Skip to contentSkip to conversation

perf: Avoid implicit function declarations in lexer/parse interface

PATCH Applied Ian Rogers 3 participants 8 messages Apr 25, 2023 – May 3, 2023 1 message not in the archive

lore
Share & tools
lore

Ian Rogers

author Acked-by
#1 lore raw
In reply to a message not in the archive
Quoted 8 lines · Florian Weimer

In future compilers, -Wno-implicit-function-declaration may not bring back support for implicit function declarations, a feature that was removed from the C language in C99. Instead of relying on implicit declarations, include the flex-generated header from the bison-generated C code.

he expr-flex.h header needs to be included later than the others

nit: s/he/The/

Quoted 4 lines

because at the early point, the definition of YYSTYPE is not yet available.

Thanks for fighting the build wrt parallel dependencies! Ian

Quoted 74 lines

---
v2: Include the flex-generated files instead of manually-written prototypes.

 tools/perf/util/Build          | 10 +++++++++-
 tools/perf/util/expr.y         |  2 ++
 tools/perf/util/parse-events.y |  1 +
 tools/perf/util/pmu.y          |  1 +
 4 files changed, 13 insertions(+), 1 deletion(-)
tools/perf/util/Build +9 -1
index 918b501f9bd8..92897068c362 100644
@@ -283,7 +283,7 @@ CFLAGS_expr-flex.o          += $(flex_flags)
 bison_flags := -DYYENABLE_NLS=0
 BISON_GE_35 := $(shell expr $(shell $(BISON) --version | grep bison | sed -e 's/.\+ \([0-9]\+\).\([0-9]\+\)/\1\2/g') \>\= 35)
 ifeq ($(BISON_GE_35),1)
-  bison_flags += -Wno-unused-parameter -Wno-nested-externs -Wno-implicit-function-declaration -Wno-switch-enum -Wno-unused-but-set-variable -Wno-unknown-warning-option
+  bison_flags += -Wno-unused-parameter -Wno-nested-externs -Wno-switch-enum -Wno-unused-but-set-variable -Wno-unknown-warning-option
 else
   bison_flags += -w
 endif
@@ -340,3 +340,11 @@ $(OUTPUT)util/vsprintf.o: ../lib/vsprintf.c FORCE
 $(OUTPUT)util/list_sort.o: ../lib/list_sort.c FORCE
        $(call rule_mkdir)
        $(call if_changed_dep,cc_o_c)
+
+# These dependencies ensure that the flex-generated .h file is
+# available at the time the bison-generated .c sources are compiled.
+# Do not depend on the generated .h file to prevent triggering
+# parallel flex invocations for the same two output files.
+$(OUTPUT)util/expr-bison.o : $(OUTPUT)util/expr-flex.c
+$(OUTPUT)util/parse-events-bison.o : $(OUTPUT)util/parse-events-flex.c
+$(OUTPUT)util/pmu-bison.o : $(OUTPUT)util/pmu-flex.c
tools/perf/util/expr.y +2 -0
index 635e562350c5..99581193ca4c 100644
@@ -53,6 +53,8 @@
 %destructor { ids__free($$.ids); } <ids>
 
 %{
+#include "expr-flex.h"
+
 static void expr_error(double *final_val __maybe_unused,
                       struct expr_parse_ctx *ctx __maybe_unused,
                       bool compute_ids __maybe_unused,
tools/perf/util/parse-events.y +1 -0
index be8c51770051..67a7f70c4767 100644
@@ -17,6 +17,7 @@
 #include "evsel.h"
 #include "parse-events.h"
 #include "parse-events-bison.h"
+#include "parse-events-flex.h"
 
 void parse_events_error(YYLTYPE *loc, void *parse_state, void *scanner, char const *msg);
 
tools/perf/util/pmu.y +1 -0
index e675d79a0274..2170f1ac7b74 100644
@@ -9,6 +9,7 @@
 #include <linux/bitmap.h>
 #include <string.h>
 #include "pmu.h"
+#include "pmu-flex.h"
 
 #define ABORT_ON(val) \
 do { \

base-commit: 173ea743bf7a9eef04460e03b00ba267cc52aee2

#2 lore raw

Em Tue, Apr 25, 2023 at 10:40:14AM -0700, Ian Rogers escreveu:

Quoted 15 lines · Florian Weimer
Quoted 8 lines

In future compilers, -Wno-implicit-function-declaration may not bring back support for implicit function declarations, a feature that was removed from the C language in C99. Instead of relying on implicit declarations, include the flex-generated header from the bison-generated C code.

he expr-flex.h header needs to be included later than the others

nit: s/he/The/

Quoted 3 lines

because at the early point, the definition of YYSTYPE is not yet available.

Quoted 6 lines
Quoted 1 line

Thanks for fighting the build wrt parallel dependencies! Ian

Thanks, applied. BTW b4 coulnd't find this message (nor the original):

⬢[acme@toolbox perf-tools-next]$ b4 am -ctsl --cc-trailers [email protected] Grabbing thread from lore.kernel.org/all/87sfcn7uot.fsf%40oldenburg.str.redhat.com/t.mbox.gz
That message-id is not known.
⬢[acme@toolbox perf-tools-next]$ b4 am -ctsl --cc-trailers CAP-5=fXZv+KCdCN05wVUcAwDCZAgXjWunoaviGBQEiUPqNwOmg@mail.gmail.com Grabbing thread from lore.kernel.org/all/CAP-5%3DfXZv%2BKCdCN05wVUcAwDCZAgXjWunoaviGBQEiUPqNwOmg%40mail.gmail.com/t.mbox.gz Analyzing 1 messages in the thread
No patches found.
⬢[acme@toolbox perf-tools-next]$

I applied it in the old fashion.

- Arnaldo

Quoted 74 lines
Quoted 74 lines

---
v2: Include the flex-generated files instead of manually-written prototypes.

 tools/perf/util/Build          | 10 +++++++++-
 tools/perf/util/expr.y         |  2 ++
 tools/perf/util/parse-events.y |  1 +
 tools/perf/util/pmu.y          |  1 +
 4 files changed, 13 insertions(+), 1 deletion(-)
tools/perf/util/Build +9 -1
index 918b501f9bd8..92897068c362 100644
@@ -283,7 +283,7 @@ CFLAGS_expr-flex.o          += $(flex_flags)
 bison_flags := -DYYENABLE_NLS=0
 BISON_GE_35 := $(shell expr $(shell $(BISON) --version | grep bison | sed -e 's/.\+ \([0-9]\+\).\([0-9]\+\)/\1\2/g') \>\= 35)
 ifeq ($(BISON_GE_35),1)
-  bison_flags += -Wno-unused-parameter -Wno-nested-externs -Wno-implicit-function-declaration -Wno-switch-enum -Wno-unused-but-set-variable -Wno-unknown-warning-option
+  bison_flags += -Wno-unused-parameter -Wno-nested-externs -Wno-switch-enum -Wno-unused-but-set-variable -Wno-unknown-warning-option
 else
   bison_flags += -w
 endif
@@ -340,3 +340,11 @@ $(OUTPUT)util/vsprintf.o: ../lib/vsprintf.c FORCE
 $(OUTPUT)util/list_sort.o: ../lib/list_sort.c FORCE
        $(call rule_mkdir)
        $(call if_changed_dep,cc_o_c)
+
+# These dependencies ensure that the flex-generated .h file is
+# available at the time the bison-generated .c sources are compiled.
+# Do not depend on the generated .h file to prevent triggering
+# parallel flex invocations for the same two output files.
+$(OUTPUT)util/expr-bison.o : $(OUTPUT)util/expr-flex.c
+$(OUTPUT)util/parse-events-bison.o : $(OUTPUT)util/parse-events-flex.c
+$(OUTPUT)util/pmu-bison.o : $(OUTPUT)util/pmu-flex.c
tools/perf/util/expr.y +2 -0
index 635e562350c5..99581193ca4c 100644
@@ -53,6 +53,8 @@
 %destructor { ids__free($$.ids); } <ids>
 
 %{
+#include "expr-flex.h"
+
 static void expr_error(double *final_val __maybe_unused,
                       struct expr_parse_ctx *ctx __maybe_unused,
                       bool compute_ids __maybe_unused,
tools/perf/util/parse-events.y +1 -0
index be8c51770051..67a7f70c4767 100644
@@ -17,6 +17,7 @@
 #include "evsel.h"
 #include "parse-events.h"
 #include "parse-events-bison.h"
+#include "parse-events-flex.h"
 
 void parse_events_error(YYLTYPE *loc, void *parse_state, void *scanner, char const *msg);
 
tools/perf/util/pmu.y +1 -0
index e675d79a0274..2170f1ac7b74 100644
@@ -9,6 +9,7 @@
 #include <linux/bitmap.h>
 #include <string.h>
 #include "pmu.h"
+#include "pmu-flex.h"
 
 #define ABORT_ON(val) \
 do { \

base-commit: 173ea743bf7a9eef04460e03b00ba267cc52aee2

Signature · 1 line
- Arnaldo
#3 lore raw

Em Fri, Apr 28, 2023 at 10:34:25PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 35 lines

Em Tue, Apr 25, 2023 at 10:40:14AM -0700, Ian Rogers escreveu:

Quoted 15 lines · Florian Weimer
Quoted 8 lines

In future compilers, -Wno-implicit-function-declaration may not bring back support for implicit function declarations, a feature that was removed from the C language in C99. Instead of relying on implicit declarations, include the flex-generated header from the bison-generated C code.

he expr-flex.h header needs to be included later than the others

nit: s/he/The/

Quoted 3 lines

because at the early point, the definition of YYSTYPE is not yet available.

Quoted 6 lines
Quoted 1 line

Thanks for fighting the build wrt parallel dependencies! Ian

Thanks, applied. BTW b4 coulnd't find this message (nor the original):

⬢[acme@toolbox perf-tools-next]$ b4 am -ctsl --cc-trailers [email protected] Grabbing thread from lore.kernel.org/all/87sfcn7uot.fsf%40oldenburg.str.redhat.com/t.mbox.gz
That message-id is not known.
⬢[acme@toolbox perf-tools-next]$ b4 am -ctsl --cc-trailers CAP-5=fXZv+KCdCN05wVUcAwDCZAgXjWunoaviGBQEiUPqNwOmg@mail.gmail.com Grabbing thread from lore.kernel.org/all/CAP-5%3DfXZv%2BKCdCN05wVUcAwDCZAgXjWunoaviGBQEiUPqNwOmg%40mail.gmail.com/t.mbox.gz Analyzing 1 messages in the thread
No patches found.
⬢[acme@toolbox perf-tools-next]$

Quoted 1 line

I applied it in the old fashion.

Not so fast, removed it for now:

CC /tmp/build/perf-tools-next/util/parse-events-bison.o
In file included from util/pmu.y:14:
/tmp/build/perf-tools-next/util/pmu-flex.h:496:1: error: unknown type name ‘YYSTYPE’
496 |

      | ^

/tmp/build/perf-tools-next/util/pmu-flex.h:498:19: error: unknown type name ‘YYSTYPE’
498 |
| ^
/tmp/build/perf-tools-next/util/pmu-flex.h:546:17: error: unknown type name ‘YYSTYPE’
546 | extern int yylex \
| ^~
util/pmu-bison.c: In function ‘perf_pmu_parse’:
/tmp/build/perf-tools-next/util/pmu-bison.c:69:25: error: implicit declaration of function ‘perf_pmu_lex’; did you mean ‘perf_pmu_free’? [-Werror=implicit-function-declaration]

   69 | #define yylex           perf_pmu_lex
      |                         ^~~~~~~~~~~~

util/pmu-bison.c:1007:16: note: in expansion of macro ‘yylex’
At top level:
cc1: note: unrecognized command-line option ‘-Wno-unknown-warning-option’ may have been intended to silence earlier diagnostics
cc1: all warnings being treated as errors
make[4]: *** [/var/home/acme/git/perf-tools-next/tools/build/Makefile.build:97: /tmp/build/perf-tools-next/util/pmu-bison.o] Error 1
make[4]: *** Waiting for unfinished jobs....
util/bpf-filter-bison.c: In function ‘perf_bpf_filter_parse’:
/tmp/build/perf-tools-next/util/bpf-filter-bison.c:69:25: error: implicit declaration of function ‘perf_bpf_filter_lex’; did you mean ‘perf_bpf_filter_parse’? [-Werror=implicit-function-declaration]

   69 | #define yylex           perf_bpf_filter_lex
      |                         ^~~~~~~~~~~~~~~~~~~

util/bpf-filter-bison.c:1283:16: note: in expansion of macro ‘yylex’
At top level:
cc1: note: unrecognized command-line option ‘-Wno-unknown-warning-option’ may have been intended to silence earlier diagnostics
cc1: all warnings being treated as errors
make[4]: *** [/var/home/acme/git/perf-tools-next/tools/build/Makefile.build:97: /tmp/build/perf-tools-next/util/bpf-filter-bison.o] Error 1 make[3]: *** [/var/home/acme/git/perf-tools-next/tools/build/Makefile.build:140: util] Error 2
make[2]: *** [Makefile.perf:676: /tmp/build/perf-tools-next/perf-in.o] Error 2
make[2]: *** Waiting for unfinished jobs....
CC /tmp/build/perf-tools-next/pmu-events/pmu-events.o
LD /tmp/build/perf-tools-next/pmu-events/pmu-events-in.o
make[1]: *** [Makefile.perf:236: sub-make] Error 2
make: *** [Makefile:113: install-bin] Error 2
make: Leaving directory '/var/home/acme/git/perf-tools-next/tools/perf'

Performance counter stats for 'make -k CORESIGHT=1 O=/tmp/build/perf-tools-next -C tools/perf install-bin':

       40330976311      cycles:u
       73953066660      instructions:u                   #    1.83  insn per cycle

       9.268030835 seconds time elapsed

       9.546339000 seconds user
       7.391844000 seconds sys

⬢[acme@toolbox perf-tools-next]$ cat /etc/redhat-release
Fedora release 37 (Thirty Seven)
⬢[acme@toolbox perf-tools-next]$ rpm -q bison flex bison-3.8.2-3.fc37.x86_64
flex-2.6.4-11.fc37.x86_64
⬢[acme@toolbox perf-tools-next]$ gcc --version
gcc (GCC) 12.2.1 20221121 (Red Hat 12.2.1-4)
Copyright (C) 2022 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

⬢[acme@toolbox perf-tools-next]$

I tried rm -rf the O= build dir, etc.

- Arnaldo

#4 lore raw

* Arnaldo Carvalho de Melo:

Quoted 1 line
Quoted 1 line

Thanks, applied. BTW b4 coulnd't find this message (nor the original):

Yes, vger drops the message after accepting it for some reason, probably something in the patch contents. I tried to resubmit from a completely separate account, no luck.

Quoted 15 lines

Not so fast, removed it for now:

CC /tmp/build/perf-tools-next/util/parse-events-bison.o
In file included from util/pmu.y:14:
/tmp/build/perf-tools-next/util/pmu-flex.h:496:1: error: unknown type name ‘YYSTYPE’
496 |

      | ^

/tmp/build/perf-tools-next/util/pmu-flex.h:498:19: error: unknown type name ‘YYSTYPE’
498 |
| ^
/tmp/build/perf-tools-next/util/pmu-flex.h:546:17: error: unknown type name ‘YYSTYPE’
546 | extern int yylex \
| ^~
util/pmu-bison.c: In function ‘perf_pmu_parse’:
/tmp/build/perf-tools-next/util/pmu-bison.c:69:25: error: implicit declaration of function ‘perf_pmu_lex’; did you mean ‘perf_pmu_free’? [-Werror=implicit-function-declaration]

Quoted 3 lines
   69 | #define yylex           perf_pmu_lex
      |                         ^~~~~~~~~~~~

util/pmu-bison.c:1007:16: note: in expansion of macro ‘yylex’

This appears to be related to some BPF filter stuff that is only in perf-next, not mainline.

Is this the right tree on which to base the patch?

<https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/>

Branch perf-tools-next?

Thanks,
Florian

#5 lore raw

* Florian Weimer:

Quoted 27 lines

* Arnaldo Carvalho de Melo:

Quoted 1 line
Quoted 1 line

Thanks, applied. BTW b4 coulnd't find this message (nor the original):

Yes, vger drops the message after accepting it for some reason, probably something in the patch contents. I tried to resubmit from a completely separate account, no luck.

Quoted 15 lines

Not so fast, removed it for now:

CC /tmp/build/perf-tools-next/util/parse-events-bison.o
In file included from util/pmu.y:14:
/tmp/build/perf-tools-next/util/pmu-flex.h:496:1: error: unknown type name ‘YYSTYPE’
496 |

      | ^

/tmp/build/perf-tools-next/util/pmu-flex.h:498:19: error: unknown type name ‘YYSTYPE’
498 |
| ^
/tmp/build/perf-tools-next/util/pmu-flex.h:546:17: error: unknown type name ‘YYSTYPE’
546 | extern int yylex \
| ^~
util/pmu-bison.c: In function ‘perf_pmu_parse’:
/tmp/build/perf-tools-next/util/pmu-bison.c:69:25: error: implicit declaration of function ‘perf_pmu_lex’; did you mean ‘perf_pmu_free’? [-Werror=implicit-function-declaration]

Quoted 3 lines
   69 | #define yylex           perf_pmu_lex
      |                         ^~~~~~~~~~~~

util/pmu-bison.c:1007:16: note: in expansion of macro ‘yylex’

Quoted 8 lines

This appears to be related to some BPF filter stuff that is only in perf-next, not mainline.

Is this the right tree on which to base the patch?

<https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/>

Branch perf-tools-next?

Sorry, it doesn't build even without my patch:

make[4]: *** No rule to make target 'zip.h', needed by '/tmp/build/libbpf/staticobjs/libbpf.o'. Stop.

Thanks,
Florian

#6 lore raw

Em Wed, May 03, 2023 at 11:40:06AM +0200, Florian Weimer escreveu:

Quoted 33 lines

* Florian Weimer:

Quoted 32 lines

* Arnaldo Carvalho de Melo:

Quoted 1 line
Quoted 1 line

Thanks, applied. BTW b4 coulnd't find this message (nor the original):

Yes, vger drops the message after accepting it for some reason, probably something in the patch contents. I tried to resubmit from a completely separate account, no luck.

Quoted 15 lines

Not so fast, removed it for now:

CC /tmp/build/perf-tools-next/util/parse-events-bison.o
In file included from util/pmu.y:14:
/tmp/build/perf-tools-next/util/pmu-flex.h:496:1: error: unknown type name ‘YYSTYPE’
496 |

      | ^

/tmp/build/perf-tools-next/util/pmu-flex.h:498:19: error: unknown type name ‘YYSTYPE’
498 |
| ^
/tmp/build/perf-tools-next/util/pmu-flex.h:546:17: error: unknown type name ‘YYSTYPE’
546 | extern int yylex \
| ^~
util/pmu-bison.c: In function ‘perf_pmu_parse’:
/tmp/build/perf-tools-next/util/pmu-bison.c:69:25: error: implicit declaration of function ‘perf_pmu_lex’; did you mean ‘perf_pmu_free’? [-Werror=implicit-function-declaration]

Quoted 3 lines
   69 | #define yylex           perf_pmu_lex
      |                         ^~~~~~~~~~~~

util/pmu-bison.c:1007:16: note: in expansion of macro ‘yylex’

This appears to be related to some BPF filter stuff that is only in perf-next, not mainline.

Is this the right tree on which to base the patch?

Quoted 7 lines
Quoted 3 lines

<https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/>

Branch perf-tools-next?

Sorry, it doesn't build even without my patch:

make[4]: *** No rule to make target 'zip.h', needed by '/tmp/build/libbpf/staticobjs/libbpf.o'. Stop.

I'm noticing this, but then I haven't merged with upstream yet, I bet this is something upstream after merging the bpf branch... Nope, I just did a test merge of what is in
git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git perf-tools with upstream and it builds without problems on:

[acme@quaco perf-tools]$ head -2 /etc/os-release
NAME="Fedora Linux"
VERSION="37 (Workstation Edition)"
[acme@quaco perf-tools]$

I'll check on f38 and rawhide.

- Arnaldo

author
#7 lore raw
Quoted 49 lines · Arnaldo Carvalho de Melo

Em Wed, May 03, 2023 at 11:40:06AM +0200, Florian Weimer escreveu:

Quoted 33 lines

* Florian Weimer:

Quoted 32 lines

* Arnaldo Carvalho de Melo:

Quoted 1 line
Quoted 1 line

Thanks, applied. BTW b4 coulnd't find this message (nor the original):

Yes, vger drops the message after accepting it for some reason, probably something in the patch contents. I tried to resubmit from a completely separate account, no luck.

Quoted 15 lines

Not so fast, removed it for now:

CC /tmp/build/perf-tools-next/util/parse-events-bison.o
In file included from util/pmu.y:14:
/tmp/build/perf-tools-next/util/pmu-flex.h:496:1: error: unknown type name ‘YYSTYPE’
496 |

      | ^

/tmp/build/perf-tools-next/util/pmu-flex.h:498:19: error: unknown type name ‘YYSTYPE’
498 |
| ^
/tmp/build/perf-tools-next/util/pmu-flex.h:546:17: error: unknown type name ‘YYSTYPE’
546 | extern int yylex \
| ^~
util/pmu-bison.c: In function ‘perf_pmu_parse’:
/tmp/build/perf-tools-next/util/pmu-bison.c:69:25: error: implicit declaration of function ‘perf_pmu_lex’; did you mean ‘perf_pmu_free’? [-Werror=implicit-function-declaration]

Quoted 3 lines
   69 | #define yylex           perf_pmu_lex
      |                         ^~~~~~~~~~~~

util/pmu-bison.c:1007:16: note: in expansion of macro ‘yylex’

This appears to be related to some BPF filter stuff that is only in perf-next, not mainline.

Is this the right tree on which to base the patch?

Quoted 7 lines
Quoted 3 lines

<https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/>

Branch perf-tools-next?

Sorry, it doesn't build even without my patch:

make[4]: *** No rule to make target 'zip.h', needed by '/tmp/build/libbpf/staticobjs/libbpf.o'. Stop.

I'm noticing this, but then I haven't merged with upstream yet, I bet this is something upstream after merging the bpf branch... Nope, I just did a test merge of what is in
git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git perf-tools with upstream and it builds without problems on:

Quoted 8 lines

[acme@quaco perf-tools]$ head -2 /etc/os-release
NAME="Fedora Linux"
VERSION="37 (Workstation Edition)"
[acme@quaco perf-tools]$

I'll check on f38 and rawhide.

- Arnaldo

The zip.c and zip.h in libbpf are new, perhaps it is a fixdeps issue? Perhaps doing a clean build and retrying will address the problem.

Thanks,
Ian

#8 lore raw

* Ian Rogers:

Quoted 55 lines · Arnaldo Carvalho de Melo
Quoted 54 lines

Em Wed, May 03, 2023 at 11:40:06AM +0200, Florian Weimer escreveu:

Quoted 33 lines

* Florian Weimer:

Quoted 32 lines

* Arnaldo Carvalho de Melo:

Quoted 1 line
Quoted 1 line

Thanks, applied. BTW b4 coulnd't find this message (nor the original):

Yes, vger drops the message after accepting it for some reason, probably something in the patch contents. I tried to resubmit from a completely separate account, no luck.

Quoted 15 lines

Not so fast, removed it for now:

CC /tmp/build/perf-tools-next/util/parse-events-bison.o
In file included from util/pmu.y:14:
/tmp/build/perf-tools-next/util/pmu-flex.h:496:1: error: unknown type name ‘YYSTYPE’
496 |

      | ^

/tmp/build/perf-tools-next/util/pmu-flex.h:498:19: error: unknown type name ‘YYSTYPE’
498 |
| ^
/tmp/build/perf-tools-next/util/pmu-flex.h:546:17: error: unknown type name ‘YYSTYPE’
546 | extern int yylex \
| ^~
util/pmu-bison.c: In function ‘perf_pmu_parse’:
/tmp/build/perf-tools-next/util/pmu-bison.c:69:25: error: implicit declaration of function ‘perf_pmu_lex’; did you mean ‘perf_pmu_free’? [-Werror=implicit-function-declaration]

Quoted 3 lines
   69 | #define yylex           perf_pmu_lex
      |                         ^~~~~~~~~~~~

util/pmu-bison.c:1007:16: note: in expansion of macro ‘yylex’

This appears to be related to some BPF filter stuff that is only in perf-next, not mainline.

Is this the right tree on which to base the patch?

Quoted 7 lines
Quoted 3 lines

<https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/>

Branch perf-tools-next?

Sorry, it doesn't build even without my patch:

make[4]: *** No rule to make target 'zip.h', needed by '/tmp/build/libbpf/staticobjs/libbpf.o'. Stop.

I'm noticing this, but then I haven't merged with upstream yet, I bet this is something upstream after merging the bpf branch... Nope, I just did a test merge of what is in
git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git perf-tools with upstream and it builds without problems on:

[acme@quaco perf-tools]$ head -2 /etc/os-release
NAME="Fedora Linux"
VERSION="37 (Workstation Edition)"
[acme@quaco perf-tools]$

Quoted 6 lines
Quoted 3 lines

I'll check on f38 and rawhide.

- Arnaldo

The zip.c and zip.h in libbpf are new, perhaps it is a fixdeps issue? Perhaps doing a clean build and retrying will address the problem.

Yeah, it's that. I tried to reproduce the issue with a make command similar to Arnaldo's, and forgot that *of course* O= overrides “git clean”.

I'm now on 5d27a645f60940fdf589e4ff5351506a7f0fdbaa, can built it cleanly, and will try to rebase my patch onto that.

Thanks,
Florian