riscv : add riscv qemu virt support and fix fs bit error in mstatus

[Jer6y]

Hi , I port threadx to qemu-riscv64-virt and run the demo

I also add toolchain cmake scripts , so we can use cmake like arm to generated libthreadx.a

To Test it

First , Maksure riscv64-unknown-elf- binutils and make is installed , and qemu-system-riscv64 is also required

cd into example_build work

cd ports/risc-v64/gnu/example_build/qemu_virt

use make to compile the demo (make clean can clean the output)

make

If done successfully, we will find A ELF FILE kernel.elf , use qemu to load it

qemu-system-riscv64 -nographic -smp 1 -bios none -m 128M -machine virt -kernel kernel.elf

If success , you will get different thread output the message like this

[UART0] : Uart Init Done, this is Test output!
[Thread] : thread_0_entry is here!
[Thread] : thread_5_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_1_entry is here!
[Thread] : thread_2_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_0_entry is here!
[Thread] : thread_5_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_0_entry is here!
[Thread] : thread_5_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!
[Thread] : thread_3_and_4_entry is here!
[Thread] : thread_6_and_7_entry is here!

PR checklist

• add riscv qemu virt support , add cmake toolchain script to generate libthreadx.a
• fix riscv64 contex_resotre bug (when f/d isa is enable)
• use weak symbol to invalidate _tx_initialize_low_level function

[fdesbiens]

Hi @Jer6y.

Thank you for your contribution! I am the new Eclipse ThreadX project lead. I am still ramping up, so I couldn't comment right away. I appreciate your patience.

I lack the knowledge and experience to review the code thoroughly myself, so I will do my best to get someone else to look at it in the upcoming weeks. In the meantime, I am curious about the history of this contribution. What led you to write this port? Feel free to contact me privately if you cannot share those details publicly.

[fdesbiens]

@eclipse-threadx/iot-threadx-committers, I need a volunteer with porting expertise to review this contribution.

[Jer6y]

Hi @Jer6y.

Thank you for your contribution! I am the new Eclipse ThreadX project lead. I am still ramping up, so I couldn't comment right away. I appreciate your patience.

I lack the knowledge and experience to review the code thoroughly myself, so I will do my best to get someone else to look at it in the upcoming weeks. In the meantime, I am curious about the history of this contribution. What led you to write this port? Feel free to contact me privately if you cannot share those details publicly.

Hi,
I am working on this porting task because I found that I can't directly use the ThreadX project on RISC-V. There are some issues with the code in terms of architecture code, and I can't find the way to run ThreadX on qemu-system-riscv64 for verification and learning.

[fdesbiens]

@rahmanih I think you mentioned you have feedback for the author of this pull request. Can you please share it with them here?

[rahmanih]

@rahmanih I think you mentioned you have feedback for the author of this pull request. Can you please share it with them here?

Hi,
I misunderstood some points, but it is clear now, LGTM, if no more comments, these can be merged.
regards
Haithem.

[fdesbiens]

@eclipse-threadx/iot-threadx-committers

@rahmanih Reviewed this contribution and determined it meets our quality standards.

I would like all the others to explicitly approve (+1) or reject (-1) this contribution. Please express your opinion by Tuesday, February 11, 2025. I will not vote myself unless there is a tie.

If the vote is positive, I will merge the code and this feature will ship with our next release.

[billlamiework]

+1

[fdesbiens]

This contribution is approved. Thanks for submitting this PR, @Jer6y!

[fdesbiens]

This feature will ship in ThreadX v6.4.2. We expect to release it by the end of February 2025.

[akifejaz]

Hi @fdesbiens, there seems to be a problem with this. I'was testing this and I found a bug, please see the details below.

The task 1,2 which are using tx_queue_send()/tx_queue_receive() are not working as expected they are failing and exiting as soon as the task enters. It is because tx_queue_create failed with status code TX_PTR_ERROR = 0x3.

see logs:

(gdb) c
Continuing.

Breakpoint 1, main () at demo_threadx.c:62
62          tx_kernel_enter();
(gdb) b demo_threadx.c:147
Breakpoint 2 at 0x80000b02: file demo_threadx.c, line 147.
(gdb) c
Continuing.

Breakpoint 2, tx_application_define (first_unused_memory=0x8000e000) at demo_threadx.c:147
147         tx_byte_allocate(&byte_pool_0, (VOID **) &pointer, DEMO_QUEUE_SIZE*sizeof(ULONG), TX_NO_WAIT);
(gdb) info reg ra
ra             0x80000b02       0x80000b02 <tx_application_define+682>
(gdb) print $ra
$1 = (void (*)()) 0x80000b02 <tx_application_define+682>
(gdb) n
150         tx_queue_create(&queue_0, "queue 0", TX_1_ULONG, pointer, DEMO_QUEUE_SIZE*sizeof(ULONG));
(gdb) s
_txe_queue_create (queue_ptr=0x80009320 <queue_0>, name_ptr=0x80008130 "queue 0", message_size=1, queue_start=0x0, queue_size=800, queue_control_block_size=104)
    at /home/akif-10xe/projects/threadx-work/threadx/common/src/txe_queue_create.c:97
97          status =  TX_SUCCESS;
(gdb) info local
interrupt_save = 0
status = 0
i = 2147540936
next_queue = 0x320
thread_ptr = 0x8000285a <_txe_byte_allocate+264>
(gdb) n
100         if (queue_ptr == TX_NULL)
(gdb) 
108         else if (queue_control_block_size != (sizeof(TX_QUEUE)))
(gdb) l
103             /* Queue pointer is invalid, return appropriate error code.  */
104             status =  TX_QUEUE_ERROR;
105         }
106
107         /* Now check for a valid control block size.  */
108         else if (queue_control_block_size != (sizeof(TX_QUEUE)))
109         {
110
111             /* Queue pointer is invalid, return appropriate error code.  */
112             status =  TX_QUEUE_ERROR;
(gdb) n
118             TX_DISABLE
(gdb) b txe_queue_create.c:235
Breakpoint 3 at 0x800031ea: file /home/akif-10xe/projects/threadx-work/threadx/common/src/txe_queue_create.c, line 237.
(gdb) c
Continuing.

Breakpoint 3, _txe_queue_create (queue_ptr=0x80009320 <queue_0>, name_ptr=0x80008130 "queue 0", message_size=1, queue_start=0x0, queue_size=800, queue_control_block_size=104)
    at /home/akif-10xe/projects/threadx-work/threadx/common/src/txe_queue_create.c:237
237         return(status);
(gdb) info local
interrupt_save = 9223372079804479616
status = 3
i = 0
next_queue = 0x0
thread_ptr = 0x8000285a <_txe_byte_allocate+264>
(gdb) n
238     }
(gdb) 
tx_application_define (first_unused_memory=0x8000e000) at demo_threadx.c:153
153         tx_semaphore_create(&semaphore_0, "semaphore 0", 1);

See the return value in _txe_queue_create.

Moreover, as per the logs we can also see the thread 1,2 only came ones which means they got exited in first iter as this condition if (status != TX_SUCCESS) break in thread 1 entry.

(gdb) b thread_1_entry
Breakpoint 1 at 0x80000c58: file demo_threadx.c, line 213.
(gdb) c
Continuing.

Breakpoint 1, thread_1_entry (thread_input=1) at demo_threadx.c:213
warning: Source file is more recent than executable.
213             puts("[Thread] : thread_1_entry is here!");
(gdb) info local
status = 0
(gdb) n
215             thread_1_counter++;
(gdb) 
218             status =  tx_queue_send(&queue_0, &thread_1_messages_sent, TX_WAIT_FOREVER);
(gdb) 
221             if (status != TX_SUCCESS)
(gdb) info local
status = 9
(gdb) l
216
217             /* Send message to queue 0.  */
218             status =  tx_queue_send(&queue_0, &thread_1_messages_sent, TX_WAIT_FOREVER);
219
220             /* Check completion status.  */
221             if (status != TX_SUCCESS)
222                 break;
223
224             /* Increment the message sent.  */
225             thread_1_messages_sent++;
(gdb) s
222                 break;
(gdb) s
227     }

let me know if someone is interested to debug this further or please assign this to me I'll look into this.

PS: I think the expected output should be like this #492