Re: drivers/scsi/pmcraid.c:3350:48: sparse: incorrect type in argument 2 (different address spaces)

[Arnd Bergmann]

On Tue, May 9, 2017 at 2:32 AM, kbuild test robot
<fengguang.wu@xxxxxxxxx> wrote:
> tree:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> head:   2d3e4866dea96b0506395b47bfefb234f2088dac
> commit: beba3a20bf90ce1b93e24592c3ebf0d0bb581bbe x86: switch to RAW_COPY_USER
> date:   6 weeks ago
> reproduce:
>         # apt-get install sparse
>         git checkout beba3a20bf90ce1b93e24592c3ebf0d0bb581bbe
>         make ARCH=x86_64 allmodconfig
>         make C=1 CF=-D__CHECK_ENDIAN__
>
>
> sparse warnings: (new ones prefixed by >>)
>
>    drivers/scsi/pmcraid.c:163:22: sparse: cast from restricted __le16
>    drivers/scsi/pmcraid.c:163:22: sparse: cast to restricted __be16
>    drivers/scsi/pmcraid.c:163:22: sparse: cast from restricted __le16
>    drivers/scsi/pmcraid.c:163:22: sparse: cast to restricted __be16
>    drivers/scsi/pmcraid.c:163:22: sparse: cast from restricted __le16
>    drivers/scsi/pmcraid.c:178:57: sparse: restricted __le16 degrades to integer
>    drivers/scsi/pmcraid.c:333:41: sparse: invalid assignment: &=
>    drivers/scsi/pmcraid.c:333:41:    left side has type restricted __le64
>    drivers/scsi/pmcraid.c:333:41:    right side has type unsigned long long
>    drivers/scsi/pmcraid.c:348:29: sparse: Using plain integer as NULL pointer
>    drivers/scsi/pmcraid.c:901:19: sparse: cast to restricted __le32
>    drivers/scsi/pmcraid.c:901:19: sparse: cast from restricted __le64
>    drivers/scsi/pmcraid.c:5727:30: sparse: incorrect type in initializer (different base types)
>    drivers/scsi/pmcraid.c:5727:30:    expected int [signed] cfg_table_size
>    drivers/scsi/pmcraid.c:5727:30:    got restricted __be32 [usertype] <noident>
>    drivers/scsi/pmcraid.c:5729:13: sparse: cast to restricted __be16
>    drivers/scsi/pmcraid.c:5729:13: sparse: cast from restricted __le16
>    drivers/scsi/pmcraid.c:5729:13: sparse: cast to restricted __be16

I'm not entirely sure why these all show up now, but my guess is that
the friendly robot got confused by my patches that I sent a few weeks
ago to fix all the sparse warnings in this driver, leading it to tag the
warnings as regressions even though they are all old.

       Arnd