Check the availability of host_get_time64 and use it to replace
host_get_time for CLOCK_REALTIME when it's available. Fall back to
host_get_time if gnumach does not support host_get_time64 but the
gnumach headers do.
host_get_time is deprecated
See
https://git.savannah.gnu.org/cgit/hurd/gnumach.g
Compared with RFC v3:
- Use #ifdef APIC in hpclock_read_counter() to avoid the runtime check
of the availability of HPET.
- Remove the redundant check in hpclock_get_counter_period_nsec().
- Fix a small bug in time_value64_add_hpc().
- Some comment improvements
Zhaoming Luo (1):
Integrate HPET
Integrate HPET so host_get_time, host_get_time64, and host_get_uptime64
are more precise. The highest precision can be 10ns when this patch is
applied.
* i386/i386/apic.c: Implement the two high-precision clock interface functions
added in this patch for i386.
* i386/i386at/model_dep.c: Initiali
Zhaoming Luo, le dim. 23 mars 2025 10:47:14 +0800, a ecrit:
> Check the availability of host_get_time64 and use it to replace
> host_get_time for CLOCK_REALTIME when it's available.
> case CLOCK_REALTIME:
>{
> - /* __host_get_time can only fail if passed an invalid host_t.
> +
Damien Zammit via Bug reports for the GNU Hurd, le dim. 23 mars 2025 10:20:37
+, a ecrit:
> You probably want to disable it entirely so it doesn't need to execute a
> branch condition every time it reads the timer, ie with #ifdef.
The if is precisely meant for the case where hpet is not avai
Hello,
Almost there!
Samuel
Zhaoming Luo, le dim. 23 mars 2025 18:13:55 +0800, a ecrit:
> +uint32_t
> +hpclock_get_counter_period_nsec(void)
> +{
> +if (hpet_addr != NULL)
> + return hpet_period_nsec;
> +else
> + return 0;
When hpet_addr is NULL, hpet_period_nsec will already be
Damien Zammit via Bug reports for the GNU Hurd, le dim. 23 mars 2025 08:22:23
+, a ecrit:
> I think we are assuming any machine with ACPI has HPET. This is true since
> ~2005.
Yes, but we currently still support the non-APIC case too.
Samuel
On Sun, Mar 23, 2025 at 10:20:37AM +, Damien Zammit wrote:
> You probably want to disable it entirely so it doesn't need to execute a
> branch condition every time it reads the timer, ie with #ifdef.
> This code path would be critical for timing.
Fair enough :/. Thanks :).
Zhaoming
>
>
> --
You probably want to disable it entirely so it doesn't need to execute a branch
condition every time it reads the timer, ie with #ifdef.
This code path would be critical for timing.
Damien
Sent from Proton Mail Android
Original Message
On 23/3/25 9:13 pm, Zhaoming Luo wrote:
Hi,
I'm wondering the code for checking whether a high-precision code is
available should be in mach_clock.c or in the machine-dependent
directory. In this patch it's in the machine-dependent directory. It can
ensure the backward compactability. Otherwise compiling a gnumach with
only '../configur
On Sun, Mar 23, 2025 at 09:03:01AM +0100, Samuel Thibault wrote:
> Zhaoming Luo, le dim. 23 mars 2025 13:29:47 +0800, a ecrit:
> > On Sun, Mar 23, 2025 at 12:48:52AM +0100, Samuel Thibault wrote:
> > > Zhaoming Luo, le sam. 22 mars 2025 15:24:50 +0800, a ecrit:
> > >
> > > > diff --git a/i386/i386
I think we are assuming any machine with ACPI has HPET. This is true since
~2005.
Damien
Sent from Proton Mail Android
Original Message
On 23/3/25 7:03 pm, Samuel Thibault wrote:
> Zhaoming Luo, le dim. 23 mars 2025 13:29:47 +0800, a ecrit:
> > On Sun, Mar 23, 2025 at 12:
Zhaoming Luo, le dim. 23 mars 2025 13:29:47 +0800, a ecrit:
> On Sun, Mar 23, 2025 at 12:48:52AM +0100, Samuel Thibault wrote:
> > Zhaoming Luo, le sam. 22 mars 2025 15:24:50 +0800, a ecrit:
> >
> > > diff --git a/i386/i386at/model_dep.c b/i386/i386at/model_dep.c
> > > index 30449c37..aa8451ac 100
13 matches
Mail list logo
