669 reputation
213
bio website
location
age
visits member for 1 year, 10 months
seen Nov 21 at 9:28

Sep
5
comment How to run a command in the background with a delay?
@gekannt: ...uh, the question is "How to run a command in the background with a delay?" - how would you achieve it otherwise? (and no, if you don't background the () group, the original shell will pause until the spawned one finishes.)
Aug
29
accepted What's with `-` in filename affecting sorting order depending on platform?
Aug
29
comment What's with `-` in filename affecting sorting order depending on platform?
Oh. After export LANG= the order switched. Must be it. (...still, what insanity possessed the developers to switch that order in the locale?)
Aug
29
comment What's with `-` in filename affecting sorting order depending on platform?
@choroba: Cygwin has pretty much all LC_ variables (except LC_ALL, empty) as "pl_PL.UTF-8". Busybox has no locale.
Aug
29
comment What's with `-` in filename affecting sorting order depending on platform?
@vfbsilva: As md5sum proves, the contents are identical, even though the file on cygwin functions on NTFS (and on busybox, on yaffs2).
Aug
29
asked What's with `-` in filename affecting sorting order depending on platform?
Aug
5
accepted condition && command + exit
Aug
5
asked condition && command + exit
Jul
2
awarded  Curious
Jun
16
awarded  Constituent
Jun
16
awarded  Caucus
Jun
9
awarded  Nice Question
May
28
comment Corruption-proof SD card filesystem for embedded Linux?
Also, while I believe that yes, Ramdisk is currently the correct solution to the problem at hand, but I'd still prefer to get the answer to the question asked - in case I do need a robust non-volatile RW filesystem on RPi. (possibly a complete, working procedure of getting yaffs2 to mtdblock.)
May
28
comment Corruption-proof SD card filesystem for embedded Linux?
@ScottSeidman: RPi was a short "experimental" period. In industrial application it fails bad - our partner company developed an industrial logger for mining machinery based on RPi and the fault rate in that environment (EMI, vibrations - the box is well sealed) - is terrible, meanwhile BeagleBones in traffic monitoring application perform very well. But these "kiosks" work in office-style environment, and while currently the failure rate is unpleasant, I believe fixing the current solution (in software) is better than developing alternative.
May
23
comment Can I achieve functionality similar to interrupts in Linux userspace?
That sounds quite awesome. If the communication is over shared memory I really could defer a realtime thread there and just use simple mechanisms like taint bits to assure synchronization with common userspace.
May
23
comment Can I achieve functionality similar to interrupts in Linux userspace?
[cont.] provide this kind of functionality. I'd love to hear more about this "priority boost", though I don't doubt the job requires saving context of currently running task and restoring context of the handler task at the very least, so the reaction will take time, no doubt. Still, considering worst-case of about 0.1s which happens if the scheduler is left untouched, it's vastly better.
May
23
comment Can I achieve functionality similar to interrupts in Linux userspace?
@ScottSeidman: We're working under an OS. Delays are unavoidable but may be quite short. Right here with me I have something a co-worker made: userspace app working under SCHED_FIFO and a kernel module that pre-empts anything that might be working currently and hands over control to that specific app when interrupt arrives. It's quite complex and had a slew of problems initially, so I'm looking for something similar but without writing a whole long kernel module which hacks the scheduler heavily - something native to Linux. I don't know if it exists, but Sim's answer with the comment seems to
May
23
accepted Can I achieve functionality similar to interrupts in Linux userspace?
May
22
comment Corruption-proof SD card filesystem for embedded Linux?
@ScottSeidman: I can tolerate it, but not at current rate. It's about 3rd time with the RPi, and we have maybe 5 of these "out in the wild", starting last year. Meanwhile the number of the EmbeddedArm.com SBCs goes into a couple hundred by now, starting good 6 years ago, and we got maybe 3 non-critical failures related to filesystem corruption (the database file with logs got corrupted, but the system never failed to boot up, binaries never got corrupted). So, reducing the fault frequency by some two orders of magnitude would suffice.
May
22
comment Corruption-proof SD card filesystem for embedded Linux?
@RBerteig: Well, for another application we use a board from EmbeddedArm, with mtd and yaffs2, and it's been extremely resilient against power-off corruption, so it's not like these two are the only options.