NetBSD-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: NetBSD on Linode kvm problems

On Fri, 9 Feb 2018, Michael Parson wrote:
On Sat, 10 Feb 2018, Michael van Elst wrote: (Michael Parson) writes:

Ran this a couple of minutes ago:

$ sudo sysctl -w kern.timecounter.hardware=clockinterrupt
kern.timecounter.hardware: hpet0 -> clockinterrupt

And my clock is already 4 minutes off :-/

Means your timer interrupt is way off, that causes all short term
delays to be off too.

Please try the other timecounter choices to see if any of these
works better on the virtual machine.

Yup, after a reboot, my Mac says it's 7:45pm, this box thinks it's
7:26pm. :(

I missed a bit there, meant to say, "a few hours after a reboot, my Mac

Trying piixpm0 as I see that in dmesg output:

$ dmesg | grep piixpm0
piixpm0 at pci0 dev 1 function 3: vendor 0x8086 product 0x7113 (rev.
timecounter: Timecounter "piixpm0" frequency 3579545 Hz quality 1000
piixpm0: 24-bit timer
piixpm0: interrupting at ioapic0 pin 9
iic0 at piixpm0: I2C bus

I now know that this wasn't really a valid test, all of the timecountes
show up in dmesg:

$ dmesg | grep timecount
timecounter: Timecounters tick every 10.000 msec
timecounter: Timecounter "i8254" frequency 1193182 Hz quality 100
timecounter: Timecounter "ACPI-Safe" frequency 3579545 Hz quality 900
timecounter: Timecounter "hpet0" frequency 100000000 Hz quality 2000
timecounter: Timecounter "piixpm0" frequency 3579545 Hz quality 1000
timecounter: Timecounter "clockinterrupt" frequency 100 Hz quality 0

In any case, I've been up for nearly 48 hours with it being set to
'piixpm0' and haven't seen it slip or stall out yet.  *fingers crossed*

It doesn't seem quite as fast as it was before the Linode Meltdown
patching, but it's not as hosed up as it was just after either.
Definitly in the usable range.

Michael Parson
Pflugerville, TX

Home | Main Index | Thread Index | Old Index