m68k: Fix virt_addr_valid() W=1 compiler warnings
authorGeert Uytterhoeven <geert@linux-m68k.org>
Fri, 5 Mar 2021 08:41:22 +0000 (09:41 +0100)
committerGeert Uytterhoeven <geert@linux-m68k.org>
Sat, 6 Mar 2021 13:15:07 +0000 (14:15 +0100)
commita65a802aadba072ca7514fc0c301fd7fdc6fc6cb
tree7944bfe1da3d75470bb8602dbf3e9209a1741e46
parenta38fd8748464831584a19438cbb3082b5a2dab15
m68k: Fix virt_addr_valid() W=1 compiler warnings

If CONFIG_DEBUG_SG=y, and CONFIG_MMU=y:

    include/linux/scatterlist.h: In function ‘sg_set_buf’:
    arch/m68k/include/asm/page_mm.h:174:49: warning: ordered comparison of pointer with null pointer [-Wextra]
      174 | #define virt_addr_valid(kaddr) ((void *)(kaddr) >= (void *)PAGE_OFFSET && (void *)(kaddr) < high_memory)
  |                                                 ^~

or CONFIG_MMU=n:

    include/linux/scatterlist.h: In function ‘sg_set_buf’:
    arch/m68k/include/asm/page_no.h:33:50: warning: ordered comparison of pointer with null pointer [-Wextra]
       33 | #define virt_addr_valid(kaddr) (((void *)(kaddr) >= (void *)PAGE_OFFSET) && \
  |                                                  ^~

Fix this by doing the comparison in the "unsigned long" instead of the
"void *" domain.

Note that for now this is only seen when compiling btrfs, due to commit
e9aa7c285d20a69c ("btrfs: enable W=1 checks for btrfs"), but as people
are doing more W=1 compile testing, it will start to show up elsewhere,
too.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Greg Ungerer <gerg@linux-m68k.org>
Link: https://lore.kernel.org/r/20210305084122.4118826-1-geert@linux-m68k.org
arch/m68k/include/asm/page_mm.h
arch/m68k/include/asm/page_no.h