[PATCH v5 5/7] null_blk: allow non power of 2 zoned devices

[Pankaj Raghav]

On 5/24/22 17:22, Nathan Chancellor wrote:
>>>         git remote add linux-review https://protect2.fireeye.com/v1/url?k=081be8db-5780d026-081a6394-000babff317b-d12fdca0fccd0493&q=1&e=5ef82219-ef70-445f-a7d0-0ae0a30be69f&u=https%3A%2F%2Fgithub.com%2Fintel-lab-lkp%2Flinux
>>>         git fetch --no-tags linux-review Pankaj-Raghav/block-make-blkdev_nr_zones-and-blk_queue_zone_no-generic-for-npo2-zsze/20220524-011616
>>>         git checkout 3d3c81da0adbd40eb0d2125327b7e227582b2a37
>>>         # save the config file
>>>         mkdir build_dir && cp config build_dir/.config
>>>         COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=hexagon SHELL=/bin/bash
>>>
>>> If you fix the issue, kindly add following tag where applicable
>>> Reported-by: kernel test robot <lkp at intel.com>
>>>
>>> All errors (new ones prefixed by >>, old ones prefixed by <<):
>>>
>>>>> ERROR: modpost: "__hexagon_umoddi3" [drivers/block/null_blk/null_blk.ko] undefined!
>>
>> I am able to apply my patches cleanly against next-20220523, build it
>> without any errors with GCC and boot into it in my x86_64 machine. Not
>> sure why this error is popping up.
> 
> Do a 32-bit build and you'll see it. With GCC 12.1.0, ARCH=i386
> defconfig + CONFIG_BLK_DEV_NULL_BLK=y + CONFIG_BLK_DEV_ZONED=y
> reproduces it for me:
> 
> $ make -skj"$(nproc)" ARCH=i386 defconfig menuconfig all
> ld: drivers/block/null_blk/zoned.o: in function `null_init_zoned_dev':
> zoned.c:(.text+0x112e): undefined reference to `__umoddi3'
> ...
> 
Ah.. I didn't see anything about 32 bit ARCH in the email so it didn't
strike to me that was the problem. It said ARCH=hexagon and I had no
idea what it was. Thanks for reproducing it.
> roundup() does a plain division with a 64-bit dividend, which will cause
> issues with 32-bit architectures. I suspect that you should really be
> using DIV_ROUND_UP_SECTOR_T() for the nr_zones calculation or maybe one
> of the other rounding macros in include/linux/math.h but I am not sure.
> 
That must be it. I will fix it. Thanks Nathan.
> Cheers,
> Nathan