Re: [iio:testing 40/43] ld.lld: error: undefined symbol: __udivdi3

[Jarkko Nikula <jarkko.nikula@linux.intel.com>]

On 9/14/21 4:52 PM, William Breathitt Gray wrote:
> On Tue, Sep 14, 2021 at 01:41:03PM +0100, Jonathan Cameron wrote:
>> On Sun, 12 Sep 2021 06:43:58 +0800
>> kernel test robot <lkp@intel.com> wrote:
>>
>>> tree:   https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git testing
>>> head:   77a4a019161bcba51af1dbbeadc04795840fb3e0
>>> commit: f7da4339ca510e4476ea35cb9fd44f3f190706c5 [40/43] counter: Internalize sysfs interface code
>>> config: i386-randconfig-r022-20210911 (attached as .config)
>>> compiler: clang version 14.0.0 (https://github.com/llvm/llvm-project 261cbe98c38f8c1ee1a482fe76511110e790f58a)
>>> reproduce (this is a W=1 build):
>>>          wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
>>>          chmod +x ~/bin/make.cross
>>>          # https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git/commit/?id=f7da4339ca510e4476ea35cb9fd44f3f190706c5
>>>          git remote add iio https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git
>>>          git fetch --no-tags iio testing
>>>          git checkout f7da4339ca510e4476ea35cb9fd44f3f190706c5
>>>          # save the attached .config to linux build tree
>>>          mkdir build_dir
>>>          COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross O=build_dir ARCH=i386 SHELL=/bin/bash
>>>
>>> If you fix the issue, kindly add following tag as appropriate
>>> Reported-by: kernel test robot <lkp@intel.com>
>>>
>>> All errors (new ones prefixed by >>):
>>>
>>>>> ld.lld: error: undefined symbol: __udivdi3
>>>     >>> referenced by intel-qep.c:290 (drivers/counter/intel-qep.c:290)
>>>     >>>               counter/intel-qep.o:(intel_qep_spike_filter_ns_write) in archive drivers/built-in.a
>> This is a result of a change to passing in a u64 instead of a string.   I've applied a fix
>> that uses do_div() and pushed out as testing again.
>>
>> William, please sanity check that fix looks right to you.
>>
>> Thanks,
>>
>> Jonathan
> 
> Your do_div() change seems sane to me; would div_u64_rem() work here as
> well, or is do_div preferred because we're not using the remainder?
> 
> I'm CCing Jarkko Nikula as well to keep them in the loop.
> 
Yeah I believe do_div() is the right fix for the "counter: Internalize 
sysfs interface code" since the calculation remains the same.

I realized now the filter length has been rounded down. E.g. values 
20-29 from user space set 20 ns long filter in HW. Perhaps better would 
be to round up or to nearest but that goes to some future patch if needed.

Jarkko