We can save execution time by
inline arm64_fullcontextrestore and arm64_switchcontext
test:
We can use qemu for testing.
compiling
make distclean -j20; ./tools/configure.sh -l qemu-armv8a:nsh_smp ;make -j20
running
qemu-system-aarch64 -cpu cortex-a53 -smp 4 -nographic -machine virt,virtualization=on,gic-version=3 -net none -chardev stdio,id=con,mux=on -serial chardev:con -mon chardev=con,mode=readline -kernel ./nuttx
Signed-off-by: hujun5 <hujun5@xiaomi.com>
This commits replicates some changes on files that were originally
written for other Espressif SoCs and inspired new implementations.
Eventually, these new implementations were improved and this commit
replicates these changes on the original sources.
Avoid using static mutex and recursive mutex as the resource to be
acquired/release. Instead, create a specific lock for each call if
it does not exist.
It seems like a wrong copy-and-paste from esp32c3.
Actually, internal memory mapping varies among processors.
esp32s3: lightly tested with wamr aot
esp32s2: not tested (i have no hardware access)
The bc022f8cd8 introduces a static way to calculate idle stack
address for risc-v platform. However, it uses the reverse order
to access idle stack, which breaks boards with smp configuration.
Correct the idle stack order of g_cpux_idlestack.
Fixes: bc022f8cd8 ("arch/risc-v: unify idle stack calculation")
Signed-off-by: Inochi Amaoto <inochiama@outlook.com>
As all the board allocate idle stack from _ebss. The idle stack
layout is fix and can be computed directly by using _ebss. There
is no need to use g_cpux_idlestack array anymore, remove it.
Signed-off-by: Inochi Amaoto <inochiama@outlook.com>
As the stack layout is unified, it is possible to
use a unify formula to calculate the top of idle
stack.
Signed-off-by: Inochi Amaoto <inochiama@outlook.com>
chip/s698pm_cpustart.c: In function 's698pm_cpu_boot':
Error: chip/s698pm_cpustart.c:74:17: error: unused variable 'tcb' [-Werror=unused-variable]
struct tcb_s *tcb = this_task();
^~~
cc1: all warnings being treated as errors
make[1]: *** [Makefile:99: s698pm_cpustart.o] Error 1
make[1]: Target 'libarch.a' not remade because of errors.
make: *** [tools/LibTargets.mk:164: arch/sparc/src/libarch.a] Error 2
make: Target 'all' not remade because of errors.
/github/workspace/sources/nuttx/tools/testbuild.sh: line 370: /github/workspace/sources/nuttx/../nuttx/nuttx.manifest: No such file or directory
Normalize s698pm-dkit/smp
Signed-off-by: hujun5 <hujun5@xiaomi.com>
Make channels synchronous (i.e. share the same timebase) with the help
of SAMV7_PWMx_CHy_SYNC defines. All the channels share the same
timebase of channel 0, so this channel must be defined too.
Signed-off-by: Stepan Pressl <pressste@fel.cvut.cz>
ESP32 SoC use a static allocated array as idle stack. To fit
the existed idle stack allocation, make idle stack allocated
from ebss for the whole esp32 series.
Signed-off-by: Inochi Amaoto <inochiama@outlook.com>
Optimal size of granule is 64B (the dcache line size).
We can use it now as we don't have max. 32 granules limitation anymore.
Signed-off-by: Jani Paalijarvi <jani.paalijarvi@unikie.com>
This adds SBI specfication v0.3 based `riscv_sbi_system_reset()` to
support SBI firmware based system reset in kernel mode.
Signed-off-by: Yanfeng Liu <yfliu2008@qq.com>
1. Adjust code to avoid PM wakelock->count less than or equal to 0.
2. Fix some document format issues.
Signed-off-by: chenwen@espressif.com <chenwen@espressif.com>
Although almost all board support allocating idle stack after ebss,
bl602 have a different memory layout for idle stack. To unify them,
make idle stack allocated from ebss for bl602.
Signed-off-by: Inochi Amaoto <inochiama@outlook.com>
NuttSBI have a simple ecall interface for the kernel, which make
it hard to add new SBI call for NuttSBI. So implement standard
ecall interface for NuttSBI and make life easier.
Signed-off-by: Inochi Amaoto <inochiama@outlook.com>
With CONFIG_MMCSD_MULTIBLOCK_LIMIT not set. (No limit)
The DMA driver would overwrite the internal buffer.
By adding CONFIG_ARCH_HAVE_SDIO_PREFLIGHT and
CONFIG_FAT_DMAMEMORY we can insure alignment and
maximize performance using no CONFIG_MMCSD_MULTIBLOCK_LIMIT
Using user allocated buffers for DMA transfers is not safe for two reasons:
- User space memory is virtual memory, DMA needs physical memory
- User memory buffer alignment cannot be guaranteed -> cache line ops
are not safe
Add a simple allocator for DMA safe memory. It will provide contiguous
blocks of memory with D-Cache line size alignment.
NOTE: The optimal granule size is the D-Cache line size (64), but due
to restrictions in the granule allocator this would result in a maximum
block size of 2K only, thus use 256B granules instead givin 8K max block
size.
Once the granule allocator is fixed this limitation can be removed.
Option CONFIG_SENSORS_QENCODER might be configured even if SAMv7 qencoder
over timer counter is not used (for example encoder over GPIO is selected
with CONFIG_SAMV7_GPIO_ENC). This can cause compile warnings, also build
of sam_qencoder.c file is unnecessary in that case.
New hidden option CONFIG_SAMV7_QENCODER is added and automatically
selected if at least one timer counter is enabled for qencoder. Build
is triggered on this option.
Signed-off-by: Michal Lenc <michallenc@seznam.cz>
PWMx_CHy_LONLY options have been added to Kconfig, too.
If LONLY is selected, it's not possible to use complementary outputs.
If LONLY is not selected, it's possible to use H or complementary
output. If configured correctly with cpol and dcpol attributes,
a H-like behaviour can be achieved. May be useful when you run out
of free MCU pins.
Signed-off-by: Stepan Pressl <pressste@fel.cvut.cz>
