Here are patches to do the following:
(1) Block MSG_SENDPAGE_* flags from leaking into ->sendmsg() from
userspace, whilst allowing splice_to_socket() to pass them in.
(2) Allow MSG_SPLICE_PAGES to be passed into tls_*_sendmsg(). Until
support is added, it will be ignored and a splice-driven sendmsg()
will be treated like a normal sendmsg(). TCP, UDP, AF_UNIX and
Chelsio-TLS already handle the flag in net-next.
(3) Replace a chain of functions to splice-to-sendpage with a single
function to splice via sendmsg() with MSG_SPLICE_PAGES. This allows a
bunch of pages to be spliced from a pipe in a single call using a
bio_vec[] and pushes the main processing loop down into the bowels of
the protocol driver rather than repeatedly calling in with a page at a
time.
(4) Provide a ->splice_eof() op[2] that allows splice to signal to its
output that the input observed a premature EOF and that the caller
didn't flag SPLICE_F_MORE, thereby allowing a corked socket to be
flushed. This attempts to maintain the current behaviour. It is also
not called if we didn't manage to read any data and so didn't called
the actor function.
This needs routing though several layers to get it down to the network
protocol.
[!] Note that I chose not to pass in any flags - I'm not sure it's
particularly useful to pass in the splice flags; I also elected
not to return any error code - though we might actually want to do
that.
(5) Provide tls_{device,sw}_splice_eof() to flush a pending TLS record if
there is one.
(6) Provide splice_eof() for UDP, TCP, Chelsio-TLS and AF_KCM. AF_UNIX
doesn't seem to pay attention to the MSG_MORE or MSG_SENDPAGE_NOTLAST
flags.
(7) Alter the behaviour of sendfile() and fix SPLICE_F_MORE/MSG_MORE
signalling[1] such SPLICE_F_MORE is always signalled until we have
read sufficient data to finish the request. If we get a zero-length
before we've managed to splice sufficient data, we now leave the
socket expecting more data and leave it to userspace to deal with it.
(8) Make AF_TLS handle the MSG_SPLICE_PAGES internal sendmsg flag.
MSG_SPLICE_PAGES is an internal hint that tells the protocol that it
should splice the pages supplied if it can. Its sendpage
implementations are then turned into wrappers around that.
I've pushed the patches here also:
https://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git/log/?h=sendpage-2-tls
David
Changes
=======
ver #5)
- In splice_to_socket(), preclear ret in case len == 0.
- Provide ->splice_eof() for UDP, TCP, Chelsio-TLS and AF_KCM.
ver #4)
- Switch to using ->splice_eof() to signal premature EOF to the splice
output[2].
ver #3)
- Include the splice-to-socket rewrite patch.
- Fix SPLICE_F_MORE/MSG_MORE signalling.
- Allow AF_TLS to accept sendmsg() with MSG_SPLICE_PAGES before it is
handled.
- Allow a zero-length send() to a TLS socket to flush an outstanding
record.
- Address TLS kselftest failure.
ver #2)
- Dropped the slab data copying.
- "rls_" should be "tls_".
- Attempted to fix splice_direct_to_actor().
- Blocked MSG_SENDPAGE_* from being set by userspace.
Link: https://lore.kernel.org/r/499791.1685485603@xxxxxxxxxxxxxxxxxxxxxx/ [1]
Link: https://lore.kernel.org/r/CAHk-=wh=V579PDYvkpnTobCLGczbgxpMgGmmhqiTyE34Cpi5Gg@xxxxxxxxxxxxxx/ [2]
Link: https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=51c78a4d532efe9543a4df019ff405f05c6157f6 # part 1
Link: https://lore.kernel.org/r/20230524153311.3625329-1-dhowells@xxxxxxxxxx/ # v1
David Howells (14):
net: Block MSG_SENDPAGE_* from being passed to sendmsg() by userspace
tls: Allow MSG_SPLICE_PAGES but treat it as normal sendmsg
splice, net: Use sendmsg(MSG_SPLICE_PAGES) rather than ->sendpage()
splice, net: Add a splice_eof op to file-ops and socket-ops
tls/sw: Use splice_eof() to flush
tls/device: Use splice_eof() to flush
ipv4, ipv6: Use splice_eof() to flush
chelsio/chtls: Use splice_eof() to flush
kcm: Use splice_eof() to flush
splice, net: Fix SPLICE_F_MORE signalling in splice_direct_to_actor()
tls/sw: Support MSG_SPLICE_PAGES
tls/sw: Convert tls_sw_sendpage() to use MSG_SPLICE_PAGES
tls/device: Support MSG_SPLICE_PAGES
tls/device: Convert tls_device_sendpage() to use MSG_SPLICE_PAGES
.../chelsio/inline_crypto/chtls/chtls.h | 1 +
.../chelsio/inline_crypto/chtls/chtls_io.c | 9 +
.../chelsio/inline_crypto/chtls/chtls_main.c | 1 +
fs/splice.c | 207 +++++++++++---
include/linux/fs.h | 3 +-
include/linux/net.h | 1 +
include/linux/socket.h | 4 +-
include/linux/splice.h | 3 +
include/net/inet_common.h | 1 +
include/net/sock.h | 1 +
include/net/tcp.h | 1 +
include/net/udp.h | 1 +
net/ipv4/af_inet.c | 18 ++
net/ipv4/tcp.c | 16 ++
net/ipv4/tcp_ipv4.c | 1 +
net/ipv4/udp.c | 16 ++
net/ipv6/af_inet6.c | 1 +
net/ipv6/tcp_ipv6.c | 1 +
net/ipv6/udp.c | 18 ++
net/kcm/kcmsock.c | 15 ++
net/socket.c | 36 +--
net/tls/tls.h | 2 +
net/tls/tls_device.c | 110 ++++----
net/tls/tls_main.c | 4 +
net/tls/tls_sw.c | 253 ++++++++++--------
25 files changed, 485 insertions(+), 239 deletions(-)
