<sb0>
maybe remote imports using asyncio will be possible...
<sturmflut2>
sb0: Oh :/ the "android-4.0" branch works quite well for me, I've only had one crash under very high load and that one might have been also related to overheating, it was over 30 degrees outside and the tablet became *very* warm.
<sb0>
3.13.11 has never crashed... even when heating a lot
<sb0>
and I did load that thing a lot - compiling fpga bitstreams, llvm, etc.
<sb0>
4.0 doesn't last one hour of light usage
<sturmflut2>
Bummer
<sb0>
there's also a gpu driver bug in some kernel versions
<sb0>
causes a complete lockup of the screen, but the machine will still respond to network
<sturmflut2>
I have the feeling that this is somewhat intentional by Intel. They want you to buy the bigger Core i3/i5 tablets for >350 €, those work perfect with Linux.
<sturmflut2>
Well, enough time spent on this. I'll look at it again in a couple of weeks probably, hacking on my Ubuntu phone seems much more worthwhile for the moment.
<sb0>
>>> bool(filter(lambda x: x>5, [2, 3, 4]))
<sb0>
True
<sb0>
>>> list(filter(lambda x: x>5, [2, 3, 4]))
<sb0>
[]
<sb0>
grmbl
<cr1901_modern>
lolpython
stekern has quit [Ping timeout: 264 seconds]
stekern has joined #m-labs
stekern has quit [Ping timeout: 264 seconds]
stekern has joined #m-labs
stekern has quit [Ping timeout: 250 seconds]
mumptai has joined #m-labs
<ysionneau>
sb0: can you give a try to artiq_storage.py over TCP? I could only test the serial comm with the pipistrello which works fine so far
<ysionneau>
you would need to use my flash_storage_for_sb branch on artiq, and my flash_storage branch on misoc
<sb0>
i'd rather call that artiq_coreconfig or similar
<ysionneau>
no need to re-synthesized though, just need to rebuild the bios
<sb0>
the bios?
<sb0>
why?
<ysionneau>
well the libbase in fact
<ysionneau>
(and regenerate the headers)
<ysionneau>
I renamed a symbol in libbase
<ysionneau>
from now on I'm using the naming "subsector" instead of "sector" to mean "a 4k chunk of flash"
<ysionneau>
because naming seems to be inconsistent between flash vendors and this seems the less ambiguous
<sb0>
does it match the "subsector" in the flash datasheet?
<ysionneau>
when the wording subsector is used: yes
<sb0>
e.g. micron nc25q has "subsectors"
<ysionneau>
let me check
<ysionneau>
what I've seen so far :
<ysionneau>
sector : 64k / block 4k
<ysionneau>
err
<ysionneau>
sector = 64k and subsector = 4k
<ysionneau>
or
<ysionneau>
block = 64k and sector = 4k
<ysionneau>
so sector is ambiguous
<sb0>
ok
<ysionneau>
and 64k seems to be too big
<ysionneau>
so I settled on using 4k and naming that subsector
<sb0>
doesn't the size of what is erased by a "subsector erase" command depend on the flash type?
<sb0>
also, 4k might be a bit small if the idle experiment becomes large, whereas wasting dozens of kilobytes of cheap flash is not a problem...
<ysionneau>
on the 3 different datasheets I looked the 4k chunk erase (whatever the naming of it) was the same 0x20 spi command
<ysionneau>
I cannot find micro nc25q :o is that its common name?