Skip to contentSkip to conversation

cmake: make libbpf's Linux UAPI headers

PATCH In review Arnaldo Carvalho de Melo 4 participants 5 messages Oct 26, 2020

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 cmake: make libbpf's Linux UAPI headers

Hi guys,

I just stumbled on this, investigating... This is with what is
in the tmp branch at
git://git.kernel.org/pub/scm/devel/pahole/pahole.git.

error: found variable in CU '/home/acme/git/linux/security/selinux/hooks.c' that has void type Encountered error while encoding BTF.

  LD      .tmp_vmlinux.kallsyms1
  KSYMS   .tmp_vmlinux.kallsyms1.S
  AS      .tmp_vmlinux.kallsyms1.S
  LD      .tmp_vmlinux.kallsyms2
  KSYMS   .tmp_vmlinux.kallsyms2.S
  AS      .tmp_vmlinux.kallsyms2.S
  LD      vmlinux
  BTFIDS  vmlinux

FAILED: load BTF from vmlinux: Unknown error -2make[1]: *** [/home/acme/git/linux/Makefile:1164: vmlinux] Error 255 make[1]: Leaving directory '/home/acme/git/build/v5.10.0-rc1+'
make: *** [Makefile:185: __sub-make] Error 2

- Arnaldo

#1 lore raw
Quoted 16 lines · Arnaldo Carvalho de Melo

Hi guys,

I just stumbled on this, investigating... This is with what is
in the tmp branch at
git://git.kernel.org/pub/scm/devel/pahole/pahole.git.

error: found variable in CU '/home/acme/git/linux/security/selinux/hooks.c' that has void type Encountered error while encoding BTF.

  LD      .tmp_vmlinux.kallsyms1
  KSYMS   .tmp_vmlinux.kallsyms1.S
  AS      .tmp_vmlinux.kallsyms1.S
  LD      .tmp_vmlinux.kallsyms2
  KSYMS   .tmp_vmlinux.kallsyms2.S
  AS      .tmp_vmlinux.kallsyms2.S
  LD      vmlinux
  BTFIDS  vmlinux
Quoted 3 lines

FAILED: load BTF from vmlinux: Unknown error -2make[1]: *** [/home/acme/git/linux/Makefile:1164: vmlinux] Error 255 make[1]: Leaving directory '/home/acme/git/build/v5.10.0-rc1+'
make: *** [Makefile:185: __sub-make] Error 2

yep, I see the same and when I use --btf_encode_force, it shows more of them:

error: found variable in CU 'arch/x86/kernel/cpu/mce/core.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'drivers/platform/x86/intel_pmc_core.c' that has void type

I have no idea how variables are encoded in dwarf, but I wonder it's another gcc bug ;-)

jirka

#2 lore raw
Quoted 34 lines · Jiri Olsa
Quoted 16 lines · Arnaldo Carvalho de Melo

Hi guys,

I just stumbled on this, investigating... This is with what is
in the tmp branch at
git://git.kernel.org/pub/scm/devel/pahole/pahole.git.

error: found variable in CU '/home/acme/git/linux/security/selinux/hooks.c' that has void type Encountered error while encoding BTF.

  LD      .tmp_vmlinux.kallsyms1
  KSYMS   .tmp_vmlinux.kallsyms1.S
  AS      .tmp_vmlinux.kallsyms1.S
  LD      .tmp_vmlinux.kallsyms2
  KSYMS   .tmp_vmlinux.kallsyms2.S
  AS      .tmp_vmlinux.kallsyms2.S
  LD      vmlinux
  BTFIDS  vmlinux
Quoted 3 lines

FAILED: load BTF from vmlinux: Unknown error -2make[1]: *** [/home/acme/git/linux/Makefile:1164: vmlinux] Error 255 make[1]: Leaving directory '/home/acme/git/build/v5.10.0-rc1+'
make: *** [Makefile:185: __sub-make] Error 2

yep, I see the same and when I use --btf_encode_force, it shows more of them:

error: found variable in CU 'arch/x86/kernel/cpu/mce/core.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'drivers/platform/x86/intel_pmc_core.c' that has void type

Quoted 5 lines

I have no idea how variables are encoded in dwarf, but I wonder it's another gcc bug ;-)

jirka

Still looking. But I found if I move

c815d266 (Hao Luo 2020-08-24 17:45:23 -0700 444)

         if (var->ip.tag.type == 0) {

2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 445)
fprintf(stderr, "error: found variable in CU '%s'
that has void type\n",
2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 446)

                         cu->name);
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 447)
                 if (force)
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 448)
                         continue;
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 449)
                 err = -1;
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 450)
                 break;
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 451)
         }

f3d9054b (Hao Luo 2020-07-08 13:44:10 -0700 452)

after

2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 453)

         type = var->ip.tag.type + type_id_off;
2e719cca (Andrii Nakryiko          2020-10-08 16:39:57 -0700 454)
         linkage = var->external ? BTF_VAR_GLOBAL_ALLOCATED :
BTF_VAR_STATIC;
2e719cca (Andrii Nakryiko          2020-10-08 16:39:57 -0700 455)
         if (!percpu_var_exists(addr, &size, &name))

2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 456)
continue; /* not a per-CPU variable */

the error is gone. I think we should check percpu_var_exists() before checking var->ip.tag.type.

Hao

#3 lore raw

Oops, sorry about my terrible formatting.

Quoted 79 lines · Hao Luo
Quoted 34 lines · Jiri Olsa
Quoted 16 lines · Arnaldo Carvalho de Melo

Hi guys,

I just stumbled on this, investigating... This is with what is
in the tmp branch at
git://git.kernel.org/pub/scm/devel/pahole/pahole.git.

error: found variable in CU '/home/acme/git/linux/security/selinux/hooks.c' that has void type Encountered error while encoding BTF.

  LD      .tmp_vmlinux.kallsyms1
  KSYMS   .tmp_vmlinux.kallsyms1.S
  AS      .tmp_vmlinux.kallsyms1.S
  LD      .tmp_vmlinux.kallsyms2
  KSYMS   .tmp_vmlinux.kallsyms2.S
  AS      .tmp_vmlinux.kallsyms2.S
  LD      vmlinux
  BTFIDS  vmlinux
Quoted 3 lines

FAILED: load BTF from vmlinux: Unknown error -2make[1]: *** [/home/acme/git/linux/Makefile:1164: vmlinux] Error 255 make[1]: Leaving directory '/home/acme/git/build/v5.10.0-rc1+'
make: *** [Makefile:185: __sub-make] Error 2

yep, I see the same and when I use --btf_encode_force, it shows more of them:

error: found variable in CU 'arch/x86/kernel/cpu/mce/core.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'drivers/platform/x86/intel_pmc_core.c' that has void type

Quoted 5 lines

I have no idea how variables are encoded in dwarf, but I wonder it's another gcc bug ;-)

jirka

Still looking. But I found if I move

c815d266 (Hao Luo 2020-08-24 17:45:23 -0700 444)

         if (var->ip.tag.type == 0) {

2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 445)
fprintf(stderr, "error: found variable in CU '%s'
that has void type\n",
2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 446)

                         cu->name);
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 447)
                 if (force)
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 448)
                         continue;
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 449)
                 err = -1;
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 450)
                 break;
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 451)
         }

f3d9054b (Hao Luo 2020-07-08 13:44:10 -0700 452)

after

2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 453)

         type = var->ip.tag.type + type_id_off;
2e719cca (Andrii Nakryiko          2020-10-08 16:39:57 -0700 454)
         linkage = var->external ? BTF_VAR_GLOBAL_ALLOCATED :
BTF_VAR_STATIC;
2e719cca (Andrii Nakryiko          2020-10-08 16:39:57 -0700 455)
         if (!percpu_var_exists(addr, &size, &name))

2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 456)
continue; /* not a per-CPU variable */

the error is gone. I think we should check percpu_var_exists() before checking var->ip.tag.type.

Hao

#4 lore raw
Quoted 82 lines · Hao Luo

Oops, sorry about my terrible formatting.

Quoted 77 lines · Hao Luo
Quoted 34 lines · Jiri Olsa
Quoted 16 lines · Arnaldo Carvalho de Melo

Hi guys,

I just stumbled on this, investigating... This is with what is
in the tmp branch at
git://git.kernel.org/pub/scm/devel/pahole/pahole.git.

error: found variable in CU '/home/acme/git/linux/security/selinux/hooks.c' that has void type Encountered error while encoding BTF.

  LD      .tmp_vmlinux.kallsyms1
  KSYMS   .tmp_vmlinux.kallsyms1.S
  AS      .tmp_vmlinux.kallsyms1.S
  LD      .tmp_vmlinux.kallsyms2
  KSYMS   .tmp_vmlinux.kallsyms2.S
  AS      .tmp_vmlinux.kallsyms2.S
  LD      vmlinux
  BTFIDS  vmlinux
Quoted 3 lines

FAILED: load BTF from vmlinux: Unknown error -2make[1]: *** [/home/acme/git/linux/Makefile:1164: vmlinux] Error 255 make[1]: Leaving directory '/home/acme/git/build/v5.10.0-rc1+'
make: *** [Makefile:185: __sub-make] Error 2

yep, I see the same and when I use --btf_encode_force, it shows more of them:

error: found variable in CU 'arch/x86/kernel/cpu/mce/core.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'security/selinux/hooks.c' that has void type
error: found variable in CU 'drivers/platform/x86/intel_pmc_core.c' that has void type

Quoted 5 lines

I have no idea how variables are encoded in dwarf, but I wonder it's another gcc bug ;-)

jirka

Still looking. But I found if I move

c815d266 (Hao Luo 2020-08-24 17:45:23 -0700 444)

         if (var->ip.tag.type == 0) {

2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 445)
fprintf(stderr, "error: found variable in CU '%s'
that has void type\n",
2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 446)

                         cu->name);
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 447)
                 if (force)
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 448)
                         continue;
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 449)
                 err = -1;
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 450)
                 break;
f3d9054b (Hao Luo                  2020-07-08 13:44:10 -0700 451)
         }

f3d9054b (Hao Luo 2020-07-08 13:44:10 -0700 452)

after

2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 453)

         type = var->ip.tag.type + type_id_off;
2e719cca (Andrii Nakryiko          2020-10-08 16:39:57 -0700 454)
         linkage = var->external ? BTF_VAR_GLOBAL_ALLOCATED :
BTF_VAR_STATIC;
2e719cca (Andrii Nakryiko          2020-10-08 16:39:57 -0700 455)
         if (!percpu_var_exists(addr, &size, &name))

2e719cca (Andrii Nakryiko 2020-10-08 16:39:57 -0700 456)
continue; /* not a per-CPU variable */

the error is gone. I think we should check percpu_var_exists() before checking var->ip.tag.type.

sounds like a plan, can you please send a patch? Thanks!

Quoted 2 lines
Quoted 2 lines

Hao