#linuxcnc-devel Logs

Mar 31 2023

#linuxcnc-devel Calendar

08:20 AM andypugh: pcw--home: Does it make sense to have the main pktUART (and plain UART) poll the status registers every thread? Currently they don’t do that. Is there a drawback other than a few extra bytes in the data that might not be used?
08:28 AM andypugh: I need to investigate what is going wrong, but I am running into an eventual segfault adding a llio->queue_read every cycle. I need to look carefully at the underlying code to see if they get deleted once actioned.
09:57 AM pcw--home: I would think the best (at least for control bandwidth) would be to poll the status registers every thread
11:52 AM JT-Cave: --readhmid what does hmid stand for?
11:56 AM Tom_L: hostmot ID
11:56 AM Tom_L: i think
11:59 AM JT-Cave: ok
01:54 PM andypugh: rmu: I am writing a new driver, raw-read / raw-write uses HAL pins.
02:20 PM pcw--home: raw-read is also difficult to use for testing hardware that has read side effects (like FIFOs)
02:44 PM andypugh: An interesting observation that I have just made is that there are two different algorithms on the internet for working out a modbus checksum, and sometimes they give different answers.
02:44 PM andypugh: One uses a lookup table, and I am wondering if an error has crept into it.
02:45 PM pcw--home: Initial conditions?
02:51 PM andypugh: I described the issue in a comment here.
02:51 PM andypugh: https://stackoverflow.com/questions/19347685/calculating-modbus-rtu-crc-16
03:03 PM pcw--home: There seem to be quite a few different versions of the non-table one
03:04 PM pcw--home: I would trust that a bit more than a table driven one as far as errors creeping in
03:10 PM andypugh: Well, what you shouldn’t trust is me :-)
03:11 PM andypugh: It turs out that I was using a signed type for the message buffer, and that was working sometimes, but not always.
04:21 PM linuxcnc-build2: Worker `checkin` is missing. It was seen last on Fri Mar 31 14:20:05 2023.
10:49 PM pere: so, someone broke the translation handling by causing a git conflict on weblate. :(
11:11 PM pere: I'm trying to clean up the mess.