t sz)
 {
 	__u32 page_sz = sysconf(_SC_PAGE_SIZE);
tools/lib/bpf/libbpf_internal.h +5 -0
index 4abdbe2fea9d..ef5d975078e5 100644
@@ -580,4 +580,9 @@ struct bpf_link * usdt_manager_attach_usdt(struct usdt_manager *man,
 					   const char *usdt_provider, const char *usdt_name,
 					   __u64 usdt_cookie);
 
+static inline bool is_pow_of_2(size_t x)
+{
+	return x && (x & (x - 1)) == 0;
+}
+
 #endif /* __LIBBPF_LIBBPF_INTERNAL_H */
tools/lib/bpf/linker.c +0 -5
index 9aa016fb55aa..85c0fddf55d1 100644
@@ -697,11 +697,6 @@ static int linker_load_obj_file(struct bpf_linker *linker, const char *filename,
 	return err;
 }
 
-static bool is_pow_of_2(size_t x)
-{
-	return x && (x & (x - 1)) == 0;
-}
-
 static int linker_sanity_check_elf(struct src_obj *obj)
 {
 	struct src_sec *sec;
Signature · 1 line
2.36.1.255.ge46751e96f-goog
author
#1 lore raw
Quoted 5 lines · Ian Rogers

From: Yuze Chi <[email protected]>

Move the correct definition from linker.c into libbpf_internal.h.

Sorry I missed this:

Fixes 0087a681fa8c ("libbpf: Automatically fix up

BPF_MAP_TYPE_RINGBUF size, if necessary")

Thanks,
Ian

Quoted 58 lines
---
 tools/lib/bpf/libbpf.c          | 5 -----
 tools/lib/bpf/libbpf_internal.h | 5 +++++
 tools/lib/bpf/linker.c          | 5 -----
 3 files changed, 5 insertions(+), 10 deletions(-)
tools/lib/bpf/libbpf.c +0 -5
index 3f4f18684bd3..346f941bb995 100644
@@ -4954,11 +4954,6 @@ bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 
 static void bpf_map__destroy(struct bpf_map *map);
 
-static bool is_pow_of_2(size_t x)
-{
-       return x && (x & (x - 1));
-}
-
 static size_t adjust_ringbuf_sz(size_t sz)
 {
        __u32 page_sz = sysconf(_SC_PAGE_SIZE);
tools/lib/bpf/libbpf_internal.h +5 -0
index 4abdbe2fea9d..ef5d975078e5 100644
@@ -580,4 +580,9 @@ struct bpf_link * usdt_manager_attach_usdt(struct usdt_manager *man,
                                           const char *usdt_provider, const char *usdt_name,
                                           __u64 usdt_cookie);
 
+static inline bool is_pow_of_2(size_t x)
+{
+       return x && (x & (x - 1)) == 0;
+}
+
 #endif /* __LIBBPF_LIBBPF_INTERNAL_H */
tools/lib/bpf/linker.c +0 -6
index 9aa016fb55aa..85c0fddf55d1 100644
@@ -697,11 +697,6 @@ static int linker_load_obj_file(struct bpf_linker *linker, const char *filename,
        return err;
 }
 
-static bool is_pow_of_2(size_t x)
-{
-       return x && (x & (x - 1)) == 0;
-}
-
 static int linker_sanity_check_elf(struct src_obj *obj)
 {
        struct src_sec *sec;
--

2.36.1.255.ge46751e96f-goog

#2 lore raw
Quoted 2 lines · Ian Rogers
Quoted 1 line · Ian Rogers

From: Yuze Chi <[email protected]>

[]

Quoted 1 line

tools/lib/bpf/libbpf_internal.h

Quoted 1 line
tools/lib/bpf/libbpf_internal.h +0 -0

[]

Quoted 10 lines

tools/lib/bpf/libbpf_internal.h struct bpf_link * usdt_manager_attach_usdt(struct usdt_manager *man, · line 588

Quoted 10 lines
tools/lib/bpf/libbpf_internal.h +5 -0
@@ -580,4 +580,9 @@ struct bpf_link * usdt_manager_attach_usdt(struct usdt_manager *man,
                                           const char *usdt_provider, const char *usdt_name,
                                           __u64 usdt_cookie);
 
+static inline bool is_pow_of_2(size_t x)
+{
+       return x && (x & (x - 1)) == 0;
+}
+
 #endif /* __LIBBPF_LIBBPF_INTERNAL_H */

If speed of execution is a potential issue, maybe:

#if __has_builtin(__builtin_popcount)
	return __builtin_popcount(x) == 1;
#else
	return x && (x & (x-1)) == 0;

#endif

#3 lore raw
Quoted 11 lines · Joe Perches
Quoted 2 lines · Ian Rogers
Quoted 1 line · Ian Rogers

From: Yuze Chi <[email protected]>

[]

Quoted 1 line

tools/lib/bpf/libbpf_internal.h

Quoted 1 line
tools/lib/bpf/libbpf_internal.h +0 -0

[]

Quoted 4 lines
Quoted 4 lines
tools/lib/bpf/libbpf_internal.h +0 -0
@@ -580,4 +580,9 @@ struct bpf_link * usdt_manager_attach_usdt(struct usdt_manager *man,
                                           const char *usdt_provider, const char *usdt_name,
                                           __u64 usdt_cookie);
 
Quoted 8 lines
Quoted 6 lines

tools/lib/bpf/libbpf_internal.h struct bpf_link * usdt_manager_attach_usdt(struct usdt_manager *man, · line 588

Quoted 6 lines
tools/lib/bpf/libbpf_internal.h +5 -0
+static inline bool is_pow_of_2(size_t x)
+{
+       return x && (x & (x - 1)) == 0;
+}
+
 #endif /* __LIBBPF_LIBBPF_INTERNAL_H */

If speed of execution is a potential issue, maybe:

It's not, as it's not in any sort of high-frequency hot path. But even if it was, we should be careful with __builtin_popcount() because depending on target CPU architecture __builtin_popcount() can be turned into a helper function call instead of using hardware instruction. But either way, keeping it simple is prefereable.

Quoted 7 lines
#if __has_builtin(__builtin_popcount)
        return __builtin_popcount(x) == 1;
#else
        return x && (x & (x-1)) == 0;

#endif

#4 lore raw
In reply to Ian Rogers
This patch was applied to bpf/bpf-next.git (master) by Andrii Nakryiko <[email protected]>:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Andrii Nakryiko <[email protected]>:

Quoted 9 lines · On Thu, 2 Jun 2022 22:51:56 -0700 you wrote:

From: Yuze Chi <[email protected]>

Move the correct definition from linker.c into libbpf_internal.h.

[...]

Here is the summary with links:
- [v2] libbpf: Fix is_pow_of_2

    https://git.kernel.org/bpf/bpf-next/c/f913ad6559e3

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
author
#5 lore raw
Quoted 15 lines · On Fri, Jun 3, 2022 at 11:30 AM <[email protected]> wrote:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Andrii Nakryiko <[email protected]>:

Quoted 8 lines · On Thu, 2 Jun 2022 22:51:56 -0700 you wrote:

From: Yuze Chi <[email protected]>

Move the correct definition from linker.c into libbpf_internal.h.

Quoted 7 lines
Quoted 1 line

[...]

Here is the summary with links:
- [v2] libbpf: Fix is_pow_of_2

    https://git.kernel.org/bpf/bpf-next/c/f913ad6559e3

You are awesome, thank you!

Will this patch get added to 5.19?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/lib/bpf/libbpf.c#n4948

Thanks,
Ian

Quoted 5 lines

--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

#6 lore raw
Quoted 19 lines · Ian Rogers
Quoted 17 lines · On Fri, Jun 3, 2022 at 11:30 AM <[email protected]> wrote:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Andrii Nakryiko <[email protected]>:

Quoted 9 lines · On Thu, 2 Jun 2022 22:51:56 -0700 you wrote:

From: Yuze Chi <[email protected]>

Move the correct definition from linker.c into libbpf_internal.h.

[...]

Quoted 9 lines
Quoted 5 lines

Here is the summary with links:
- [v2] libbpf: Fix is_pow_of_2

    https://git.kernel.org/bpf/bpf-next/c/f913ad6559e3

You are awesome, thank you!

Will this patch get added to 5.19?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/lib/bpf/libbpf.c#n4948

I've applied it to bpf-next, so as it stands right now - no. Do you need this for perf?

Quoted 8 lines

Thanks,
Ian

Quoted 5 lines

--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

author
#7 lore raw
Quoted 25 lines · Andrii Nakryiko
Quoted 23 lines · Ian Rogers
Quoted 21 lines · On Fri, Jun 3, 2022 at 11:30 AM <[email protected]> wrote:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Andrii Nakryiko <[email protected]>:

Quoted 9 lines · On Thu, 2 Jun 2022 22:51:56 -0700 you wrote:

From: Yuze Chi <[email protected]>

Move the correct definition from linker.c into libbpf_internal.h.

[...]

Here is the summary with links:
- [v2] libbpf: Fix is_pow_of_2

    https://git.kernel.org/bpf/bpf-next/c/f913ad6559e3
Quoted 8 lines
Quoted 5 lines
Quoted 1 line

You are awesome, thank you!

Will this patch get added to 5.19?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/lib/bpf/libbpf.c#n4948

I've applied it to bpf-next, so as it stands right now - no. Do you need this for perf?

Nope. We carry it as a patch against 5.19 in Google and was surprised to see I didn't need to drop the patch. Our internal code had encountered the bug, hence needing the fix. I'd expect others could encounter it, but I'm unaware of an issue with it and perf.

Thanks,
Ian

Quoted 8 lines
Quoted 8 lines

Thanks,
Ian

Quoted 5 lines

--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

#8 lore raw
Quoted 34 lines · Ian Rogers
Quoted 31 lines · Andrii Nakryiko
Quoted 28 lines · Ian Rogers
Quoted 15 lines · On Fri, Jun 3, 2022 at 11:30 AM <[email protected]> wrote:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Andrii Nakryiko <[email protected]>:

Quoted 8 lines · On Thu, 2 Jun 2022 22:51:56 -0700 you wrote:

From: Yuze Chi <[email protected]>

Move the correct definition from linker.c into libbpf_internal.h.

Quoted 7 lines
Quoted 1 line

[...]

Here is the summary with links:
- [v2] libbpf: Fix is_pow_of_2

    https://git.kernel.org/bpf/bpf-next/c/f913ad6559e3

You are awesome, thank you!

Will this patch get added to 5.19?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/lib/bpf/libbpf.c#n4948

Quoted 8 lines
Quoted 2 lines

I've applied it to bpf-next, so as it stands right now - no. Do you need this for perf?

Nope. We carry it as a patch against 5.19 in Google and was surprised to see I didn't need to drop the patch. Our internal code had encountered the bug, hence needing the fix. I'd expect others could encounter it, but I'm unaware of an issue with it and perf.

So the fix is in Github mirror ([0]) and it is expected that everyone is using libbpf based on Github repo, so not sure why you'd care about this fix in bpf tree. I somehow assumed that you need it due to perf, but was a bit surprised that perf is affected because I don't think it's using BPF ringbuf.

So I guess the question I have is why you don't use libbpf from [0] and what can be done to switch to the official libbpf repo?

[0] https://github.com/libbpf/libbpf

Quoted 11 lines

Thanks,
Ian

Quoted 8 lines
Quoted 8 lines

Thanks,
Ian

Quoted 5 lines

--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

author
#9 lore raw
Quoted 51 lines · Andrii Nakryiko
Quoted 34 lines · Ian Rogers
Quoted 31 lines · Andrii Nakryiko
Quoted 28 lines · Ian Rogers
Quoted 15 lines · On Fri, Jun 3, 2022 at 11:30 AM <[email protected]> wrote:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Andrii Nakryiko <[email protected]>:

Quoted 8 lines · On Thu, 2 Jun 2022 22:51:56 -0700 you wrote:

From: Yuze Chi <[email protected]>

Move the correct definition from linker.c into libbpf_internal.h.

Quoted 7 lines
Quoted 1 line

[...]

Here is the summary with links:
- [v2] libbpf: Fix is_pow_of_2

    https://git.kernel.org/bpf/bpf-next/c/f913ad6559e3

You are awesome, thank you!

Will this patch get added to 5.19?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/lib/bpf/libbpf.c#n4948

Quoted 8 lines
Quoted 2 lines

I've applied it to bpf-next, so as it stands right now - no. Do you need this for perf?

Nope. We carry it as a patch against 5.19 in Google and was surprised to see I didn't need to drop the patch. Our internal code had encountered the bug, hence needing the fix. I'd expect others could encounter it, but I'm unaware of an issue with it and perf.

So the fix is in Github mirror ([0]) and it is expected that everyone is using libbpf based on Github repo, so not sure why you'd care about this fix in bpf tree. I somehow assumed that you need it due to perf, but was a bit surprised that perf is affected because I don't think it's using BPF ringbuf.

Quoted 4 lines

So I guess the question I have is why you don't use libbpf from [0] and what can be done to switch to the official libbpf repo?

[0] https://github.com/libbpf/libbpf

Agreed on not following Linus' tree for libbpf. Not sure about having such an obvious bug in Linus' tree.

Thanks,
Ian

Quoted 11 lines
Quoted 11 lines

Thanks,
Ian

Quoted 8 lines
Quoted 8 lines

Thanks,
Ian

Quoted 5 lines

--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html