[BUG] Unusable HRTimer

[Fix-Point]

Description / Steps to reproduce the issue

I fully-agree @anchao said in #17517 (comment)

There are various implementation approaches for hrtimer, but the benefit it delivers to business users is a high-precision timer. We need to examine this issue from the perspective of applications and business stakeholders—otherwise, any implementation would merely be an exercise in self-indulgence for developers.

HRtimer is intended to provide users with high-precision timer APIs. However, this HRtimer implemantation is completely unusable for users. In the words of @anchao, it is purely a self-indulgent exercise for developers.

The HRTimer I used is from #17517.
My configuration is rv-virt:smp, with CONFIG_HRTIMER and CONFIG_SCHED_TICKLESS enabled.

The test case used is as follows. This test case simulates a very typical driver application scenario: a shared device driver uses a common timeout timer, which may be accessed by multiple threads and therefore should be protected by a spinlock or a big kernel lock.

I attempted three approaches, and either the timer would crash inside the critical section, or it would sleep inside the critical section.
The first approach: hrtimer_cancel → hrtimer_start. This failed because if the callback function is executing, the timer cannot be started, and the system crashes.
The second approach: hrtimer_cancel → loop hrtimer_start. This was expected to succeed, but the system still crashed. Even if it succeeded, there would be a spinning issue within the critical section.
The third approach: hrtimer_cancel_sync → hrtimer_start. This succeeded, but it calls sleep within the critical section, which is unacceptable.

This problem stems from a fundamental flaw in the hrtimer state machine design. Minor modifications cannot fix it. Therefore, I strongly suggest this implementation should be removed and replaced with a better and completely different implementation (#17556).

The test code and error report is showed as below.

/****************************************************************************
 * apps/examples/hello/hello_main.c
 *
 * SPDX-License-Identifier: Apache-2.0
 *
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.  The
 * ASF licenses this file to you under the Apache License, Version 2.0 (the
 * "License"); you may not use this file except in compliance with the
 * License.  You may obtain a copy of the License at
 *
 *   http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
 * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.  See the
 * License for the specific language governing permissions and limitations
 * under the License.
 *
 ****************************************************************************/

/****************************************************************************
 * Included Files
 ****************************************************************************/

#include <nuttx/config.h>
#include <stdio.h>

#include <nuttx/hrtimer.h>

#define HRTIMER_TEST_THREAD_NR (CONFIG_SMP_NCPUS * 8)
/****************************************************************************
 * Public Functions
 ****************************************************************************/


static uint32_t test_callback(FAR struct hrtimer_s *hrtimer)
{
  printf(".");
  return 0;
}

static void* test_thread(void *arg)
{
  irqstate_t  flags; 
  hrtimer_t   timer;
  spinlock_t  lock = SP_UNLOCKED;
  hrtimer_init(&timer, test_callback, NULL);
  while (1)
    {
      uint64_t delay = rand() % NSEC_PER_MSEC;
      int ret;

      /* Simulate the usage of driver->wait_dog. */

      flags = spin_lock_irqsave(&lock);

      /* The driver lock acquired */

      /* First try, failed. Because hrtimer_start can not ensure the timer being started. */

      ret = hrtimer_cancel(&timer);
      // ret = hrtimer_start(&timer, 10 * NSEC_PER_USEC, HRTIMER_MODE_REL); /* May fail */
  
      /* This try-loop start should be OK. But it failed again.
       * Besides, we can not sleep or spin in the critical sections.
       */

      while (hrtimer_start(&timer, 10 % NSEC_PER_USEC, HRTIMER_MODE_REL) != OK);
      ret = OK;

      /* Second try, Success, but we can not sleep or spin in the critical section. */

      // ret = hrtimer_cancel_sync(&timer); /* Sleep in critical sections */
      // ret = hrtimer_start(&timer, delay, HRTIMER_MODE_REL); 
  

      spin_unlock_irqrestore(&lock, flags);

      if (ret != OK)
        {
          printf("hrtimer_start failed\n");
        }
      up_ndelay(delay);
    }
  return NULL;
}
/****************************************************************************
 * hello_main
 ****************************************************************************/

int main(int argc, FAR char *argv[])
{
  unsigned int   thread_id;
  pthread_attr_t attr;
  pthread_t      pthreads[HRTIMER_TEST_THREAD_NR];

  printf("hrtimer_test start...\n");

  ASSERT(pthread_attr_init(&attr) == 0);

  /* Create wdog test thread */

  for (thread_id = 0; thread_id < HRTIMER_TEST_THREAD_NR; thread_id++)
    {
      ASSERT(pthread_create(&pthreads[thread_id], &attr,
                            test_thread, NULL) == 0);
    }

  for (thread_id = 0; thread_id < HRTIMER_TEST_THREAD_NR; thread_id++)
    {
      pthread_join(pthreads[thread_id], NULL);
    }

  ASSERT(pthread_attr_destroy(&attr) == 0);

  printf("hrtimer_test end...\n");
  return 0;
}

➜   qemu-system-riscv32 -semihosting -M virt,aclint=on -cpu rv32 -smp 8 -bios none -nographic -kernel nuttx -s

NuttShell (NSH) NuttX-12.10.0
nsh> hello
hrtimer_test start...
................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................[CPU0] EXCEPTION: Load access fault. MCAUSE: 00000005, EPC: 80005960, MTVAL: 00000000
[CPU0] PANIC!!! Exception = 00000005
[CPU0] Current Version: NuttX  12.10.0 3663e78f82e-dirty Dec 19 2025 10:56:02 risc-v
[CPU0] Assertion failed panic: at file: :0 task(CPU0): hello process: hello 0x80016b7c
[CPU0] EPC: 80005960
[CPU0] A0: 8002be34 A1: 80032c78 A2: 00000000 A3: 80032c78
[CPU0] A4: 00000001 A5: 00000000 A6: ee6b2800 A7: 00000002
[CPU0] T0: 80029530 T1: 80028d30 T2: 00000000 T3: 00000000
[CPU0] T4: 00000000 T5: 00000000 T6: 00000000
[CPU0] S0: 80032c78 S1: 80032c78 S2: 80033e38 S3: 8002be34
[CPU0] S4: 00000000 S5: 00000001 S6: 00000002 S7: 00000008
[CPU0] S8: 00000003 S9: acd2225e S10: 00000000 S11: 00000000
[CPU0] SP: 80029488 FP: 80032c78 TP: 00000000 RA: 80005a62
[CPU0] IRQ Stack:
[CPU0]   base: 0x80028d30
[CPU0]   size: 00002048
[CPU0]     sp: 0x80029488
[CPU0] 0x80029468: 8002c000 0000002e 8003114c 2e007d90 8002c000 0000002e 80031148 8001b4ce
[CPU0] 0x80029488: 00000000 80007880 acd2696c 8002c000 8002c000 80007880 80033e38 800048cc
[CPU0] 0x800294a8: 800253fc 00000100 00000000 00000017 8002bd54 00000000 00000000 00000000
[CPU0] 0x800294c8: 80025448 3597c6e4 00000000 00000017 8002bd54 00000000 80029bd0 80009e30
[CPU0] 0x800294e8: 8002bd54 00000000 80029bd0 80000f64 00000000 00000000 80029bd0 80036084
[CPU0] 0x80029508: 8000000b 0000000a 00000002 80036210 80000007 80009de4 80029530 80000196
[CPU0] 0x80029528: 80036084 80006574 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] ERROR: Stack pointer 80029488 is not within the stack
[CPU0] User Stack:
[CPU0]   base: 0x80035ac0
[CPU0]   size: 00002032
[CPU0] 0x80036084: 80009de4 80006574 80036210 deadbeef 00000000 00000000 000b3650 00000000
[CPU0] 0x800360a4: 00000000 fffe712c 01ba5bb3 00000000 00000017 00000000 0200bffc 0200bff8
[CPU0] 0x800360c4: ee6b2800 00000002 8002be3c 00000002 00000002 3597c6e4 80025448 00000000
[CPU0] 0x800360e4: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] 0x80036104: 80007880 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] 0x80036124: 00000000 00000000 00000000 00000000 80007880 00000000 00000000 00000000
[CPU0] 0x80036144: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] 0x80036164: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] 0x80036184: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] 0x800361a4: 00003880 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] 0x800361c4: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] 0x800361e4: 00000000 00000000 00000000 00000000 00000000 00000000 8002c000 80006080
[CPU0] 0x80036204: acc0949e 80036248 80004a30 00000000 00000000 00000002 358b009e 00000000
[CPU0] 0x80036224: 00000000 00000008 80006088 00000001 000f4240 000b3650 80016bee 00000000
[CPU0] 0x80036244: 00000000 00000000 00000000 80032398 00000000 00000000 80016bf6 00000000
[CPU0] 0x80036264: 00000000 acc0949e 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] 0x80036284: 00000000 800359d8 8001ab54 00000000 00000000 00000000 8001bec8 00000000
[CPU0] 0x800362a4: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[CPU0] Dump CPU1: RUNNING
[CPU0] Dump CPU2: RUNNING
[CPU0] Dump CPU3: RUNNING
[CPU0] Dump CPU4: RUNNING
[CPU0] Dump CPU5: RUNNING
[CPU0] Dump CPU6: RUNNING
[CPU0] Dump CPU7: RUNNING
[CPU0]    PID GROUP   CPU PRI POLICY   TYPE    NPX STATE   EVENT      SIGMASK          STACKBASE  STACKSIZE      USED   FILLED    COMMAND
[CPU0]   ----   ---     0 --- -------- ------- --- ------- ---------- ---------------- 0x80028d30      2048       120     5.8%    irq
[CPU0]   ----   ---     1 --- -------- ------- --- ------- ---------- ---------------- 0x80028530      2048       120     5.8%    irq
[CPU0]   ----   ---     2 --- -------- ------- --- ------- ---------- ---------------- 0x80027d30      2048       120     5.8%    irq
[CPU0]   ----   ---     3 --- -------- ------- --- ------- ---------- ---------------- 0x80027530      2048       120     5.8%    irq
[CPU0]   ----   ---     4 --- -------- ------- --- ------- ---------- ---------------- 0x80026d30      2048       120     5.8%    irq
[CPU0]   ----   ---     5 --- -------- ------- --- ------- ---------- ---------------- 0x80026530      2048       120     5.8%    irq
[CPU0]   ----   ---     6 --- -------- ------- --- ------- ---------- ---------------- 0x80025d30      2048       120     5.8%    irq
[CPU0]   ----   ---     7 --- -------- ------- --- ------- ---------- ---------------- 0x80025530      2048       120     5.8%    irq
[CPU0]       0     0     0   0 FIFO     Kthread -   Assigned           0000000000000000 0x8002bef0      2032       492    24.2%    CPU0 IDLE
[CPU0]       1     0     1   0 FIFO     Kthread -   Assigned           0000000000000000 0x8002c6f0      2032       828    40.7%    CPU1 IDLE
[CPU0]       2     0     2   0 FIFO     Kthread -   Assigned           0000000000000000 0x8002cef0      2032       828    40.7%    CPU2 IDLE
[CPU0]       3     0     3   0 FIFO     Kthread -   Assigned           0000000000000000 0x8002d6f0      2032       828    40.7%    CPU3 IDLE
[CPU0]       4     0     4   0 FIFO     Kthread -   Assigned           0000000000000000 0x8002def0      2032       828    40.7%    CPU4 IDLE
[CPU0]       5     0     5   0 FIFO     Kthread -   Assigned           0000000000000000 0x8002e6f0      2032       828    40.7%    CPU5 IDLE
[CPU0]       6     0     6   0 FIFO     Kthread -   Assigned           0000000000000000 0x8002eef0      2032       860    42.3%    CPU6 IDLE
[CPU0]       7     0     7   0 FIFO     Kthread -   Assigned           0000000000000000 0x8002f6f0      2032       828    40.7%    CPU7 IDLE
[CPU0]       8     8     0 100 RR       Task    -   Waiting Semaphore  0000000000000000 0x800305d0      2000      1004    50.2%    nsh_main
[CPU0]       9     9     0 100 RR       Task    -   Waiting Semaphore  0000000000000000 0x80031348      2008       924    46.0%    hello
[CPU0]      10     9     1 100 RR       pthread -   Running            0000000000000000 0x80031c18      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      11     9     2 100 RR       pthread -   Running            0000000000000000 0x800324f8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      12     9     3 100 RR       pthread -   Running            0000000000000000 0x80032dd8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      13     9     4 100 RR       pthread -   Running            0000000000000000 0x800336b8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      14     9     5 100 RR       pthread -   Running            0000000000000000 0x80033f98      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      15     9     6 100 RR       pthread -   Running            0000000000000000 0x80034878      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      16     9     7 100 RR       pthread -   Running            0000000000000000 0x80035158      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      17     9     0 100 RR       pthread -   Running            0000000000000000 0x80035ac0      2032       556    27.3%    hello 0x80016b7c 0
[CPU0]      18     9     0 100 RR       pthread -   Ready              0000000000000000 0x800363a0      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      19     9     0 100 RR       pthread -   Ready              0000000000000000 0x80036c80      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      20     9     0 100 RR       pthread -   Ready              0000000000000000 0x80037560      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      21     9     0 100 RR       pthread -   Ready              0000000000000000 0x80037e40      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      22     9     0 100 RR       pthread -   Ready              0000000000000000 0x80038720      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      23     9     0 100 RR       pthread -   Ready              0000000000000000 0x80039000      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      24     9     0 100 RR       pthread -   Ready              0000000000000000 0x800398e0      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      25     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003a1c0      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      26     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003aaa0      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      27     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003b380      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      28     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003bc60      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      29     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003c540      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      30     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003ce20      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      31     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003d700      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      32     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003dfe0      2032       396    19.4%    hello 0x80016b7c 0
[CPU0]      33     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003e9c8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      34     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003f2a8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      35     9     0 100 RR       pthread -   Ready              0000000000000000 0x8003fb88      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      36     9     0 100 RR       pthread -   Ready              0000000000000000 0x80040468      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      37     9     0 100 RR       pthread -   Ready              0000000000000000 0x80040d48      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      38     9     0 100 RR       pthread -   Ready              0000000000000000 0x80041628      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      39     9     0 100 RR       pthread -   Ready              0000000000000000 0x80041f08      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      40     9     0 100 RR       pthread -   Ready              0000000000000000 0x800427e8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      41     9     0 100 RR       pthread -   Ready              0000000000000000 0x800430c8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      42     9     0 100 RR       pthread -   Ready              0000000000000000 0x800439a8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      43     9     0 100 RR       pthread -   Ready              0000000000000000 0x80044288      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      44     9     0 100 RR       pthread -   Ready              0000000000000000 0x80044b68      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      45     9     0 100 RR       pthread -   Ready              0000000000000000 0x80045448      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      46     9     0 100 RR       pthread -   Ready              0000000000000000 0x80045d28      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      47     9     0 100 RR       pthread -   Ready              0000000000000000 0x80046608      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      48     9     0 100 RR       pthread -   Ready              0000000000000000 0x80046ee8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      49     9     0 100 RR       pthread -   Ready              0000000000000000 0x800477c8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      50     9     0 100 RR       pthread -   Ready              0000000000000000 0x800480a8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      51     9     0 100 RR       pthread -   Ready              0000000000000000 0x80048988      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      52     9     0 100 RR       pthread -   Ready              0000000000000000 0x80049268      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      53     9     0 100 RR       pthread -   Ready              0000000000000000 0x80049b48      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      54     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004a428      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      55     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004ad08      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      56     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004b5e8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      57     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004bec8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      58     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004c7a8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      59     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004d088      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      60     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004d968      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      61     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004e248      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      62     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004eb28      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      63     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004f408      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      64     9     0 100 RR       pthread -   Ready              0000000000000000 0x8004fce8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      65     9     0 100 RR       pthread -   Ready              0000000000000000 0x800506f8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      66     9     0 100 RR       pthread -   Ready              0000000000000000 0x80050fd8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      67     9     0 100 RR       pthread -   Ready              0000000000000000 0x800518b8      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      68     9     0 100 RR       pthread -   Ready              0000000000000000 0x80052198      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      69     9     0 100 RR       pthread -   Ready              0000000000000000 0x80052a78      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      70     9     0 100 RR       pthread -   Ready              0000000000000000 0x80053358      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      71     9     0 100 RR       pthread -   Ready              0000000000000000 0x80053c38      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      72     9     0 100 RR       pthread -   Ready              0000000000000000 0x80054518      2024       396    19.5%    hello 0x80016b7c 0
[CPU0]      73     9     0 100 RR       pthread -   Ready              0000000000000000 0x80054df8      2024       396    19.5%    hello 0x80016b7c 0

Additionally, I have to made some code changes to get it to compile. I suspect @wangchdo may not have tested or compiled his code.

diff --git a/sched/Makefile b/sched/Makefile
index 92989e81253..2d743dfd3bb 100644
--- a/sched/Makefile
+++ b/sched/Makefile
@@ -43,6 +43,7 @@ include timer/Make.defs
 include tls/Make.defs
 include wdog/Make.defs
 include wqueue/Make.defs
+include hrtimer/Make.defs
 
 CFLAGS += ${INCDIR_PREFIX}$(TOPDIR)$(DELIM)sched
 
diff --git a/sched/hrtimer/hrtimer_cancel.c b/sched/hrtimer/hrtimer_cancel.c
index bf974d240aa..a33bd241d4c 100644
--- a/sched/hrtimer/hrtimer_cancel.c
+++ b/sched/hrtimer/hrtimer_cancel.c
@@ -28,6 +28,8 @@
 #include <nuttx/arch.h>
 #include <nuttx/clock.h>
 
+#include "sched/sched.h"
+
 #include <errno.h>
 
 #include "hrtimer/hrtimer.h"
diff --git a/sched/sched/sched_processhrtimer.c b/sched/sched/sched_processhrtimer.c
index 2f04a8da9f9..713498af950 100644
--- a/sched/sched/sched_processhrtimer.c
+++ b/sched/sched/sched_processhrtimer.c
@@ -96,7 +96,7 @@ static uint32_t nxsched_hrtimer_callback(FAR hrtimer_t *hrtimer)
 int nxsched_hrtimer_start(clock_t ticks)
 {
   return hrtimer_start(&g_nxsched_hrtimer,
-                       hrtimer->expired + ticks * NSEC_PER_TICK,
+                       (&g_nxsched_hrtimer)->expired + ticks * NSEC_PER_TICK,
                        HRTIMER_MODE_ABS);
 }
 #endif

On which OS does this issue occur?

[OS: Linux]

What is the version of your OS?

Ubuntu 22.04

NuttX Version

3663e78

Issue Architecture

[Arch: risc-v]

Issue Area

[Area: OS Components]

Host information

No response

Verification

• I have verified before submitting the report.

[Fix-Point]

@anchao @wangchdo @GUIDINGLI

[anchao]

@Fix-Point

This is my final response to you,

1. The current HRTIMER implementation is based on the AMP mode. While it is not perfect, it is indeed functional.
   We can collaborate on optimizing this implementation. If you identify any flaws, please contribute to improving it.

2. Your words and deeds have proven that you are engaging in factional disputes. I have exercised great restraint in responding to your remarks. Similar issues arose during the previous optimization of the tick-to-counter mechanism.

3. Test cases have already been provided in sched/sched: Part 1: add high resolution timer (hrtimer) only (without os tick support with hrtimer) for NuttX #17517. If you believe there are defects, please enhance the test cases so that we can collectively refine them, instead of creating a dedicated thread to disparage others' work and tout your own.

I will not respond to someone as unprofessional as you anymore. Please refrain from @mentioning me in the future. Thank you.

[Fix-Point]

@Fix-Point

This is my final response to you,

1. The current HRTIMER implementation is based on the AMP mode. While it is not perfect, it is indeed functional.
   We can collaborate on optimizing this implementation. If you identify any flaws, please contribute to improving it.
2. Your words and deeds have proven that you are engaging in factional disputes. I have exercised great restraint in responding to your remarks. Similar issues arose during the previous optimization of the tick-to-counter mechanism.
3. Test cases have already been provided in sched/sched: Part 1: add high resolution timer (hrtimer) only (without os tick support with hrtimer) for NuttX #17517. If you believe there are defects, please enhance the test cases so that we can collectively refine them, instead of creating a dedicated thread to disparage others' work and tout your own.

I will not respond to someone as unprofessional as you anymore. Please refrain from @mentioning me in the future. Thank you.

1. I have not seen any mention in the PR, commits, or documentation that this is designed exclusively for AMP. The claim that HRTIMER is only suitable for AMP is merely your own wishful thinking—even @wangchdo has never made any such statement. If I have misunderstood, please add the constraints to the documentation and include depends !SMP for hrtimer to prevent misuse. Even so, I believe such code is not mature enough to be merged into the upstream. This is because NuttX must consider all application scenarios, including CONFIG_SMP.

2. How can you ignore the facts and make such a statement? I have no intention of arguing with you. My position has always been: if your implementation is functionally correct and well-designed, I would certainly support it.

3. I do not intend to boast about my own work. All I wish is to make NuttX faster and robust. Currently, there is a clearly problematic PR sched/sched: Part 1: add high resolution timer (hrtimer) only (without os tick support with hrtimer) for NuttX #17517 here. I have proposed my alternative solution, yet you are disregarding it and pushing for the flawed PR to be merged. I believe this is not good.

Your assertion that I am unprofessional without providing any facts or evidence is, in my opinion, incorrect. Please provide evidence of my lack of professionalism; otherwise, it will only make you appear unprofessional.

[wangchdo]

This is not a bug.

I have provided the hrtimer_cancel_sync API. If you want to cancel an hrtimer synchronously, you should call hrtimer_cancel_sync. Please refer to the hrtimer documentation for details.

.. c:function:: int hrtimer_cancel_sync(FAR hrtimer_t *hrtimer)

  Cancel a high-resolution timer and wait synchronously until the timer
  becomes inactive.

  This function first calls hrtimer_cancel() to request cancellation of
  the timer.  If the timer callback is currently executing, this function
  will wait until the callback has completed and the timer state has
  transitioned to HRTIMER_STATE_INACTIVE.

  This function may sleep and must not be called from interrupt context.

  :param hrtimer: Timer instance to cancel

  :return: ``OK`` on success; negated errno on failure.

  **POSIX Compatibility:** This is a NON-POSIX interface.

[Fix-Point]

I provided hrtimer_cancel_sync api， you need to call hrtimer_cancel_sync

I mentioned it:

The third approach: hrtimer_cancel_sync → hrtimer_start. This succeeded, but it calls sleep within the critical section, which is unacceptable.

You mean we can sleep or spin in the spin-lock critical-section? How could we ensure the determinism in this case?
My suggestion is that you would better carefully review your state-machine design and refactor it.

I fully-agree @anchao . We should consider things from the user's perspective, rather than focusing on the simplicity of implementing hrtimer. Handling concurrency correctly is hard, and we need to maintain respect for it. That is why the state-machine in #17556 is much more complex.