Add words to explain how zero copy feature works, and why it has to be trusted for handling IO read command. Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxx> --- Documentation/block/ublk.rst | 28 ++++++++++++++++++++-------- 1 file changed, 20 insertions(+), 8 deletions(-) diff --git a/Documentation/block/ublk.rst b/Documentation/block/ublk.rst index 1e0e7358e14a..33efff25b54d 100644 --- a/Documentation/block/ublk.rst +++ b/Documentation/block/ublk.rst @@ -309,18 +309,30 @@ with specified IO tag in the command data: ``UBLK_IO_COMMIT_AND_FETCH_REQ`` to the server, ublkdrv needs to copy the server buffer (pages) read to the IO request pages. -Future development -================== - Zero copy --------- -Zero copy is a generic requirement for nbd, fuse or similar drivers. A -problem [#xiaoguang]_ Xiaoguang mentioned is that pages mapped to userspace -can't be remapped any more in kernel with existing mm interfaces. This can -occurs when destining direct IO to ``/dev/ublkb*``. Also, he reported that -big requests (IO size >= 256 KB) may benefit a lot from zero copy. +ublk zero copy relies on io_uring's fixed kernel buffer, which provides +two APIs: `io_buffer_register_bvec()` and `io_buffer_unregister_bvec`. + +ublk adds IO command of `UBLK_IO_REGISTER_IO_BUF` to call +`io_buffer_register_bvec()` for ublk server to register client request +buffer into io_uring buffer table, then ublk server can submit io_uring +IOs with the registered buffer index. IO command of `UBLK_IO_UNREGISTER_IO_BUF` +calls `io_buffer_unregister_bvec` to unregister the buffer. + +ublk server implementing zero copy has to be CAP_SYS_ADMIN and be trusted, +because it is ublk server's responsibility to make sure IO buffer filled +with data, and ublk server has to handle short read correctly by returning +correct bytes filled to io buffer. Otherwise, uninitialized kernel buffer +will be exposed to client application. + +ublk server needs to align the parameter of `struct ublk_param_dma_align` +with backend for zero copy to work correctly. +For reaching best IO performance, ublk server should align its segment +parameter of `struct ublk_param_segment` with backend for avoiding +unnecessary IO split. References ========== -- 2.47.0
