On 10/8/21 2:19 AM, Geert Uytterhoeven wrote:
> On Fri, Oct 8, 2021 at 12:12 AM kernel test robot <lkp(a)intel.com> wrote:
>> tree:
https://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux-block.git
nvme-passthru-wip
>> head: 9c18980ac90053bcdb21594eae48935d89bf389c
>> commit: 5cc445dd8df6e06f3482711aa590170450364393 [10/19] nvme: wire-up support
for async-passthru on char-device.
>> config: m68k-allmodconfig (attached as .config)
>> compiler: m68k-linux-gcc (GCC) 11.2.0
>> 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/axboe/linux-block.git/com...
>> git remote add axboe-block
https://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux-block.git
>> git fetch --no-tags axboe-block nvme-passthru-wip
>> git checkout 5cc445dd8df6e06f3482711aa590170450364393
>> # save the attached .config to linux build tree
>> COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-11.2.0 make.cross
ARCH=m68k
>>
>> If you fix the issue, kindly add following tag as appropriate
>> Reported-by: kernel test robot <lkp(a)intel.com>
>>
>> All errors (new ones prefixed by >>):
>>
>> drivers/nvme/host/ioctl.c: In function 'nvme_pt_task_cb':
>>>> drivers/nvme/host/ioctl.c:47:19: error: cast to pointer from integer of
different size [-Werror=int-to-pointer-cast]
>> 47 | ptcmd64 = (void __user *) bcmd->unused2[0];
>> | ^
>> drivers/nvme/host/ioctl.c:62:58: error: cast to pointer from integer of
different size [-Werror=int-to-pointer-cast]
>> 62 | struct nvme_passthru_cmd __user *ptcmd = (void
*)bcmd->unused2[0];
>> | ^
>> drivers/nvme/host/ioctl.c: In function 'nvme_ns_async_ioctl':
>> drivers/nvme/host/ioctl.c:472:29: error: cast to pointer from integer of
different size [-Werror=int-to-pointer-cast]
>> 472 | void __user *argp = (void __user *) bcmd->unused2[0];
>> | ^
>> cc1: all warnings being treated as errors
>>
>>
>> vim +47 drivers/nvme/host/ioctl.c
>>
>> 39
>> 40 static void nvme_pt_task_cb(struct io_uring_cmd *ioucmd)
>> 41 {
>> 42 struct uring_cmd_data *ucd;
>> 43 struct nvme_passthru_cmd64 __user *ptcmd64 = NULL;
>> 44 struct block_uring_cmd *bcmd;
>> 45
>> 46 bcmd = (struct block_uring_cmd *) &ioucmd->pdu;
>> > 47 ptcmd64 = (void __user *) bcmd->unused2[0];
>
> Casts from u64 to a pointer on 32-bit need an intermediate cast to uintptr_t:
>
> ptcmd64 = (void __user *)(uintptr_t)bcmd->unused2[0];
>
> Note that you can improve on the naming, as people may be surprised
> discovering "unused2" is actually used ;-)
There's a reason the branch is called -wip, it's just a hodge podge of
stuff. Not sure why the kernel test robot bothers reporting failures
publicly for that...