The delayed ACK logic previously sent an ACK for at least every second
received segment (hard-coded threshold of 2 per RFC 1122). Add the
NET_TCP_ACK_FREQUENCY Kconfig option (range 1-255, default 2) to make
this threshold configurable at build time.
The delayed ACK timer still forces an ACK after at most 0.5 seconds, so
RFC 1122 timing compliance is preserved regardless of the configured
threshold. The default value of 2 keeps the exact current behavior:
the new condition rx_unackseg >= FREQ - 1 is equivalent to the previous
rx_unackseg > 0, and the counter increment degenerates to the previous
rx_unackseg = 1 assignment.
Signed-off-by: zhekunren <zhekunren@qq.com>
Assisted-by: GLM-5.2 <noreply@z.ai>
The condition work_available(&conn->work) && tx_unacked != 0
prevented tcp_update_retrantimer from being called when the work
queue was still busy, leaving conn->timer stale or zero on
subsequent sends. This caused the RTT estimation to compute a
false RTT (m = rto - 0 = rto), creating a positive feedback loop
that inflated the RTO to extreme values (e.g., 232 half-seconds
= ~116 seconds).
Fix: remove the work_available check so that tcp_update_retrantimer
is always called when there is unacknowledged data. The decision to
re-queue the work is handled internally by tcp_update_timer.
Signed-off-by: zhekunren <zhekunren@qq.com>
tcp_start_monitor() is called from accept() (net/inet/inet_sockif.c) for
each newly accepted connection. When the peer had already closed the
connection before accept() ran, the monitor takes an early-return path so
that any read-ahead data buffered on the connection can still be drained;
it returns OK in that case. accept() (net/socket/accept.c) then marks the
new socket _SF_CONNECTED unconditionally.
If the peer aborts the connection with an RST immediately after the
three-way handshake completes (for example any close with SO_LINGER
{1, 0}), the connection is moved to TCP_CLOSED with no buffered data, yet
accept() still hands back a socket that reports _SS_ISCONNECTED. A
subsequent blocking send() on that socket passes the connected check,
registers a send callback and waits on its semaphore forever: the only
TCP_ABORT event was delivered before the callback existed, and no further
ACK, POLL or disconnect event is generated for a closed connection, so the
waiter is never woken.
Any server that writes before reading can hit this; the telnet daemon
(netutils/telnetd) is one example, where the accepted session task blocks
in send() and never completes.
Only return OK from the already-closed path when there is actually
read-ahead data to drain. Otherwise the connection is dead, so fall
through to the -ENOTCONN return: accept() then fails cleanly instead of
handing back a socket wedged on a connection that will never make progress.
The graceful-close-with-pending-data case (the reason the OK path exists)
is preserved by the conn->readahead check.
Signed-off-by: Ricard Rosson <ricard@groundbits.com>
Assisted-by: Claude (Anthropic Claude Code)
POSIX leaves the signedness of time_t and clock_t unspecified, but
mainstream implementations (Linux glibc/musl, the BSDs, macOS, RTEMS,
Zephyr's POSIX layer, Windows _time64) expose time_t as signed 64-bit.
NuttX has historically used uint64_t only because it was tied to the
CONFIG_SYSTEM_TIME64 knob; with that gone, switch:
time_t : uint64_t -> int64_t
clock_t : uint64_t -> int64_t
CLOCK_MAX: UINT64_MAX -> INT64_MAX
This lets (time_t)-1 sentinels, negative tick deltas, and host-side
headers behave as on every other POSIX system without source churn.
Headers updated:
- include/sys/types.h, include/limits.h, include/nuttx/clock.h
- include/nuttx/fs/hostfs.h (nuttx_time_t alias)
- include/nuttx/{mqueue.h,wdog.h,wqueue.h,timers/clkcnt.h}
Because clock_t is now signed 64-bit, the NuttX-internal sclock_t
alias becomes redundant: every sclock_t/SCLOCK_MAX use is folded
back to clock_t/CLOCK_MAX (notably in sched/wdog, sched/mqueue,
sched/sched, sched/clock, sched/timer, libs/libc/time, fs/vfs and
the drivers/arch consumers below).
Tick/period constants (NSEC_PER_SEC, USEC_PER_SEC, MSEC_PER_SEC,
SEC_PER_MIN, ...) in include/nuttx/clock.h are retyped from "long"
literals to INT64_C(...) so that 64-bit arithmetic no longer
depends on the host's long width.
Strip now-redundant (time_t)/(clock_t)/(unsigned long) casts and
unsigned-only branches across the tree:
- arch RTC / oneshot / tickless lowerhalfs:
arm: cxd56xx, efm32, imxrt, lc823450, max326xx, sam34, sama5,
samd5e5, samv7, stm32, stm32f7, stm32h7, stm32l4, stm32wb,
xmc4
mips: pic32mz sparc: bm3803 x86_64: intel64
risc-v/xtensa: espressif (esp_i2c[_slave], esp_rtc,
esp32c3{_i2c,_rtc,_wifi_adapter}, esp32{,s2,s3}_*),
mpfs_perf
- drivers: audio/tone, input/aw86225, power/pm/{activity,
stability}_governor, rpmsg/rpmsg_ping,
timers/{ds3231,mcp794xx,pcf85263,rx8010},
wireless/ieee80211/bcm43xxx, wireless/spirit/spirit_spi
- core: fs/vfs/{fs_poll,fs_timerfd}, mm/iob/iob_alloc,
libs/libc/{netdb/lib_dnscache,time/{lib_calendar2utc,
lib_time}}, net/icmp/icmp_pmtu, net/icmpv6/icmpv6_pmtu,
net/ipfrag, net/tcp/{tcp.h,tcp_timer},
net/utils/net_snoop, net/mld/mld_query (drop the now-dead
mld_mrc2mrd helper since signed math handles it directly),
sched/clock/{clock,clock_initialize},
sched/sched/{sched_profil,sched_setparam,sched_setscheduler},
sched/pthread/pthread_create,
sched/wdog/{wd_gettime,wd_start,wdog.h},
sched/timer/timer_gettime, sched/mqueue/*
Flip the few in-tree printf format strings that assumed an
unsigned 64-bit tv_sec:
* drivers/rpmsg/rpmsg_ping.c PRIu64 -> PRId64
* arch/xtensa/src/esp32{,s2,s3}/esp32*_oneshot_lowerhalf.c
PRIu32 (already wrong) -> PRId64
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
debug.h is a NuttX-specific, non-POSIX header. Placing it in the
top-level include/ directory creates naming conflicts with external
projects that define their own debug.h.
This commit moves the canonical header to include/nuttx/debug.h,
following the NuttX convention for non-POSIX/non-standard headers,
and updates all in-tree references.
A backward-compatibility shim is left at include/debug.h that
emits a deprecation #warning and re-includes <nuttx/debug.h>,
allowing out-of-tree code to continue building while migrating.
Signed-off-by: Piyush Patle <piyushpatle228@gmail.com>
Fix 269 occurrences of duplicate "the" word typo found in 209 files
across source code, header files, and configuration.
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
This reverts commit fa652f9c24.
When testing iperf on boards such as ESP32-C6, ESP32-C3, and ESP32,
blocking issues are encountered, so the patch is reverted.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
The main content of this submission is to limit both the TX/RX buffers
of TCP/UDP to throttled IOBs, avoiding impacts on the sending and
receiving of control-type messages.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
The RFC requires sending an TCP_RST packet in this scenario, so to better
comply with the standard definition, the sending of TCP_RST is added.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
urgent data needs to be treated as normal data when
CONFIG_NET_TCPURGDATA disable. some test sets will verify this
behavior, correct the processing logic.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
optimize the current code format according to the previous net_xxx_wait
implementation to reduce multiple calls of similar code
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
reduce the execution consumption of irrelevant code
testing the TX rates of TCP and UDP based on the Infineon board can
increase them by 13%.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
According rfc793 p69, from state SYN-RECEIVED to state TIME-WAIT, if the incoming segment is not acceptable, an acknowledgment should be sent in reply (unless the RST bit is set):
<SEQ=SND.NXT><ACK=RCV.NXT><CTL=ACK>
Signed-off-by: wenquan1 <wenquan1@xiaomi.com>
According rfc793 p69, In SYN-RCVD state, if the incoming segment is not acceptable, an acknowledgment should be sent in reply (unless the RST bit is set):
<SEQ=SND.NXT><ACK=RCV.NXT><CTL=ACK>
Signed-off-by: wenquan1 <wenquan1@xiaomi.com>
According to RFC793, the flags of TCP reply packet should be as follows:
1. `RST` if flags of request packet has `ACK`
2. `RST|ACK` if flags of request packet has no `ACK`, at the same time, the sequence should be set to zero.
Signed-off-by: wenquan1 <wenquan1@xiaomi.com>
in multi-core cpu, tcp_free and tcp_timer_expiry->tcp_timer will work in parallel, after tcp_free call work_cancle, tcp->timer will call tcp_update_timer to re-add work to worker queue, then tcp_free free conn, in this condition, it will result use after free.
Signed-off-by: wenquan1 <wenquan1@xiaomi.com>
According RFC793,p72~p73 In states from ESTABLISHED to LASTACK:"If the ACK acks something not yet sent (SEG.ACK > SND.NXT) then send an ACK, drop the segment, and return"
Signed-off-by: wenquan1 <wenquan1@xiaomi.com>
in the current process, if tcpstate is at TCP_FIN_WAIT_1, TCP_CLOSING,
and TCP_LAST_ACK, and the peer does not reply TCP_ACK, it will be
retransmitted permanently. to avoid continuous occupation of resources,
now when the retransmission equals TCP_MAXRTX, a reset message will be
sent and the conn resources will be reclaimed in tcp_close_work.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
since we added TCP_TXCLOSE and TCP_RXCLOSE, only RST-like exception
flows now trigger TCP_CLOSE events, so replace the remaining
TCP_CLOSE with TCP_ABORT.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
notify the local waiting recv process; otherwise, after the peer TX
is closed, the wait of the recv process will continue to wait.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
shutdown should send TCP_FIN packet.
close should send TCP_RST packet when the data in readahead has not been
read and NEW_DATA has arrived.
send TCP_FIN packet when in other cases.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
after the listener fd is closed, the half-open connections associated
with the listener are not reclaimed. According to the protocol standard,
they should be reclaimed in a timely manner.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
UDP and PKT will hold the dev lock and conn lock when calling back the
protocol stack from the network dirver. At this time, if the poll_notify
notifies usrsock_server and executes poll_teardown, a conn double lock
will appear. so we changed the lock to a recursive lock
sched_dumpstack
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/Libs/libc/sched/sched_dumpstack.c:71
__assert
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/libs/libc/assert/lib_assert.c:49
nxsem_wait_slow
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/sched/semaphore/sem_wait.c:102 (discriminator 1)
nxmutex_lock
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/include/nuttx/mutex.h:514
udp_pollteardown
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/net/udp/udp_netpoll.c:278
usrsock_rpmsg_poll_setup
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/drivers/usrsock/usrsock_rpmsg_server.c:1232
usrsock_rpmsg_poll_cb
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/drivers/usrsock/usrsock_rpmsg_server.c:1299
poll_notify
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/fs/vfs/fs_poll.c:296
udp_poll_eventhandler
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/net/udp/udp_netpoll.c:105
devif_conn_event
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/net/devif/devif_callback.c:481
udp_callback
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/net/udp/udp_callback.c:324
udp_input_conn
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/net/udp/udp_input.c:150
ipv4_in
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/net/devif/ipv4_input.c:425
ipv4_input
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/net/devif/ipv4_input.c:533
tun_net_receive_tun
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/drivers/net/tun.c:574
tun_net_receive_tun
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/drivers/net/tun.c:574
file_writev_compat
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/fs/vfs/fs_write.c:89
rpmsgdev_write_handler
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/drivers/misc/rpmsgdev_server.c:354
rpmsg_virtio_process_rx_buffer
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/drivers/rpmsg/rpmsg_virtio.c:256
rpmsg_virtio_rx_worker
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/drivers/rpmsg/rpmsg_virtio.c:291
work_dispatch
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/sched/wqueue/kwork_thread.c:233
work_thread
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/sched/wqueue/kwork_thread.c:293
nxtask_start
/home/work/ssd1/workspace/MiRTOS-062-dev-system-CI/nuttx/sched/task/task_start.c:99
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
add net_lock in the tcp_callback_cleanup & Set the devif_callback_s of
tcp_callback_s to a two-dimensional pointer.
Signed-off-by: wangchen <wangchen41@xiaomi.com>
use netdev_list_lock to protect all network card traversal, registration,
and deregistration operations.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
the maximum retransmission interval allowed for car projects
cannot exceed 6 seconds, and allows for fixed retransmission intervals.
according to the previous algorithm, the retransmission interval
may be 0.5 * 2 ^ 4 = 8 seconds, which does not meet the requirements.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
CLOSE-WAIT - represents waiting for a connection termination request
from the local user.
TCP A TCP B
1. ESTABLISHED ESTABLISHED
2. (Close)
FIN-WAIT-1 --> <SEQ=100><ACK=300><CTL=FIN,ACK> --> CLOSE-WAIT
3. FIN-WAIT-2 <-- <SEQ=300><ACK=101><CTL=ACK> <-- CLOSE-WAIT
4. (Close)
TIME-WAIT <-- <SEQ=300><ACK=101><CTL=FIN,ACK> <-- LAST-ACK
5. TIME-WAIT --> <SEQ=101><ACK=301><CTL=ACK> --> CLOSED
6. (2 MSL)
CLOSED
in the current state, we can continue to send data until the user
calls shutdown or close, then directly enter the TCP_LAST_ACK state
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
when dev is down, tcp_pollsetup returns the value 'OK' and sets the connection status to 'down' and eventset to POLLERR & POLLUP
Signed-off-by: wangchen <wangchen41@xiaomi.com>
After first connection refused happend,The user use the same connection failed,becasue the tcpstateflags is not TCP_ALLOCATED
Signed-off-by: wangchen <wangchen41@xiaomi.com>
When triggering TCP retransmission through a timer, conn->unacked_q will save more packets, which may result in sndseq being equal to ackseq. Previous implementations may cause conn->nrtx to not be cleared, leading to an increase in conn->rto errors and a longer timer time
Signed-off-by: wangchen <wangchen41@xiaomi.com>
The modification is to solve the problem that the opposite end cannot receive the retransmission message, and the protocol stack no longer sends retransmission messages
Signed-off-by: wangchen <wangchen41@xiaomi.com>
