On 9/26/21 4:36 AM, Hao Xu wrote:
> 在 2021/9/25 上午4:59, Pavel Begunkov 写道:
>> Add a helper io_free_batch_list(), which takes a single linked list and
>> puts/frees all requests from it in an efficient manner. Will be reused
>> later.
>>
>> Signed-off-by: Pavel Begunkov <asml.silence@xxxxxxxxx>
>> ---
>> fs/io_uring.c | 34 +++++++++++++++++++++-------------
>> 1 file changed, 21 insertions(+), 13 deletions(-)
>>
>> diff --git a/fs/io_uring.c b/fs/io_uring.c
>> index 205127394649..ad8af05af4bc 100644
>> --- a/fs/io_uring.c
>> +++ b/fs/io_uring.c
>> @@ -2308,12 +2308,31 @@ static void io_req_free_batch(struct req_batch *rb, struct io_kiocb *req,
>> wq_stack_add_head(&req->comp_list, &state->free_list);
>> }
>> +static void io_free_batch_list(struct io_ring_ctx *ctx,
>> + struct io_wq_work_list *list)
>> + __must_hold(&ctx->uring_lock)
>> +{
>> + struct io_wq_work_node *node;
>> + struct req_batch rb;
>> +
>> + io_init_req_batch(&rb);
>> + node = list->first;
>> + do {
>> + struct io_kiocb *req = container_of(node, struct io_kiocb,
>> + comp_list);
>> +
>> + node = req->comp_list.next;
>> + if (req_ref_put_and_test(req))
>> + io_req_free_batch(&rb, req, &ctx->submit_state);
>> + } while (node);
>> + io_req_free_batch_finish(ctx, &rb);
>> +}
> Hi Pavel, Why not we use the wq_list_for_each here as well?
Because we do wq_stack_add_head() in the middle of the loop and
there is also refcounting adding potential races. We'd need
some for_each_*_safe version of it.
--
Pavel Begunkov
