[ofa-general] RE: sparse output

Sean Hefty sean.hefty at intel.com
Wed Aug 8 16:32:34 PDT 2007


>/home/mst/build/x86_64/linux-2.6/drivers/infiniband/core/ucm.c:625:53: warning:
>incorrect type in argument 3 (different signedness)
>/home/mst/build/x86_64/linux-2.6/drivers/infiniband/core/ucm.c:625:53:
>expected int *qp_attr_mask
>/home/mst/build/x86_64/linux-2.6/drivers/infiniband/core/ucm.c:625:53:    got
>unsigned int *<noident>

>/home/mst/build/x86_64/linux-2.6/drivers/infiniband/core/ucma.c:781:49:
>warning: incorrect type in argument 3 (different signedness)
>/home/mst/build/x86_64/linux-2.6/drivers/infiniband/core/ucma.c:781:49:
>expected int *qp_attr_mask
>/home/mst/build/x86_64/linux-2.6/drivers/infiniband/core/ucma.c:781:49:    got
>unsigned int *<noident>

Both of the above issues relate to exporting the qp attribute mask to user
space.  The userspace ABI uses u32, but the kernel interface is int.  I would
say that we just cast these two warnings away.

- Sean



More information about the general mailing list