Skip to contentSkip to conversation

package/crucible: disable cgo to fix 32-bit build failures

PATCH In review Fabio Estevam 2 participants 6 messages Dec 27, 2025 – Oct 11

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 package/crucible: disable cgo to fix 32-bit build failures

The Buildroot autobuilder fails when building crucible on 32-bit architectures (notably i386) due to unresolved _cgo_* symbols during linking.

Crucible does not require cgo, and building it with CGO_ENABLED=0 avoids toolchain and linker issues while producing a fully functional binary.

Disable cgo explicitly to ensure reliable builds across all supported architectures.

Fixes:

https://autobuild.buildroot.org/results/0b59fb382f7a1d9c3644da71e7bd53b6a19418ce/build-end.log

---
 package/crucible/crucible.mk | 3 +++
 1 file changed, 3 insertions(+)
package/crucible/crucible.mk +3 -0
index 9b34cc52d950..6f6d1dfab94b 100644
@@ -10,4 +10,7 @@ CRUCIBLE_LICENSE = BSD-3-Clause
 CRUCIBLE_LICENSE_FILES = LICENSE
 CRUCIBLE_GOMOD = ./cmd/crucible
 
+# Disable cgo to avoid toolchain/linker issues on 32-bit architectures
+CRUCIBLE_GO_ENV = CGO_ENABLED=0
+
 $(eval $(golang-package))
Signature · 6 lines
2.34.1

_______________________________________________
buildroot mailing list
[email protected]
https://lists.buildroot.org/mailman/listinfo/buildroot
#1 lore raw

Hello Fabio,

Thanks for the patch. Adding Christian to the loop. Christian, could you comment on the below patch and help us understand whether this is the right fix?

Fabio, one comment/request below.

Quoted 8 lines · On Sat, 27 Dec 2025 18:22:44 -0300 Fabio Estevam <[email protected]> wrote:

The Buildroot autobuilder fails when building crucible on 32-bit architectures (notably i386) due to unresolved _cgo_* symbols during linking.

Crucible does not require cgo, and building it with CGO_ENABLED=0 avoids toolchain and linker issues while producing a fully functional binary.

Quoted 8 lines

Disable cgo explicitly to ensure reliable builds across all supported architectures.

Fixes:

https://autobuild.buildroot.org/results/0b59fb382f7a1d9c3644da71e7bd53b6a19418ce/build-end.log

In commit logs, we like to have some details on what initially introduced the problem, so that we can understand whether the patch is a fix that should be backported to our LTS release. Ideally a pointer to the Buildroot commit introducing the problem. If not, at least some indication that the problem exists since <some time> (to be specified, with an old autobuilder failure). In our case, we're mostly interested in understanding whether that fix should be backported to Buildroot 2025.02.x. So if the problem already existed prior to February 2025, we will want to backport it.

Thanks!

Thomas

Signature · 7 lines
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
[email protected]
https://lists.buildroot.org/mailman/listinfo/buildroot
Reviewed-by Christian Stewart
#2 lore raw

Hi all,

Quoted 29 lines · Thomas Petazzoni

Hello Fabio,

Thanks for the patch. Adding Christian to the loop. Christian, could you comment on the below patch and help us understand whether this is the right fix?

Fabio, one comment/request below.

Quoted 16 lines · On Sat, 27 Dec 2025 18:22:44 -0300 Fabio Estevam <[email protected]> wrote:

The Buildroot autobuilder fails when building crucible on 32-bit architectures (notably i386) due to unresolved _cgo_* symbols during linking.

Crucible does not require cgo, and building it with CGO_ENABLED=0 avoids toolchain and linker issues while producing a fully functional binary.

Disable cgo explicitly to ensure reliable builds across all supported architectures.

Fixes:

https://autobuild.buildroot.org/results/0b59fb382f7a1d9c3644da71e7bd53b6a19418ce/build-end.log

If Cgo is not required, then disabling it should be OK.

Looks good to me!

_______________________________________________
buildroot mailing list
[email protected]
https://lists.buildroot.org/mailman/listinfo/buildroot

#3 lore raw
In reply to Thomas Petazzoni via buildroot

Hi Thomas,

Quoted 9 lines · Thomas Petazzoni

In commit logs, we like to have some details on what initially introduced the problem, so that we can understand whether the patch is a fix that should be backported to our LTS release. Ideally a pointer to the Buildroot commit introducing the problem. If not, at least some indication that the problem exists since <some time> (to be specified, with an old autobuilder failure). In our case, we're mostly interested in understanding whether that fix should be backported to Buildroot 2025.02.x. So if the problem already existed prior to February 2025, we will want to backport it.

This failure has been present since the introduction of the crucible support in Buildroot.

It is OK to backport this fix to older releases.

Thanks
_______________________________________________
buildroot mailing list
[email protected]
https://lists.buildroot.org/mailman/listinfo/buildroot

#4 lore raw
Quoted 16 lines · Fabio Estevam

Hi Thomas,

Quoted 9 lines · Thomas Petazzoni

In commit logs, we like to have some details on what initially introduced the problem, so that we can understand whether the patch is a fix that should be backported to our LTS release. Ideally a pointer to the Buildroot commit introducing the problem. If not, at least some indication that the problem exists since <some time> (to be specified, with an old autobuilder failure). In our case, we're mostly interested in understanding whether that fix should be backported to Buildroot 2025.02.x. So if the problem already existed prior to February 2025, we will want to backport it.

Quoted 4 lines

This failure has been present since the introduction of the crucible support in Buildroot.

It is OK to backport this fix to older releases.

Ping. I keep receiving emails of crucible build failures.
_______________________________________________
buildroot mailing list
[email protected]
https://lists.buildroot.org/mailman/listinfo/buildroot

#5 lore raw
In reply to Fabio Estevam
Quoted 7 lines · Fabio Estevam

The Buildroot autobuilder fails when building crucible on 32-bit architectures (notably i386) due to unresolved _cgo_* symbols during linking.

Crucible does not require cgo, and building it with CGO_ENABLED=0 avoids toolchain and linker issues while producing a fully functional binary.

This has been fixed in the infra with

ec8f1b03e6 "package/go: decrease debug level for CGO linking"

Regards,
Arnout

Quoted 18 lines

Disable cgo explicitly to ensure reliable builds across all supported architectures.

Fixes:

https://autobuild.buildroot.org/results/0b59fb382f7a1d9c3644da71e7bd53b6a19418ce/build-end.log

---
  package/crucible/crucible.mk | 3 +++
  1 file changed, 3 insertions(+)
package/crucible/crucible.mk +0 -0
index 9b34cc52d950..6f6d1dfab94b 100644

package/crucible/crucible.mk CRUCIBLE_LICENSE = BSD-3-Clause · line 16

Quoted 7 lines
package/crucible/crucible.mk +3 -0
@@ -10,4 +10,7 @@ CRUCIBLE_LICENSE = BSD-3-Clause
  CRUCIBLE_LICENSE_FILES = LICENSE
  CRUCIBLE_GOMOD = ./cmd/crucible
 
+# Disable cgo to avoid toolchain/linker issues on 32-bit architectures
+CRUCIBLE_GO_ENV = CGO_ENABLED=0
+
  $(eval $(golang-package))

_______________________________________________
buildroot mailing list
[email protected]
https://lists.buildroot.org/mailman/listinfo/buildroot