[GIT PULL] More Kbuild updates for v5.19-rc1

[GIT PULL] More Kbuild updates for v5.19-rc1

[Masahiro Yamada]

Hi Linus,

Please pull some more Kbuild updates for v5.19-rc1.
Thanks.




The following changes since commit 7e284070abe53d448517b80493863595af4ab5f0:

  Merge tag 'for-5.19/dm-changes' of
git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm
(2022-05-26 21:13:45 -0700)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild.git
tags/kbuild-v5.19-2

for you to fetch changes up to 37447c4c4735ea34c7388217da2f4c160b4366ca:

  scripts: objdiff: fix stale comments about .tmp_objdiff (2022-06-04
01:30:10 +0900)

----------------------------------------------------------------
Kbuild updates for v5.19 (2nd)

 - Fix build regressions for parisc, csky, nios2, openrisc

 - Simplify module builds for CONFIG_LTO_CLANG and CONFIG_X86_KERNEL_IBT

 - Remove arch/parisc/nm, which was presumably a workaround for old tools

 - Check the odd combination of EXPORT_SYMBOL and 'static' precisely

 - Make external module builds robust against "too long argument error"

 - Fix modpost to detect EXPORT_SYMBOL marked as __init or__exit
   (This uncovered a couple of potential issues of exported symbols)

 - Support j, k keys for moving the cursor in nconfig

----------------------------------------------------------------
Alexander Lobakin (1):
      modpost: fix removing numeric suffixes

Isak Ellmer (1):
      scripts: kconfig: nconf: make nconfig accept jk keybindings

Jing Leng (1):
      kbuild: Fix include path in scripts/Makefile.modpost

Masahiro Yamada (22):
      modpost: fix undefined behavior of is_arm_mapping_symbol()
      modpost: remove the unused argument of check_sec_ref()
      modpost: reuse ARRAY_SIZE() macro for section_mismatch()
      modpost: squash if...else-if in find_elf_symbol2()
      kbuild: replace $(if A,A,B) with $(or A,B) in scripts/Makefile.modpost
      kbuild: do not try to parse *.cmd files for objects provided by compiler
      kbuild: replace $(linked-object) with CONFIG options
      kbuild: do not create *.prelink.o for Clang LTO or IBT
      parisc: remove arch/parisc/nm
      kbuild: check static EXPORT_SYMBOL* by script instead of modpost
      kbuild: make built-in.a rule robust against too long argument error
      kbuild: make *.mod rule robust against too long argument error
      kbuild: add cmd_and_savecmd macro
      kbuild: rebuild multi-object modules when objtool is updated
      kbuild: remove redundant cleanups in scripts/link-vmlinux.sh
      kbuild: clean .tmp_* pattern by make clean
      kbuild: move vmlinux.o link to scripts/Makefile.vmlinux_o
      kbuild: factor out the common objtool arguments
      modpost: fix section mismatch check for exported init/exit sections
      modpost: simplify mod->name allocation
      modpost: use fnmatch() to simplify match()
      scripts: objdiff: fix stale comments about .tmp_objdiff

Schspa Shi (1):
      kbuild: Allow to select bash in a modified environment

Yuntao Wang (1):
      scripts/kallsyms: update usage message of the kallsyms program

 Makefile                   |   4 +-
 arch/parisc/Makefile       |   1 -
 arch/parisc/nm             |   6 ---
 scripts/Kbuild.include     |  10 ++++-
 scripts/Makefile.build     | 110
++++++++++++++++++----------------------------------
 scripts/Makefile.lib       |  31 ++++++++++++---
 scripts/Makefile.modfinal  |   5 +--
 scripts/Makefile.modpost   |  12 ++----
 scripts/Makefile.vmlinux_o |  87 ++++++++++++++++++++++++++++++++++++++++++
 scripts/check-local-export |  65 +++++++++++++++++++++++++++++++
 scripts/kallsyms.c         |   2 +-
 scripts/kconfig/nconf.c    |   8 +++-
 scripts/link-vmlinux.sh    | 126
+++++-------------------------------------------------------
 scripts/mod/file2alias.c   |   2 -
 scripts/mod/modpost.c      | 163
++++++++++++++++++------------------------------------------------------------
 scripts/mod/modpost.h      |   3 ++
 scripts/objdiff            |   6 +--
 17 files changed, 290 insertions(+), 351 deletions(-)
 delete mode 100644 arch/parisc/nm
 create mode 100644 scripts/Makefile.vmlinux_o
 create mode 100755 scripts/check-local-export


-- 
Best Regards
Masahiro Yamada

Re: [GIT PULL] More Kbuild updates for v5.19-rc1

[Linus Torvalds]

On Fri, Jun 3, 2022 at 9:42 AM Masahiro Yamada <masahiroy@kernel.org> wrote:
>
>  - Fix modpost to detect EXPORT_SYMBOL marked as __init or__exit
>    (This uncovered a couple of potential issues of exported symbols)

I pulled, and then unpulled again, since this introduces new warnings.

If you introduce new checks, you fix what it checks.

And if the warnings didn't get sorted out, then it wasn't ready to be merged.

There's a  good *reason* I require a clean build: non-clean builds
just mean that people will ignore new warnings because they think they
are the old ones that they have already seen.

So no, it is *NOT* ok to introduce new warnings and think "people will
fix these up because now there are warnings about it".

Quite the reverse. People not only *won't* fix things up, and it will
just hide new warnings from *other* things instead.

                   Linus

Re: [GIT PULL] More Kbuild updates for v5.19-rc1

[Masahiro Yamada]

On Sun, Jun 5, 2022 at 5:25 AM Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
> On Fri, Jun 3, 2022 at 9:42 AM Masahiro Yamada <masahiroy@kernel.org> wrote:
> >
> >  - Fix modpost to detect EXPORT_SYMBOL marked as __init or__exit
> >    (This uncovered a couple of potential issues of exported symbols)
>
> I pulled, and then unpulled again, since this introduces new warnings.
>
> If you introduce new checks, you fix what it checks.
>
> And if the warnings didn't get sorted out, then it wasn't ready to be merged.
>
> There's a  good *reason* I require a clean build: non-clean builds
> just mean that people will ignore new warnings because they think they
> are the old ones that they have already seen.
>
> So no, it is *NOT* ok to introduce new warnings and think "people will
> fix these up because now there are warnings about it".
>
> Quite the reverse. People not only *won't* fix things up, and it will
> just hide new warnings from *other* things instead.
>
>                    Linus


OK, I will drop the offending commit
and send v2.


-- 
Best Regards
Masahiro Yamada