On 09/01/2021 20:09, Chaitanya Kulkarni wrote:
> On 1/9/21 07:59, Pavel Begunkov wrote:
>> iov_iter_bvec() initialises iterators well, no need to pre-zero it
>> beforehand as done in fd_execute_rw_aio(). Compilers can't optimise it
>> out and generate extra code for that (confirmed with assembly).
> It will be great if we can quantify this optimization with the actual
> performance
> numbers.
I expect you won't find any, but such little things can pile up
into a not-easy-to-spot overhead over time.
In any case, I don't think this requires performance justification
because it neither makes it less safe or uglier. Those iov_iter*()
are there to handle initialisation, that's a part of the iter API.
>> Signed-off-by: Pavel Begunkov <asml.sile...@gmail.com>
>> ---
>> drivers/target/target_core_file.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/target/target_core_file.c
>> b/drivers/target/target_core_file.c
>> index cce455929778..5a66854def95 100644
>> --- a/drivers/target/target_core_file.c
>> +++ b/drivers/target/target_core_file.c
>> @@ -267,7 +267,7 @@ fd_execute_rw_aio(struct se_cmd *cmd, struct scatterlist
>> *sgl, u32 sgl_nents,
>> struct fd_dev *fd_dev = FD_DEV(dev);
>> struct file *file = fd_dev->fd_file;
>> struct target_core_file_cmd *aio_cmd;
>> - struct iov_iter iter = {};
>> + struct iov_iter iter;
>> struct scatterlist *sg;
>> ssize_t len = 0;
>> int ret = 0, i;
>
--
Pavel Begunkov
