In sys/kern/subr_disk.c, diskerr() uses %d to display block numbers.

For block numbers > 2^32, the output is not going to be correct, but
if I look at daddr_t, it may not always be int64_t.

Is there a reason why this shouldn't be %qd ?

On a related, but different note, does anyone know if the printf format
style of "%I64d" is part of Posix and/or likely to be supported by NetBSD
anytime soon ?  This capability is already present in Visual Studio .NET.

Darren