Hi all,
After merging the sound-asoc tree, today's linux-next build (x86_64
allmodconfig) failed like this:
sound/soc/codecs/snd-soc-wm8750: struct spi_device_id is 40 bytes. The last of 2 is:
0x77 0x6d 0x38 0x39 0x38 0x37 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
FATAL: sound/soc/codecs/snd-soc-wm8750: struct spi_device_id is not terminated with a NULL entry!
Caused by commit 40045a85df0e ("ASoC: Fix SPI driver binding for WM8987").
I have used the sound-asoc tree from next-20110809 again for today.
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/
On Thu, Aug 11, 2011 at 01:11:29PM +1000, Stephen Rothwell wrote:
sound/soc/codecs/snd-soc-wm8750: struct spi_device_id is 40 bytes. The last of 2 is:
0x77 0x6d 0x38 0x39 0x38 0x37 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
FATAL: sound/soc/codecs/snd-soc-wm8750: struct spi_device_id is not terminated with a NULL entry!
Caused by commit 40045a85df0e ("ASoC: Fix SPI driver binding for WM8987").
No it wasn't, it was there already. Anyway, should be fixed now.
Do you have any idea what you're doing to turn on this detection? It
really shouldn't be something that's specific to the x86 build but
there's nothing obvious I can see that does it.
Out of interest you used to carry reverts for commits when you'd
identified them but that seems not to happen any more?
Hi Mark,
On Thu, 11 Aug 2011 13:06:23 +0900 Mark Brown [off-list ref] wrote:
On Thu, Aug 11, 2011 at 01:11:29PM +1000, Stephen Rothwell wrote:
quoted
Caused by commit 40045a85df0e ("ASoC: Fix SPI driver binding for WM8987").
No it wasn't, it was there already. Anyway, should be fixed now.
Thanks.
Do you have any idea what you're doing to turn on this detection? It
really shouldn't be something that's specific to the x86 build but
there's nothing obvious I can see that does it.
My x86_64 allmodconfig build is (effectively) done like this (I am using
gcc v4.5.2):
echo CONFIG_PROFILE_ALL_BRANCHES=n >../x86_64_allmodconfig.except
echo CONFIG_DEBUG_INFO=n >>../x86_64_allmodconfig.except
echo CONFIG_X86_DECODER_SELFTEST=n >>../x86_64_allmodconfig.except
KCONFIG_ALLCONFIG=../x86_64_allmodconfig.except \
make O=../x86_64_allmodconfig allmodconfig
make O=../x86_64_allmodconfig -j48 -s
Out of interest you used to carry reverts for commits when you'd
identified them but that seems not to happen any more?
It depends on when I find a problem. If I find a build problem
immediately after mergeing a particular tree, I reset linux-next and use
the previous version of the tree instead. If I do not find the problem
until after I have completely merged all the trees, I may revert a
problem commit, or add a patch (is there is an interaction between two
trees). In the latter case I will carry the revert or patch as long as
needed.
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/