Previously "ifconfig eth0 up" fell through to the host-IP parsing
branch, where inet_addr("up") returns INADDR_NONE, so the address was
silently set to 255.255.255.255 and the interface was never brought
up. Users had to run the separate "ifup"/"ifdown" commands.
Recognize the "up" and "down" tokens explicitly and apply them once
the rest of the configuration is in place, so up/down is just another
keyword in the argument list as it is in Linux ifconfig.
A bare "ifconfig <iface> up|down" needs no special case: the preceding
patch made the address, netmask, gateway, DNS and DHCP blocks check
whether they were asked for, so the command falls through them without
touching anything.
The up/down handling calls netlib_ifup()/netlib_ifdown() from netlib, so
it stays available regardless of CONFIG_NSH_DISABLE_IFUPDOWN (which only
strips the standalone ifup/ifdown commands).
Assisted-by: GitHubCopilot:claude-opus-5
Signed-off-by: wangjianyu3 <wangjianyu3@xiaomi.com>
cmd_ifconfig() returns early for argc <= 2, so by the time the argument
loop is reached argc > 2 always holds. Remove the dead condition and
unindent the loop.
No functional change.
Assisted-by: GitHubCopilot:claude-opus-5
Signed-off-by: wangjianyu3 <wangjianyu3@xiaomi.com>
cmd_ifconfig() pushed every setting to the device on each invocation,
whether or not the command line carried it, and stopped parsing as soon
as it saw "mtu". So a command meant to touch one thing quietly rewrote
the rest of the interface configuration:
ifconfig eth0 hw 00:11:22:33:44:55 # clears the address, netmask,
# gateway and resolver, and kicks
# off a DHCP request
ifconfig eth0 mtu 1500 # same, plus the default route
ifconfig eth0 mtu 1500 dns 8.8.8.8 # dns silently ignored
In detail:
- the address was written unconditionally. For IPv6 that read the
uninitialized addr6 off the stack and pushed whatever it held into
the device; for IPv4 it forced 0.0.0.0.
- the netmask fell back to a hard-coded 255.255.255.0, or to
ffff:ffff:ffff:ffff:: for IPv6.
- the IPv4 gateway was always written, unlike the IPv6 one, so it fell
back to INADDR_ANY.
- the resolver fell back to that gateway, i.e. to 0.0.0.0.
- the DHCP client was started whenever gip was left at zero instead of
when "dhcp" was asked for, and gip only becomes non-zero when an
address or a gateway is parsed.
- the "mtu" branch returned as soon as netlib_set_mtu() succeeded, so
every argument behind it was dropped.
Write each setting only when the user provided it, or when an address
is being assigned and the setting belongs to it. Keeping the address
case is deliberate: a freshly assigned address still needs a mask and
a route, so dropping the derived netmask and "x.x.x.1" gateway there
would be a regression of its own. DHCP now triggers on the "dhcp"
keyword, which is the only way to request it. And with nothing left
to clobber, the "mtu" branch no longer has to bail out early.
This is how ifconfig behaves elsewhere: Linux net-tools walks the
argument vector in a single loop where every keyword handler ends in
"continue" and none of them returns, so each keyword is an independent,
idempotent operation applied in the order it was written.
As a side effect mip is now only read along the paths that assign it,
since the IPv4 gateway fallback only derives an address from mip when
gip came from hostip.
Assisted-by: GitHubCopilot:claude-opus-5
Signed-off-by: wangjianyu3 <wangjianyu3@xiaomi.com>
A bare `boot` (no argument, argc == 1) passes argv[1] == NULL.
nsh_getfullpath(NULL) returns strdup(g_home) == "/" instead of NULL,
which turned the default-boot path (NULL -> board default image)
into "/". board_boot_image("/") then failed with -EINVAL (-22),
preventing the board from booting.
Guard the nsh_getfullpath() call so NULL passes through unchanged,
restoring the original default-boot behavior while still resolving
relative paths when an argument is given.
This fixes the regression introduced by the relative-path support
```
commit fabafbc361 (origin/master, origin/HEAD)
Author: Junbo Zheng <zhengjunbo1@xiaomi.com>
Date: Fri Jul 24 23:47:44 2026 +0800
nshlib: add relative image path support in boot command
cmd_boot passed the image path straight to boardctl(), which resolves
it in a context that does not inherit the NSH shell cwd, so relative
paths failed and only absolute paths worked. Use nsh_getfullpath() to
resolve relative paths against the cwd before calling boardctl().
Signed-off-by: Junbo Zheng <zhengjunbo1@xiaomi.com>
```
Signed-off-by: Junbo Zheng <zhengjunbo1@xiaomi.com>
Both read a keyboard as a stream of characters: NSH uses a USB HID
keyboard for stdin, and the raw mode of the Microwindows keyboard driver
decodes the stream with the codec. That stream now comes from
INPUT_KEYBOARD_BYTESTREAM rather than from the USB HID driver itself.
Say so, so that Kconfig refuses a combination that cannot work instead of
letting the application read events and treat them as text. No in-tree
configuration selects either option.
Signed-off-by: Jorge Guzman <jorge.gzm@gmail.com>
Require a cryptodev backend for FSUTILS_PASSWD and tie NSH console/telnet
login to ROMFS passwd generation when ETC_ROMFS is enabled.
Fixes#19573
Signed-off-by: Abhishek Mishra <mishra.abhishek2808@gmail.com>
Use getresuid() and getresgid() for Linux-style id output with name
resolution and a groups= list based on the effective GID.
Signed-off-by: Abhishek Mishra <mishra.abhishek2808@gmail.com>
cmd_boot passed the image path straight to boardctl(), which resolves
it in a context that does not inherit the NSH shell cwd, so relative
paths failed and only absolute paths worked. Use nsh_getfullpath() to
resolve relative paths against the cwd before calling boardctl().
Signed-off-by: Junbo Zheng <zhengjunbo1@xiaomi.com>
Add a du command to NSH that recursively summarizes the size of each
path argument in 1K-blocks, or human-readable form with -h.
Supports -s (summary only), -a (all files), -d N (max-depth), and -h,
matching GNU du semantics.
Testing:
Built and ran on sim:nsh with:
```bash
cmake -B out/nuttx_sim_nsh -S nuttx -DBOARD_CONFIG=sim:nsh -GNinja
ninja -C out/nuttx_sim_nsh
./out/nuttx_sim_nsh/nuttx
nsh> du
397449 /data/test/elf
71993 /data/test/coredump
5 /data/test/log2
3251 /data/test/log1
472700 /data/test
nsh> du -h
388.1M /data/test/elf
70.3M /data/test/coredump
4.1K /data/test/log2
3.1M /data/test/log1
461.6M /data/test
```
Host Ubuntu22.04 du on the same directory:
```bash
$ du
397456 ./elf
72000 ./coredump
8 ./log2
3252 ./log1
472720 .
$ du -h
389M ./elf
71M ./coredump
8.0K ./log2
3.2M ./log1
462M .
```
Signed-off-by: Junbo Zheng <zhengjunbo1@xiaomi.com>
readline cached the current prompt only when TAB completion was enabled,
even though line-editing redraws also depend on that cached prompt.
With TAB completion disabled, full-line redraws erased the prompt and
repainted only the command buffer. Make the prompt cache available
whenever line editing is enabled.
Signed-off-by: Alan C. Assis <acassis@gmail.com>
This commit adds support to nsh> command line editing. This is one
of the most missing feature of the NuttShell. If you typed a long
command line and make a mistake you need to press Backspace and
remove everything until reach that typo. Only the basic editing
feature is enabled by default (left/right keys movement, Home/End
to move to the beginning or ending of the command line).
More advanced features are available when CONFIG_READLINE_EDIT_EMACS
is selected. It enables the Emacs-style control keys (Ctrl+A/B/D/E/F/K/U/W)
that allow more flexible line editing.
Other more advanced feature is enabled when CONFIG_READLINE_CMD_HISTORY
and CONFIG_READLINE_EDIT_EMACS_REVERSE_SEARCH are enabled. It allows the
user to press Ctrl+R to do reverse search in the command line history.
Signed-off-by: Alan C. Assis <acassis@gmail.com>
Assisted-by: Claude Code
Report password policy failures clearly from useradd and passwd when a
password does not meet complexity requirements.
Signed-off-by: Abhishek Mishra <mishra.abhishek2808@gmail.com>
Migrate passwd encrypt/verify to PBKDF2 modular crypt format using
kernel cryptodev (CRYPTO_PBKDF2_HMAC_SHA256 via /dev/crypto). Add
passwd_pbkdf2 wrapper, base64url helpers, complexity validation, and
pbkdf2_test for RFC 6070 vector coverage. FSUTILS_PASSWD selects
CRYPTO, ALLOW_BSD_COMPONENTS, and CRYPTO_CRYPTODEV so existing sim
defconfigs keep building. Change NSH_LOGIN_USERNAME default to root and
remove fixed-login password defaults.
BREAKING CHANGE: TEA-encoded /etc/passwd entries no longer verify.
Regenerate each entry after upgrading. Pair with the nuttx host mkpasswd
changes in apache/nuttx#19209. Boards must enable the appropriate
software or hardware crypto backend for PBKDF2 at runtime. When
CONFIG_NSH_LOGIN_FIXED=y, set CONFIG_NSH_LOGIN_PASSWORD in the board
defconfig or menuconfig; there is no default password.
Signed-off-by: Abhishek Mishra <mishra.abhishek2808@gmail.com>
Print a summary header before the task list in the style of the
Linux top command: uptime, task counts (total/running/sleeping),
CPU busy/idle and memory usage, all gathered via sysinfo().
Refresh the screen in place (cursor home + erase-line per row +
erase-below) instead of clearing the whole screen every cycle,
which removes the flicker.
Poll stdin during the update interval and exit on 'q', ESC or
Ctrl-C. Previously the only exit path was the SIGINT handler,
which requires both CONFIG_ENABLE_ALL_SIGNALS (handler) and
CONFIG_TTY_SIGINT (serial console converts Ctrl-C into SIGINT,
default n); on configurations missing either option top could
not be terminated at all.
Tested on linum-stm32h753bi:nsh.
Signed-off-by: Jorge Guzman <jorge.gzm@gmail.com>
NSH error shouldn't be disabled by default, even when in SMALL
Memory because it will make difficult to discover why things are
not working. It should be disabled manually by experienced dev.
Signed-off-by: Alan C. Assis <acassis@gmail.com>
Fix two potential NULL pointer dereferences in nsh_fscmds.c:
1. fdinfo_callback: asprintf() failure left filepath potentially
NULL, which was then passed to nsh_catfile(). Add early return
on asprintf failure.
2. cmd_cat: malloc(BUFSIZ) for stdin reading was used without
checking the return value. Add NULL check with -ENOMEM return.
Also fix a typo in error message: 'nsh_catfaile' -> 'nsh_catfile'.
Signed-off-by: hanzhijian <hanzhijian@zepp.com>
Fixes#17063
Add modification time display to ls -l long format output.
The timestamp is shown as 'YYYY-MM-DD HH:MM' between the permission
string and the file size, following the ISO long-iso format.
Files with st_mtime == 0 (e.g., procfs, tmpfs pseudo-entries) skip
the timestamp display to avoid showing a meaningless epoch time.
Signed-off-by: hanzhijian <hanzhijian@zepp.com>
The O_RDOK/O_WROK aliases have been removed from fcntl.h. Replace
all remaining O_RDOK usage with O_RDONLY in the apps repository.
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
This commit fixes the previous behavior where an echo with a
single string and its default new line is send as single write().
This issue came from: https://github.com/apache/nuttx-apps/pull/1559
Signed-off-by: Alan C. Assis <acassis@gmail.com>
Integrated SSH daemon authenticating against FSUTILS_PASSWD, with an
ECDSA P-256 host key and an NSH session over a PTY per connection. Built
from the upstream Dropbear tarball (pinned commit) and patched for
NuttX, using Dropbear's bundled libtomcrypt for all crypto. setsid()
(apache/nuttx#19184) and link() now come from NuttX, not local stubs.
Signed-off-by: Felipe Moura <moura.fmo@gmail.com>
When stdout is broken (e.g. PTY master closed), nsh_catfile could spin
forever: write() fails, error logging generates syslog output to
/dev/log, next read() picks it up, write() fails again, ad infinitum.
Two fixes:
- nsh_console: drop the _err() in nsh_consolewrite entirely. The risk
is errno-agnostic (EPIPE / EIO / ENOSPC / ...): any failure logged
here can be re-injected by the syslog backend when OUTFD is bound to
/dev/log. Callers already see the failure via the negative return
value and errno, so on-failure logging at this layer is redundant.
- nsh_fsutils: jump straight from the inner write-failure path to a
single errout: cleanup label instead of using a two-level break +
flag check, so a failed write cannot fall through to another read().
Assisted-by: GitHubCopilot:claude-4.7-opus
Signed-off-by: wangjianyu3 <wangjianyu3@xiaomi.com>
Add missing failure handling for direct strdup() and asprintf() calls where the allocated result is consumed locally before any NULL/error check.
This keeps the scope to the functions named in #1727 and avoids changing pass-through return sites where callers already receive NULL on allocation failure.
Signed-off-by: Nightt <87569709+nightt5879@users.noreply.github.com>
Pass explicit open flags for the NSH script redirect file so the mode argument is used as file creation permissions.
Add the missing mode argument to AIO ostest open() calls that create the test file.
Signed-off-by: Nightt <87569709+nightt5879@users.noreply.github.com>
Now that time_t is unconditionally 64-bit (signed int64_t) and the
struct timespec fields tv_sec / tv_nsec are wide enough on their own,
the explicit (uint64_t)/(int64_t)/(int) casts that used to guard the
multiplications and subtractions in *_us / *_ms / *_ns helpers are no
longer needed. Drop them to keep the timekeeping math readable.
In the same spirit, this commit also normalises the printf-style format
specifiers and casts used to print tv_sec / tv_nsec / tv_usec values.
The prior code was a mix of "%d"/"%u"/"%ld"/"%lu"/"%lld" with matching
(int)/(unsigned long)/(long long) casts; some formats truncated time_t
on 32-bit hosts, others mismatched signedness or width. Replace all
such cases with the portable POSIX-recommended forms:
- tv_sec (time_t, signed, impl-defined width) -> %jd + (intmax_t)
- tv_nsec (long, signed) -> %ld (no cast)
- tv_usec (suseconds_t / long) -> %ld (no cast)
Also drop two stale `(FAR const time_t *)&ts.tv_sec` casts that are
unnecessary now that ts.tv_sec is plain time_t.
Arithmetic-cleanup files (existing scope):
- benchmarks/cyclictest/cyclictest.c: timediff_us()
- benchmarks/sd_bench/sd_bench_main.c: get_time_delta_us()
- examples/oneshot/oneshot_main.c: maxus computation
- examples/watchdog/watchdog_main.c: current_time_ms (x2)
- industry/nxmodbus/nxmb_internal.h: nxmb_util_clock_ms()
- netutils/ntpclient/ntpclient.c: timespec2ntp()
- netutils/ptpd/ptpd.c: ptp_adjtime()
- system/dd/dd_main.c: elapsed accounting
- testing/drivers/drivertest/drivertest_posix_timer.c:
get_timestamp()
- testing/drivers/sd_stress/sd_stress_main.c:get_time_delta()
- testing/sched/getprime/getprime_main.c: elapsed accounting
- testing/sched/pthread_mutex_perf/pthread_mutex_perf.c:
timespec_avg()
Printf-format-fix files (new in this revision):
- examples/adjtime/adjtime_main.c
- examples/charger/charger_main.c
- examples/netpkt/netpkt_ethercat.c
- fsutils/mkfatfs/mkfatfs.c
- graphics/tiff/tiff_initialize.c
- netutils/ptpd/ptpd.c
- nshlib/nsh_timcmds.c
- system/coredump/coredump.c
- system/ptpd/ptpd_main.c
- testing/drivers/drivertest/drivertest_oneshot.c
- testing/mm/kasantest/kasantest.c
- testing/ostest/semtimed.c
- testing/sched/pthread_mutex_perf/pthread_mutex_perf.c
- testing/sched/timerjitter/timerjitter.c
- testing/testsuites/kernel/time/cases/clock_test_clock01.c
- testing/testsuites/kernel/time/cases/clock_test_smoke.c
No behavioural change.
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Add minimal chmod/chown support to NSH using the
existing libc/VFS syscall interfaces.
Supported forms:
- chmod <octal-mode> <path>
- chown <uid>[:gid] <path>
The chown implementation supports the numeric
ownership forms already handled by the underlying
NuttX chown() implementation, including unchanged
uid/gid semantics via omitted fields.
Only numeric permission modes and numeric uid/gid
forms are supported in this initial implementation.
This adds interactive filesystem permission and
ownership management support to NSH and aligns the
shell more closely with standard POSIX environments.
Signed-off-by: Abhishek Mishra <mishra.abhishek2808@gmail.com>
BREAKING: In an effort to simplify board initialization logic for NuttX,
NSH will no longer support architecture initialization. This will happen
during boot via the BOARD_LATE_INITIALIZE option. The boardctl command
BOARDIOC_INIT is also no longer available from user-space.
Quick fix:
Any application relying on BOARDIOC_INIT should now enable
BOARD_LATE_INITIALIZE to have initialization performed by the kernel in
advance of the application running. If control over initialization is
still necessary, BOARDIOC_FINALINIT should be implemented and used.
Boards relying on NSH for initialization should also enable
BOARD_LATE_INITIALIZE instead.
Signed-off-by: Matteo Golin <matteo.golin@gmail.com>
Value zero in tm_isdst means daylight saving time (or summer
time) is not in effect. This is obviously not true for half of the year
in most timezones and keeping it at zero leads to incorrect time
being set. This commit sets tm_isdst to -1, which means the information
is not available and it should be handled according to timezone rules.
Signed-off-by: Michal Lenc <michallenc@seznam.cz>
Replace app-side includes of <debug.h> with <nuttx/debug.h> to use the
header from the NuttX tree explicitly after the header move.
Signed-off-by: Piyush Patle <piyushpatle228@gmail.com>
removes redundant type casts in ARP-related network library functions.
The casts from req.arp_dev (already a char array) to (FAR char *)
are unnecessary and can be safely removed.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
This commit implements stderr redirection in HSN, support
Bash-like syntax for redirecting standard error output.
Support both foreground and background commands.
example:
- nsh> ls noexists 2> err.log
- nsh> ls noexists 2> err.log &
- nsh> ls noexists 2>> err.log
- nsh> sh < /dev/ttyS0 > /dev/ttyS0 2> /dev/ttyS1 &
- nsh> sh < /dev/ttyS0 > /dev/ttyS0 2>&1 &
Signed-off-by: fangpeina <fangpeina@xiaomi.com>
Fix dependency issue when signals are partially or fully enabled
Co-authored-by: guoshichao <guoshichao@xiaomi.com>
Signed-off-by: Chengdong Wang <wangchengdong@lixiang.com>
Now we allow setting a default PCP when creating VLAN like:
`vconfig add iface-name vlan-id [pcp]`
e.g.
`vconfig add eth0 10` will create eth0.10 with VID=10 and no PCP(0)
`vconfig add eth0 10 3` will create eth0.10 with VID=10 and PCP=3
Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
Replace floating-point arithmetic with fixed-point integer math to avoid
linking soft-float library (~2-3KB Flash) when displaying human-readable
file sizes (ls -lh command).
Changes:
- Use integer division and modulo to calculate size components
- Calculate decimal part: (remainder * 10) / unit for one decimal place
- Refactor duplicated code: consolidate GB/MB/KB logic into single path
- Remove CONFIG_HAVE_FLOAT dependency
This eliminates calls to __aeabi_f2d, __aeabi_fmul, __aeabi_i2f and other
ARM EABI floating-point helpers, reducing Flash footprint for systems
compiled with -mfloat-abi=soft.
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
This fixes heap corruption when deleting a folder containing other folders
or files. The issue appeared at commit 131d50ae9d, which removed the
stack-based temporary buffer.
unlink_recursive requires that the path is provided in PATH_MAX sized
buffer. It concatenates sub-folder or file names to the same buffer.
nsh_getfullpath just allocates a buffer using strdup, so there is no room
for concatenating more data to it.
To keep the stack usage smaller, instead of reverting the breaking commit,
allocate the temporary buffer with lib_get_pathbuffer instead.
Signed-off-by: Jukka Laitinen <jukka.laitinen@tii.ae>
ID previously marked as PID was in fact TID. PID was hidden under
GROUP column. This fixes the description. The order of TID, PPID, PID
is also changed to more logical TID, PID, PPID.
Signed-off-by: Michal Lenc <michallenc@seznam.cz>
The NSH exits when a command exits with a non-zero status, even if the "e" flag is not set.
This error does not exist in NSH scripts.
Without this patch:
nsh> sh -c "set -e; mkdir /test; echo $?"
nsh: /test: mkdir failed: 17
nsh> sh -c "set +e; mkdir /test; echo $?"
nsh: /test: mkdir failed: 17
nsh> rm /test
nsh> sh -c "set +e; mkdir /test; echo $?"
0
With this patch:
nsh> sh -c "set -e; mkdir /test; echo $?"
nsh: /test: mkdir failed: 17
nsh> sh -c "set +e; mkdir /test; echo $?"
nsh: /test: mkdir failed: 17
1
nsh> rm /test
nsh> sh -c "set +e; mkdir /test; echo $?"
0
Signed-off-by: wangjianyu3 <wangjianyu3@xiaomi.com>
Fix format string not appropriate warning for cause.flag.
nsh_syscmds.c: In function 'cmd_reset_cause':
nsh_syscmds.c:513:24: error: format '%lu' expects argument of type 'long unsigned int', but argument 4 has type 'uint32_t' {aka 'unsigned int'} [-Werror=format=]
513 | nsh_output(vtbl, "%s(%lu)\n",
| ^~~~~~~~~~~
514 | g_resetcause[cause.cause], cause.flag);
| ~~~~~~~~~~
| |
| uint32_t {aka unsigned int}
Signed-off-by: wangjianyu3 <wangjianyu3@xiaomi.com>
