When supporting high-priority interrupts, updating the
g_running_tasks within a high-priority interrupt may be
cause problems. The g_running_tasks should only be updated
when it is determined that a task context switch has occurred.
Signed-off-by: zhangyuan21 <zhangyuan21@xiaomi.com>
Instead of clearing the fields individually, just wipe the whole register.
This can be done because flags and rm are just parts of the fcsr.
31 8 5 0
+--------------+--------+-----------+
| | | |
| RESERVED | FRM | FSTATUS |
| | | |
+--------------+--------+-----------+
FCSR
- Save the FPU registers into the tcb so they don't get lost if the stack
frame for xcp.regs moves (as it does)
- Handle interger and FPU register save/load separately
- Integer registers are saved/loaded always, like before
- FPU registers are only saved during a context switch:
- Save ONLY if FPU is dirty
- Restore always if FPU has been used (not in FSTATE_OFF, FSTATE_INIT)
- Remove all lazy-FPU related logic from the macros, it is not needed
Why? The tcb can contain info that is needed by the context switch
routine. One example is lazy-FPU handling; the integer registers can
be stored into the stack, because they are always stored & restored.
Lazy-FPU however needs a non-volatile location to store the FPU registers
as the save feature will skip saving a clean FPU, but the restore must
always restore the FPU registers if the thread uses FPU.
Basically, it reserves an area of the RTC memory to preserve the
RTC timer.
Please refer to:
fa76c82a5b
This commit also removes the rtc.dummy section because C3, C6 and
H2 don't need to skip it once the region is accessed by the same
address space using the instruction and data bus.
1. add lib_fork api in libs/libc, we need a fork() api to implement the
fork relative method, such as pthread_atfork
2. rename the assembly fork entry function name to up_fork(), and rename
the up_fork() to arch specific name, such as
sim_fork()/arm_fork()/mips_fork() etc.
Signed-off-by: guoshichao <guoshichao@xiaomi.com>
Clone the esp-hal-3rdparty repository and, then, checkout to a
specific version (usually indicated by the commit SHA) instead of
using a branch as the HEAD.
Also switch to automatic one by default, as it saves a lot of flash. The manual training
code is left in for now to be able to use it as an option if there are problems with automatic one
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
The implementation of up_perf_*() is in a different static library in nuttx:
Hardware: libarch.a
Software: libdrivers.a (weak function)
Since functions with weak attributes cannot be correctly replaced in multiple static libraries,
this PR will use macros to replace whether the arch supports hardware perf events
Signed-off-by: chao an <anchao@xiaomi.com>
If -fstack-protector-all is enabled, gcc linker will need GCC
SSP(Stack Smashing Protector) support, Since the implement of SSP
is related to the OS, most of embedded toolchain does not provide
ssp support, so an error will be reported when linking:
enable CONFIG_LTO_FULL && CONFIG_STACK_CANARIES
arm-none-eabi/bin/ld: cannot find -lssp_nonshared: No such file or directory
arm-none-eabi/bin/ld: cannot find -lssp: No such file or directory
https://github.com/gcc-mirror/gcc/blob/master/gcc/gcc.cc#L983-L985
Since nuttx has already implemented SSP related hook functions,
so in this PR, we filter out this option in the link phase to ensure that
the implementation of lssp/lssp_nonshared will not be referenced
Signed-off-by: chao an <anchao@xiaomi.com>
Writes to MPFS_CFG_DDR_SGMII_PHY_EXPERT_DFI_STATUS_OVERRIDE register were not done properly. Use correct address for writes.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Reading the CSD field misses 3 bytes as the residual bytes
are not carried over properly. Fix this by adding the missing
bytes due to shifting.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
Remove TABs
Fix indentation
Fix Multi-line comments
Fix Comments to the Right of Statements.
Fix nuttx coding style
Fix Comments to the Right of Statements.
1. as we can use fork to implement vfork, so we rename the vfork to
fork, and use the fork method as the base to implement vfork method
2. create the vfork function as a libc function based on fork
function
Signed-off-by: guoshichao <guoshichao@xiaomi.com>
Reference:https://xpack.github.io/blog/2022/05/18/riscv-none-elf-gcc-v12-1-0-2-released/
RISC-V ISA updates
Compared to previous releases, starting from 12.x, the compiler implements the new RISC-V ISA, which introduces an incompatibility issue, and builds might throw error messages like unrecognized opcode csrr.
The reason is that csr read/write (csrr*/csrw*) instructions and fence.i instruction were separated from the I extension, becoming two standalone extensions: Zicsr and Zifencei.
The solution is to add _zicsr and/or _zifencei to the -march option, e.g. -march=rv32imac becomes -march=rv32imac_zicsr_zifencei.
Initialize ICACHE way with correct mask.
Initialize scratchpad with constant g_init_marker as it has been done in HSS
Signed-off-by: Jani Paalijarvi <jani.paalijarvi@unikie.com>
Changes the executable type built against the `make export` target fully linked by default. This greatly improves performance when loading applications, as relocations no longer need to be processed.
In the default configuration, the Litex ethernet peripheral contains two RX and two TX buffers. The active buffer for the peripheral should be swapped as soon as possible, in order to reduce packet loss.
This modification acknowledges the receive buffer as soon as the pending data is copied into the NuttX device data buffer. Improving reliability under heavy load.
- Save the FPU registers into the tcb so they don't get lost if the stack
frame for xcp.regs moves (as it does)
- Handle interger and FPU register save/load separately
- Integer registers are saved/loaded always, like before
- FPU registers are only saved during a context switch:
- Save ONLY if FPU is dirty
- Restore always if FPU has been used (not in FSTATE_OFF, FSTATE_INIT)
- Remove all lazy-FPU related logic from the macros, it is not needed
Why? The tcb can contain info that is needed by the context switch
routine. One example is lazy-FPU handling; the integer registers can
be stored into the stack, because they are always stored & restored.
Lazy-FPU however needs a non-volatile location to store the FPU registers
as the save feature will skip saving a clean FPU, but the restore must
always restore the FPU registers if the thread uses FPU.
This adds support for the CoreMMC v3.1 FPGA driver as described
in Microchip Handbook HB0510. The driver doesn't support DMA.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
If a kernel stack exists, use that whenever the user process is in
privileged mode, i.e. running an exception or in system call. Previously
the exception context was stored into the user's stack, which is not ideal.
Why?
1. Because the exception entry status (REG_INT_CTX) is needed by the
kernel, and this is now in user memory which requires that the correct
user mappings are active when it is accessed.
2. The user must currently account for the exception stack frame (which
is BIG) in its own stack allocation. Moving the exception context save
to the kernel stack offloads this responsibility from the user to the
kernel, which is IMO the correct behavior.
3. The kernel access to user memory is currently allowed without condition,
however this is not ideal either. The privileged mode status CSR allows
blocking access to user memory via the STATUS_SUM-bit, which should be
disabled by default and only enabled when access to user space is really
needed. This patch allows implementing such features.
This is preparation to use kernel stack for everything when the user
process enters the kernel. Now the user stack is in use when the user
process runs a system call, which might not be the safest option.
This is a minimalistic SBI implementation for NuttX.
Provides a single service for now:
- Access to machine timer
Provides a start trampoline to start NuttX in S-mode:
- Exceptions / faults are delegated to S-mode.
- External interrupts are delegated to S-mode.
Machine mode timer is used as follows:
- The timer compare match register reload happens in M-mode, via
call gate "riscv_sbi_set_timer"
- The compare match event is dispatched to S-mode ISR, which will
notify the kernel to advance time
- Clearing the STIP interrupt does not work from S-mode,
so the call gate does this from M-mode
The only supported (tested) target for now is MPFS.
Remove unnecessary reading of the status register when loading / unloading
the FIFOs. Reading from the IP block is slow due to BUS synchronization and
this basically makes the SPI busy loop for no reason at all, destroying the
CPU usage.
The overall benefit of these changes is approx. 25%-points, which is a
MASSIVE improvement.
Adds a driver for an FPGA fabric / CoreSPI implementation.
Supports multiple instances, assuming they reside in some base address,
offsettable by a constant value.
External function to query vbus status. Reading from the block requires
the clock, but if no devices are open -> vbus detect does not work.
This creates a chicken / egg problem, if vbus detect is used to start
the usb device.
bmp388 works poorly as the system fires STOPs even in
a beginning of a transaction. Don't let unrelated STOPs
to distort the data flow.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
This adds 2 more FPGA I2Cs. Also rework the indexing
so that it matches the earlier work without major changes.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
This fixes the following issue:
- After sending the address, the driver writes an extra zero
Without this patch, the extra write causes an extra ACK that would
terminate the sequence prematurely. This is observed as data read
corruption.
With this fix, the condition is detected precisely. That being the
case, the sequence is continued with a repeated start, after which
the read continues normally.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
This incorporates an fpga i2c driver into the existing i2c driver.
This fpga i2c works almost 100% as the MSS i2c, but the difference
is that the fpga driver terminates all transactions with a stop
sent -interrupt. That needs to be handled.
Fpga clock source is also different, act accordingly.
SEC2TICK(10) is an overkill to any app, use just one second instead.
modifyreg32s are simplified as well, no need to clear and set
as set is enough.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
I2C status register reset value (0xf8) was not handled properly causing unnecessary bus resets.
Added critical section to mpfs_i2c_reset() and removed unnecessary interrupt disabling elsewhere.
to avoid the infinite recusive dispatch:
*0 myhandler (signo=27, info=0xf3e38b9c, context=0x0) at ltp/testcases/open_posix_testsuite/conformance/interfaces/sigqueue/7-1.c:39
*1 0x58f1c39e in nxsig_deliver (stcb=0xf4e20f40) at signal/sig_deliver.c:167
*2 0x58fa0664 in up_schedule_sigaction (tcb=0xf4e20f40, sigdeliver=0x58f1bab5 <nxsig_deliver>) at sim/sim_schedulesigaction.c:88
*3 0x58f19907 in nxsig_queue_action (stcb=0xf4e20f40, info=0xf4049334) at signal/sig_dispatch.c:115
*4 0x58f1b089 in nxsig_tcbdispatch (stcb=0xf4e20f40, info=0xf4049334) at signal/sig_dispatch.c:435
*5 0x58f31853 in nxsig_unmask_pendingsignal () at signal/sig_unmaskpendingsignal.c:104
*6 0x58f1ca09 in nxsig_deliver (stcb=0xf4e20f40) at signal/sig_deliver.c:199
*7 0x58fa0664 in up_schedule_sigaction (tcb=0xf4e20f40, sigdeliver=0x58f1bab5 <nxsig_deliver>) at sim/sim_schedulesigaction.c:88
*8 0x58f19907 in nxsig_queue_action (stcb=0xf4e20f40, info=0xf4049304) at signal/sig_dispatch.c:115
*9 0x58f1b089 in nxsig_tcbdispatch (stcb=0xf4e20f40, info=0xf4049304) at signal/sig_dispatch.c:435
*10 0x58f31853 in nxsig_unmask_pendingsignal () at signal/sig_unmaskpendingsignal.c:104
*11 0x58f1ca09 in nxsig_deliver (stcb=0xf4e20f40) at signal/sig_deliver.c:199
*12 0x58fa0664 in up_schedule_sigaction (tcb=0xf4e20f40, sigdeliver=0x58f1bab5 <nxsig_deliver>) at sim/sim_schedulesigaction.c:88
*13 0x58f19907 in nxsig_queue_action (stcb=0xf4e20f40, info=0xf40492d4) at signal/sig_dispatch.c:115
*14 0x58f1b089 in nxsig_tcbdispatch (stcb=0xf4e20f40, info=0xf40492d4) at signal/sig_dispatch.c:435
*15 0x58f31853 in nxsig_unmask_pendingsignal () at signal/sig_unmaskpendingsignal.c:104
*16 0x58f1ca09 in nxsig_deliver (stcb=0xf4e20f40) at signal/sig_deliver.c:199
*17 0x58fa0664 in up_schedule_sigaction (tcb=0xf4e20f40, sigdeliver=0x58f1bab5 <nxsig_deliver>) at sim/sim_schedulesigaction.c:88
*18 0x58f19907 in nxsig_queue_action (stcb=0xf4e20f40, info=0xf40492a4) at signal/sig_dispatch.c:115
*19 0x58f1b089 in nxsig_tcbdispatch (stcb=0xf4e20f40, info=0xf40492a4) at signal/sig_dispatch.c:435
*20 0x58f31853 in nxsig_unmask_pendingsignal () at signal/sig_unmaskpendingsignal.c:104
*21 0x58f1ca09 in nxsig_deliver (stcb=0xf4e20f40) at signal/sig_deliver.c:199
*22 0x58fa0664 in up_schedule_sigaction (tcb=0xf4e20f40, sigdeliver=0x58f1bab5 <nxsig_deliver>) at sim/sim_schedulesigaction.c:88
*23 0x58f19907 in nxsig_queue_action (stcb=0xf4e20f40, info=0xf4049274) at signal/sig_dispatch.c:115
*24 0x58f1b089 in nxsig_tcbdispatch (stcb=0xf4e20f40, info=0xf4049274) at signal/sig_dispatch.c:435
*25 0x58f31853 in nxsig_unmask_pendingsignal () at signal/sig_unmaskpendingsignal.c:104
*26 0x58f1ca09 in nxsig_deliver (stcb=0xf4e20f40) at signal/sig_deliver.c:199
*27 0x58fa0664 in up_schedule_sigaction (tcb=0xf4e20f40, sigdeliver=0x58f1bab5 <nxsig_deliver>) at sim/sim_schedulesigaction.c:88
*28 0x58f19907 in nxsig_queue_action (stcb=0xf4e20f40, info=0xf4049244) at signal/sig_dispatch.c:115
*29 0x58f1b089 in nxsig_tcbdispatch (stcb=0xf4e20f40, info=0xf4049244) at signal/sig_dispatch.c:435
*30 0x58f31853 in nxsig_unmask_pendingsignal () at signal/sig_unmaskpendingsignal.c:104
*31 0x58f1ca09 in nxsig_deliver (stcb=0xf4e20f40) at signal/sig_deliver.c:199
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Move the mapping functionality from up_shmat/shmdt into two generic
mapping functions. This makes it possible to do other mappings besides
user shared memory area mappings.
During the serial reconfiguration from bootloader to the
NuttX a trash character "?" (Unicode replacement U+FFFD)
was printed in the screen.
This fix was discovered by Sylvio Alves from Espressif!
Instead of using Espressif's emulated NVS to save Wi-Fi data, use
`wapi`s wireless configure initialization mechanism for saving
Wi-Fi data. It 1) avoids creating a specific storage partition
just to save Wi-Fi data (ESP32-C3's storage partition is used
instead); 2) avoids initialization problems of the emulated NVS
when SMP is enabled (the Wi-Fi driver tries to initialize it before
the actual partition is initialized); and 3) enables reconnecting
using `wapi reconnect` command and connect the device automatically
on bringup if `CONFIG_NETUTILS_NETINIT` is selected.
- putreg32() is used the wrong way around (reg, val) instead of (val, reg)
- MPFS_SPI_FRAMESIZE is not a register, FSIZE is the name
- Clear all interrupts _before_ writing the FIFO, this prevents a race
condition where a short transmission completes before the interrupt is
enabled.
in SMP, signal processing cannot be nested, we use xcp.sigdeliver to identify whether there is currently a signal being processed, but this state does not match the actual situation
One possible scenario is that signal processing has already been completed, but an interrupt occurs, resulting in xcp.sigdeliver not being correctly set to NULL,
At this point, a new signal arrives, which can only be placed in the queue and cannot be processed immediately
Our solution is that signal processing and signal complete status are set in the same critical section, which can ensure status synchronization
Signed-off-by: hujun5 <hujun5@xiaomi.com>
1. Get the value of sp from dump regs when an exception occurs,
to avoid getting the value of fp from up_getsp and causing
incomplete stack printing.
2. Determine which stack the value belongs to based on the value
of SP to avoid false reports of stack overflow
Signed-off-by: zhangyuan21 <zhangyuan21@xiaomi.com>
CURRENT_REGS may change during assert handling, so pass
in the 'regs' parameter at the entry point of _assert.
Signed-off-by: zhangyuan21 <zhangyuan21@xiaomi.com>
minidump will backtrace failure when use C code to save user context,
because the stack push operation in C code can disrupt the stack information.
Signed-off-by: zhangyuan21 <zhangyuan21@xiaomi.com>
- Fix DMA addressing issues within litex_sendsetup/litex_recvsetup
- Extend with handling specific to eMMC commands during init & use.
- Cleanup of 4-bit BUS handling for SD and eMMC
- For eMMC, Send CMD0 during init as per JEDEC v4.41 for pre-idle
Store the old environment in a local context so another temporary address
environment can be selected. This can happen especially when a process
is being loaded (the new process's mappings are temporarily instantiated)
and and interrupt occurs.
- Remove the temporary "saved" variable when temporarily changing MMU
mappings to access another process's memory. The fact that it has an
address environment is enough to make the choice
- Restore nxflat_addrenv_restore-macro. It was accidentally lost when
the address environment handling was re-factored.
Summary:
- Support arm64 pmu api, Currently only the cycle counter function is supported.
- Using ARM64 PMU hardware capability to implement perf interface, modify all
perf interface related code.
- Support for pmu init under smp.
Signed-off-by: wangming9 <wangming9@xiaomi.com>
Summary:
- I noticed that 'sleep 1' on nsh took 10 seconds on QEMU-6.1,
though the old version (e.g. QEMU-5.2) works correctly.
- I think we should implement PLL for the QEMU environment.
However, this fix works as a tentative solution.
Impact:
- K210 on QEMU only
Tested
- Tested with QEMU-7.1
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
RISCV has a modular instruction set. It's hard to define cpu-model to support all toolchain.
For Zig, cpu model is this formal: generic_rv[32|64][i][m][a][f][d][c]
For Rust, cpu model is this formal: riscv[32|64][i][m][a][f][d][c]
So, it's better to map the NuttX config to LLVM builtin cpu model, these models supported by
all LLVM based toolchain.
Refer to : https://github.com/llvm/llvm-project/blob/release/15.x/llvm/lib/Target/RISCV/RISCV.td
These models can't cover all implementation of RISCV, but it's enough for most cases.
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
This has been a long issue for me as it results in random crashes when
asynchronous events occur when the idle process is active.
The problem is that the kernel cannot access user memory, because the CPU
status prevents it.
Current Toolchain.defs set the compile flags directly, it's OK for
target specified gcc toolchain.
But some LLVM based toolchains (Rust/Zig etc) use single toolchain to handle all supported paltform.
In this patch, arch level Toolchain.defs export standard LLVM style arch flags, and let <Lang>.defs to map them into internal style,
This will simplify the intergration of non-c language.
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
* Remove the 'COEXISTENCE' suffix from Wi-Fi's STA + SoftAP mode.
Coexistence usually refers to Wi-Fi + BLE, instead of Wi-Fi's
operation mode;
* Remove commented debug code;
* Remove outdate function descriptions;
Initially supporting ESP32-C3 chip, to be followed by other RISC-V-based
chips from Espressif.
Signed-off-by: Gustavo Henrique Nihei <gustavo.nihei@espressif.com>
If transfer is restarted in irq handler the interrupts shall be
cleared before the start bit is set in control register. This is
to avoid ints being accidentally cleared before they are handled leading
to timeout error.
In spi_irq handler the data is written into txfifo and transfer
is started before the TXDONE interrupt is cleared. If the bus/memory
access is in some cases delayed, the spi transfer may have been
finished already before the interrupt register is cleaned for the
transfer. This leads the early arrived interrupt to be just removed
and never handled, which would cause a timeout error.
This patch moves the clearing of the interrupt to the place before
the tx is started, so the interrupt is not missed in above cases.
