CRYPTO_CHACHA20 implements the RFC 8439/IETF parameterization (32-bit
counter + 96-bit nonce). SSH's chacha20-poly1305@openssh.com uses the
original DJB construction instead: a 64-bit block counter in state
words 12..13 and a 64-bit nonce in words 14..15 (libtomcrypt's
chacha_ivctr64). The two layouts produce different keystreams for the
same key, so an SSH server cannot interoperate with OpenSSH clients
through the IETF variant.
Signed-off-by: Felipe Moura <moura.fmo@gmail.com>
Wire the STM32H7 family to the shared boards/arm/common/stm32 board-common
directory.
Drop the local copies that are already provided by the shared directory:
- stm32_reset.c from nucleo-h723zg, nucleo-h743zi, nucleo-h743zi2,
nucleo-h753zi, stm32h745i-disco and stm32h750b-dk
- stm32_romfs_initialize.c and stm32_romfs.h from nucleo-h743zi (identical
to the shared versions)
Enable CONFIG_ARCH_BOARD_COMMON=y in the configs that use CONFIG_BOARDCTL_RESET
or CONFIG_STM32_ROMFS so board_reset()/romfs are pulled from the shared
directory.
Signed-off-by: raiden00pl <raiden00@railab.me>
uart_putxmitchar() manipulates dev->xmit.head/buffer directly with
no internal locking - every other caller (uart_write()) takes
dev->xmit.lock first. uart_readv()'s ECHO handling (both the
backspace/delete erase sequence and the normal character echo)
calls uart_putxmitchar() without taking that lock, so a concurrent
uart_write() and a local echo can race on the same circular buffer
state, corrupting it.
Take dev->xmit.lock around each echo's uart_putxmitchar() calls,
matching what uart_write() already does. uart_readv() holds
dev->recv.lock for its own duration, but no other code path ever
acquires recv.lock while holding xmit.lock, so nesting xmit.lock
inside the existing recv.lock scope here doesn't introduce a new
lock-ordering cycle.
Fixes#14845
Signed-off-by: yi chen <94xhn1@gmail.com>
The pico-sdk-derived rp23xx linker scripts place .got/.got.plt as orphan
sections after the .data output section, so the boot-time .data copy
(which runs to _edata) never copies the GOT into RAM. Any position-
independent (-fPIC) object linked into an application then reads an
uninitialised GOT full of NULLs: the first GOT-indirected access
branches to 0 and the board hardfaults before main(). A PIC static
library (for example liboqs, whose SHA3 dispatch reads its function
pointers through the GOT) triggers this deterministically.
Fold *(.got)/*(.got.plt) into the .data{} section, before _edata, in all
three boot-mode scripts (memmap_default/copy_to_ram/no_flash) for every
rp23xx board, so the GOT is initialised in RAM by the standard startup
copy. No effect on non-PIC images (the GOT is empty).
Reproduced and fixed on a Raspberry Pi Pico 2 W: an app linking a -fPIC
liboqs.a wedged the board on its first crypto call before the change and
runs correctly after it.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: Ricard Rosson <ricard@groundbits.com>
Reading /proc/<pid>/group/env for another task dereferenced tg_envp
under the caller's own address environment instead of the target
task's, since tg_envp lives in the target's user heap. Switch to the
target's address environment around the traversal, and to the
caller's own environment only around the copy into the caller's
buffer.
Signed-off-by: liang.huang <liang.huang@houmo.ai>
STM32_ETHMAC selects NETDEVICES, but NETDEVICES depends on NET. Enabling
ETHMAC without NET produced an invalid Kconfig
Signed-off-by: raiden00pl <raiden00@railab.me>
Expose the ChaCha20 stream cipher and the ChaCha20-Poly1305 AEAD through
the OCF crypto framework (/dev/crypto) so applications such as an SSH
server (chacha20-poly1305) can use them directly, and fix the underlying
ChaCha nonce/counter layout so both match RFC 8439.
RFC 8439 nonce layout fix
-------------------------
chacha_ivsetup() previously used the original DJB layout: a 64-bit block
counter (input[12..13]) followed by a 64-bit nonce (input[14..15]). RFC
8439 defines a 32-bit block counter (input[12]) and a 96-bit / 12-byte
nonce (input[13..15]). With the old layout the existing ChaCha20-Poly1305
AEAD could not reproduce the RFC 8439 test vectors (the last 4 bytes of a
12-byte nonce were consumed as the high half of the counter). This
commit switches chacha_ivsetup() to the RFC 8439 layout and updates the
ChaCha20-Poly1305 one-shot helpers to pass a 12-byte nonce accordingly.
Standalone ChaCha20 on the unified enc path
-------------------------------------------
Instead of introducing a separate multi-buffer stream path (parallel
encrypt_multi/decrypt_multi callbacks), extend the existing enc_xform
encrypt/decrypt callback signature with a length argument:
void (*encrypt)(caddr_t, FAR uint8_t *, size_t len);
void (*decrypt)(caddr_t, FAR uint8_t *, size_t len);
With that single change every cipher, block or stream, flows through the
same swcr_encdec path. swcr_encdec already handles a short final block
via buflen = MIN(i, blocksize), so arbitrary-length data works without a
second code path. This is exactly how the existing stream ciphers
(AES-CTR/OFB/CFB) already behave: the cipher keeps its own counter in the
context and swcr_encdec feeds it whole blocks (only the last one may be
shorter). chacha20_crypt likewise relies on the underlying chacha state
block counter (input[12]) to continue the keystream across calls, so no
per-call keystream caching is needed.
* chacha_private.h: chacha_ivsetup uses a 4-byte counter and a 12-byte
nonce (RFC 8439).
* chachapoly.c / chachapoly.h: split reinit into chacha20_reinit (raw,
counter 0) and chachapoly_reinit (AEAD, counter 1); chacha20_crypt
takes a length and encrypts it in one pass, mirroring aes_ctr_crypt;
12-byte nonce for the one-shot AEAD helpers.
* xform.h / xform.c: add size_t len to encrypt/decrypt; add
enc_xform_chacha20 (blocksize 64, 12-byte IV).
* cryptodev.c / cryptosoft.c: register CRYPTO_CHACHA20 as a txform
cipher, route new sessions to enc_xform_chacha20, feed the AEAD AAD
through crp_aad/crp_aadlen, and handle the short final block in
swcr_encdec.
* cryptodev.h: add CRYPTO_CHACHA20; bump EALG_MAX_BLOCK_LEN to 64.
This keeps all ciphers on one uniform path instead of maintaining two,
and any future stream cipher drops in with just an xform table entry.
Impact: extends an internal kernel callback signature (enc_xform
encrypt/decrypt). All in-tree implementations are updated in the same
commit and the user-facing /dev/crypto ABI is unchanged, so this is
self-contained and not a breaking change for existing configurations.
Testing:
Build host: Ubuntu Linux x86_64, GCC (host sim toolchain)
Target: sim:crypto (CONFIG_ARCH=sim)
Ran the crypto test apps. ChaCha20 uses RFC 8439 2.4.2 vectors
(including a 375-byte multi-block vector exercising cross-block counter
continuity); ChaCha20-Poly1305 uses the RFC 8439 2.8.2 AEAD vector.
A full regression of the other ciphers was run to confirm the extended
encrypt/decrypt signature does not change their behaviour:
nsh> chacha20
chacha20: 2/2 vectors passed
nsh> chachapoly
OK test vector 0
chachapoly: 1/1 vectors passed
nsh> des3cbc -> all vectors OK
nsh> aescbc -> all vectors OK
nsh> aesctr -> all vectors OK
nsh> aesxts -> 14 vectors OK (encrypt + decrypt)
nsh> hmac -> md5 / sha1 / sha256 all success
Signed-off-by: makejian <makejian@xiaomi.com>
The top command was added to nshlib in 2024 but was never documented.
Add a section to the NSH commands page covering syntax, options,
example output and configuration dependencies, including the
Linux-like summary header.
Also enable the command in the linum-stm32h753bi:nsh config: procfs,
CPU load measurement, stack coloration and task names, so that top
and ps are fully functional out of the box. This requires a dedicated
interrupt stack and larger IDLE/init task stacks: with SCHED_CPULOAD
the per-tick accounting runs in interrupt context, and with
ARCH_INTERRUPTSTACK=0 it lands on the stack of the interrupted task,
overflowing the 1 KiB IDLE stack and corrupting the adjacent heap.
Also add the VT100 escape sequences used by the top command screen
refresh as string literals in include/nuttx/vt100.h (VT100_STR_*),
next to the existing VT100_FMT_* definitions.
Tested on hardware.
Signed-off-by: Jorge Guzman <jorge.gzm@gmail.com>
Document how to transfer files with scp on the Dropbear server port,
which requires the legacy protocol (scp -O) since the port has no SFTP
server. Recent OpenSSH clients default to SFTP and fail otherwise.
Signed-off-by: Felipe Moura <moura.fmo@gmail.com>
riscv_fillpage() unconditionally panics the kernel on an unmappable
or invalid access, even when the fault is caused by a user task. A
user-space fault should terminate that task with SIGSEGV instead of
taking down the whole system, matching the existing behavior for
other exceptions in riscv_exception().
Signed-off-by: liang.huang <liang.huang@houmo.ai>
scan_dir() checked name_len before assigning it. Directory entries already compute and validate strlen(a_dirent->d_name) before allocating the dir item, while the root path passed to scan_dir() is not encoded as a directory entry name.
Remove the stale pre-opendir check so scan_dir() no longer branches on an uninitialized local variable.
Generated-by: OpenAI Codex
Signed-off-by: aineoae86-sys <ai.neo.ae86@gmail.com>
coredump_dump_tcb() copied tcb->name unconditionally, but the tcb_s name
member only exists when CONFIG_TASK_NAME_SIZE > 0
Signed-off-by: raiden00pl <raiden00@railab.me>
Run update_romfs_password.sh, check_passwd_keys.sh, and
gen_passwd_keys.sh through an explicit sh/bash interpreter so CMake
configure works on Linux, macOS, MSYS2, and Cygwin instead of relying
on direct script execution.
Signed-off-by: Abhishek Mishra <mishra.abhishek2808@gmail.com>
Support NUTTX_ROMFS_PASSWD_PASSWORD via update_romfs_password.sh for
configs that enable ROMFS passwd without a defconfig password (sim/login
CI). Enable RANDOMIZE_KEYS in sim/login defconfig. Update mkpasswd.c
header, platform docs, and the mkpasswd_autogen guide.
Signed-off-by: Abhishek Mishra <mishra.abhishek2808@gmail.com>
Remove implicit default credentials and add build-time validation.
Add check_passwd_keys.sh and gen_passwd_keys.sh; run key setup via
passwd_keys.mk before config.h is generated. Mirror the same logic in
cmake/nuttx_add_romfs.cmake for CMake builds.
BREAKING CHANGE: Builds with CONFIG_BOARD_ETC_ROMFS_PASSWD_ENABLE=y now
require an explicit admin password and non-default TEA keys. The
Kconfig default password "Administrator" and default TEA keys are no
longer accepted. Fix: run make menuconfig, set Admin password under
Board Selection -> Auto-generate /etc/passwd, enable random TEA keys or
set CONFIG_FSUTILS_PASSWD_KEY1..4 manually, and use NSH login with
Encrypted password file verification.
Signed-off-by: Abhishek Mishra <mishra.abhishek2808@gmail.com>
switch stm32f7 board common to boards/arm/common/stm32.
Move the STM32F7-only board-common features into the shared directory
so they are available to every STM32 family:
- bmi270, cs4344, can_setup, cansock_setup, reset, romfs_initialize and
spitest
- the duplicate stm32_bh1750 is removed (already present in common).
Signed-off-by: raiden00pl <raiden00@railab.me>
Enabling STM32_FOC without MOTOR_FOC failed to build with
"unknown type name 'foc_current_t'". Add the dependency on MOTOR_FOC.
Signed-off-by: raiden00pl <raiden00@railab.me>
stm32h5/stm32_i2c.c selected the I2C2/3/4 kernel clock with
RCC_CCIPR4_I2CnSEL_PCLKx, but the H5 RCC header names those values
RCC_CCIPR4_I2CnSEL_RCCPCLKx.
Signed-off-by: raiden00pl <raiden00@railab.me>
Add ReStructuredText documentation for the MPU6050 6-axis IMU uORB driver under Special Sensor Drivers. Includes device registration, Kconfig options, uORB topic descriptions, and bring-up examples.
Signed-off-by: Shriyans S Sahoo <shriyans.s.sahoo@gmail.com>
Add lower-half uORB sensor driver for the InvenSense MPU6050 over I2C. Supports accelerometer and gyroscope sensor types, register read/write, fetch, and control operations.
All internal register definitions and full-scale range constants are placed in the private driver C file (drivers/sensors/mpu6050_uorb.c) for clean encapsulation. The device struct uses explicit named members ('accel' and 'gyro') for high readability.
Add Kconfig option CONFIG_SENSORS_MPU6050 under Sensor Drivers, and add build integration to drivers/sensors/Make.defs and CMakeLists.txt.
Signed-off-by: Shriyans S Sahoo <shriyans.s.sahoo@gmail.com>
Added a Mixed case exclusion for MQTTErrors defined in
apps/netutils/mqttc/MQTT-C/include/mqtt.h
Signed-off-by: Mihai Pacuraru <mpacuraru@protonmail.com>
Update the DNS stream send and receive helpers to advance their buffers
by the number of bytes actually transferred. Advancing by the original
remaining length skips data when send() or recv() completes partially.
Signed-off-by: Old-Ding <ai.neo.ae86@gmail.com>
This commit fixes the incorrect casting of signed types to unsigned
types in the I2S character driver.
NOTE: the I2S character driver IOCTLs retain their original argument
types. This is fine, because errors are reported through errno by the
ioctl call. The returned value in the 'arg' parameter is to be ignored
when the call results in an error. Outside of error codes, all i2s
interfaces return unsigned values.
Signed-off-by: Matteo Golin <matteo.golin@gmail.com>
Enable the OTG FS host on the Linum board so an external USB HID keyboard
can be used: configure the USB VBUS, power-switch and over-current GPIOs
at board bring-up (the power switch on PI12 is active low) and start the
USB host from stm32_bringup() through stm32_usbhost_initialize().
Add the lvglterm_kbda configuration - the LVGL terminal driven by a USB
HID keyboard on /dev/kbda - with the microSD card enabled alongside the
USB host. They coexist once CONFIG_MMCSD_MMCSUPPORT is left disabled, so
an SD card is not probed as an MMC device.
Also switch esp32s3-m5-cardputer:lvglterm to the renamed matrix-keyboard
input option (EXAMPLES_LVGLTERM_INPUT_KBD_MATRIX) to match the reworked
lvglterm example Kconfig.
Update the documentation accordingly: the LVGL Terminal example page for
the three-way input choice (touch, matrix, USB HID) with cursor-key
scrolling and per-variant configuration plus a photo of the USB variant
running on the Linum, the lvglterm_kbda entry on the Linum board page,
and the matrix-keyboard variant name on the M5Stack Cardputer board page.
Signed-off-by: Jorge Guzman <jorge.gzm@gmail.com>
riscv_fillpage() is the LOADPF/STOREPF handler used under
CONFIG_PAGING. It checked whether intermediate page table levels were
already allocated, but never checked the final leaf PTE before
installing a new mapping.
RISC-V raises the same LOADPF/STOREPF cause both when a leaf PTE is
absent (a real fault) and when it is present but its permission bits
don't satisfy the access, e.g. a store to a .text page whose write
access was revoked after ELF loading. The two cases are
indistinguishable from mcause alone.
Treating both cases as "page missing" let riscv_fillpage silently
allocate a fresh, zeroed physical page over an existing mapping,
discarding the old page (a leak) and defeating whatever permission
that mapping was enforcing. Reproduced on real hardware: a user-space
store to an already-loaded .text page got a fresh writable page
instead of being rejected.
Check the leaf PTE's valid bit before allocating; if a mapping already
exists, panic instead of overwriting it.
Signed-off-by: liang.huang <liang.huang@houmo.ai>
The CI refresh step removes the explicit LVGL terminal keyboard input selection from the esp32s3-m5-cardputer lvglterm defconfig.
Update the saved defconfig so unrelated pull requests do not fail the normalize check on this board.
Signed-off-by: Old-Ding <ai.neo.ae86@gmail.com>
This board config is reaching the 64KB Flash limit, so disable some
not used interface to reduce size.
Signed-off-by: Alan C. Assis <acassis@gmail.com>
Rewrite the RTL8721Dx / PKE8721DAF and RTL8720F / RTL8720F-EVB pages against
the official Realtek product specifications, from an application-developer
point of view:
- Describe the application core NuttX runs on -- an Arm Cortex-M55-compatible
core, up to 345 MHz on RTL8721Dx (KM4) and up to 320 MHz on RTL8720F (KM4TZ)
-- and the on-chip memory, instead of the internal multi-core / IPC
arrangement that is not relevant to application developers.
- Fix the RTL8721Dx SRAM size: 512 KB (was incorrectly documented as 288 KB).
- Add per-part memory (RTL8721DAF and RTL8720FBF: 4 MB NOR flash, 512 KB
SRAM), wireless (RTL8721Dx: Wi-Fi 4 dual-band + BLE; RTL8720F: Wi-Fi 6
2.4 GHz + BLE + Thread), peripheral and security highlights.
Signed-off-by: raul_chen <raul_chen@realsil.com.cn>
Simplify complex ifdefs by using UNUSED.
The current conditions do not cover all cases anyway,
and adding more conditions only complicates the code.
Signed-off-by: raiden00pl <raiden00@railab.me>
rtl8720f_wifi_initialize() spawned a kthread to run the blocking WHC host
bring-up and register wlan0, so the netdev appeared only after netinit had
already run its one-shot associate -- the WiFi never auto-connected at
boot. Do the bring-up synchronously on the board bring-up path (mirroring
rtl8721dx_wifi_initialize) so wlan0 exists before netinit runs.
Enable the standard netinit auto-connect machinery in both board
defconfigs: NETINIT_THREAD (runs associate + DHCP off the NSH init path so
the prompt is not blocked), NETINIT_DHCPC, and placeholder
NETINIT_WAPI_SSID/PASSPHRASE the user replaces with their own credentials
(WIRELESS_WAPI + the wapi cmdtool are already enabled).
Signed-off-by: raul_chen <raul_chen@realsil.com.cn>
Rework the Ameba WHC WiFi driver from the legacy flat net_driver_s model
to the netdev lower-half framework (netdev_lowerhalf_s + netdev_ops_s with
.transmit/.receive, IOB-based quota backpressure and a dedicated
NETDEV_RX_THREAD). This provides real TX backpressure and moves RX off
the IPC callback onto a bounded rx queue.
Raise the host TX skb pool (skb_num_ap) to a minimum of 16. The SDK
default of 4 is too small for sustained TCP TX: the host send runs out of
buffers (-2), which exhausts retransmits and tears the connection down.
Enable CONFIG_NET_TCP_SELECTIVE_ACK (which selects OUT_OF_ORDER) in both
board defconfigs. WiFi's occasional loss otherwise stalls TCP RX on RTO
timeouts (~2 Mbit/s); with SACK + out-of-order queuing RX recovers to
~12-13 Mbit/s.
Signed-off-by: raul_chen <raul_chen@realsil.com.cn>
CCM_CR_COUNT is defined as the last CCM root index in the non-SCMI tables, and CCM_CR_PALCAMESCAN also uses that value. Use a strict greater-than check so root 94 reaches the CCM register helpers.
Generated-by: OpenAI Codex
Signed-off-by: Old-Ding <35417409+Old-Ding@users.noreply.github.com>
Replace the TODO placeholder with a description of the two-pass build
mechanism (CONFIG_BUILD_2PASS). pass1/ supports two-pass builds where
application-generated source files are compiled and linked into the
kernel address space during pass 2, primarily for kernel symbol table
generation required by loadable kernel modules.
Signed-off-by: hanzhijian <hanzhijian@zepp.com>
Author: hanzhijian <hanzhijian@zepp.com>
