CFQ - "In case of asynchronous requests, all the requests from all the processes are batched together according to their process's I/O priority"

Question

I think the following statements about CFQ contradict each other:

1. "In case of asynchronous requests, all the requests from all the processes are batched together according to their process's I/O priority."
2. The only requests that the IO scheduler considers to be asynchronous are regular "buffered" writes.
3. "ionice does nothing to de-prioritize asynchronous write IO."

Which statements are incorrect? Or if all of them are correct, why do they not contradict each other?

Answer 1

I believe statements 2 and 3 are incorrect.

It is possible for a user process to generate asynchronous write IOs using sync_file_range(). Since these IOs are submitted directly by the calling process, they could be affected by the process' IO priority.

sync_file_range() is designed to allow triggering async writeback. The current implementation (Linux v4.20) avoids setting REQ_SYNC, by using the writeback mode WB_SYNC_NONE. This is true even when your program waits for the writeback by including the flag SYNC_FILE_RANGE_WAIT_AFTER. (However, kernels between v2.6.29 and v4.4 incorrectly used WB_SYNC_ALL and hence REQ_SYNC for all writeback triggered by sync_file_range()).

I guess the process IO priority might also get used if it had to perform "direct reclaim" during an async write() call. That was allegedly gotten rid of in 2011.