On Mon, Oct 31, 2011 at 8:25 AM, Zhi Yong Wu <zwu.ker...@gmail.com> wrote: > On Mon, Oct 31, 2011 at 4:10 PM, Kevin Wolf <kw...@redhat.com> wrote: >> Am 31.10.2011 04:01, schrieb Zhi Yong Wu: >>> It is better to use qiov.size in qed-table.c to get nb_sectors than >>> iov.iov_len. >>> >>> Signed-off-by: Zhi Yong Wu <wu...@linux.vnet.ibm.com> >> >> The commit message should probably say why it's better. Not saying >> otherwise, but I can't see the different at the first sight. > They are equal, but if the number of iov isn't ONE, they will be not > equal. qiov.size contains the sum of all iov's size while iov.iov_len > is only the size of one iov. Although in current scenario, they are > equal, but i think that it is better if qiov.size is used.
I see your reasoning. Especially in qed_read_table_cb() it's nice to use qiov->size because that function doesn't obviously use a single struct iovec. If you want to change it I agree but please send a patch with a proper explanation that mentions that this is purely a refactoring (does not change behavior) and why. Stefan