#linuxcnc-devel Logs

Apr 05 2019

01:03 AM -!- #linuxcnc-devel mode set to +v by ChanServ
01:33 AM -!- #linuxcnc-devel mode set to +v by ChanServ
01:45 AM -!- #linuxcnc-devel mode set to +v by ChanServ
07:58 AM -!- #linuxcnc-devel mode set to +v by ChanServ
07:58 AM -!- #linuxcnc-devel mode set to +v by ChanServ
08:41 AM -!- #linuxcnc-devel mode set to +v by ChanServ
08:41 AM -!- #linuxcnc-devel mode set to +v by ChanServ
05:04 PM -!- #linuxcnc-devel mode set to +v by ChanServ
07:01 PM skunkworks: pcw_home: the encoder counter works :) (don't know if it counts up to 10k yet - maybe tomorrow
07:04 PM rene_dev_: andypugh I see that you wrote the mesa uart driver
07:04 PM skunkworks: http://electronicsam.com/images/greenmachine/20190405_185736.jpg
07:04 PM andypugh: Yes.
07:05 PM andypugh: I would do it differently now.
07:05 PM rene_dev_: andypugh pepole want to use it with vfd comps, and modbus
07:05 PM andypugh: Hmm, should be possible.
07:05 PM andypugh: Would need a sub-driver.
07:05 PM rene_dev_: and instead of modifying them, why not create a pseodo tty
07:06 PM andypugh: (All it does now is give you hooks into the hostmot2 registers)
07:06 PM rene_dev_: then you could just point the driver to the virtual serial port, and everything should just work as before
07:06 PM rene_dev_: https://github.com/LinuxCNC/linuxcnc/blob/0f91c553a238c3f5e8a52285044761c2dcfd7de5/src/hal/drivers/mesa-hostmot2/uart.c#L265
07:06 PM andypugh: Would that not lose realtime?
07:07 PM rene_dev_: this line bothers me
07:07 PM rene_dev_: realtime is not relevant, pepole currently use usb adaptors
07:07 PM rene_dev_: and no, why would it
07:09 PM andypugh: I assume that a psudo-tty would have to pass through other kernel code
07:09 PM andypugh: (But I know next to nothing about it)
07:09 PM rene_dev_: none of the vfd drivers are realtime
07:09 PM rene_dev_: nor is the modbus driver
07:09 PM andypugh: No, bit they would work better if they were.
07:10 PM andypugh: (but)
07:10 PM andypugh: And as for using llio.. Well, if you can find a way to register reads with the TRAM at run-time...
07:11 PM rene_dev_: but they are not :D at 9600 baud, which my vfd uses, its not fast anyway
07:11 PM rene_dev_: ah, it doesnt use tram reads
07:11 PM andypugh: Does hm2-eth use TRAM? I think that, actually, everything is still llio read/write as the TRAM stuff was never implemented.
07:11 PM rene_dev_: yes
07:12 PM rene_dev_: does it really hurt performance? have you tried?
07:14 PM rene_dev_: are there any known issues with the uart driver? I want to look into it, need it for at least 2 machines.
07:15 PM andypugh: I didn’t write that comment. It is one of many added by the hm2_eth author. Many misguided and wrong (such as the use of llio reads and writes in config and setup sections)
07:16 PM rene_dev_: whats wrong with that?
07:17 PM rene_dev_: Im not sure I understand the tx and rx modes...
07:17 PM andypugh: The only issue I know of is that I wrote it a long time ago, before sserial made me have to find a way to configure HAL pins at run-time and before the abdolute encoders. If I was doing it now UART and BSPI would use a config string like the absolute encoders do.
07:18 PM andypugh: As far as I am concerned llio reads and writes are the only practical way to do config.
07:18 PM rene_dev_: how does a config string help? I would just change it, so it creates a device, and you can use any software/driver whatever that expects a normal serial port
07:18 PM andypugh: So having someone sprinking my code with comments that say I am doing it wrong (but then don’t fix what they pervieve as the problem) annoy me.
07:18 PM rene_dev_: I see
07:19 PM andypugh: Especially as they never told me, I just keep finding them.
07:21 PM rene_dev_: I dont see why extra packages hurt
07:22 PM rene_dev_: network is designed to handle this
07:23 PM rene_dev_: the exported function is called from a comp, which is not executed at the same time as hm2_read or hm2_write
07:25 PM rene_dev_: so yes, it does generate a packet, but I dont expect it to affect performance without measurements
07:26 PM rene_dev_: now the only thing left to do is to figure out the txen for the rs485
07:29 PM andypugh: The idea of a config string was to make it unnecessary to write a sub-driver. You could just specify what the bits mean, and what HAL pins to make for them.
07:29 PM andypugh: (like the absolute Hostmot2 encoders are)
07:32 PM rene_dev_: unfortunately uart is not that easy. modbus and almost all of the vfd drivers require a lot more
07:33 PM rene_dev_: http://wiki.linuxcnc.org/uploads/vlt5000_vfd.comp
07:33 PM rene_dev_: no way you put that in a config string :D
07:41 PM sync: we tried yesterday to patch in the correct ioctl stuff for rs485 mode but it turns out the cheap adapters lack driver support in linux
07:41 PM sync: and rene_dev_ was not motivated to work on the kernel module for some reason
08:08 PM andypugh: I haven’t really gone back to UART because I don’t think it is at all heavily used.
08:08 PM andypugh: pcw_home: might have more idea, because I guess UART users probably need special firmware.
08:16 PM linuxcnc-build: build #3256 of 4016.deb-wheezy-i386 is complete: Failure [4failed apt-get-update shell_2] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4016.deb-wheezy-i386/builds/3256 blamelist: Rene Hopf <renehopf@mac.com>
08:20 PM linuxcnc-build: build #2515 of 4014.deb-wheezy-rtpreempt-i386 is complete: Failure [4failed apt-get-update shell_2] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4014.deb-wheezy-rtpreempt-i386/builds/2515 blamelist: Rene Hopf <renehopf@mac.com>
08:30 PM linuxcnc-build: build #2549 of 4015.deb-wheezy-rtpreempt-amd64 is complete: Failure [4failed apt-get-update shell_2] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4015.deb-wheezy-rtpreempt-amd64/builds/2549 blamelist: Rene Hopf <renehopf@mac.com>
08:33 PM linuxcnc-build: build #2947 of 4018.deb-wheezy-rtai-i386 is complete: Failure [4failed apt-get-update shell_2] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4018.deb-wheezy-rtai-i386/builds/2947 blamelist: Rene Hopf <renehopf@mac.com>
08:35 PM linuxcnc-build: build #3259 of 4017.deb-wheezy-amd64 is complete: Failure [4failed apt-get-update shell_2] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4017.deb-wheezy-amd64/builds/3259 blamelist: Rene Hopf <renehopf@mac.com>