Skip to contentSkip to conversation

ib_umem: fix type mismatch

PATCH In review Arnd Bergmann 2 participants 3 messages Sep 6, 2019 – Sep 8, 2019 InfiniBand

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ib_umem: fix type mismatch

On some 32-bit architectures, size_t is defined as 'int' rather than 'long', causing a harmless warning:

drivers/infiniband/core/umem_odp.c:220:7: error: comparison of distinct pointer types ('typeof (umem_odp->umem.address) *' (aka 'unsigned long *') and 'typeof (umem_odp->umem.length) *' (aka 'unsigned int *')) [-Werror,-Wcompare-distinct-pointer-types]

                if (check_add_overflow(umem_odp->umem.address,
                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/overflow.h:59:15: note: expanded from macro 'check_add_overflow'
        (void) (&__a == &__b);                  \
                ~~~~ ^  ~~~~

As size_t is always the same length as unsigned long in all supported architectures, change the structure definition to use the unsigned long type for both.

Fixes 204e3e5630c5 ("RDMA/odp: Check for overflow when computing the umem_odp end")
---
 include/rdma/ib_umem.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/rdma/ib_umem.h +1 -1
index a91b2af64ec4..5dffe05402ef 100644
@@ -44,7 +44,7 @@ struct ib_umem_odp;
 struct ib_umem {
 	struct ib_device       *ibdev;
 	struct mm_struct       *owning_mm;
-	size_t			length;
+	unsigned long		length;
 	unsigned long		address;
 	u32 writable : 1;
 	u32 is_odp : 1;
Signature · 1 line
2.20.0
#1 lore raw
Quoted 14 lines · Arnd Bergmann

On some 32-bit architectures, size_t is defined as 'int' rather than 'long', causing a harmless warning:

drivers/infiniband/core/umem_odp.c:220:7: error: comparison of distinct pointer types ('typeof (umem_odp->umem.address) *' (aka 'unsigned long *') and 'typeof (umem_odp->umem.length) *' (aka 'unsigned int *')) [-Werror,-Wcompare-distinct-pointer-types]

                if (check_add_overflow(umem_odp->umem.address,
                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/overflow.h:59:15: note: expanded from macro 'check_add_overflow'
        (void) (&__a == &__b);                  \
                ~~~~ ^  ~~~~

As size_t is always the same length as unsigned long in all supported architectures, change the structure definition to use the unsigned long type for both.

Quoted 6 lines
Fixes 204e3e5630c5 ("RDMA/odp: Check for overflow when computing the umem_odp end")
---
 include/rdma/ib_umem.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Hi Arnd,

I had slightly different fix in my submission queue, which I think is better because it leaves length to be size_t.

https://lore.kernel.org/linux-rdma/[email protected]/T/#u

Thanks

#2 lore raw
Quoted 6 lines · Leon Romanovsky

On Fri, Sep 06, 2019 at 05:42:37PM +0200, Arnd Bergmann wrote:

I had slightly different fix in my submission queue, which I think is better because it leaves length to be size_t.

https://lore.kernel.org/linux-rdma/[email protected]/T/#u

Ok, I'm dropping my patch from my local test tree in favor of that one then.

Arnd