Re: [PATCH mach v4 1/1] Integrate HPET so the functions used for getting time can have a higher accuracy

[Damien Zammit]

You might need to run qemu with -smp 1 even though your gnumach is not SMP 
enabled, so that the acpi tables and apic works properly.

I haven't tried UP with --enable-apic and no smp flag in qemu, I should try 
that to confirm.

Damien

Sent from Proton Mail Android


-------- Original Message --------
On 13/6/25 3:14 am, Guy-Fleury Iteriteka <gfleury@disroot.org> wrote:

>  Le 12 juin 2025 18:49:22 GMT+02:00, Michael Banck via Bug reports for the 
> GNU Hurd <bug-hurd@gnu.org> a écrit :
>  >Hi,
>  >
>  >On Thu, Jun 12, 2025 at 09:54:22PM +0800, Zhaoming Luo wrote:
>  >> On Wed, Jun 11, 2025 at 09:07:10PM +0200, Michael Banck wrote:
>  >> > On Mon, Mar 24, 2025 at 12:25:51PM +0800, Zhaoming Luo wrote:
>  >> > > 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.
>  >> >
>  >> > Is there something else besides this patch that needs to be
>  >> > done/patched/configured to get higher resolution timers?
>  >> >
>  >> > I am running the current gnumach Debian package which (to the best of my
>  >> > knowledge) includes this patch, but clock_getres() still reports
>  >> > 10000000 ns timer resolution and applications like PostgreSQL that rely
>  >> > on timing still have 10ms resolution.
>  >> >
>  >> The current higher resolution depends on HPET, and the HPET will be
>  >> enabled when APIC is enabled[0]. Therefore, please try to compile a
>  >> gnumach kernel with --enable-apic in the ./configure stage.
>  >
>  >Alright, I tried that (rebuilding the current Debian unstable gnumach 
> package),
>  >but (at least on i486) it hangs early on boot in KVM:
>  
>  With which qemu command?
>  
>  There is a bug when you use qemu -smp 2
>  
>  Compile gnumach with --disable-linux-groups --enable-apic --enable-ncpus=4 
> or 8 and in qemu use -smp 1 or 3 or 4 expect 2
>  
>  >
>  >|Loading GNU Mach ...
>  >|Loading the Hurd ...
>  >|ow:13148 high:21913
>  >|vm_page: HIGHMEM: pages: 301024 (1175M), free: 0 (0M)
>  >|vm_page: HIGHMEM: min:15051 low:18061 high:30102
>  >|ACPI v1.0
>  >|HPET at physical address 0xfed00000
>  >|ACPI:
>  >| rsdp = 0x47ffe1ac6
>  >| rsdt/xsdt = 0xf8e90ac6 (n = 4)
>  >|8 bit APIC ids
>  >|APIC entry=0xf8e92a1a end=0xf8e92a66
>  >|APIC entry=0xf8e92a1a end=0xf8e92a66
>  >|APIC entry=0xf8e92a22 end=0xf8e92a66
>  >|APIC entry=0xf8e92a2e end=0xf8e92a66
>  >|APIC entry=0xf8e92a38 end=0xf8e92a66
>  >|APIC entry=0xf8e92a42 end=0xf8e92a66
>  >|APIC entry=0xf8e92a4c end=0xf8e92a66
>  >|APIC entry=0xf8e92a56 end=0xf8e92a66
>  >|APIC entry=0xf8e92a60 end=0xf8e92a66
>  >|Unhandled APIC entry type 0x4
>  >|CPUS:
>  >| CPU 0 - APIC ID 0 - addr=0xf8e96000
>  >|IOAPICS:
>  >| IOAPIC 0 - APIC ID 0 - addr=0xf8e97000
>  >|IOAPIC version 0x11
>  >|IRQ override: pin=0 gsi=2 trigger=EDGE polarity=HIGH
>  >|IRQ override: pin=5 gsi=5 trigger=LEVEL polarity=HIGH
>  >|IRQ override: pin=9 gsi=9 trigger=LEVEL polarity=HIGH
>  >|IRQ override: pin=10 gsi=10 trigger=LEVEL polarity=HIGH
>  >|IRQ override: pin=11 gsi=11 trigger=LEVEL polarity=HIGH
>  >|IOAPIC 0 configured with GSI 0-23
>  >
>  >
>  >Michael
>  >
>  
>  Hi,
>  
>