Skip to contentSkip to conversation

libbpf build fixes

GIT PULL In review Arnaldo Carvalho de Melo 5 participants 7 messages Jul 19, 2019 – Jul 22, 2019

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

Hi Daniel,

Please consider pulling or applying from the patches, if someone
has any issues, please holler,

- Arnaldo

Arnaldo Carvalho de Melo (2):
libbpf: Fix endianness macro usage for some compilers
libbpf: Avoid designated initializers for unnamed union members

 tools/lib/bpf/btf.c    |  5 +++--
 tools/lib/bpf/libbpf.c | 19 ++++++++++---------
 2 files changed, 13 insertions(+), 11 deletions(-)
Signature · 22 lines
2.21.0

The following changes since commit 9b3d15e6b05e0b916be5fbd915f90300a403098b:

  bnxt_en: Fix VNIC accounting when enabling aRFS on 57500 chips. (2019-07-18 16:33:27 -0700)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git libbpf

for you to fetch changes up to 5443fc23eb80e7cd03f9111fb16fb9aaa76fedc9:

  libbpf: Avoid designated initializers for unnamed union members (2019-07-19 11:21:33 -0300)

----------------------------------------------------------------
Arnaldo Carvalho de Melo (2):
      libbpf: Fix endianness macro usage for some compilers
      libbpf: Avoid designated initializers for unnamed union members

 tools/lib/bpf/btf.c    |  5 +++--
 tools/lib/bpf/libbpf.c | 19 ++++++++++---------
 2 files changed, 13 insertions(+), 11 deletions(-)
#1 lore raw

PATCH 1/2 libbpf: Fix endianness macro usage for some compilers

From: Arnaldo Carvalho de Melo <[email protected]>

Using endian.h and its endianness macros makes this code build in a wider range of compilers, as some don't have those macros (__BYTE_ORDER__, __ORDER_LITTLE_ENDIAN__, __ORDER_BIG_ENDIAN__), so use instead endian.h's macros (__BYTE_ORDER, __LITTLE_ENDIAN, __BIG_ENDIAN) which makes this code even shorter :-)

Cc Adrian Hunter <[email protected]>
Cc Alexei Starovoitov <[email protected]>
Cc Daniel Borkmann <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Fixes 12ef5634a855 ("libbpf: simplify endianness check")
Fixes e6c64855fd7a ("libbpf: add btf__parse_elf API to load .BTF and .BTF.ext")
---
 tools/lib/bpf/btf.c    | 5 +++--
 tools/lib/bpf/libbpf.c | 5 +++--
 2 files changed, 6 insertions(+), 4 deletions(-)
tools/lib/bpf/btf.c +3 -2
index 467224feb43b..d821107f55f9 100644
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause)
 /* Copyright (c) 2018 Facebook */
 
+#include <endian.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -419,9 +420,9 @@ struct btf *btf__new(__u8 *data, __u32 size)
 
 static bool btf_check_endianness(const GElf_Ehdr *ehdr)
 {
-#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+#if __BYTE_ORDER == __LITTLE_ENDIAN
 	return ehdr->e_ident[EI_DATA] == ELFDATA2LSB;
-#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
+#elif __BYTE_ORDER == __BIG_ENDIAN
 	return ehdr->e_ident[EI_DATA] == ELFDATA2MSB;
 #else
 # error "Unrecognized __BYTE_ORDER__"
tools/lib/bpf/libbpf.c +3 -2
index 794dd5064ae8..b1dec5b1de54 100644
@@ -20,6 +20,7 @@
 #include <inttypes.h>
 #include <string.h>
 #include <unistd.h>
+#include <endian.h>
 #include <fcntl.h>
 #include <errno.h>
 #include <asm/unistd.h>
@@ -612,10 +613,10 @@ static int bpf_object__elf_init(struct bpf_object *obj)
 
 static int bpf_object__check_endianness(struct bpf_object *obj)
 {
-#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+#if __BYTE_ORDER == __LITTLE_ENDIAN
 	if (obj->efile.ehdr.e_ident[EI_DATA] == ELFDATA2LSB)
 		return 0;
-#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
+#elif __BYTE_ORDER == __BIG_ENDIAN
 	if (obj->efile.ehdr.e_ident[EI_DATA] == ELFDATA2MSB)
 		return 0;
 #else
Signature · 1 line
2.21.0
#4 lore raw

PATCH 2/2 libbpf: Avoid designated initializers for unnamed union members

In reply to Arnaldo Carvalho de Melo

From: Arnaldo Carvalho de Melo <[email protected]>

As it fails to build in some systems with:

libbpf.c: In function 'perf_buffer__new':
libbpf.c:4515: error: unknown field 'sample_period' specified in initializer libbpf.c:4516: error: unknown field 'wakeup_events' specified in initializer

Doing as:

    attr.sample_period = 1;

I.e. not as a designated initializer makes it build everywhere.

Cc Andrii Nakryiko <[email protected]>
Cc Adrian Hunter <[email protected]>
Cc Daniel Borkmann <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Fixes fb84b8224655 ("libbpf: add perf buffer API")
---
 tools/lib/bpf/libbpf.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
tools/lib/bpf/libbpf.c +7 -7
index b1dec5b1de54..aaca132def74 100644
@@ -4508,13 +4508,13 @@ struct perf_buffer *perf_buffer__new(int map_fd, size_t page_cnt,
 				     const struct perf_buffer_opts *opts)
 {
 	struct perf_buffer_params p = {};
-	struct perf_event_attr attr = {
-		.config = PERF_COUNT_SW_BPF_OUTPUT,
-		.type = PERF_TYPE_SOFTWARE,
-		.sample_type = PERF_SAMPLE_RAW,
-		.sample_period = 1,
-		.wakeup_events = 1,
-	};
+	struct perf_event_attr attr = { 0, };
+
+	attr.config = PERF_COUNT_SW_BPF_OUTPUT,
+	attr.type = PERF_TYPE_SOFTWARE;
+	attr.sample_type = PERF_SAMPLE_RAW;
+	attr.sample_period = 1;
+	attr.wakeup_events = 1;
 
 	p.attr = &attr;
 	p.sample_cb = opts ? opts->sample_cb : NULL;
Signature · 1 line
2.21.0
#2 lore raw

Re: [PATCH 1/2] libbpf: Fix endianness macro usage for some compilers

In reply to Arnaldo Carvalho de Melo
Quoted 8 lines · Arnaldo Carvalho de Melo

From: Arnaldo Carvalho de Melo <[email protected]>

Using endian.h and its endianness macros makes this code build in a wider range of compilers, as some don't have those macros (__BYTE_ORDER__, __ORDER_LITTLE_ENDIAN__, __ORDER_BIG_ENDIAN__), so use instead endian.h's macros (__BYTE_ORDER, __LITTLE_ENDIAN, __BIG_ENDIAN) which makes this code even shorter :-)

gcc 4.6.0 starts to support __BYTE_ORDER__, __ORDER_LITTLE_ENDIAN__, etc. I guess those platforms with failing compilation have gcc < 4.6.0. Agree that for libbpf, which will be used outside kernel bpf selftest should try to compile with lower versions of gcc.

Quoted 68 lines
Cc Adrian Hunter <[email protected]>
Cc Alexei Starovoitov <[email protected]>
Cc Daniel Borkmann <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Fixes 12ef5634a855 ("libbpf: simplify endianness check")
Fixes e6c64855fd7a ("libbpf: add btf__parse_elf API to load .BTF and .BTF.ext")
---
 tools/lib/bpf/btf.c    | 5 +++--
 tools/lib/bpf/libbpf.c | 5 +++--
 2 files changed, 6 insertions(+), 4 deletions(-)
tools/lib/bpf/btf.c +3 -2
index 467224feb43b..d821107f55f9 100644
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause)
 /* Copyright (c) 2018 Facebook */
 
+#include <endian.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -419,9 +420,9 @@ struct btf *btf__new(__u8 *data, __u32 size)
 
 static bool btf_check_endianness(const GElf_Ehdr *ehdr)
 {
-#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+#if __BYTE_ORDER == __LITTLE_ENDIAN
        return ehdr->e_ident[EI_DATA] == ELFDATA2LSB;
-#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
+#elif __BYTE_ORDER == __BIG_ENDIAN
        return ehdr->e_ident[EI_DATA] == ELFDATA2MSB;
 #else
 # error "Unrecognized __BYTE_ORDER__"
tools/lib/bpf/libbpf.c +3 -3
index 794dd5064ae8..b1dec5b1de54 100644
@@ -20,6 +20,7 @@
 #include <inttypes.h>
 #include <string.h>
 #include <unistd.h>
+#include <endian.h>
 #include <fcntl.h>
 #include <errno.h>
 #include <asm/unistd.h>
@@ -612,10 +613,10 @@ static int bpf_object__elf_init(struct bpf_object *obj)
 
 static int bpf_object__check_endianness(struct bpf_object *obj)
 {
-#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+#if __BYTE_ORDER == __LITTLE_ENDIAN
        if (obj->efile.ehdr.e_ident[EI_DATA] == ELFDATA2LSB)
                return 0;
-#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
+#elif __BYTE_ORDER == __BIG_ENDIAN
        if (obj->efile.ehdr.e_ident[EI_DATA] == ELFDATA2MSB)
                return 0;
 #else
--

2.21.0

Acked-by on v1 2/2
#5 lore raw

Re: [PATCH 2/2] libbpf: Avoid designated initializers for unnamed union members

In reply to Arnaldo Carvalho de Melo
Quoted 23 lines · Arnaldo Carvalho de Melo

From: Arnaldo Carvalho de Melo <[email protected]>

As it fails to build in some systems with:

libbpf.c: In function 'perf_buffer__new':
libbpf.c:4515: error: unknown field 'sample_period' specified in initializer libbpf.c:4516: error: unknown field 'wakeup_events' specified in initializer

Doing as:

    attr.sample_period = 1;

I.e. not as a designated initializer makes it build everywhere.

Cc Andrii Nakryiko <[email protected]>
Cc Adrian Hunter <[email protected]>
Cc Daniel Borkmann <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Fixes fb84b8224655 ("libbpf: add perf buffer API")
Quoted 1 line

---

Thanks, Arnaldo!

Quoted 31 lines
 tools/lib/bpf/libbpf.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
tools/lib/bpf/libbpf.c +7 -8
index b1dec5b1de54..aaca132def74 100644
@@ -4508,13 +4508,13 @@ struct perf_buffer *perf_buffer__new(int map_fd, size_t page_cnt,
                                     const struct perf_buffer_opts *opts)
 {
        struct perf_buffer_params p = {};
-       struct perf_event_attr attr = {
-               .config = PERF_COUNT_SW_BPF_OUTPUT,
-               .type = PERF_TYPE_SOFTWARE,
-               .sample_type = PERF_SAMPLE_RAW,
-               .sample_period = 1,
-               .wakeup_events = 1,
-       };
+       struct perf_event_attr attr = { 0, };
+
+       attr.config = PERF_COUNT_SW_BPF_OUTPUT,
+       attr.type = PERF_TYPE_SOFTWARE;
+       attr.sample_type = PERF_SAMPLE_RAW;
+       attr.sample_period = 1;
+       attr.wakeup_events = 1;
 
        p.attr = &attr;
        p.sample_cb = opts ? opts->sample_cb : NULL;
--

2.21.0

#3 lore raw

Re: [PATCH 1/2] libbpf: Fix endianness macro usage for some compilers

In reply to Y Song

Em Fri, Jul 19, 2019 at 09:25:07AM -0700, Y Song escreveu:

Quoted 11 lines · Arnaldo Carvalho de Melo
Quoted 5 lines

Using endian.h and its endianness macros makes this code build in a wider range of compilers, as some don't have those macros (__BYTE_ORDER__, __ORDER_LITTLE_ENDIAN__, __ORDER_BIG_ENDIAN__), so use instead endian.h's macros (__BYTE_ORDER, __LITTLE_ENDIAN, __BIG_ENDIAN) which makes this code even shorter :-)

gcc 4.6.0 starts to support __BYTE_ORDER__, __ORDER_LITTLE_ENDIAN__, etc. I guess those platforms with failing compilation have gcc < 4.6.0. Agree that for libbpf, which will be used outside kernel bpf selftest should try to compile with lower versions of gcc.

Yeah, I wouldn't mind at all if the answer to this patch was hey, the minimal version for building libbpf is X.Y, no problem, its just that in this case there is an alternative that works everywhere and the resulting source code is even more compact :-)

- Arnaldo

Quoted 67 lines
Quoted 67 lines
Cc Adrian Hunter <[email protected]>
Cc Alexei Starovoitov <[email protected]>
Cc Daniel Borkmann <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Fixes 12ef5634a855 ("libbpf: simplify endianness check")
Fixes e6c64855fd7a ("libbpf: add btf__parse_elf API to load .BTF and .BTF.ext")
---
 tools/lib/bpf/btf.c    | 5 +++--
 tools/lib/bpf/libbpf.c | 5 +++--
 2 files changed, 6 insertions(+), 4 deletions(-)
tools/lib/bpf/btf.c +3 -2
index 467224feb43b..d821107f55f9 100644
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause)
 /* Copyright (c) 2018 Facebook */
 
+#include <endian.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -419,9 +420,9 @@ struct btf *btf__new(__u8 *data, __u32 size)
 
 static bool btf_check_endianness(const GElf_Ehdr *ehdr)
 {
-#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+#if __BYTE_ORDER == __LITTLE_ENDIAN
        return ehdr->e_ident[EI_DATA] == ELFDATA2LSB;
-#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
+#elif __BYTE_ORDER == __BIG_ENDIAN
        return ehdr->e_ident[EI_DATA] == ELFDATA2MSB;
 #else
 # error "Unrecognized __BYTE_ORDER__"
tools/lib/bpf/libbpf.c +3 -3
index 794dd5064ae8..b1dec5b1de54 100644
@@ -20,6 +20,7 @@
 #include <inttypes.h>
 #include <string.h>
 #include <unistd.h>
+#include <endian.h>
 #include <fcntl.h>
 #include <errno.h>
 #include <asm/unistd.h>
@@ -612,10 +613,10 @@ static int bpf_object__elf_init(struct bpf_object *obj)
 
 static int bpf_object__check_endianness(struct bpf_object *obj)
 {
-#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+#if __BYTE_ORDER == __LITTLE_ENDIAN
        if (obj->efile.ehdr.e_ident[EI_DATA] == ELFDATA2LSB)
                return 0;
-#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
+#elif __BYTE_ORDER == __BIG_ENDIAN
        if (obj->efile.ehdr.e_ident[EI_DATA] == ELFDATA2MSB)
                return 0;
 #else
--

2.21.0

Signature · 1 line
- Arnaldo
maintainer Applied
#6 lore raw
In reply to Arnaldo Carvalho de Melo
Quoted 7 lines · Arnaldo Carvalho de Melo

Hi Daniel,

Please consider pulling or applying from the patches, if someone
has any issues, please holler,

- Arnaldo

Quoted 8 lines

Arnaldo Carvalho de Melo (2):
libbpf: Fix endianness macro usage for some compilers
libbpf: Avoid designated initializers for unnamed union members

 tools/lib/bpf/btf.c    |  5 +++--
 tools/lib/bpf/libbpf.c | 19 ++++++++++---------
 2 files changed, 13 insertions(+), 11 deletions(-)

Applied, thanks!