* [DM/PCI] Validate MSI-X table BIR and offset against BAR
BIR and the table offset in the MSI-X capability are controlled by the
device. msix_table_remap() used BIR directly as an index into
pdev->resource[] and never checked the table against the BAR size, so a
malicious device could make the kernel map and write to an arbitrary
physical address during probe.
Reject out-of-range BIR values, unassigned or zero-sized BARs, and
tables that do not fit completely inside the assigned BAR. Also verify
that the requested vector count does not exceed the table size read
from the MSI-X control register.
Fixes#11985
* [DM/PCI] Fix cppcheck warnings in MSI-X code
- Restore the default IRQ only when an MSI descriptor exists, avoiding
a potential NULL pointer dereference in rt_pci_msi_shutdown
- Initialize the error code in msix_setup_msi_descs so that an empty
vector list returns a defined value
is_pte_valid() accepted a GPT entry as long as both LBAs were inside the
disk, so a crafted image could set ending_lba < starting_lba. The size
computed in efi_partition() then underflowed to a huge value and was
registered as the logical partition size.
Reject entries whose end LBA is below the start LBA, and compute the
partition start/size only after the entry passed validation. This avoids
evaluating the underflow at all and reuses the validated start value for
the size calculation.
FixesRT-Thread/rt-thread#11271.
Co-authored-by: DeepSeek 4.1 Flash <noreply@deepseek.com>
sys_getaddrinfo() validated the top-level res buffer, then wrote the
resolved socket address through res->ai_addr, a user-controlled nested
pointer that was never checked. On Smart/MMU builds a user program could
therefore turn getaddrinfo() into an arbitrary kernel write. The hints
fields were also read directly from user memory after the access check,
leaving a TOCTOU window.
Copy res and hints into kernel temporaries with lwp_get_from_user() so
all nested pointers come from one validated snapshot. Write the nested
sockaddr and the outer addrinfo back with lwp_put_to_user(), which
validates the destination range and never dereferences an unchecked user
pointer in kernel context.
Also fix the servname copy check (a size_t result can never be < 0),
allocate the kernel addrinfo with sizeof(*k_hints), and release k_res on
the new error paths.
Co-Authored-By: Claude Code <noreply@anthropic.com>
- register SCMI reset and power-domain drivers during subsystem initialization
- resolve SCMI reset state through the reset-controller provider
- use the parsed DT reset domain ID when querying reset attributes
- retain the SCMI device reference in the reset provider instance
Signed-off-by: GuEe-GUI <2991707448@qq.com>
* lwp: fix pid leak on exec failure before task startup
When msh tries to execute a non-ELF path, lwp_execve() may allocate a PID
before lwp_load() fails. The old error path only dropped the LWP reference,
leaving the PID tree entry pointing to a freed LWP.
In an init-less boot flow, this can poison pid 1 after a failed command from
msh. A later LWP launch may then treat the stale pid 1 entry as a valid parent
LWP, resulting in invalid pgrp/session state and a job-control assertion during
process exit.
Add lwp_pid_rollback() for exec/spawn failures before the process becomes
runnable. Unlike lwp_pid_put(), it always releases the PID lock and does not
enter the "no more pid allocation" state when the PID tree becomes empty.
Use the rollback helper in lwp_execve() failure paths after PID allocation.
Signed-off-by: zhangyang <gaoshanliukou@163.com>
* [drivers][sdio] avoid SDHCI device name stack overflow
rt_sdhci_init_host() allocated a fixed 32-byte buffer for device-derived object names, but sdio_host_set_name() may copy up to RT_NAME_MAX bytes into the output buffer. When RT_NAME_MAX is larger than 32, this can overwrite the stack and corrupt nearby locals.
Size the temporary name buffer with RT_NAME_MAX, force it to be NUL-terminated, and derive the append offset from the actual string length. Also replace the suffix rt_sprintf() calls with bounded rt_snprintf() so timer and workqueue suffixes cannot overflow the buffer when the base name is already close to RT_NAME_MAX.
Signed-off-by: zhangyangysu <zhangyangysu0928@gmail.com>
---------
Signed-off-by: zhangyang <gaoshanliukou@163.com>
Signed-off-by: zhangyangysu <zhangyangysu0928@gmail.com>
Co-authored-by: yang.zhang <zhangyang01@kylinos.cn>
- move syscall implementations from lwp_syscall.c into syscall/
- group memory, file, socket, process, IPC and time syscalls
- add a shared syscall dispatch table and internal header
- register the new directory from the LWP build script
- no functional change
- place the TIOCLINUX case brace on its own line
- use sizeof(tty_ctrl_charset) directly as the memcpy size
- fix the cppcheck sizeofDivisionMemfunc warning
- return zero keyboard modifiers for TIOCLINUX instead of failing
- keep terminal applications that probe console state working
- initialize the TTY window size to 24x80 when none is reported
Validate option headers and declared lengths before walking or returning options. Check minimum value sizes at the MESSAGE_TYPE and REQUESTED_IP call sites and bound parsing by the received packet length.
Fixes: #11913
- read reg-shift and reg-io-width from the device tree
- select the MMIO access width from reg-io-width
- reject unsupported register widths during early setup
- keep the generic 8250 path for other compatibles
- resolve fops from dfs_file instead of vnode directly
- keep compatibility with both DFSv1 and DFSv2
- validate fops before dereferencing to avoid NULL access
* [net][sal] validate RAW socket protocol
- replace the IPPROTO_RAW-only check with VALID_PROTOCOL()
- allow RAW sockets for protocols such as ICMP
- keep the existing domain and type validation unchanged
* [net][sal] fix clang-format alignment
- align the macro continuation after the RAW protocol check
- no functional change
- UTIME_OMIT means the corresponding timestamp must not be changed
- the inverted condition overwrote mtime and ignored a valid timestamp
- correct the comparison so UTIME_OMIT leaves the attribute untouched
- add RT_LWIP_USING_LOOPBACK as a configurable option
- register lo/lo0 as a netdev and bring the link up
- make loopback the default interface when none is configured
- increase the RAW receive mailbox size for loopback echo
rt_console_set_device() is typically called in rt_hw_board_init() right
after rt_components_board_init(). null was INIT_DEVICE (level 3), so it
was not registered yet and could not be RT_CONSOLE_DEVICE_NAME.
Raise to INIT_BOARD_EXPORT; null is a pure software sink with no HW deps.
Co-Authored-By: Grok <seteiro@mail.grokbot.com>
The touch framework has no private headers in its source directory, so
adding the directory to CPPPATH is unnecessary. Remove it to avoid
polluting the global header search path, as discussed in #10980.
* [libcpu][risc-v] Refactor the RISC-V architecture port
Introduce an opt-in unified RISC-V common layer for RV32 and RV64 while keeping existing BSPs on the legacy implementation during migration.
- use ARCH_RISCV and ARCH_CPU_64BIT for architecture and XLEN selection
- unify context, trap, MMU, cache, timer, SBI, SMP and ISA handling
- add PLIC, CLIC, ACLINT, APLIC and IMSIC support
- map physical hart IDs to contiguous logical CPU IDs
- add T-Head vendor overrides while preserving legacy ports
- migrate qemu-virt64-riscv with RV32 and RV64 support
- update clock_time, MM, LWP, OFW, PCI and Rust integration
Existing BSPs continue using legacy_common or common64 unless they select RT_USING_RISCV_NEW_COMMON.
Signed-off-by: GuEe-GUI <2991707448@qq.com>
* [libcpu][risc-v] skip disabled CPU nodes in early FDT walk
Rbb666 noted that rt_hw_fdt_install_early() walked every /cpus child
without checking status, so disabled cores still joined hart registration
and MMU capability probing. Skip nodes that are not okay/ok via
rt_fdt_device_is_available() before reading reg or mmu-type.
Signed-off-by: seteiro <chenmoushun@163.com>
* [libcpu][risc-v] Fix per-CPU page table mapping
- reuse existing non-leaf page table entries when mapping per-CPU areas
- reject leaf entry conflicts instead of overwriting mappings
Signed-off-by: GuEe-GUI <2991707448@qq.com>
---------
Signed-off-by: GuEe-GUI <2991707448@qq.com>
Signed-off-by: seteiro <chenmoushun@163.com>
Co-authored-by: seteiro <chenmoushun@163.com>
The 'index == times' condition in the delay_tick calculation is never true,
so the delay was still applied after the last ping, contrary to the comment.
Skip the delay on the final ping by checking 'index != times - 1'.
Also only call rt_thread_delay() when delay_tick > 0, since passing 0
returns -RT_EINVAL.
Reject full configuration descriptors with more interfaces than the instance can hold. Clean up using the fixed interface-array capacity so an attach failure cannot leave detach trusting an oversized device-supplied count.
Fixes#11288
netdev_cmd_ping() computed avg_time / received before checking that any
reply arrived. When every request times out, received is 0 and the
integer division traps (SIGFPE on the simulator, a fault on Cortex-M),
taking the shell thread down. Only compute the average inside the
existing received > 0 branch where it is printed.
Fixes#11852
The '#define assert ASSERT' in ulog_def.h conflicts with the 'assert'
used in reset.c (rstc->rstcer->ops->assert), causing unintended macro expansion.
Signed-off-by: Jiang Yu <yu.jiang@rock-chips.com>
With RT_USB_DEVICE_CDC and RT_USING_SERIAL_V2 both enabled, _function_enable
assigned the embedded rx_ringbuffer to serial.serial_rx. Serial V2 reads that
pointer as a struct rt_serial_rx_fifo, whose first member is a rt_ringbuffer,
so ring-buffer access happened to work while rx_cpt, rx_cpt_index and
rx_timeout landed on the following members of struct vcom, corrupting tx_rbp.
The layout shifts again with RT_SERIAL_USING_DMA, which inserts dma_ping_rb
ahead of those fields.
Serial V2 already owns the receive FIFO: dev_serial_v2.c allocates it with
rt_malloc on open and releases it with rt_free on close, so the assignment
also made rt_free run on memory inside struct vcom.
Drop the assignment and read the FIFO Serial V2 owns. rt_hw_serial_isr
reports RX_IND from serial_rx->rb, so the endpoint handler now fills that
ring buffer and _vcom_getc drains it. The Serial V1 path is unchanged.
Fixes#11739
Signed-off-by: Jaeyeong Lee <lee@jaeyeong.cc>
Newer GCC emits C++ global constructors into .init_array, which many
linker scripts keep outside [__ctors_start__, __ctors_end__), so they
never ran. Walk __init_array_start/__init_array_end (weak) as well,
skipping entries already covered by the ctors range.
Add RT_USING_CPP_CRT_INIT for ports whose startup runtime already runs
.init_array, default on for the simulator. With it set, only the
.init_array walk is skipped; the ctors range is still walked.
Renesas RA/RA_WIRELESS/RZ families select it when C++ is enabled: FSP
SystemInit() already walks .init_array, so boards whose linker script
only provides __init_array_* would run constructors twice. Boards that
fold .init_array into the ctors range (ra6m3-hmi-board, ra8d1, rz)
still run them once via RT-Thread.
Also source env.sh and restore RTT_EXEC_PATH in ToolsCI project generate
and dist steps so scons is on PATH (exit 127) and the ARM toolchain path
is not overwritten by env.sh.
Fixes#11778
The option loop advanced by the option length without checking the
number of bytes actually received by recvfrom(), so a malformed or
truncated DHCP packet could make the server read past the end of the
receive buffer.
Check that the option type, length and payload stay inside the received
data before reading them, and stop parsing when a requested IP option
is shorter than 4 bytes or a message type option is shorter than 1 byte.
Fixes#11323