"Page fault while in kernel mode" on APU2 after bios/coreboot upgrade
-
we have a saying in Italy, literally translated as ‘not all donuts come out with a hole’ meaning ‘not everything turns out as planned’
it's called "silicon lottery", not all cpu are the same, there is ample opportunity for some microscopic part of a CPU, which works fine at a certain speed/voltage combination, to no work if the speed or voltage is increased. -
@CS said in "Page fault while in kernel mode" on APU2 after bios/coreboot upgrade:
Could anyone share their APU2 loader.config.local file for reference? I'm wondering if I'm missing something obvious, I haven't done any tuning for years because it has been running smoothly with no issues.
Hi, here the content of my loader.config.local:
legal.intel_ipw.license_ack=1
legal.intel_iwi.license_ack=1
debug.acpi.avoid="_SB_.PCI0.GPIO" (necessary for loading apuled.ko)if you still have "hint.acpi_perf.0.disabled=1" in your loader.conf.local you will see those increased frecv. in sysctl dev.cpu even when you have disabled CPB in BIOS.
Regards,
fireodo -
@CS said in "Page fault while in kernel mode" on APU2 after bios/coreboot upgrade:
other folks have it enabled on APU2 without experiencing any issues.
I confirm this
we have lot of such units at end users, they are "run" with CPB without any problems
we basically configure these "routers / NGFWs" + pfSense with CPBCPB as I wrote above has been enabled in the Coreboot BIOS, but can only be interpreted on 1 core with a frequency of 1,400 instead of 1,000 this is good for OpenVPN stuff, for example...
@CS I think don't look for the rabbit in the bush...
this is not an issue whic is caused by CPB or pfSenseI think the APU2 MOBO is damaged somewhere, cold soldering or something like that
which causes a malfunction in the BUS or RAM operation due to the elevated clock....???
maybe try a CPU shock test under linux and insulate the APU2 housing to warm up .....Voilà, maybe there will be results
@kiokoman anyway, this is an AMD embedded series CPU can not really be overdriven, designed for low-power devices
either it works or it doesn't, there is no overclocking it only the CPB allows for a small tuning... -
@DaddyGo @fireodo I won't continue troubleshooting this honestly, the board works fine for me with CPB disabled and I still get the boosted CPU frequency by having the right settings in my loader.conf.local. I'm not even sure if my performance would get any better! Actually, I'm now wondering, just out of curiosity, if this happens when you have both, the CPU boost settings in loader.conf.local and the BIOS setting enabled.
-
@CS , every hardware has a different outcome in Q.C. Even with the same parts, but a different batch.
Rule of thumb, 4 years lifespan (4 is death in Chinese). Nowadays you should be happy if your electronic works for more than 4 years.I am not sure how handy you are but you could try heating up the cpu (without thermal paste) with the heat gun on a flat surface. Keep around 10 cm distance with circular motion for around 10-15 mins. But be warn, you could burn the cpu.
-
@AKEGEC said in "Page fault while in kernel mode" on APU2 after bios/coreboot upgrade:
but you could try heating up the cpu (without thermal paste) with the heat gun on a flat surface.
It's a very bad idea.
This AMD CPU reaches its maximum TDP in about 40 seconds without a cooling surface (heat shrink) and dies...
(moreover as I wrote it is an embedded CPU, soldered to the PCB)
(earlier than the said Chinese 4-year death)the pcEngines stuff is stable and we have several pieces of it that has been working for 6 years (from ALIX and APU series)
The ALIXs works as a radio os WISP PtP and AP and is constantly exposed to the weather.
So these are not subject to your Chinese rule -
@DaddyGo said in "Page fault while in kernel mode" on APU2 after bios/coreboot upgrade:
It's a very bad idea.
This AMD CPU reaches its maximum TDP in about 40 seconds without a cooling surface (heat shrink) and dies...
(moreover as I wrote it is an embedded CPU, soldered to the PCB)
(earlier than the said Chinese 4-year death)the pcEngines stuff is stable and we have several pieces of it that has been working for 6 years (from ALIX and APU series)
The ALIXs works as a radio os WISP PtP and AP and is constantly exposed to the weather.
So these are not subject to your Chinese ruleWell to solder embedded cpu you need a temperature between 200-400°c.
Anyway I was talking about heating it up a bit. As long you are not reaching 90°c you will be fine. But if it already passed the 4 years mark, then I would leave it as it is.
I don't know why manufactures are shortening their products lifespan. It used to be 15-30 years quality guarantee. -
@AKEGEC said in "Page fault while in kernel mode" on APU2 after bios/coreboot upgrade:
Well to solder embedded cpu you need a temperature between 200-400°c.
Yes, this soldering temperature is a separate data in the catalog and the automatic production lines (soldering machines) solder to the permitted seconds. (2 - 3s)
The temperature of the inner silicon layer of the CPU does not tolerate this temperature (cca. 150 max).I don't know, if you already had an APU MOBO in your hand?
So this is what it looks like...
(the metal router housing itself cools the CPU and connects to the metal surface with a bit of heat transfer)
https://www.pcengines.ch/apucool.htm
As I wrote above, the test method for the following may be, CPU shock test under f.e. Linux and while covering the house.
Once we launched such a MOBO without its metal housing for testing and it "boiled" quickly under load.
You're right, half of today's stuff can't stand it until then 25-30 years.
Anno, I even repaired 30-year-old cathode ray tubes and black and white televisions and continued to operate for another 10 years.Welcome to today's money-hungry world. hahahahaha
-
@DaddyGo , did we just revealed our age? hahahahahaa
Oh well age is just a number.
-
@AKEGEC said in "Page fault while in kernel mode" on APU2 after bios/coreboot upgrade:
did we just revealed our age?
Not a shame ...
the age, I think, brings wisdomyea and everyone is as old as he/she feel
droll - quadragenarian, hihihihi
-
@cs Thank you CS, I found this post after a couple of weeks of trying to sort out the exact same issue. I cant count the different firmwares I tried and reinstalls of PFsense. Even tried OPNsense. Long story short updated to v4.13.0.3 BIOS disabled CPB and my firewall has been stable since. Uptime 18 Hours 56 Minutes 32 Seconds at present, but I am quietly confident that the issue is resolve.
Once again thank you.
-
@cs Thank you very much for this post, it seems I have the same issue... Is this due to a defective board/cpu - has your firewall been running stable since disabling SPB?
-
@kuser Super stable, I have had no issues at all since then!
-
@cs Thanks for the feedback. Looks like they've found a bug.
I reported this issue a couple of weeks ago:
https://github.com/pcengines/coreboot/issues/469
