> Ummm, in isolation that change is technically correct, but when you > look at what bufblks contains it is clearly wrong. nbblks is an > int, so "1 << ffs(nbblks)" should not be larger than an int. > i.e. bufblks is simply a count of blocks in the log, which by > definition cannot be more than an int (in fact, 2^31 / 2^9 is the > largest legal value it can have). Hence it can't be larger than an > int, and all the functions it is passed to expect it to be an > int... > >Hence the use of xfs_daddr_t is wrong, and that's the first thing >that needs fixing.... thanks for review! I will resend updated patch. thanks! _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
