fix: make rt_tick_t 64-bit on ARCH_CPU_64BIT and improve clock time precision#10749
Open
lygstate wants to merge 2 commits into
Open
fix: make rt_tick_t 64-bit on ARCH_CPU_64BIT and improve clock time precision#10749lygstate wants to merge 2 commits into
lygstate wants to merge 2 commits into
Conversation
github-actions
Bot
added
BSP
Kernel
|
👋 感谢您对 RT-Thread 的贡献！Thank you for your contribution to RT-Thread! 为确保代码符合 RT-Thread 的编码规范，请在你的仓库中执行以下步骤运行代码格式化工作流。 🛠 操作步骤 | Steps
完成后，提交将自动更新至 如有问题欢迎联系我们，再次感谢您的贡献！💐 |
📌 Code Review Assignment🏷️ Tag: componentsReviewers: @Maihuanyi Changed Files (Click to expand)
🏷️ Tag: components_libcReviewers: @GorrayLi @mysterywolf Changed Files (Click to expand)
🏷️ Tag: documentationReviewers: @CXSforHPU @GorrayLi @lianux-mm Changed Files (Click to expand)
📊 Current Review Status (Last Updated: 2026-07-14 20:32 CST)
📝 Review Instructions
|
lygstate
force-pushed
the
rt_tick_t_64_bit
branch
3 times, most recently
from
0bc44ca to
5839f13
Compare
|
|
lygstate
force-pushed
the
rt_tick_t_64_bit
branch
2 times, most recently
from
c448a01 to
264c203
Compare
lygstate
force-pushed
the
rt_tick_t_64_bit
branch
3 times, most recently
from
28b2e48 to
d8c1321
Compare
Contributor
There was a problem hiding this comment.
Pull Request Overview
This PR fixes overflow issues in clock_gettime and rt_ktime_boottime_get_ns by improving time conversion precision and enabling 64-bit tick counts on 64-bit architectures. The changes replace the previous resolution-based approach with direct frequency-based calculations to ensure more accurate timing operations.
Key changes:
- Introduces precision-preserving multiplication/division functions (
rt_muldiv_u64,rt_muldiv_u32) - Makes
rt_tick_t64-bit on 64-bit architectures to support long-running programs - Replaces resolution-based time calculations with frequency-based conversions throughout the timing subsystem
Reviewed Changes
Copilot reviewed 12 out of 12 changed files in this pull request and generated no comments.
Show a summary per file
| File | Description |
|---|---|
src/rttypes.c |
Adds new multiplication/division functions for precision-preserving arithmetic |
include/rttypes.h |
Declares the new rt_muldiv_u64 and rt_muldiv_u32 functions |
include/rtdef.h |
Makes RT_TICK_MAX 64-bit on 64-bit architectures |
components/libc/compilers/common/ctime.c |
Updates timer implementations to use frequency-based calculations |
components/drivers/rtc/dev_soft_rtc.c |
Updates RTC resolution calculation |
components/drivers/ktime/src/ |
Refactors timing subsystem to use frequency instead of resolution |
bsp/rockchip/rk3500/driver/hwtimer/ |
Updates hardware timer interface |
Comments suppressed due to low confidence (2)
src/rttypes.c:1
- Corrected spelling of 'loseless' to 'lossless'.
/*
src/rttypes.c:1
- Corrected spelling of 'loseless' to 'lossless'.
/*
d8c1321 to
9c73d71
Compare
Contributor
Author
|
@Rbb666 ping |
lygstate
force-pushed
the
rt_tick_t_64_bit
branch
from
9c73d71 to
e13dc95
Compare
10 tasks
lygstate
force-pushed
the
rt_tick_t_64_bit
branch
from
e13dc95 to
ed41d9e
Compare
lygstate
force-pushed
the
rt_tick_t_64_bit
branch
from
ed41d9e to
679bfb5
Compare
lygstate
force-pushed
the
rt_tick_t_64_bit
branch
from
679bfb5 to
5ee5f1d
Compare
10 tasks
lygstate
changed the title
Contributor
Author
|
ping for review, and avoid newly conflict like #11111 After this, unlock #10735 would be strait-forward |
lygstate
force-pushed
the
rt_tick_t_64_bit
branch
from
5ee5f1d to
a6b5b05
Compare
Rbb666
requested changes
Jun 17, 2026
…recision - Make rt_tick_t and RT_TICK_MAX 64-bit when ARCH_CPU_64BIT is defined - Add rt_muldiv_u64() and rt_muldiv_u32() for lossless counter and nanosecond conversions - Convert boottime, clock_gettime, POSIX timers, and hrtimer paths with clock source frequency instead of scaled resolution - Rewrite rt_clock_time_counter_to_ns() and rt_clock_time_ns_to_counter() with rt_muldiv_u64() for the VDSO path - Remove RT_CLOCK_TIME_RESMUL, res_scale, and scaled-resolution APIs: - res_scale was only a fixed-point multiplier for resolution math (res_scaled = 1e9 * res_scale / freq); it did not customize clock frequency - All in-tree devices always used the default RT_CLOCK_TIME_RESMUL value, so the field provided no real per-board tuning - Converting through res_scaled introduced extra rounding error compared with using get_freq() directly - Custom frequency remains available through rt_clock_time_ops.get_freq, rt_clock_time_source_init(), and rt_clock_time_device_register() - Drop RT_CLOCK_TIME_RESMUL, struct rt_clock_time_device.res_scale, rt_clock_time_get_res_scaled(), rt_clock_time_get_event_res_scaled(), and rt_clock_hrtimer_getres() - Port the original ktime fixes to the clock_time API - Keep soft-timer fallback rounding in _hrtimer_cnt_to_tick() so sub-tick timeouts schedule at least one OS tick - Update clock_time core documentation
…hecks - Move rt_muldiv_u64 out of rttypes into private clock_time_internal with truncated overflow and an optional overflow out-parameter - Name the helper rt_clock_time_muldiv_u64 and include its .c from clock_time_core.c so IDE projects do not need a new source entry - Add rt_clock_time_get_res() and reject zero-frequency paths in clock_getres, timer_gettime, timer_settime, and soft RTC - Clamp or error on overflowed/out-of-range delay conversions in hrtimer and POSIX timer reload paths - Keep one-shot timer intervals from being rounded up to a 1-tick reload, and clamp timer_gettime remain time after expiry - Update clock_time core documentation for resolution and overflow
lygstate
force-pushed
the
rt_tick_t_64_bit
branch
from
a6b5b05 to
f66ccdb
Compare
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
拉取/合并请求描述：(PR description)
continue work of #9008
Fixes overflow of
clock_gettimert_ktime_boottime_get_nsthoroughlyCurrently the return value of clock_gettime and rt_ktime_boottime_get_ns are restricted by
rt_ktime_cputimer_getres, as the value ofrt_ktime_cputimer_getresare not exactly represent thefrequency of the clock, directly usingrt_ktime_cputimer_getfrqwill ensure the return value of rt_ktime_boottime_get_ns be more exact.And also add function rt_muldiv_u64 rt_muldiv_u32 for not losing precision when convert between
cputimerandhrtimerandnanoseconds.Because
rt_ktime_boottime_get_nsdepends onrt_ktime_cputimer_getcntandrt_ktime_cputimer_getcntdependsrt_tick_getSo the rt_tick_t should be 64bit for long running program.
Use
#if defined(ARCH_CPU_64BIT)is forfuture allow rt_tick_t can be 64bit on 32bit CPU.
Tick count should not be restricted to the CPU
[
为什么提交这份PR (why to submit this PR)
你的解决方案是什么 (what is your solution)
请提供验证的bsp和config (provide the config and bsp)
]
当前拉取/合并请求的状态 Intent for your PR
必须选择一项 Choose one (Mandatory):
代码质量 Code Quality：
我在这个拉取/合并请求中已经考虑了 As part of this pull request, I've considered the following:
#if 0代码，不包含已经被注释了的代码 All redundant code is removed and cleaned up