[PATCH] glibc: always use bfd linker

[PATCH] glibc: always use bfd linker

[Martin Jansa]

* Work around broken ld-2.29.so when gold is used
  causing qemu-arm to segfault during e.g. gobject-introspection
  or postinst at do_rootfs time, more details in:
  http://lists.openembedded.org/pipermail/openembedded-devel/2019-March/198937.html
  https://sourceware.org/bugzilla/show_bug.cgi?id=24148
  https://sourceware.org/bugzilla/show_bug.cgi?id=10937
  https://sourceware.org/bugzilla/show_bug.cgi?id=18103

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
---
 meta/recipes-core/glibc/glibc_2.29.bb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/meta/recipes-core/glibc/glibc_2.29.bb b/meta/recipes-core/glibc/glibc_2.29.bb
index 9184c5932c..073d1533e3 100644
--- a/meta/recipes-core/glibc/glibc_2.29.bb
+++ b/meta/recipes-core/glibc/glibc_2.29.bb
@@ -110,7 +110,7 @@ do_configure () {
 
 do_compile () {
 	# -Wl,-rpath-link <staging>/lib in LDFLAGS can cause breakage if another glibc is in staging
-	unset LDFLAGS
+	LDFLAGS="-fuse-ld=bfd"
 	base_do_compile
 	echo "Adjust ldd script"
 	if [ -n "${RTLDLIST}" ]
-- 
2.17.1

Re: [PATCH] glibc: always use bfd linker

[Martin Jansa]

[-- Attachment #1: Type: text/plain, Size: 2161 bytes --]

On Thu, Mar 28, 2019 at 10:21:39AM +0000, Martin Jansa wrote:
> * Work around broken ld-2.29.so when gold is used
>   causing qemu-arm to segfault during e.g. gobject-introspection
>   or postinst at do_rootfs time, more details in:
>   http://lists.openembedded.org/pipermail/openembedded-devel/2019-March/198937.html
>   https://sourceware.org/bugzilla/show_bug.cgi?id=24148
>   https://sourceware.org/bugzilla/show_bug.cgi?id=10937
>   https://sourceware.org/bugzilla/show_bug.cgi?id=18103

ping.

RP: I know you said on IRC that it's late for warrior and you were
hoping for khem's opinion, is it now acceptable at least for master?

Here is Khem's reply in the other thread about this:
http://lists.openembedded.org/pipermail/openembedded-core/2019-March/280546.html

and Trevor's testing confirms that it's not only issue when
gold-built-libc is used by qemu, but it kills init in runtime as well:

[    4.140000] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b

http://lists.openembedded.org/pipermail/openembedded-core/2019-March/280605.html
http://lists.openembedded.org/pipermail/openembedded-core/2019-March/280610.html

Unusable libc on arm is a bit sad start for warrior build :). Using bfd
like it was in thud seems like reasonable work around even when we don't
fully understand the internal details why it fails with gold.

> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
> ---
>  meta/recipes-core/glibc/glibc_2.29.bb | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/meta/recipes-core/glibc/glibc_2.29.bb b/meta/recipes-core/glibc/glibc_2.29.bb
> index 9184c5932c..073d1533e3 100644
> --- a/meta/recipes-core/glibc/glibc_2.29.bb
> +++ b/meta/recipes-core/glibc/glibc_2.29.bb
> @@ -110,7 +110,7 @@ do_configure () {
>  
>  do_compile () {
>  	# -Wl,-rpath-link <staging>/lib in LDFLAGS can cause breakage if another glibc is in staging
> -	unset LDFLAGS
> +	LDFLAGS="-fuse-ld=bfd"
>  	base_do_compile
>  	echo "Adjust ldd script"
>  	if [ -n "${RTLDLIST}" ]
> -- 
> 2.17.1
> 

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 201 bytes --]