CONFIG_ARCH_PERF_EVENTS is on by default for ARMv7-M, but imxrt never
called up_perf_init(), so perf_gettime() read zero and anything timed
with it, rpmsg_ping for one, reported 0 ns. Initialise the counter with
BOARD_CPU_FREQUENCY after the MPU is set up, as samv7 and stm32h7 do.
Assisted-by: Claude:claude-fable-5-1
Signed-off-by: Lourens Naude <lourens@bearmetal.eu>
P1.0 powers the switched core off and keeps the XIP cache and SRAM.
Every peripheral loses its registers, so the chip comes back through
the bootrom and the ordinary boot, not from the WFI. The idle governor
never selects it: an application asks for it with the new
BOARDIOC_RP23XX_SUSPEND boardctl() command (arch/chip/pm.h, handled by
the common rp23xx board_ioctl()).
rp23xx_pm_suspend():
- saves the NVIC, writes a marker to POWMAN SCRATCH0, and arms the
wake: the RP23XX_PM_WAKEUP_GPIO pin in a POWMAN power-up detector,
and the always-on timer alarm for a timed wake. An armed RTC alarm
that comes first powers the chip up itself, so an application can
set the wake with RTC_SET_ALARM. Otherwise the RTC alarm is saved
with the new rp23xx_rtc_savealarm() and given back after the wake
with rp23xx_rtc_restorealarm().
- cleans the XIP cache, with the RP2350-E11 workaround, because the
resume discards it and PSRAM can hold task stacks.
- requests P1.0 and waits in WFI.
On the next boot, __start asks rp23xx_pm_resume_pending() (the marker
and CHIP_RESET.HAD_SWCORE_PD) before .bss and .data are touched. For a
resume it moves to its own stack, because the idle thread still runs on
the idle stack, sets up the hardware with the cold boot code (now
rp23xx_hwinit()), and longjmps back to the suspended thread. The UARTs
are set up from the driver state in RAM, the PSRAM format is applied
again without detection (the part is still in quad mode), the dormant-
wake GPIOs are armed again (the IO bank lost them, and the next dormant
period could then never end), and the time of day is taken from the
always-on timer.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
Map the NuttX PM states onto the RP2350 low-power modes:
- PM_STANDBY is the RP2350 SLEEP state: a WFI with the clocks of the
blocks that have no driver in the configuration gated (SLEEP_EN0/1).
- PM_SLEEP is the DORMANT state: clk_sys moves to the crystal
oscillator, the PLLs and then the oscillator stop, and the clock tree
is restored after the wake, as pico-extras does. Only the GPIO
dormant-wake detector can wake the chip, so PM_SLEEP gives the
standby state when no wake GPIO is configured. The wake GPIO also
raises a one-shot interrupt, so that the core leaves its WFI and
restores the PLLs at once. The time of day is taken back from the
always-on timer after the wake.
The dormant state stops the UARTs. A PM_STANDBY wakelock
(pm_staytimeout()) is held for RP23XX_PM_WAKE_HOLD_MS after a dormant
wake and after each character a UART receives. The serial driver
refuses PM_SLEEP in its prepare callback while a UART transmits. A
dormant chip does not answer SWD, so the pm configurations use
PM_GOVERNOR_EXPLICIT_RELAX to stay out of it for a time after boot.
With RP23XX_PM_QUIESCE_PADS, arm_pminitialize() also isolates the
unused pads and holds the blocks with no driver in reset. A floating
bank 0 input settles near 2.2V (erratum RP2350-E9) and its input
buffer then draws a static current in every state.
Also select ARCH_HAVE_PM, and fix the LED PM callbacks of three boards,
which used BOARD_LED where the boards define BOARD_LED1. They did not
build with CONFIG_PM.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
Add the blank line after the declarations in up_putc() and in
rp23xx_led_pminitialize() of three boards. Whitespace only.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
pgalloc() grows a process heap for sbrk(). It extended the address
environment of the running task (addrenv_own). While exec() sets up a
new process, the caller selects the new address environment and
allocates the new process's stack from its heap. When that stack does
not fit in the initial heap, the heap must grow, but the running task is
the caller. For the kernel thread that starts init this was an
assertion; for a user task it would have grown the caller's heap.
Use the selected address environment (addrenv_curr). For a normal sbrk()
it is the same as addrenv_own.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
This allows using >25MHz bus speeds by adding a configration option
to set SDIO_CAPS_SD_HS_MODE capability.
Signed-off-by: Jukka Laitinen <jukka.laitinen@tii.ae>
- Add the missing clock gating macros for imxrt118x
- Change sw_cd_gpio from int32_t to gpio_pinset_t (64-bits on 118x)
- Invalidate cache again after the data has been received. Cache might
be refilled by a prefetch or, in theory, by cpu read touching the same
cache line - even though the latter should not happen.
- Map any DMA accesses to/from DTCM to the shadow address window, which
gives the uSDHC DMA access to the DTCM.
Signed-off-by: Jukka Laitinen <jukka.laitinen@tii.ae>
A JRCR reset only flushes and halts the job ring; a second reset, once
JRINT reports the halt done, completes it, as Linux's caam_reset_hw_jr()
does. With the single write the ring stayed halted, so the RNG
instantiate job never completed and every boot logged "job ring did not
answer" before the settle loop's second ring init finished the reset.
Signed-off-by: Royyan Zahir <royzah@gmail.com>
The completion handler did:
dmach->callback(...);
dmach->callback = NULL; /* after the call */
so a callback that registers itself again -- which is the only way for a
driver to keep a channel running continuously -- has that registration
wiped the moment it returns. The channel transfers one more block and
then goes deaf, with no error raised anywhere and nothing in the
registers to say why. A PWM audio driver hit this and worked around it
by restarting from a thread instead, which put a millisecond of silence
into every buffer boundary.
Lift the callback and its argument out first and clear them before the
call. One-shot behaviour is unchanged for every existing user, since
none of them re-register from inside the call; the difference is only
that one which does now survives.
Tested on a Pimoroni Pico Plus 2 W (pimoroni-pico-plus-2-w:nsh) with a
local test: a memory-to-memory transfer whose callback starts the next
one, 100 times. Master gives 1 completion of 100; this change gives 100.
The existing users (SPI, I2S, CYW43439, WS2812) do not start a transfer
from inside the callback, so they do not change.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
nxstyle reports "Missing blank line after declarations" in
rp23xx_dmachannel(). Add the blank line, because CI checks every file
that a change touches.
No functional change. The change is whitespace only.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
Select STM32_HAVE_IP_USART_M33_V3 and drop the family serial, USART
header, and low-level console sources. Provide the USART clock and RCC
gate definitions for all thirteen ports in stm32_rcc_m33.h and the RX
DMA request numbers in the family DMA signal map. Include the family
DMA header from stm32.h so the common driver reaches the DMA API.
Extend the common Cortex-M33 v3 USART support with the STM32H5
features: USART6, UART7-9, USART10-11, and UART12 ports, the LPUART
prescaler for low baud rates, RX DMA through the family-provided
request number, wakeup-from-stop CR3 definitions, and per-port
descriptors gated by the family peripheral options.
Enable the USART FIFO for every Cortex-M33 family. The low-level
console uses the per-port kernel clock and RCC gate, which corrects
the UART9 and UART12 enable register. The RX DMA callback delivers
data only while reception is enabled.
The non-BSD break ioctl now uses the send-break request register
instead of a CR1 bit this USART IP does not have. Correct the UART12
gates and the RXDMA and console undef lists in the USART driver
header. Raise STM32_NUSART to six on the H56x and H57x parts so UART12
fits the device table, add the missing UART9 buffers, write LPUART CR1
only on the USART path, and build device names with snprintf so minors
above nine are valid.
Keep the termios flow control fields writable so TCSETS compiles with
input flow control, let up_putc emit a bare newline since syslog adds
the carriage return, and make the unconfigure-on-close options
available to every Cortex-M33 family.
Signed-off-by: raiden00pl <raiden00@railab.me>
Assisted-by: Claude Code
Select STM32_HAVE_IP_GPIO_M33_V1 and STM32_HAVE_IP_EXTI_M33_V1 and
drop the family GPIO and EXTI sources and headers in favor of the
common Cortex-M33 v1 implementation. Add the RM0481 line count for the
H56x and H57x parts to the common EXTI line inventory.
Add GPIO_PORTI to the common Cortex-M33 pin encoding for the parts
with a ninth GPIO port.
Signed-off-by: raiden00pl <raiden00@railab.me>
Assisted-by: Claude Code
Enable STM32_COMMON_M33 for STM32H5 and drop the family reset, NVIC,
SysTick, and idle sources in favor of the common Cortex-M33 v1
implementation. The common heap allocator replaces the generic ARM one
through a STM32_PRIMARY_SRAM_SIZE spanning the contiguous SRAM banks
and skips the SRAM2 region when the primary heap already covers it.
Rename the family RCC header to stm32_rcc_m33.h for the common RCC
dispatch.
Add the SRAM3 parity initialization block and the CONFIG_STM32_ICACHE
enable call to the common reset handler, taken from the STM32H5 start
logic. Without CONFIG_STM32_ICACHE the reset handler disables an
ICACHE left enabled by a bootloader on STM32_HAVE_ICACHE families.
Move the flat-build MPU initialization to stm32_mpuinit_m33_v1.c in
common/stm32, built for every Cortex-M33 family with ARM_MPU, and
declare stm32_mpuinitialize for ARM_MPU as well as protected builds.
Declare stm32_board_initialize in the common start header.
Signed-off-by: raiden00pl <raiden00@railab.me>
Assisted-by: Claude Code
Expose the RTL8730E I2C controllers through the shared Ameba I2C driver
(arch/arm/src/common/ameba/ameba_i2c.c) by adding the chip-specific
glue, build wiring and a board bus table. The change is gated by
CONFIG_AMEBA_I2C (default disabled).
Chip glue (ameba_i2c_chip.h) supplies the three controller register
bases, the APB function/clock masks -- amebasmart encodes these in
bit 25/26/27 with the group selector bit30=0, unlike the (bit30 |
bit10/11) layout of the KM4-based parts -- and the pad-mux code.
amebasmart has a single generic PINMUX_FUNCTION_I2C shared by every I2C
pad instead of per-signal SCL/SDA crossbar codes, and its
I2C_InitTypeDef omits the DMA request-level fields, so
AMEBA_I2C_HAS_DMA_FIELDS stays undefined. The whole I2C fwlib lives in
ram_common/ameba_i2c.c rather than lib_rom.a, so it is compiled into
libameba_fwlib.a when I2C is enabled.
The shared driver gains an optional per-controller IP-clock hook,
AMEBA_I2C_IPCLK_FN(bus). amebasmart needs it because its
I2C_StructInit() fills in a 10 MHz placeholder rather than the real
reference clock (the other Ameba parts fill in a correct XTAL_ClkGet()
/ PLL_GetHBUSClk() / HPERI_ClkGet()), which makes I2C_SetSpeed()
miscompute the SCL counts. The hook resolves the rate at run time: the
two HS controllers sit on HS_AHB, i.e. NP_PLL divided by
REG_LSYS_CKD_GRP0.CKD_HBUS, so it is a PLL/board setting and not a
constant -- an EVB measured CKD_HBUS=8 (div9, 88.9 MHz) while the SDK's
own I2CCLK_TABLE claims a flat 100 MHz and the register reset value
would imply 80 MHz. The LP-domain I2C0 keeps its fixed 20 MHz. Chips
that leave the macro undefined use the fwlib default and are
unaffected; regression-tested on pke8721daf:i2c.
All three controllers are registered: /dev/i2c0 on PA9/PA10, /dev/i2c1
on PA3/PA4 and /dev/i2c2 on PB10/PB11. Pad choice is constrained on
this chip -- a pad reaches exactly one controller (SDA fixed per
controller, SCL the next pad up), and the pads inside the analogue
audio ranges (PA20-PA29, PA30-PB2, PB3-PB6) are driven by the codec and
leave the bus stuck idle. Both rules are noted in the table's
comment.
Also add a weak DiagVprintf stub: the I2C fwlib logs through RTK_LOGx()
-> rtk_log_write(), and unlike DiagPrintf that symbol is not in the AP
ROM symbol table. It is weak so the strong definition in
rtl8730e_wifi_stubs.c still wins when WiFi is enabled; both route to
vprintf.
Hardware-verified on an RTL8730E EVB against a second board running an
I2C slave: register read, write, write/read round-trip, multi-byte dump
and a full address scan, on all three buses at 100 kHz.
Signed-off-by: dechao_gong <dechao_gong@realsil.com.cn>
Assisted-by: Claude <noreply@anthropic.com>
With CONFIG_ARCH_HIPRI_INTERRUPT enabled, arm_doirq() dispatches PendSV
for deferred processing. The STM32H7 debug handler unconditionally panics
on that valid interrupt, preventing normal operation with debug enabled.
Guard the panic with CONFIG_ARCH_HIPRI_INTERRUPT, allowing the common
interrupt path to finish signal delivery and context switching. Preserve
the diagnostic when high-priority interrupts are disabled.
Assisted-by: Codex:GPT-6
Signed-off-by: jsanchez-2g <jsanchez@2g-eng.com>
Change the imxrt1180-evk M7 NSH configuration to use the SoC memories
more efficiently:
- Place .data, .bss, idle stack and primary heap into DTCM
- Allocate available OCRAM as a secondary heap
- Add a separate .dmamemory section in OCRAM for USB device DMA
allocations
- Place .ramfunc into ITCM, together with hand-picked "hot" functions.
The section is copied to ITCM at boot by the ramfunc copy.
The eDMA accesses DTCM through the SoC's dedicated bus window.
This configuration acts as an example of performance optimization for
imxrt1180 based boards.
Signed-off-by: Jukka Laitinen <jukka.laitinen@tii.ae>
Cortex-M7 core's DTCM is available for other peripherals via a
dedicated address space. If transfers are done to or from the DTCM,
translate addresses to work on this shadow memory region instead, via
which the eDMA can access the DTCM.
This allows using the existing imxrt peripherals, which use DMA, to work
directly even if .data/.bss are located in DTCM.
Signed-off-by: Jukka Laitinen <jukka.laitinen@tii.ae>
Add support for Cortex-M33 code to initialize the M7 TCM memories before
releasing it to run. The TCM has ECC, which needs to be initialized before
the memory is usable for M7.
Specifically, the TCM needs to be initialized sequentially in 64-bit writes.
Use eDMA4 for this; this is the same mechanism which the NXP MCUXpresso SDK
code does.
Split imxrt118x_release_cm7() into imxrt118x_prepare_cm7() and
imxrt118x_start_cm7(). The TCM ECC initialization is done in
imxrt118x_prepare_cm7(), after the M7 has been released from reset and
before the M7 is started. Also reset M7_CFG[TCM_SIZE] to the default
256 KiB ITCM / 256 KiB DTCM layout.
Co-Authored-By: Jukka Laitinen <jukka.laitinen@tii.ae>
Signed-off-by: Jukka Laitinen <jukka.laitinen@tii.ae>
Move the imxrt118x SRC register defintions to an own file. They differ
from the other imxrt chips, and were also scattered between blockctrl
and a common imxrt_src headers.
Signed-off-by: Jukka Laitinen <jukka.laitinen@tii.ae>
RP23XX_FLASH_MTD_OFFSET and RP23XX_FLASH_MTD_SIZE come from Kconfig.
If the region ends past the end of the flash, the flash wraps the
address around, and an erase or program hits the start of the flash,
where the NuttX image is. For example, a 4M region at 1M does not
fit on the 4M flash of a Raspberry Pi Pico 2.
Read the JEDEC ID at initialization, in QMI direct mode as the Pico
SDK flash_do_cmd() does, and refuse a region that does not fit. The
capacity byte is log2 of the size in bytes. If the ID does not look
valid, warn and do not check.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
While the bootrom erases or programs the flash, the QMI is in direct
mode, and an access to the XIP space (flash or PSRAM) gives a bus
fault. The flash MTD driver accessed it in two cases:
- The data to program was in flash or PSRAM. flash_range_program()
read it during the operation. Now the driver copies each such page
to an SRAM buffer first.
- The caller's stack was in PSRAM. This is the normal case with
RP23XX_PSRAM_HEAP_USER, and possible with RP23XX_PSRAM_HEAP_SINGLE.
The operation pushed to that stack. Now the driver switches to a
small SRAM stack for the operation if the stack is in the XIP space.
The operation data is static (SRAM) since the previous commit.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
The flash MTD driver disabled interrupts for a whole request. A
multi-block erase or a large write kept them off for seconds.
Erase one 64K block (or one 4K sector where the range is not block
aligned) and program one 256 byte page per step. Enable interrupts and
release the other core between steps. A single block erase is still
long, but that is the limit of the flash.
Also, on SMP:
- Do not send the pause call to the CPU that does the operation.
nxsched_smp_call_single_async() runs it at once on that CPU.
- Keep the isolation data in a static, not on the stack. The other
CPU spins on it while the flash is busy.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
After a flash operation the driver called flash_select_xip_read_mode()
with a fixed EBh quad mode and clock divisor 4, and called
flash_enter_cmd_xip() if it "failed". But that ROM function returns
void, so the check read a random r0. The fixed mode and divisor can
also be different from the ones the bootrom found at boot.
The datasheet (5.2.7, 5.4.8.10) and the Pico SDK use a different
method: after a flash boot the bootrom leaves an XIP setup function in
the first 256 bytes of boot RAM. It restores the read mode and clock
divisor found at boot. Boot RAM is not executable, so copy the
function to SRAM once at initialization, and call the copy.
If boot RAM is empty (no flash boot), use flash_enter_cmd_xip(), as
RP23XX_FLASH_MTD_SAFE_XIP does.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
The bootrom flash functions reset the QMI window 1 (chip select 1)
registers and the QSPI pads. flash_flush_cache() also discards dirty
XIP cache lines. The flash MTD driver did not save anything, so after
the first erase or program the PSRAM on chip select 1 read garbage,
and PSRAM writes still in the cache were lost.
Do what the Pico SDK hardware_flash library does:
- Clean the XIP cache before the operation. Clean by set/way through
the top of the maintenance window, to avoid erratum RP2350-E11.
- Save the QSPI pads and the five QMI M1 registers before, and write
them back after XIP is restored. Also keep XIP_CTRL.WRITABLE_M1.
rp23xx_psram_restore() was the earlier fix for this, but nothing called
it. Remove it.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
The always-on timer has an alarm comparator, but the RTC driver did not
use it: rp23xx_rtc.c implemented only up_rtc_initialize(),
up_rtc_time() and up_rtc_settime().
Add the alarm and an RTC lower half for /dev/rtc0:
- rp23xx_rtc_setalarm(), rp23xx_rtc_cancelalarm() and
rp23xx_rtc_rdalarm() on the ALARM_TIME registers and the POWMAN
timer interrupt.
- An RTC lower half with rdtime, settime, setalarm, setrelative,
cancelalarm and rdalarm, registered by the common board bringup.
The comparator asserts while the time is past the alarm time, not on
a transition. So the interrupt handler disables the alarm before it
does anything else; clearing only the status makes the interrupt
repeat. The arming sequence is the one of
powman_timer_enable_alarm_at_ms() in the Pico SDK.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
setup_period() printed the uint8_t slice number with %d, and the
uint32_t frequency, the uint16_t top and the uint32_t divisor with %lu.
Use %u for the two small fields and PRIu32 for the two uint32_t fields.
No build warns about this today, because GCC does not check syslog
format strings. It shows with CONFIG_DEBUG_PWM_INFO only.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
Indent the three else blocks and the switch in rp23xx_pwm_ioctl() as
nxstyle wants. Whitespace only; git diff -w is empty.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
With -mfdpic and -mlong-calls, GCC turns a call to an imported function
in tail position into "ldr r3, [r9, #off]; bx r3". The GOT slot holds
the address of the function descriptor, so the branch goes to the
descriptor in RAM instead of through it, and the core faults. A normal
call loads the code address and the data base from the descriptor
first. GCC 13.2 and 15.3 both do this.
Only an optimized build makes tail calls. The C++ library of
apps/testing/fs/xipfs then faults in its constructor, which ends in a
call to syslog(), and the test stops at "stage the C++ module".
Pass -fno-optimize-sibling-calls with the other FDPIC flags, in the make
build and the CMake build.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
GCC before 14 does not pass --fdpic to the assembler when it compiles
with -mfdpic. The assembler then rejects every FDPIC relocation with
"Relocation supported only in FDPIC mode". The NuttX CI image has GCC
13.2, so the crt0.o of an FDPIC configuration does not build there.
Pass -Wa,--fdpic with -mfdpic, in the make build and the CMake build. A
newer GCC passes the same option itself.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
With CONFIG_FDPIC, the CMake build gave the FDPIC options to a loadable
module only. A shared library (DYNLIB) got neither -mfdpic nor the FDPIC
link, but the "-r" link of the non-FDPIC case. So the library was a
relocatable object, and a module that named it in DT_NEEDED did not link:
"multiple definition" and "dangerous relocation".
Give a shared library the same options as a module, as LDMODULEFLAGS and
CMODULEFLAGS in common/Toolchain.defs already do for the make build.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
The i.MX95 has five GPIO instances, but g_gpio_base[] only listed the
first four. IMX9_GPIO_BASE(n) indexes this table directly, so any
access to port GPIO5 read one element past the end of the array and
caused a crash.
The table was previously selected on CONFIG_ARCH_CHIP_IMX9_CORTEX_M,
which covers both the i.MX93 M33 and the i.MX95 M7. Since the i.MX93
only has four GPIO instances and does not define IMX9_GPIO5_BASE, key
the five entry table off CONFIG_ARCH_CHIP_IMX95_M7 and keep the four
entry table for the i.MX93 (both the Cortex-A CONFIG_ARCH_CHIP_IMX93
and the Cortex-M CONFIG_ARCH_CHIP_IMX93_M33 variants).
Signed-off-by: Peter van der Perk <peter.vanderperk@nxp.com>
RP23XX_PSRAM_M1_TIMING was the constant 0x61a07102. The comment said
it matched the Pico SDK, but it does not. The SDK computes the timing
from clk_sys and the APS6404 limits (133 MHz SCK, 8 us maximum select,
18 ns minimum deselect). At 150 MHz it gives 0x60242202:
field old SDK
clkdiv 2 2
rxdelay 1 2
max_select 16 18
min_deselect 7 2
select_hold 3 0
The old RX delay samples the read data half a clk_sys cycle earlier
than the SDK does. The constant is also wrong for any other clk_sys.
Compute the fields from BOARD_SYS_FREQ at build time, with the SDK
formula, and stop the build if a field is out of range. The result
is identical to the SDK value at 48, 125, 150, 200, 266 and 300 MHz.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
The RP2350 datasheet (12.14.5) says: set DIRECT_CSR.EN, then poll BUSY
until it is low, before the first direct-mode transfer. BUSY stays high
while an XIP transfer is in its cooldown. The Pico SDK waits here too.
rp23xx_psram_detect() did not wait at the two places where it enables
direct mode. It worked on the boards we tested because the cooldown
ended before the first chip select. Add the two waits, and put the
BUSY loop in one RAM-resident helper.
Assisted-by: Claude Code:claude-opus-5-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
An enabled ICACHE does not manage write transactions: it flags cacheable
writes as errors (ICACHE_SR.ERRF), and RM0481 8.4.5 recommends modifying
the memory with the ICACHE disabled. On an STM32H563, erasing and
programming a flash block with the ICACHE enabled leaves ICACHE_SR at
0x6 (BSYENDF and ERRF set).
The ICACHE also keeps serving lines cached before the change. When the
block had been read through the ICACHE just before it was programmed,
up_progmem_write() failed its read-back check with -EIO: the flash held
the new data, with no flash or ECC error flagged, but the read-back hit
the cached erased data. up_progmem_eraseblock() fails its erased-range
check the same way when programmed data of the block is cached.
Disable the ICACHE for the duration of up_progmem_eraseblock() and
up_progmem_write(), and enable it again afterwards if it was enabled on
entry. Disabling it invalidates it, so the refill after re-enabling it
sees the new flash content.
If the ICACHE cannot be re-enabled because its invalidate times out, it
is left disabled, which is safe but slower, and an error is logged.
Assisted-by: Claude:claude-sonnet-5-5
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Ricardo Maurizio Paul <ricardopaul@geotab.com>
A bootloader may hand over with the ICACHE enabled, and nothing waits for
the invalidate that runs when the ICACHE is disabled. The driver also
polled BUSYF without a bound, so a stuck flag would hang the boot, and it
ignored an invalidate that never finished.
- stm32_enable_icache(): on first use, disable and invalidate the ICACHE
before the associativity and region registers are written (they are
only writable while EN=0), and wait for any pending invalidate before
enabling it (RM0481 8.4.5).
- stm32_disable_icache(): wait for the invalidate that EN=0 starts and
clear BSYENDF and ERRF.
- Bound every BUSYF wait with STM32_ICACHE_BUSY_TIMEOUT. RM0481 gives no
invalidate duration, so the value is a margin, not a measured limit.
- stm32_enable_icache() now returns OK or -ETIMEDOUT instead of void. On
a timeout the ICACHE is left disabled: with BUSYF stuck it would not
cache anything anyway (RM0481 8.4.5). Existing callers ignore the result
and keep working.
- __start: when CONFIG_STM32_ICACHE is not set, disable an ICACHE left on
by a bootloader, so reads of the OTP and UID cannot fault. On an
STM32H563 with the ICACHE left enabled this way, a 16-bit read of the
UID raised a precise bus fault and up_progmem_write() failed its
read-back check with -EIO.
Assisted-by: Claude:claude-sonnet-5-5
Signed-off-by: Ricardo Maurizio Paul <ricardopaul@geotab.com>
stm32_get_uniqueid() and flash_read_eccsafe16() (OTP and EDATA word
reads) disabled the ICACHE around their reads and enabled it again
afterwards, so that the read did not go through the ICACHE. The MPU
region added by the previous commit makes these areas non-cacheable, so
the reads bypass the ICACHE anyway.
Remove the disable/enable pairs. Each pair also invalidated the whole
ICACHE, and flash_read_eccsafe16() did it with interrupts disabled, twice
for every 32-bit OTP read.
The MPU is not applied in the HardFault and NMI handlers (HFNMIENA=0),
so a UID read from those handlers is no longer protected. Nothing in the
tree does that.
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Ricardo Maurizio Paul <ricardopaul@geotab.com>
The OTP, read-only (UID, flash size, package) and high-cycle data (EDATA)
flash areas only accept 16/32-bit accesses and return a bus error
otherwise (RM0481 Table 77). The manual requires the MPU to disable local
cacheability for them (RM0481 7.3.2); with the ICACHE enabled and no such
region, reading them raises a precise bus error.
Until now this was worked around piecemeal: the driver disabled the
ICACHE around stm32_get_uniqueid() and the OTP and EDATA word reads, and
nucleo-h563zi mapped the 4 KB OTP/RO area non-cacheable in its board
code. Other reads, for example stm32_otp_read() or an application
reading the OTP on another board, still raised a precise bus error when
the ICACHE was enabled.
Map 0x08fff000-0x09017fff, which covers the three contiguous areas, as
Normal non-cacheable and execute-never with a single MPU region before
the ICACHE is enabled. STM32_ICACHE now selects ARM_MPU so that
stm32_mpuinitialize() has reset and enabled the MPU by then.
Remove the nucleo-h563zi OTP region in the same commit: the Armv8-M MPU
faults on an address that matches more than one region, so keeping both
would make OTP and UID reads fault on that board.
Assisted-by: Claude:claude-sonnet-5-5
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Ricardo Maurizio Paul <ricardopaul@geotab.com>
An ICACHE build with no ICACHE region configured warns about two unused
symbols:
- stm32_icache_setup_region() is only called when one of the
CONFIG_STM32_ICACHE_REGION0..3 options is set, so build it only then.
- 'regval' in stm32_icache_initialize() is only used with
CONFIG_STM32_ICACHE_DIRECT, so declare it only then. The interrupt
block gets its own local variable.
No functional change.
Assisted-by: Claude:claude-sonnet-5-5
Signed-off-by: Ricardo Maurizio Paul <ricardopaul@geotab.com>
Adds the Data Transfer Controller (DTC) as a generic register-triggered
DMA engine (ra_dtc.c/.h, hardware/ra8m1_dtc.h, CONFIG_RA_DTC, normal
mode only), and generalizes SCI_B's 16-stage FIFO and DTC support from
SCI9-only to every instance (SCI0-4, SCI9): CONFIG_RA_SCIn_FIFO and
CONFIG_RA_SCIn_TXDTC/RXDTC are mutually exclusive, not combinable --
TX DTC moves queued ring-buffer data in the background through
CONFIG_SERIAL_TXDMA, RX DTC moves each byte as it arrives, and measured
on hardware the FIFO's standing multi-byte buffer gives far better
high-baud overrun margin than this one-byte-per-activation RX DTC
design, which also had a real arming bug fixed here (up_rxint()'s old
"first enable" check could never fire, so the DTC was never armed),
alongside a FIFO register-clear/ordering fix. up_ioctl() also gains
TCGETS/TCSETS (baud rate, parity, stop bits), board.h gains SCI0's
GPIO pins (P609/P610) to exercise it on this board, and comments
throughout ra_serial.c/Kconfig are condensed from the narrative
debugging form they accumulated down to the essential fact each one
needs. Tested on hardware: SCI0 and SCI9 simultaneously, FIFO and DTC
independently and mutually exclusive; FIFO (TTRG=15/RTRG=0) holds up
cleanly to 3 Mbps for transfers within the RX ring buffer, well past
RX DTC's ceiling (see below).
A separate, more serious bug was also found and fixed while hardware-
testing the overrun path at high baud: up_erinterrupt() never cleared
its ICU IELSRn.IR flag (RA8M1 User's Manual section 13.5.1 is explicit
that this causes the NVIC to re-enter the handler indefinitely), which
livelocked the whole system on any receive error, in every
configuration, not something specific to this series. Fixed by
clearing it like the other two SCI interrupt handlers already did, by
draining RDR/the FIFO on an error before clearing it (SCIn_ERI fires
instead of SCIn_RXI for every receive error, so data already received
was otherwise never picked up), and by having the RX DTC path also
recognize a byte stuck via CSR.RDRF directly, not just its own
bookkeeping, since its event can be silently dropped by the ICU while
an earlier byte's IR is still pending. Also added CONFIG_SERIAL_
TIOCGICOUNT (frame/overrun/parity counters via the standard
TIOCGICOUNT ioctl), matching stm32's precedent, since nothing in the
generic NuttX serial core surfaces a receive error to an application
otherwise. Confirmed on hardware: nsh on SCI9 stays fully responsive
through a sustained 3 Mbps overrun on SCI0 that previously froze the
whole board.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Assisted-by: Claude:claude-sonnet-5
Signed-off-by: leocafonso <leocafonso@gmail.com>
Add an opt-in STM32N6_DEBUG setting to reopen the debug access port and
secure/non-secure debug at the current BSEC protection level. Enable the
BSEC clock and configure access before clock and memory initialization so
a debugger can attach to a flash-booted development image.
Assisted-by: Codex:GPT-6
Signed-off-by: raiden00pl <raiden00@railab.me>
In the nominal ROM clock configuration, flash boot leaves PLL1 driving
the CPU at 400 MHz. Skipping clock setup when PLL1 is already selected
makes SysTick run twice as fast as the board's 200 MHz configuration
expects.
Switch CPU and system clocks to HSI before reconfiguring PLL1, then apply
the board clock tree.
Remove the incorrect comments claiming CFGR1 and CFGR2 lock after the first
clock switch.
Assisted-by: Codex:GPT-6
Signed-off-by: raiden00pl <raiden00@railab.me>
