So is the problem with the kernel or is it with the SETI ARM code?

Depends on your point of view. The 3.10.x kernel works fine with the present ARM Seti v8 app,

Note also that i was having this problem with the v7 app at r1924, this is before the ARM function choices were added,
so it's a problem with the generic code and not any specific ARM code.

Claggy

So is the problem with the kernel or is it with the SETI ARM code?

Depends on your point of view. The 3.10.x kernel works fine with the present ARM Seti v8 app,

Note also that i was having this problem with the v7 app at r1924, this is before the ARM function choices were added,
so it's a problem with the generic code and not any specific ARM code.

Claggy

That is interesting. Does any know what this issue is with the current kernels? Is anyone working on a fix? I've stopped using my Pi B because of the hassle of having to restart Boinc all the time.

The workaround would be to observe what function choices your device uses usually then to rebuild SETI MB stock code with hardwired function choices. That way hang in initial benchmark could be avoided. Also worth to completely disable BOINC's benchmark via cc_config.xml file.

---

News about SETI opt app releases: https://twitter.com/Raistmer

The workaround would be to observe what function choices your device uses usually then to rebuild SETI MB stock code with hardwired function choices. That way hang in initial benchmark could be avoided. Also worth to completely disable BOINC's benchmark via cc_config.xml file.

What hang in initial function choices? There might have been one once, but Eric did changes, function choices work now.

Armv6 devices on later kernels suffer from a kernel panic mid computation, This was even happening at r1924 when there were no Arm function choices as they hadn't been added yet.

Claggy

The workaround would be to observe what function choices your device uses usually then to rebuild SETI MB stock code with hardwired function choices. That way hang in initial benchmark could be avoided. Also worth to completely disable BOINC's benchmark via cc_config.xml file.

What hang in initial function choices? There might have been one once, but Eric did changes, function choices work now.

Armv6 devices on later kernels suffer from a kernel panic mid computation, This was even happening at r1924 when there were no Arm function choices as they hadn't been added yet.

Claggy

Ok, some misunderstanding perhaps. But FYI recently there was log with error definitely inside function choices. On Linux.

---

News about SETI opt app releases: https://twitter.com/Raistmer

So is the problem with the kernel or is it with the SETI ARM code?

After posting on Github's Kernel bug running BOINC tasks, 3.12.18/19 #600 thread, I was directed to a patch that fixed a similar/same issue:

Repeatable kernel crash in #737 #859

the patch had been posted upstream twice for discussion six plus months ago, and had no response, looking though GitHub, I saw it hadn't been committed there,
so Cloned the Raspberry Pi kernel repository, applied it, and pushed it upstream, it has been committed atm to the rpi-4.4.y, 4.9.y and 4.10.y branches,
and applied to the rpi-update test kernel, If anyone with a Original Pi wishes to try it will need to do:

sudo apt-get install rpi-update
sudo rpi-update

If you get 4.4.48+ #964 or 4.9.9+ #965 or later then you'll have a kernel with the fix.

8.02 apps at the Main project should work no problem now, and 8.03 here should too, 8.04 don't work on the Pi full stop, expect 8.05 soon.

Claggy

The Linux ARM 8.06 app is now available here on Beta and works on Raspberry Pi 1's (original Pi B's, all Pi A's, and the Pi Zero) with the updated kernel. It includes updated code that uses the VFP chirp function for a little extra performance.

The Pi2 and Pi3 really struggle with the vlar work units.

I am getting around 36+ hours on the Pi2 and 17 hours on the Pi3 using the 8.06 app. That's at stock clocks.

---

BOINC blog