[GIT PULL] kbuild misc fixes for v4.2-rc1

[GIT PULL] kbuild misc fixes for v4.2-rc1

[Michal Marek]

Hi Linus,

this is the non-critical part of kbuild:
- usual make deb-pkg/rpm improvements
- .gitignore additions
- grammar cleansing in scripts/coccinelle
- new coccinelle semantic patch
Michal

The following changes since commit b787f68c36d49bb1d9236f403813641efa74a031:

  Linux 4.1-rc1 (2015-04-26 17:59:10 -0700)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild.git misc

for you to fetch changes up to 21a59991ce0cd9a0b54b135305e3fcf880f2aaf1:

  scripts/package/Makefile: rpmbuild is needed for rpm targets (2015-06-15 16:08:05 +0200)

----------------------------------------------------------------
Andi Kleen (1):
      Kbuild: Add ID files to .gitignore

Andrey Skvortsov (1):
      builddeb: fix stripped module signatures if CONFIG_DEBUG_INFO and CONFIG_MODULE_SIG_ALL are set

Arnaud Patard (1):
      deb-pkg: Add device tree blobs to the package

Borislav Petkov (1):
      kbuild/mkspec: Simplify vmlinux.bz2 creation

Dmitry Kalinkin (1):
      coccinelle: api: add vma_pages.cocci

Fabio Estevam (8):
      coccinelle: pm_runtime: Insert blank line
      coccinelle: returnvar: Use imperative mood
      coccinelle: ifaddr: Fix the sentence
      coccinelle: simple_open: Use imperative mood
      scripts/coccinelle/misc/semicolon.cocci: Use imperative mood
      scripts/coccinelle/misc/irqf_oneshot.cocci: Fix grammar
      coccinelle: irqf_oneshot.cocci: Improve the generated commit log
      coccinelle: simple_return: Add a blank line

Florian Fainelli (1):
      gitignore: Add MIPS vmlinux.32 to the list

Jim Davis (1):
      scripts/package/Makefile: rpmbuild is needed for rpm targets

Riku Voipio (1):
      deb-pkg: move setting debarch for a separate function

 .gitignore                                  |  4 ++
 scripts/coccinelle/api/pm_runtime.cocci     |  2 +-
 scripts/coccinelle/api/simple_open.cocci    |  4 +-
 scripts/coccinelle/api/vma_pages.cocci      | 60 +++++++++++++++++++++++++++++
 scripts/coccinelle/misc/ifaddr.cocci        |  3 +-
 scripts/coccinelle/misc/irqf_oneshot.cocci  |  7 +++-
 scripts/coccinelle/misc/returnvar.cocci     |  2 +-
 scripts/coccinelle/misc/semicolon.cocci     |  2 +-
 scripts/coccinelle/misc/simple_return.cocci |  2 +-
 scripts/package/Makefile                    |  8 +---
 scripts/package/builddeb                    | 31 +++++++++++----
 scripts/package/mkspec                      |  4 +-
 12 files changed, 103 insertions(+), 26 deletions(-)
 create mode 100644 scripts/coccinelle/api/vma_pages.cocci

Re: [GIT PULL] kbuild misc fixes for v4.2-rc1

[Linus Torvalds]

So with all these changes to the build system fro 4.2, I'm *still*
getting that annoying

   X.509 certificate list changed

issue. Which apparently people don't normally see, because it does to
stdout rather than to stderr, so it's hidden by all the other random
build output.

Making it show more information gives me

   X.509 certificate list changed to "signing_key.x509" from
"./signing_key.x509"

so it's still that "./" issue.

This is *trivial* to reproduce for me using

    git clean -dqfx
    make oldconfig
    make -j16
    make -j16

where that second "make" invocation will do it.

Not only is that "./" wrong, but the whole rule to check equality is
bad because the certificate list is still generated the wrong way
around, and then randomly deleted without proper rules for it.

I sent a patch to do it right (using filechk), but it apparently never
got merged? Instead, we continue to have that racy "remove the
.x509.list file at a random time" thing.

Do I need to fish out my patch and apply it directly? I hate just
bypassing the maintainership chain for stupid things like this, but
this has been pending for way too long. At least that fixes the
re-generation of that file, even if it doesn't fix the problem with
the spurious extra ./ at the head.

                  Linus

Re: [GIT PULL] kbuild misc fixes for v4.2-rc1

[Michal Marek]

On 2015-07-03 00:25, Linus Torvalds wrote:
> So with all these changes to the build system fro 4.2, I'm *still*
> getting that annoying
> 
>    X.509 certificate list changed
> 
> issue. Which apparently people don't normally see, because it does to
> stdout rather than to stderr, so it's hidden by all the other random
> build output.

There is a fix for that in in David Howels's tree, the topmost commit being

commit 9c71c950793b1b8c23c6d945b31f6545f82adced
Author: David Woodhouse <David.Woodhouse@intel.com>
Date:   Thu May 21 12:23:55 2015 +0100

    modsign: Add explicit CONFIG_SYSTEM_TRUSTED_KEYS option

I though it was meant for 4.2, but I do not even see it in linux-next.
David, what's your plan with the patches by David W?

Anyway, none of the kbuild branches was *meant* to fix the issue.

Thanks,
Michal

Re: [GIT PULL] kbuild misc fixes for v4.2-rc1

[Florian Fainelli]

On 03/07/15 03:50, Michal Marek wrote:
> On 2015-07-03 00:25, Linus Torvalds wrote:
>> So with all these changes to the build system fro 4.2, I'm *still*
>> getting that annoying
>>
>>    X.509 certificate list changed
>>
>> issue. Which apparently people don't normally see, because it does to
>> stdout rather than to stderr, so it's hidden by all the other random
>> build output.
> 
> There is a fix for that in in David Howels's tree, the topmost commit being
> 
> commit 9c71c950793b1b8c23c6d945b31f6545f82adced
> Author: David Woodhouse <David.Woodhouse@intel.com>
> Date:   Thu May 21 12:23:55 2015 +0100
> 
>     modsign: Add explicit CONFIG_SYSTEM_TRUSTED_KEYS option
> 
> I though it was meant for 4.2, but I do not even see it in linux-next.
> David, what's your plan with the patches by David W?
> 
> Anyway, none of the kbuild branches was *meant* to fix the issue.

I cannot seem to find any of these changes in 4.2-rc2, was there
anything else that needed to be fixed in Michal's pull request?

Thanks
-- 
Florian

Re: [GIT PULL] kbuild misc fixes for v4.2-rc1

[Florian Fainelli]

On 03/07/15 03:50, Michal Marek wrote:
> On 2015-07-03 00:25, Linus Torvalds wrote:
>> So with all these changes to the build system fro 4.2, I'm *still*
>> getting that annoying
>>
>>    X.509 certificate list changed
>>
>> issue. Which apparently people don't normally see, because it does to
>> stdout rather than to stderr, so it's hidden by all the other random
>> build output.
> 
> There is a fix for that in in David Howels's tree, the topmost commit being
> 
> commit 9c71c950793b1b8c23c6d945b31f6545f82adced
> Author: David Woodhouse <David.Woodhouse@intel.com>
> Date:   Thu May 21 12:23:55 2015 +0100
> 
>     modsign: Add explicit CONFIG_SYSTEM_TRUSTED_KEYS option
> 
> I though it was meant for 4.2, but I do not even see it in linux-next.
> David, what's your plan with the patches by David W?
> 
> Anyway, none of the kbuild branches was *meant* to fix the issue.

Michal, it does not look like this pull request has been merged into any
4.2-rc release, but I have not seen a resent either, what do we do with
these changes now?
-- 
Florian

Re: [GIT PULL] kbuild misc fixes for v4.2-rc1

[Michal Marek]

On 2015-08-28 00:16, Florian Fainelli wrote:
> On 03/07/15 03:50, Michal Marek wrote:
>> On 2015-07-03 00:25, Linus Torvalds wrote:
>>> So with all these changes to the build system fro 4.2, I'm *still*
>>> getting that annoying
>>>
>>>    X.509 certificate list changed
>>>
>>> issue. Which apparently people don't normally see, because it does to
>>> stdout rather than to stderr, so it's hidden by all the other random
>>> build output.
>>
>> There is a fix for that in in David Howels's tree, the topmost commit being
>>
>> commit 9c71c950793b1b8c23c6d945b31f6545f82adced
>> Author: David Woodhouse <David.Woodhouse@intel.com>
>> Date:   Thu May 21 12:23:55 2015 +0100
>>
>>     modsign: Add explicit CONFIG_SYSTEM_TRUSTED_KEYS option
>>
>> I though it was meant for 4.2, but I do not even see it in linux-next.
>> David, what's your plan with the patches by David W?
>>
>> Anyway, none of the kbuild branches was *meant* to fix the issue.
> 
> Michal, it does not look like this pull request has been merged into any
> 4.2-rc release, but I have not seen a resent either, what do we do with
> these changes now?

It will be part of the 4.3-rc1 pull request.

Michal