new file mode 100755
@@ -0,0 +1,38 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: GPL-2.0
+
+"""Sets sysctl values and writes a file that restores them.
+
+The arguments are of the form "<proc-file>=<val>" separated by spaces.
+The program first reads the current value of the proc-file and creates
+a shell script named "/tmp/sysctl_restore_${PACKETDRILL_PID}.sh" which
+restores the values when executed. It then sets the new values.
+
+PACKETDRILL_PID is set by packetdrill to the pid of itself, so a .pkt
+file could restore sysctls by running `/tmp/sysctl_restore_${PPID}.sh`
+at the end.
+"""
+
+import os
+import subprocess
+import sys
+
+filename = '/tmp/sysctl_restore_%s.sh' % os.environ['PACKETDRILL_PID']
+
+# Open file for restoring sysctl values
+restore_file = open(filename, 'w')
+print('#!/bin/bash', file=restore_file)
+
+for a in sys.argv[1:]:
+ sysctl = a.split('=')
+ # sysctl[0] contains the proc-file name, sysctl[1] the new value
+
+ # read current value and add restore command to file
+ cur_val = subprocess.check_output(['cat', sysctl[0]], universal_newlines=True)
+ print('echo "%s" > %s' % (cur_val.strip(), sysctl[0]), file=restore_file)
+
+ # set new value
+ cmd = 'echo "%s" > %s' % (sysctl[1], sysctl[0])
+ os.system(cmd)
+
+os.system('chmod u+x %s' % filename)
new file mode 100644
@@ -0,0 +1,55 @@
+// SPDX-License-Identifier: GPL-2.0
+// basic zerocopy test:
+//
+// send a packet with MSG_ZEROCOPY and receive the notification ID
+// repeat and verify IDs are consecutive
+
+`./defaults.sh`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+ +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+ +0 bind(3, ..., ...) = 0
+ +0 listen(3, 1) = 0
+
+ +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+ +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
+ +0 < . 1:1(0) ack 1 win 257
+
+ +0 accept(3, ..., ...) = 4
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 1:4001(4000) ack 1
+ +0 < . 1:1(0) ack 4001 win 257
+
+ +0 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=0}}
+ ]}, MSG_ERRQUEUE) = 0
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 4001:8001(4000) ack 1
+ +0 < . 1:1(0) ack 8001 win 257
+
+ +0 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=1,
+ ee_data=1}}
+ ]}, MSG_ERRQUEUE) = 0
new file mode 100644
@@ -0,0 +1,41 @@
+// SPDX-License-Identifier: GPL-2.0
+// batch zerocopy test:
+//
+// send multiple packets, then read one range of all notifications.
+
+`./defaults.sh`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+ +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+ +0 bind(3, ..., ...) = 0
+ +0 listen(3, 1) = 0
+
+ +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+ +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
+ +0 < . 1:1(0) ack 1 win 257
+
+ +0 accept(3, ..., ...) = 4
+ +0 setsockopt(4, SOL_SOCKET, SO_MARK, [666], 4) = 0
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 1:4001(4000) ack 1
+ +0 < . 1:1(0) ack 4001 win 257
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 4001:8001(4000) ack 1
+ +0 < . 1:1(0) ack 8001 win 257
+
+ +0 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=1}}
+ ]}, MSG_ERRQUEUE) = 0
new file mode 100644
@@ -0,0 +1,31 @@
+// SPDX-License-Identifier: GPL-2.0
+// Minimal client-side zerocopy test
+
+`./defaults.sh`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
+ +0 setsockopt(4, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+ +0...0 connect(4, ..., ...) = 0
+
+ +0 > S 0:0(0) <mss 1460,sackOK,TS val 0 ecr 0,nop,wscale 8>
+ +0 < S. 0:0(0) ack 1 win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+ +0 > . 1:1(0) ack 1
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 1:4001(4000) ack 1
+ +0 < . 1:1(0) ack 4001 win 257
+
+ +0 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=0}}
+ ]}, MSG_ERRQUEUE) = 0
+
new file mode 100644
@@ -0,0 +1,45 @@
+// SPDX-License-Identifier: GPL-2.0
+// send with MSG_ZEROCOPY on a non-established socket
+//
+// verify that a send in state TCP_CLOSE correctly aborts the zerocopy
+// operation, specifically it does not increment the zerocopy counter.
+//
+// First send on a closed socket and wait for (absent) notification.
+// Then connect and send and verify that notification nr. is zero.
+
+`./defaults.sh`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
+ +0 setsockopt(4, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = -1 EPIPE (Broken pipe)
+
+ +0.1 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[]}, MSG_ERRQUEUE) = -1 EAGAIN (Resource temporarily unavailable)
+
+ +0...0 connect(4, ..., ...) = 0
+
+ +0 > S 0:0(0) <mss 1460,sackOK,TS val 0 ecr 0,nop,wscale 8>
+ +0 < S. 0:0(0) ack 1 win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+ +0 > . 1:1(0) ack 1
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 1:4001(4000) ack 1
+ +0 < . 1:1(0) ack 4001 win 257
+
+ +0 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=0}}
+ ]}, MSG_ERRQUEUE) = 0
+
new file mode 100644
@@ -0,0 +1,61 @@
+// SPDX-License-Identifier: GPL-2.0
+// epoll zerocopy test:
+//
+// EPOLLERR is known to be not edge-triggered unlike EPOLLIN and EPOLLOUT but
+// it is not level-triggered either.
+//
+// fire two sends with MSG_ZEROCOPY and receive the acks. confirm that EPOLLERR
+// is correctly fired only once, when EPOLLET is set. send another packet with
+// MSG_ZEROCOPY. confirm that EPOLLERR is correctly fired again only once.
+`./defaults.sh`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+ +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+ +0 bind(3, ..., ...) = 0
+ +0 listen(3, 1) = 0
+
+ +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+ +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
+ +0 < . 1:1(0) ack 1 win 257
+
+ +0 accept(3, ..., ...) = 4
+
+ +0 fcntl(4, F_SETFL, O_RDWR|O_NONBLOCK) = 0
+ +0 epoll_create(1) = 5
+ +0 epoll_ctl(5, EPOLL_CTL_ADD, 4, {events=EPOLLOUT|EPOLLET, fd=4}) = 0
+ +0 epoll_wait(5, {events=EPOLLOUT, fd=4}, 1, 0) = 1
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 1:4001(4000) ack 1
+ +0 < . 1:1(0) ack 4001 win 257
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 4001:8001(4000) ack 1
+ +0 < . 1:1(0) ack 8001 win 257
+
+// receive only one EPOLLERR for the two sends above.
+ +0 epoll_wait(5, {events=EPOLLERR|EPOLLOUT, fd=4}, 1, 0) = 1
+ +0 epoll_wait(5, {events=0, ptr=0}, 1, 0) = 0
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 8001:12001(4000) ack 1
+ +0 < . 1:1(0) ack 12001 win 257
+
+// receive only one EPOLLERR for the third send above.
+ +0 epoll_wait(5, {events=EPOLLERR|EPOLLOUT, fd=4}, 1, 0) = 1
+ +0 epoll_wait(5, {events=0, ptr=0}, 1, 0) = 0
+
+ +0 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=2}}
+ ]}, MSG_ERRQUEUE) = 0
new file mode 100644
@@ -0,0 +1,63 @@
+// SPDX-License-Identifier: GPL-2.0
+// epoll zerocopy test:
+//
+// EPOLLERR is known to be not edge-triggered unlike EPOLLIN and EPOLLOUT but
+// it is not level-triggered either. this tests verify that the same behavior is
+// maintained when we have EPOLLEXCLUSIVE.
+//
+// fire two sends with MSG_ZEROCOPY and receive the acks. confirm that EPOLLERR
+// is correctly fired only once, when EPOLLET is set. send another packet with
+// MSG_ZEROCOPY. confirm that EPOLLERR is correctly fired again only once.
+`./defaults.sh`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+ +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+ +0 bind(3, ..., ...) = 0
+ +0 listen(3, 1) = 0
+
+ +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+ +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
+ +0 < . 1:1(0) ack 1 win 257
+
+ +0 accept(3, ..., ...) = 4
+
+ +0 fcntl(4, F_SETFL, O_RDWR|O_NONBLOCK) = 0
+ +0 epoll_create(1) = 5
+ +0 epoll_ctl(5, EPOLL_CTL_ADD, 4,
+ {events=EPOLLOUT|EPOLLET|EPOLLEXCLUSIVE, fd=4}) = 0
+ +0 epoll_wait(5, {events=EPOLLOUT, fd=4}, 1, 0) = 1
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 1:4001(4000) ack 1
+ +0 < . 1:1(0) ack 4001 win 257
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 4001:8001(4000) ack 1
+ +0 < . 1:1(0) ack 8001 win 257
+
+// receive only one EPOLLERR for the two sends above.
+ +0 epoll_wait(5, {events=EPOLLERR|EPOLLOUT, fd=4}, 1, 0) = 1
+ +0 epoll_wait(5, {events=0, ptr=0}, 1, 0) = 0
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 8001:12001(4000) ack 1
+ +0 < . 1:1(0) ack 12001 win 257
+
+// receive only one EPOLLERR for the third send above.
+ +0 epoll_wait(5, {events=EPOLLERR|EPOLLOUT, fd=4}, 1, 0) = 1
+ +0 epoll_wait(5, {events=0, ptr=0}, 1, 0) = 0
+
+ +0 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=2}}
+ ]}, MSG_ERRQUEUE) = 0
new file mode 100644
@@ -0,0 +1,66 @@
+// SPDX-License-Identifier: GPL-2.0
+// epoll zerocopy test:
+//
+// This is a test to confirm that EPOLLERR is only fired once for an FD when
+// EPOLLONESHOT is set.
+//
+// fire two sends with MSG_ZEROCOPY and receive the acks. confirm that EPOLLERR
+// is correctly fired only once, when EPOLLONESHOT is set. send another packet
+// with MSG_ZEROCOPY. confirm that EPOLLERR is not fired. Rearm the FD and
+// confirm that EPOLLERR is correctly set.
+`./defaults.sh`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+ +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+ +0 bind(3, ..., ...) = 0
+ +0 listen(3, 1) = 0
+
+ +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+ +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
+ +0 < . 1:1(0) ack 1 win 257
+
+ +0 accept(3, ..., ...) = 4
+
+ +0 fcntl(4, F_SETFL, O_RDWR|O_NONBLOCK) = 0
+ +0 epoll_create(1) = 5
+ +0 epoll_ctl(5, EPOLL_CTL_ADD, 4,
+ {events=EPOLLOUT|EPOLLET|EPOLLONESHOT, fd=4}) = 0
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 1:4001(4000) ack 1
+ +0 < . 1:1(0) ack 4001 win 257
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 4001:8001(4000) ack 1
+ +0 < . 1:1(0) ack 8001 win 257
+
+// receive only one EPOLLERR for the two sends above.
+ +0 epoll_wait(5, {events=EPOLLERR|EPOLLOUT, fd=4}, 1, 0) = 1
+ +0 epoll_wait(5, {events=0, ptr=0}, 1, 0) = 0
+
+ +0 send(4, ..., 4000, MSG_ZEROCOPY) = 4000
+ +0 > P. 8001:12001(4000) ack 1
+ +0 < . 1:1(0) ack 12001 win 257
+
+// receive no EPOLLERR for the third send above.
+ +0 epoll_wait(5, {events=0, ptr=0}, 1, 0) = 0
+
+// rearm the FD and verify the EPOLLERR is fired again.
+ +0 epoll_ctl(5, EPOLL_CTL_MOD, 4, {events=EPOLLOUT|EPOLLONESHOT, fd=4}) = 0
+ +0 epoll_wait(5, {events=EPOLLERR|EPOLLOUT, fd=4}, 1, 0) = 1
+ +0 epoll_wait(5, {events=0, ptr=0}, 1, 0) = 0
+
+ +0 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=2}}
+ ]}, MSG_ERRQUEUE) = 0
new file mode 100644
@@ -0,0 +1,56 @@
+// SPDX-License-Identifier: GPL-2.0
+// Fastopen client zerocopy test:
+//
+// send data with MSG_FASTOPEN | MSG_ZEROCOPY and verify that the
+// kernel returns the notification ID.
+//
+// Fastopen requires a stored cookie. Create two sockets. The first
+// one will have no data in the initial send. On return 0 the
+// zerocopy notification counter is not incremented. Verify this too.
+
+`./defaults.sh`
+
+// Send a FastOpen request, no cookie yet so no data in SYN
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0
+ +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+ +0 sendto(3, ..., 500, MSG_FASTOPEN|MSG_ZEROCOPY, ..., ...) = -1 EINPROGRESS (Operation now in progress)
+ +0 > S 0:0(0) <mss 1460,sackOK,TS val 1000 ecr 0,nop,wscale 8,FO,nop,nop>
+ +.01 < S. 123:123(0) ack 1 win 14600 <mss 940,TS val 2000 ecr 1000,sackOK,nop,wscale 6, FO abcd1234,nop,nop>
+ +0 > . 1:1(0) ack 1 <nop,nop,TS val 1001 ecr 2000>
+
+// Read from error queue: no zerocopy notification
+ +1 recvmsg(3, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[]}, MSG_ERRQUEUE) = -1 EAGAIN (Resource temporarily unavailable)
+
+ +.01 close(3) = 0
+ +0 > F. 1:1(0) ack 1 <nop,nop,TS val 1002 ecr 2000>
+ +.01 < F. 1:1(0) ack 2 win 92 <nop,nop,TS val 2001 ecr 1002>
+ +0 > . 2:2(0) ack 2 <nop,nop,TS val 1003 ecr 2001>
+
+// Send another Fastopen request, now SYN will have data
+ +.07 `sysctl -q net.ipv4.tcp_timestamps=0`
+ +.1 socket(..., SOCK_STREAM, IPPROTO_TCP) = 5
+ +0 fcntl(5, F_SETFL, O_RDWR|O_NONBLOCK) = 0
+ +0 setsockopt(5, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+ +0 sendto(5, ..., 500, MSG_FASTOPEN|MSG_ZEROCOPY, ..., ...) = 500
+ +0 > S 0:500(500) <mss 1460,nop,nop,sackOK,nop,wscale 8,FO abcd1234,nop,nop>
+ +.05 < S. 5678:5678(0) ack 501 win 14600 <mss 1460,nop,nop,sackOK,nop,wscale 6>
+ +0 > . 501:501(0) ack 1
+
+// Read from error queue: now has first zerocopy notification
+ +0.5 recvmsg(5, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=0}}
+ ]}, MSG_ERRQUEUE) = 0
new file mode 100644
@@ -0,0 +1,44 @@
+// SPDX-License-Identifier: GPL-2.0
+// Fastopen server zerocopy test:
+//
+// send data with MSG_FASTOPEN | MSG_ZEROCOPY and verify that the
+// kernel returns the notification ID.
+
+`./defaults.sh
+ ./set_sysctls.py /proc/sys/net/ipv4/tcp_fastopen=0x207`
+
+// Set up a TFO server listening socket.
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +.1 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+ +0 bind(3, ..., ...) = 0
+ +0 listen(3, 1) = 0
+ +0 setsockopt(3, SOL_TCP, TCP_FASTOPEN, [2], 4) = 0
+ +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+
+// Client sends a SYN with data.
+ +.1 < S 0:1000(1000) win 32792 <mss 1460,sackOK,nop,nop>
+ +0 > S. 0:0(0) ack 1001 <mss 1460,nop,nop,sackOK>
+
+// Server accepts and replies with data.
++.005 accept(3, ..., ...) = 4
+ +0 read(4, ..., 1024) = 1000
+ +0 sendto(4, ..., 1000, MSG_ZEROCOPY, ..., ...) = 1000
+ +0 > P. 1:1001(1000) ack 1001
+ +.05 < . 1001:1001(0) ack 1001 win 32792
+
+// Read from error queue: now has first zerocopy notification
+ +0.1 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=0}}
+ ]}, MSG_ERRQUEUE) = 0
+
+`/tmp/sysctl_restore_${PPID}.sh`
new file mode 100644
@@ -0,0 +1,119 @@
+// SPDX-License-Identifier: GPL-2.0
+// tcp_MAX_SKB_FRAGS test
+//
+// Verify that sending an iovec of tcp_MAX_SKB_FRAGS + 1 elements will
+// 1) fit in a single packet without zerocopy
+// 2) spill over into a second packet with zerocopy,
+// because each iovec element becomes a frag
+// 3) the PSH bit is set on an skb when it runs out of fragments
+
+`./defaults.sh`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+ +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+
+ // Each pinned zerocopy page is fully accounted to skb->truesize.
+ // This test generates a worst case packet with each frag storing
+ // one byte, but increasing truesize with a page (64KB on PPC).
+ +0 setsockopt(3, SOL_SOCKET, SO_SNDBUF, [2000000], 4) = 0
+
+ +0 bind(3, ..., ...) = 0
+ +0 listen(3, 1) = 0
+
+ +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+ +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
+ +0 < . 1:1(0) ack 1 win 257
+
+ +0 accept(3, ..., ...) = 4
+
+ // send an iov of 18 elements: just becomes a linear skb
+ +0 sendmsg(4, {msg_name(...)=...,
+ msg_iov(18)=[{..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}],
+ msg_flags=0}, 0) = 18
+
+ +0 > P. 1:19(18) ack 1
+ +0 < . 1:1(0) ack 19 win 257
+
+ // send a zerocopy iov of 18 elements:
+ +1 sendmsg(4, {msg_name(...)=...,
+ msg_iov(18)=[{..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}],
+ msg_flags=0}, MSG_ZEROCOPY) = 18
+
+ // verify that it is split in one skb of 17 frags + 1 of 1 frag
+ // verify that both have the PSH bit set
+ +0 > P. 19:36(17) ack 1
+ +0 < . 1:1(0) ack 36 win 257
+
+ +0 > P. 36:37(1) ack 1
+ +0 < . 1:1(0) ack 37 win 257
+
+ +1 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=0}}
+ ]}, MSG_ERRQUEUE) = 0
+
+ // send a zerocopy iov of 64 elements:
+ +0 sendmsg(4, {msg_name(...)=...,
+ msg_iov(64)=[{..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1},
+ {..., 1}, {..., 1}, {..., 1}, {..., 1}],
+ msg_flags=0}, MSG_ZEROCOPY) = 64
+
+ // verify that it is split in skbs with 17 frags
+ +0 > P. 37:54(17) ack 1
+ +0 < . 1:1(0) ack 54 win 257
+
+ +0 > P. 54:71(17) ack 1
+ +0 < . 1:1(0) ack 71 win 257
+
+ +0 > P. 71:88(17) ack 1
+ +0 < . 1:1(0) ack 88 win 257
+
+ +0 > P. 88:101(13) ack 1
+ +0 < . 1:1(0) ack 101 win 257
+
+ +1 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=1,
+ ee_data=1}}
+ ]}, MSG_ERRQUEUE) = 0
+
new file mode 100644
@@ -0,0 +1,58 @@
+// SPDX-License-Identifier: GPL-2.0
+// small packet zerocopy test:
+//
+// verify that SO_EE_CODE_ZEROCOPY_COPIED is set on zerocopy
+// packets of all sizes, including the smallest payload, 1B.
+
+`./defaults.sh`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+ +0 setsockopt(3, SOL_SOCKET, SO_ZEROCOPY, [1], 4) = 0
+ +0 bind(3, ..., ...) = 0
+ +0 listen(3, 1) = 0
+
+ +0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+ +0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 8>
+ +0 < . 1:1(0) ack 1 win 257
+
+ +0 accept(3, ..., ...) = 4
+
+ // send 1B
+ +0 send(4, ..., 1, MSG_ZEROCOPY) = 1
+ +0 > P. 1:2(1) ack 1
+ +0 < . 1:1(0) ack 2 win 257
+
+ +1 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=0,
+ ee_data=0}}
+ ]}, MSG_ERRQUEUE) = 0
+
+ // send 1B again
+ +0 send(4, ..., 1, MSG_ZEROCOPY) = 1
+ +0 > P. 2:3(1) ack 1
+ +0 < . 1:1(0) ack 3 win 257
+
+ +1 recvmsg(4, {msg_name(...)=...,
+ msg_iov(1)=[{...,0}],
+ msg_flags=MSG_ERRQUEUE,
+ msg_control=[
+ {cmsg_level=CMSG_LEVEL_IP,
+ cmsg_type=CMSG_TYPE_RECVERR,
+ cmsg_data={ee_errno=0,
+ ee_origin=SO_EE_ORIGIN_ZEROCOPY,
+ ee_type=0,
+ ee_code=SO_EE_CODE_ZEROCOPY_COPIED,
+ ee_info=1,
+ ee_data=1}}
+ ]}, MSG_ERRQUEUE) = 0
+