[Buildroot] [PATCH] package/uclibc: backport static linking fixes

[Petr Vorel]

Hi Thomas,

> Hello,

> On Tue,  8 Sep 2020 16:48:37 +0200
> Petr Vorel <petr.vorel at gmail.com> wrote:

> > Reported-by: Waldemar Brodkorb <wbx at openadk.org>
> > Signed-off-by: Petr Vorel <petr.vorel at gmail.com>
> > ---
> > Hi,

> > not sure if the second commit, which changes Rules.mak does not require
> > also some change on our side (it shouldn't, but better ask).

> The commit log should ideally have a link to the autobuilder failures
> that this is fixing (we have some I believe).
Sure, found several, added in v2.

> > diff --git a/package/uclibc/0001-Revert-Fix-static-linking-with-GCC-10.patch b/package/uclibc/0001-Revert-Fix-static-linking-with-GCC-10.patch
> > new file mode 100644
> > index 0000000000..be18b32f79
> > --- /dev/null
> > +++ b/package/uclibc/0001-Revert-Fix-static-linking-with-GCC-10.patch
> > @@ -0,0 +1,43 @@
> > +From e80384786d7b7f895c97f2447d4b91af8eb5f0f4 Mon Sep 17 00:00:00 2001
> > +From: Waldemar Brodkorb <wbx at openadk.org>
> > +Date: Wed, 2 Sep 2020 09:45:17 +0200
> > +Subject: [PATCH 1/2] Revert "Fix static linking with GCC-10"

> Could you generate the patches with "git format-patch -N". This is
> pointed out by "make check-package".

Sure, updated in v2. And thanks for a tip. FYI it didn't find a problem in v1
patch:

$ make check-package
find ../buildroot -type f \( -name '*.mk' -o -name '*.hash' -o -name 'Config.*' \) \
	-exec ./utils/check-package {} +
48551 lines processed
0 warnings generated
37530 lines processed
0 warnings generated
30032 lines processed
0 warnings generated
37002 lines processed
0 warnings generated
38287 lines processed
0 warnings generated


> > +
> > +This reverts commit 5b58a1ebd89a4f05778441814e81817c82193fa3.
> > +
> > +This breaks all static builds earlier to gcc 10 :(
> > +Bad testing on my side.

> We need you to sign-off on the patches.

> Same comment for the second patch of course.
Also fixed.

> > +diff --git a/Rules.mak b/Rules.mak
> > +index f270b7bee..55e88d462 100644
> > +--- a/Rules.mak
> > ++++ b/Rules.mak
> > +@@ -288,7 +288,7 @@ OPTIMIZATION += $(CFLAG_-fstrict-aliasing)
> > + 
> > + # Why -funsigned-char: I hunted a bug related to incorrect
> > + # sign extension of 'char' type for 10 hours straight. Not fun.
> > +-CPU_CFLAGS-y := -funsigned-char -fno-builtin
> > ++CPU_CFLAGS-y := -funsigned-char -fno-builtin -fcommon

> Obviously, this is papering over the problem a bit, but OK, if that's
> the upstream fix for the moment, let's go with it.
Good.

> Thanks!
Thanks a lot for your review.

Kind regards,
Petr