Re: [9front] qemu clock/timer issue

[Jacob Moody <moody@posixcafe.org>]

On 4/16/24 12:07, Douglas Guptill wrote:
> Hello moody:
> 
> On Tue, Apr 16, 2024 at 08:32:42AM -0500, Jacob Moody wrote:
> 
>> It is understood that this bug only happens with intel kvm virtual machines.
>>  From all that has been tested it was currently understood to affect newer intel machines.
>> It seems however that your machine is a bit older(10+ years) and still triggering this.
>> I've certainly run 9front in kvm on intel machines before so it is likely either
>> something that changed with us or something that has changed with QEMU. My money
>> is on some QEMU change.
> 
> Ahhh.. That is a big help.  
> 
>> If you do not feel comfortable debugging
> 
> I'm a programmer. :)

Excellent news. Thank you. :)

> 
> I can look at the code, build and insert write statements, in qemu and
> the kvm module of linux.  Cannot yet do anything like that with
> 9front.

We can help with understanding any of the 9front bits.

> 
>> but would still be interested in helping
> 
> For sure.
> 
>> perhaps you can help bisect 9front and qemu (one side at a time) to see if you can
>> narrow down what has changed to get more information to fix these things.
>> I would start with the nightly iso just to double check,
> 
> Assuming that you mean a nightly iso of 9front, I just did that with
> 9front-10501.amd64.iso.gz.  No noticable change; but then I don't yet
> understand 9front.  I stopped it quickly.  Attached is a jpg of what I
> saw.

Got it, thanks for checking that. Also uramekus on the grid had some
additional context about this. He said that he personally has experienced
this syslog spam but there was otherwise no adverse effects. He mentioned
that: "options kvm min_timer_period_us=10" in modprobe.conf about kvm makes it shut up.

However there was _also_ the issue that mcf originally reported regarding the fast
running clock, just to clarify are you also observing this behavior? It seems perhaps
this syslog may not be directly related to the fast running clock. So perhaps there
are two different issues that we're observing.

In either case it would be nice to at least understand what is going on, and then perhaps
fix these things.

> 
>> then move back through the releases going back.
> 
> Sorry, I don't understand.  Releases of...qemu?  9front?
> 
> There is a newer qemu (9.0.0-rc3) out; I will make a build of that,
> and then start looking at its code, and the kvm module code, to see if
> I can find, and maybe understand, the offending code.

Yes I initially meant going back through 9front releases, however with
uramekus saying he has observed this for a long time I think it might
be better use of time to debug this with the latest code and just dig in.

I will say that your attached image does not having anything out of order
beside those mpintrassign lines, not entirely sure what that's about.


Thank you,
moody