Re: (tiny) patch review: return if RTC diag fails
Emile `iMil' Heitor <[email protected]>
| Newsgroups | gmane.os.netbsd.devel.kernel |
|---|---|
| Message-ID | <[email protected]> |
On Fri, 14 Feb 2025, Manuel Bouyer wrote:
>> (also it should be specific to the broken hypervisor concerned and not
>> applied to all hypervisors)
>
> that's what VM_GUEST_* is for
This behavior is specific to the microvm machine type in QEMU, which
sets rtc=off by default, we can't just have a condition on the
hypervisor.
As a matter of fact we do have a validity/presence test in
sys/arch/x86/isa/rtc.c rtcget():
if ((mc146818_read(NULL, MC_REGD) & MC_REGD_VRT) == 0)
return (-1);
ref https://www.scs.stanford.edu/23wi-cs212/pintos/specs/mc146818a.pdf
page 16 (thanks khorben@ for the hint).
Can't we just move this test to startrtclock()?
------------------------------------------------------------------------
Emile `iMil' Heitor <imil@{home.imil.net,NetBSD.org}> | https://imil.net