Re: stable-rc build: 72 warnings 1 failures (stable-rc/v4.4.30-35-gf821e08)

[Arnd Bergmann <arnd@arndb.de>]

On Tuesday, November 8, 2016 3:26:52 PM CET Olof Johansson wrote:
> On Tue, Nov 8, 2016 at 2:17 PM, Olof Johansson <olof@lixom.net> wrote:
> > On Tue, Nov 8, 2016 at 2:14 PM, Arnd Bergmann <arnd@arndb.de> wrote:
> >> On Tuesday, November 8, 2016 9:16:28 AM CET Olof's autobuilder wrote:
> >>> Here are the build results from automated periodic testing.
> >>>
> >>> The tree being built was stable-rc, found at:
> >>>
> >>> https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable-rc.git/
> >>>
> >>> Topmost commit:
> >>>
> >>> f821e08 Linux 4.4.31-rc1
> >>>
> >>> Build logs (stderr only) can be found at the following link (experimental):
> >>>
> >>> http://arm-soc.lixom.net/buildlogs/stable-rc/v4.4.30-35-gf821e08/
> >>
> >> These seem to be largely caused by building with gcc-6. It's probably
> >> a good idea to keep supporting that configuration though and
> >> backport the fixes. Here are the upstream commit IDs I've found.
> >
> > That's a lot of noise. I'll move back to build with gcc 4.9.2 instead, for now.
> >
> > It's not entirely reasonable to expect older releases to build with
> > new toolchains without warnings, and I'm not sure if it makes sense to
> > bring back those fixes to -stable (unless they fix real bugs, of
> > course).
> 
> A lot less noise with older gcc:
> 
> http://arm-soc.lixom.net/buildlogs/stable-rc/v4.4.30-44-gc5be865/

Out of the ones I mentioned earlier, that leaves

aaaab56dba9a ("of: silence warnings due to max() usage")
badbda53e505 ("mm/cma: silence warnings due to max() usage")

but adds a couple of others:

| arch/powerpc/relocs_check.sh: line 46: ccache /usr/local/cross-4.9.2/bin/powerpc64-linux-objdump: No such file or directory
| arch/arm64/Makefile:25: LSE atomics not supported by binutils

These seem to be caused by an older binutils release in your cross-4.9.2 build

| In file included from /work/build/batch/drivers/isdn/hardware/eicon/message.c:30:0:
| /work/build/batch/drivers/isdn/hardware/eicon/message.c: In function 'mixer_notify_update':
| /work/build/batch/drivers/isdn/hardware/eicon/platform.h:333:18: warning: array subscript is above array bounds [-Warray-bounds]
|  *(__le16 *)addr = cpu_to_le16(v);

This looks like a false-positive bug caused by 4.9 that is no longer
present in newer compilers.

I assume this is the line

PUT_WORD(&(((CAPI_MSG *) msg)->info.facility_req.structs[1]), LI_REQ_SILENT_UPDATE);

	Arnd