Message ID | 20230623225513.2732256-4-dhowells@redhat.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | None | expand |
On Sat, Jun 24, 2023 at 12:55 AM David Howells <dhowells@redhat.com> wrote: > > Use sendmsg() and MSG_SPLICE_PAGES rather than sendpage in ceph when > transmitting data. For the moment, this can only transmit one page at a > time because of the architecture of net/ceph/, but if > write_partial_message_data() can be given a bvec[] at a time by the > iteration code, this would allow pages to be sent in a batch. > > Signed-off-by: David Howells <dhowells@redhat.com> > cc: Ilya Dryomov <idryomov@gmail.com> > cc: Xiubo Li <xiubli@redhat.com> > cc: Jeff Layton <jlayton@kernel.org> > cc: "David S. Miller" <davem@davemloft.net> > cc: Eric Dumazet <edumazet@google.com> > cc: Jakub Kicinski <kuba@kernel.org> > cc: Paolo Abeni <pabeni@redhat.com> > cc: Jens Axboe <axboe@kernel.dk> > cc: Matthew Wilcox <willy@infradead.org> > cc: ceph-devel@vger.kernel.org > cc: netdev@vger.kernel.org > --- > > Notes: > ver #5) > - Switch condition for setting MSG_MORE in write_partial_message_data() > > net/ceph/messenger_v1.c | 60 ++++++++++++++--------------------------- > 1 file changed, 20 insertions(+), 40 deletions(-) > > diff --git a/net/ceph/messenger_v1.c b/net/ceph/messenger_v1.c > index d664cb1593a7..814579f27f04 100644 > --- a/net/ceph/messenger_v1.c > +++ b/net/ceph/messenger_v1.c > @@ -74,37 +74,6 @@ static int ceph_tcp_sendmsg(struct socket *sock, struct kvec *iov, > return r; > } > > -/* > - * @more: either or both of MSG_MORE and MSG_SENDPAGE_NOTLAST > - */ > -static int ceph_tcp_sendpage(struct socket *sock, struct page *page, > - int offset, size_t size, int more) > -{ > - ssize_t (*sendpage)(struct socket *sock, struct page *page, > - int offset, size_t size, int flags); > - int flags = MSG_DONTWAIT | MSG_NOSIGNAL | more; > - int ret; > - > - /* > - * sendpage cannot properly handle pages with page_count == 0, > - * we need to fall back to sendmsg if that's the case. > - * > - * Same goes for slab pages: skb_can_coalesce() allows > - * coalescing neighboring slab objects into a single frag which > - * triggers one of hardened usercopy checks. > - */ > - if (sendpage_ok(page)) > - sendpage = sock->ops->sendpage; > - else > - sendpage = sock_no_sendpage; > - > - ret = sendpage(sock, page, offset, size, flags); > - if (ret == -EAGAIN) > - ret = 0; > - > - return ret; > -} > - > static void con_out_kvec_reset(struct ceph_connection *con) > { > BUG_ON(con->v1.out_skip); > @@ -464,7 +433,6 @@ static int write_partial_message_data(struct ceph_connection *con) > struct ceph_msg *msg = con->out_msg; > struct ceph_msg_data_cursor *cursor = &msg->cursor; > bool do_datacrc = !ceph_test_opt(from_msgr(con->msgr), NOCRC); > - int more = MSG_MORE | MSG_SENDPAGE_NOTLAST; > u32 crc; > > dout("%s %p msg %p\n", __func__, con, msg); > @@ -482,6 +450,10 @@ static int write_partial_message_data(struct ceph_connection *con) > */ > crc = do_datacrc ? le32_to_cpu(msg->footer.data_crc) : 0; > while (cursor->total_resid) { > + struct bio_vec bvec; > + struct msghdr msghdr = { > + .msg_flags = MSG_SPLICE_PAGES, Hi David, This appears to be losing MSG_DONTWAIT | MSG_NOSIGNAL flags which were set previously? > + }; > struct page *page; > size_t page_offset; > size_t length; > @@ -493,10 +465,13 @@ static int write_partial_message_data(struct ceph_connection *con) > } > > page = ceph_msg_data_next(cursor, &page_offset, &length); > - if (length == cursor->total_resid) > - more = MSG_MORE; > - ret = ceph_tcp_sendpage(con->sock, page, page_offset, length, > - more); > + if (length != cursor->total_resid) > + msghdr.msg_flags |= MSG_MORE; > + > + bvec_set_page(&bvec, page, length, page_offset); > + iov_iter_bvec(&msghdr.msg_iter, ITER_SOURCE, &bvec, 1, length); > + > + ret = sock_sendmsg(con->sock, &msghdr); > if (ret <= 0) { And this is losing munging -EAGAIN -> 0? > if (do_datacrc) > msg->footer.data_crc = cpu_to_le32(crc); > @@ -526,7 +501,10 @@ static int write_partial_message_data(struct ceph_connection *con) > */ > static int write_partial_skip(struct ceph_connection *con) > { > - int more = MSG_MORE | MSG_SENDPAGE_NOTLAST; > + struct bio_vec bvec; > + struct msghdr msghdr = { > + .msg_flags = MSG_SPLICE_PAGES | MSG_MORE, > + }; > int ret; > > dout("%s %p %d left\n", __func__, con, con->v1.out_skip); > @@ -534,9 +512,11 @@ static int write_partial_skip(struct ceph_connection *con) > size_t size = min(con->v1.out_skip, (int)PAGE_SIZE); > > if (size == con->v1.out_skip) > - more = MSG_MORE; > - ret = ceph_tcp_sendpage(con->sock, ceph_zero_page, 0, size, > - more); > + msghdr.msg_flags &= ~MSG_MORE; > + bvec_set_page(&bvec, ZERO_PAGE(0), size, 0); > + iov_iter_bvec(&msghdr.msg_iter, ITER_SOURCE, &bvec, 1, size); > + > + ret = sock_sendmsg(con->sock, &msghdr); > if (ret <= 0) Same here... I would suggest that you keep ceph_tcp_sendpage() function and make only minimal modifications to avoid regressions. Thanks, Ilya
Ilya Dryomov <idryomov@gmail.com> wrote: > Same here... I would suggest that you keep ceph_tcp_sendpage() function > and make only minimal modifications to avoid regressions. This is now committed to net-next. I can bring ceph_tcp_sendpage() back into existence or fix it in place for now if you have a preference. Note that I'm working on patches to rework the libceph transmission path so that it isn't dealing with transmitting a single page at a time, but it's not ready yet. David
Ilya Dryomov <idryomov@gmail.com> wrote:
> > - int flags = MSG_DONTWAIT | MSG_NOSIGNAL | more;
Btw, why are you setting MSG_DONTWAIT? If you're in the middle of
transmitting a message on a TCP socket, surely you can't just switch to
transmitting a different message on the same socket without doing some sort of
reframing?
David
On Mon, Jun 26, 2023 at 4:00 PM David Howells <dhowells@redhat.com> wrote: > > Ilya Dryomov <idryomov@gmail.com> wrote: > > > Same here... I would suggest that you keep ceph_tcp_sendpage() function > > and make only minimal modifications to avoid regressions. > > This is now committed to net-next. This needs to be dropped from linux-next because both this and especially the other (net/ceph/messenger_v2.c) patch introduce regressions. > I can bring ceph_tcp_sendpage() back into > existence or fix it in place for now if you have a preference. I already mentioned that I would prefer if ceph_tcp_sendpage() was brought back into existence. > > Note that I'm working on patches to rework the libceph transmission path so > that it isn't dealing with transmitting a single page at a time, but it's not > ready yet. That is a worthwhile improvement now that sock_sendmsg() can take advantage of multiple pages! It would be pretty invasive though so I think ceph_tcp_sendpage() is better to remain in place until then. Thanks, Ilya
On Mon, Jun 26, 2023 at 5:12 PM David Howells <dhowells@redhat.com> wrote: > > Ilya Dryomov <idryomov@gmail.com> wrote: > > > > - int flags = MSG_DONTWAIT | MSG_NOSIGNAL | more; > > Btw, why are you setting MSG_DONTWAIT? If you're in the middle of > transmitting a message on a TCP socket, surely you can't just switch to > transmitting a different message on the same socket without doing some sort of > reframing? We don't want to hog kworker threads. You are correct that we can't switch to transmitting a different message on the same socket but Ceph is massively parallel and there can be dozens or even hundreds of other sockets to work on. Thanks, Ilya
Ilya Dryomov <idryomov@gmail.com> wrote: > > This is now committed to net-next. > > This needs to be dropped from linux-next because both this and > especially the other (net/ceph/messenger_v2.c) patch introduce > regressions. net-next, not linux-next. I'm not sure they drop things from there rather than reverting them. David
diff --git a/net/ceph/messenger_v1.c b/net/ceph/messenger_v1.c index d664cb1593a7..814579f27f04 100644 --- a/net/ceph/messenger_v1.c +++ b/net/ceph/messenger_v1.c @@ -74,37 +74,6 @@ static int ceph_tcp_sendmsg(struct socket *sock, struct kvec *iov, return r; } -/* - * @more: either or both of MSG_MORE and MSG_SENDPAGE_NOTLAST - */ -static int ceph_tcp_sendpage(struct socket *sock, struct page *page, - int offset, size_t size, int more) -{ - ssize_t (*sendpage)(struct socket *sock, struct page *page, - int offset, size_t size, int flags); - int flags = MSG_DONTWAIT | MSG_NOSIGNAL | more; - int ret; - - /* - * sendpage cannot properly handle pages with page_count == 0, - * we need to fall back to sendmsg if that's the case. - * - * Same goes for slab pages: skb_can_coalesce() allows - * coalescing neighboring slab objects into a single frag which - * triggers one of hardened usercopy checks. - */ - if (sendpage_ok(page)) - sendpage = sock->ops->sendpage; - else - sendpage = sock_no_sendpage; - - ret = sendpage(sock, page, offset, size, flags); - if (ret == -EAGAIN) - ret = 0; - - return ret; -} - static void con_out_kvec_reset(struct ceph_connection *con) { BUG_ON(con->v1.out_skip); @@ -464,7 +433,6 @@ static int write_partial_message_data(struct ceph_connection *con) struct ceph_msg *msg = con->out_msg; struct ceph_msg_data_cursor *cursor = &msg->cursor; bool do_datacrc = !ceph_test_opt(from_msgr(con->msgr), NOCRC); - int more = MSG_MORE | MSG_SENDPAGE_NOTLAST; u32 crc; dout("%s %p msg %p\n", __func__, con, msg); @@ -482,6 +450,10 @@ static int write_partial_message_data(struct ceph_connection *con) */ crc = do_datacrc ? le32_to_cpu(msg->footer.data_crc) : 0; while (cursor->total_resid) { + struct bio_vec bvec; + struct msghdr msghdr = { + .msg_flags = MSG_SPLICE_PAGES, + }; struct page *page; size_t page_offset; size_t length; @@ -493,10 +465,13 @@ static int write_partial_message_data(struct ceph_connection *con) } page = ceph_msg_data_next(cursor, &page_offset, &length); - if (length == cursor->total_resid) - more = MSG_MORE; - ret = ceph_tcp_sendpage(con->sock, page, page_offset, length, - more); + if (length != cursor->total_resid) + msghdr.msg_flags |= MSG_MORE; + + bvec_set_page(&bvec, page, length, page_offset); + iov_iter_bvec(&msghdr.msg_iter, ITER_SOURCE, &bvec, 1, length); + + ret = sock_sendmsg(con->sock, &msghdr); if (ret <= 0) { if (do_datacrc) msg->footer.data_crc = cpu_to_le32(crc); @@ -526,7 +501,10 @@ static int write_partial_message_data(struct ceph_connection *con) */ static int write_partial_skip(struct ceph_connection *con) { - int more = MSG_MORE | MSG_SENDPAGE_NOTLAST; + struct bio_vec bvec; + struct msghdr msghdr = { + .msg_flags = MSG_SPLICE_PAGES | MSG_MORE, + }; int ret; dout("%s %p %d left\n", __func__, con, con->v1.out_skip); @@ -534,9 +512,11 @@ static int write_partial_skip(struct ceph_connection *con) size_t size = min(con->v1.out_skip, (int)PAGE_SIZE); if (size == con->v1.out_skip) - more = MSG_MORE; - ret = ceph_tcp_sendpage(con->sock, ceph_zero_page, 0, size, - more); + msghdr.msg_flags &= ~MSG_MORE; + bvec_set_page(&bvec, ZERO_PAGE(0), size, 0); + iov_iter_bvec(&msghdr.msg_iter, ITER_SOURCE, &bvec, 1, size); + + ret = sock_sendmsg(con->sock, &msghdr); if (ret <= 0) goto out; con->v1.out_skip -= ret;
Use sendmsg() and MSG_SPLICE_PAGES rather than sendpage in ceph when transmitting data. For the moment, this can only transmit one page at a time because of the architecture of net/ceph/, but if write_partial_message_data() can be given a bvec[] at a time by the iteration code, this would allow pages to be sent in a batch. Signed-off-by: David Howells <dhowells@redhat.com> cc: Ilya Dryomov <idryomov@gmail.com> cc: Xiubo Li <xiubli@redhat.com> cc: Jeff Layton <jlayton@kernel.org> cc: "David S. Miller" <davem@davemloft.net> cc: Eric Dumazet <edumazet@google.com> cc: Jakub Kicinski <kuba@kernel.org> cc: Paolo Abeni <pabeni@redhat.com> cc: Jens Axboe <axboe@kernel.dk> cc: Matthew Wilcox <willy@infradead.org> cc: ceph-devel@vger.kernel.org cc: netdev@vger.kernel.org --- Notes: ver #5) - Switch condition for setting MSG_MORE in write_partial_message_data() net/ceph/messenger_v1.c | 60 ++++++++++++++--------------------------- 1 file changed, 20 insertions(+), 40 deletions(-)