| Age | Commit message (Collapse) | Author |
|---|
|
Changes:
- create new folder name `common64` and save all common codes of rv64
inside
Signed-off-by: Shell <[email protected]>
|
|
DMA and CPU address view is different, we need to convert them:
+--------+ +--------+ +---------+ +--------+
| | | | | | | |
| CPUs | | DEV0 | | IOMMU <----+ | DEV1 |
| | | | | | | | |
+----+---+ +----+---+ +----+----+ | +----+---+
| | | | |
0x200000 | 0x1000 | 0x1000 | | 0x8000 |
| | | | |
+-------------+-----------+ +----------+
|
|
+--------v----------------------------------------------------+
| |
| Address BUS |
| |
+-------------------------------------------------------------+
Signed-off-by: GuEe-GUI <[email protected]>
|
|
Maybe use for memory/DMA buffer init before
ofw_node unflatten.
Signed-off-by: GuEe-GUI <[email protected]>
|
|
The ofw parse should:
1. Check obj_name EQU the current node's rt_data.
2. Find the next object name.
3. goto "2" until obj_name EQU the cmp_cell's obj_name.
Signed-off-by: GuEe-GUI <[email protected]>
|
|
Signed-off-by: GuEe-GUI <[email protected]>
|
|
Signed-off-by: GuEe-GUI <[email protected]>
|
|
rt_ofw_append_child append a new child node and return,
ref_get is necessary.
Signed-off-by: GuEe-GUI <[email protected]>
|
|
Signed-off-by: GuEe-GUI <[email protected]>
|
|
netdev的函数不会在中断中调用，无需关中断
|
|
|
|
|
|
|
|
此目录中为rv64通用代码
|
|
Signed-off-by: Yilin Sun <[email protected]>
|
|
* [action][cppcheck] add summary
添加summary 方便查看出错信息
|
|
'--check-level=exhaustive'
这个选项可以更深入的在分支中检测，但是会耗时比较久一些。
|
|
This patch consolidates the separated architecture-specific code for
rv64 (virt64 and c906) under a more unified approach. The changes
aim to enhance maintainability and code reuse, reducing duplication
between these two architectures while adding small improvements in
porting compatibility.
Changes:
- Modified build scripts (SConscript) for both virt64 and c906 to
remove ASID and vector dependencies when not required.
- Updated c906's sbi.c and sbi.h to use standard integer types
(uint32_t) and include the missing <stdint.h> header.
- Unified inline function declaration for `sbi_call` across both
c906 and virt64 using `rt_inline`.
- Disabled FPU and vector in c906's startup assembly file, aligning it
with the virt64 handling.
- Corrected syscall handler type definitions in c906 for consistency.
Signed-off-by: Shell <[email protected]>
|
|
This patch aims to unify the two currently separated RISC-V 64-bit
architecture ports, 'virt64' and 'c906', into a single generic
'common64' port. The changes include renaming files and updating
includes to use a unified 'interrupt.h' header, as well as making
adjustments to IO and trap handling to be more consistent between the
two architectures.
Changes:
- Renamed 'rt_interrupt.h' to 'interrupt.h' and updated includes accordingly.
- Unified IO register access functions in 'riscv_io.h'.
- Added 'opcode.h' for portable assembly support.
- Updated 'plic.c' and 'plic.h' to handle interrupts in a unified manner.
- Modified 'trap.c' to handle exceptions and interrupts consistently for 'rv64'.
Signed-off-by: Shell <[email protected]>
|
|
These changes are designed to standardize the memory management across
'virt64' and 'c906', ensuring efficient handling of address spaces and
page tables.
Changes:
- Creation of ASID management files (`asid.c`) for both 'c906' and
'virt64' architectures, which is essential for maintaining stability.
- Extensive updates to the MMU configuration and handling in `mmu.c` and `mmu.h` files.
- Addition of functions to manage ASID allocation and switching of page tables.
- For c906, accommodated the early memory setup to the one from virt64.
Signed-off-by: Shell <[email protected]>
|
|
Changes:
- added config for NEW_CTX_SWITCH
- used new context on c906 sched switch routine
- update _rt_hw_stack_init to fit into NEW_CTX_SWITCH
- separated vector ctx from the generic
Signed-off-by: Shell <[email protected]>
|
|
|
|
|
|
|
|
check log
|
|
When the driver request a API gets RT_NULL which return value is ptr,
they could not know why get a RT_NULL.
some API return RT_NULL, is not error internal maybe, it just not
supported for this platform, but the driver still could work ok,
the API can return (RT_NULL + -RT_EEMPTY) to driver.
on the other hand, the driver can do more behaviors by error no.
When the API return the -RT_EBUSY, driver can wait for a moment and retry.
When the API return the -RT_ENOSYS, driver can try the next mode or request's name.
Signed-off-by: GuEe-GUI <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
udp一个socket跟多个地址通讯bug修复、DNS加锁防止多线程错乱、实现at_gethostbyname_r、稳定性优化
|
|
|
|
|
|
PCI/PCIe have better performance and more devices support, such as
NVMe, GPU, Powerful NIC (Like RDMA). PCI/PCIe can access control by
IOMMU that the virtualiztion and userspace driver will more safety.
PCI/PCIe device could hot plugging, no design modifications SoC required,
PCI/PCIe on Embedded SoC is popular now.
We make a simple framework to support them.
Feature Lists:
1.PCI INTx: the INT[A-D] pin IRQ for legacy PCI, work with platform PIC.
2.MSI/MSI-X: the message write IRQ for PCIe, work with platform's PIC.
3.PME: we only support the D0, D1, D2, D3HOT, D3COLD init by framework.
4.Endpoint: a simple EP framework for PCI FPGA or NTB function.
5.OFW: we only support work on OFW SoC, ACPI support in the future maybe.
Host controller:
1. Common PCI host controller on ECAM.
2. Generic PCI host controller on ECAM.
Signed-off-by: GuEe-GUI <[email protected]>
|
|
1. RT_FIELD_PREP: prepare a bitfield element.
2. RT_FIELD_GET: extract a bitfield element.
3. rt_offsetof: member offset of a struct
4. rt_upper_32_bits: high 32 bits of value.
5. rt_lower_32_bits: lower 32 bits of value.
6. rt_upper_16_bits: high 16 bits of value.
7. rt_lower_16_bits: lower 16 bits of value.
8. rt_max_t: fix type of max(...).
9. rt_ilog2: integer logarithm base 2.
Signed-off-by: GuEe-GUI <[email protected]>
|
|
* add mutex_pi
* fix assertion error when thread wakeup while waiting mutex
|
|
|
|
CV18xx RISC-V C906L.
Configuration value for CPU and ARCH are already defined in rtconfig.py.
Remove these duplication from SConstruct.
Signed-off-by: Chen Wang <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
为什么提交这份PR (why to submit this PR)
Issue: using iperf server with sshd together and the kernel will get stuck occasionally.
taken_list is accessed from other mutex when updating priority.
你的解决方案是什么 (what is your solution)
protect taken_list on shared lock, but not on mutex local spinlock.
Signed-off-by: Shell <[email protected]>
|
|
After first called from uart open, the dw8250_uart_configure
API cannot be called again. Otherwise, this will mess up the
device, and uart will not act properly to interrupt again.
Analysis: Configure uart device will close recive interrupte,
causing uart device to malfunction.
Solution: After configure uart device, enable the device's
recive interrupte.
Signed-off-by: Shicheng Chu <[email protected]>
Reviewed-by: Chen Wang <[email protected]>
|
|
|
|
Signed-off-by: GuEe-GUI <[email protected]>
|
|
The bus should call driver's `remove` callback not
instead of `shutdown` in rt_bus_remove_device.
Signed-off-by: GuEe-GUI <[email protected]>
|
