Current Server Time: 23:34 (Central Europe)

#apertus IRC Channel Logs

2020/08/17

Timezone: UTC


02:25
intrac
left the channel
02:27
intrac
joined the channel
04:03
Bertl_oO
off to bed now ... have a good one everyone!
04:03
Bertl_oO
changed nick to: Bertl_zZ
06:52
BAndiT1983|away
changed nick to: BAndiT1983
06:55
BAndiT1983
changed nick to: BAndiT1983|away
06:57
BAndiT1983|away
changed nick to: BAndiT1983
07:18
mumptai
joined the channel
07:34
BAndiT1983
changed nick to: BAndiT1983|away
08:29
Spirit532
left the channel
08:36
BAndiT1983|away
changed nick to: BAndiT1983
08:39
Spirit532
joined the channel
08:49
BAndiT1983
changed nick to: BAndiT1983|away
08:51
BAndiT1983|away
changed nick to: BAndiT1983
09:11
_florent_
left the channel
09:12
_florent_
joined the channel
11:49
Bertl_zZ
changed nick to: Bertl
11:49
Bertl
morning folks!
12:03
illwieckz
joined the channel
12:05
BAndiT1983
changed nick to: BAndiT1983|away
13:12
Bertl
off for now ... bbl
13:12
Bertl
changed nick to: Bertl_oO
15:55
Bertl_oO
meeting in five minutes!
15:55
Bertl_oO
changed nick to: Bertl
16:00
Bertl
meeting time!
16:00
Bertl
se6ast1an is a little late today, so I'll moderate the meeting
16:00
Bertl
please /msg me in case you want to report (which I hope you want to do :)
16:01
metal_dent[m]
hi everyone!
16:01
Bertl
metal_dent[m], the channel is yours!
16:02
metal_dent[m]
i was trying to read back the hex data already present in the key managers using the ICSProgrammer class in the bootloader
16:02
metal_dent[m]
we decided to first read the data and if that works perfectly then write the data ':D
16:03
metal_dent[m]
i tried separately (like it was implemented as a demo in the bootloader main) and it was working
16:04
metal_dent[m]
but when i tried to trigger using the python script and selected the key manager east i encountered a problem with my script, it's a small one regarding the crc8 (the checksum)
16:04
metal_dent[m]
note: by selecting the other chips (pic32 and key manager west) it works
16:05
metal_dent[m]
after i solve that little obstacle, the data will be read back with km_east too
16:05
metal_dent[m]
that's it from me! thanks!
16:05
Bertl
great! thanks!
16:06
Bertl
apoorva_arora: please go ahead, channel is yours!
16:06
apoorva_arora
Hello everyone, I finished the final scheduler code. link: https://github.com/Apoorva-ar/GSOC_2020/blob/master/Scheduler_Master.vhd, testbench link: https://github.com/Apoorva-ar/GSOC_2020/blob/master/Scheduler_tb.vhd.
16:07
apoorva_arora
The code incorporates all the three layers: namely scheduling, packet and PHY.
16:07
se6ast1an
good evening, sorry for the delay, here now
16:08
apoorva_arora
The code accepts command from user concurrently and offers bidirectional service based on the priority of the task.
16:09
apoorva_arora
Hence, high priority tasks are scheduled first.
16:10
apoorva_arora
The coming days will be hardware testing using AXI Lite GP0 and GP1 interfaces of ZYNQ for controlling SPI master on MACHXO2
16:10
apoorva_arora
Thats all from my side
16:10
Bertl
okay, thanks!
16:11
Bertl
bluez_[m]: please go ahead and report your progress!
16:11
bluez_[m]
tx!, hi everyone
16:11
bluez_[m]
https://github.com/Swaraj1998/axiom-beta-rfdev/commits/master
16:12
bluez_[m]
please look at the commits since 4 days ago
16:12
Bertl
(well, that's going to be complicated for folks who read up on the logs later :)
16:12
bluez_[m]
the major one being the very latest one...which i am still testing
16:13
bluez_[m]
so this week i started with making a basic skeleton ioctl interface..
16:13
bluez_[m]
/dev/rfe and /dev/rfw can be used for ioctl commands
16:14
bluez_[m]
<Bertl "(well, that's going to be compli"> true xD.. last four commits starting today actually..
16:14
bluez_[m]
then i spent much of time studying that previously mentioned jtag patchset
16:15
bluez_[m]
and i shortlisted 3 important commands: set state, get state and xfer
16:16
bluez_[m]
as i currently dont have a way to keep track of states... set state can only be used to goto reset or 'idle through reset'
16:17
bluez_[m]
the latter would be a way to initialize to idle as 'xfer' also assumes the current state as idle.. and also goes back to idle after each xfer command
16:18
bluez_[m]
xfer command here is used to send/receive arbitrary length data... providing the standard SDR and SIR functionality
16:18
bluez_[m]
which i guess initially would be quite enough for most things
16:19
bluez_[m]
so in the latest commit i just added the xfer command... going to test it now
16:20
Bertl
okay, great! anything else?
16:21
bluez_[m]
also added rf_tdi_out and rf_tdo_in functions for more abstraction... so tidied up code from other parts as well
16:21
bluez_[m]
thats it form me!
16:21
Bertl
excellent! preetimenghwani[, please go ahead with your report!
16:22
preetimenghwani[
Hi Everyone
16:22
preetimenghwani[
This week I spent time on understanding a few more modules and ended up with an idea on how throughput can be doubled.
16:23
preetimenghwani[
I discussed it with Bertl and he has suggested me a few steps to follow
16:23
preetimenghwani[
Currently I am extending the framework to four writers
16:24
preetimenghwani[
I am facing some issues so spending time on debugging it
16:24
preetimenghwani[
Thats it from me! Thanks!
16:24
Bertl
thanks for the report!
16:25
Bertl
se6ast1an: you got a few bits of information for us as well?
16:25
se6ast1an
quick updates from me: the hackathon with vup and anu3jn ended this weekend
16:25
se6ast1an
it concludes two great weeks with lots of progress and work done
16:25
se6ast1an
maybe anu3jn and vup can share a bit of details themselves?
16:25
vup
is here in theory
16:26
vup
I can share a bit
16:26
se6ast1an
I had a meeting with teles new external hardware production person last week to get the axiom beta panel take the next preparation steps
16:26
se6ast1an
and our OLPF filter samples are unfortunately affected by the USPS trouble in the USA right now
16:27
se6ast1an
our forwarder will or has already posted them as priority mail so the hope is that it leaves the country before more infrastructure is removed...
16:27
se6ast1an
done from my side
16:28
Bertl
great!
16:28
Bertl
vup: please go ahead and share a bit (or maybe even more :)
16:28
vup
sure
16:28
vup
so continuing from where we left of last week, we mainly worked on two fronts
16:28
vup
Using our work on gateware and software for the ft60x (the usb3 fifo chip on the usb3 plugin module), we started working on integration with the main beta gateware. Bertl modified the hdmi output gateware to instead transmit the data to the usb3 plugin module.
16:30
vup
On the usb3 plugin module side anuejn implemented a receiver for the data. To debug this receiver we also wanted to implement a way to read and write registers on the machxo2 using jtag.
16:30
vup
The previous week we had implemented such capabilities for the hdmi digitizer board (which has a artix-7 fpga) and we hoped implementing the same on the machxo2 would not be too hard.
16:31
vup
However it turns out that the jtag primitive of the machxo2 differs in a subtle but very significant way from the artix-7 one
16:32
vup
It has a 1 bit register on the tdi pin, which the artix-7 doesn't have, so a lot of time was spent debugging that (the documentation for the machxo2 primitive is basically non-existant)
16:33
vup
The other front we worked on was firmware for the remote
16:34
vup
The remote is using a pic32mz microcontroller, for which microchip provides a gcc-fork.
16:35
vup
Microchip's gcc-fork makes you pay for optimizations and is based on a really old gcc version (4.8)
16:35
vup
so we tried to make the mainline gcc work with the pic32mz.
16:36
vup
In theory this is not too hard, as the pic32mz is simply a mips m14k core, which is already supported by the mainline gcc
16:37
vup
However especially their linker contains a bunch of special handling for the pic32mz, like placing the stack somehow and generating some `data initialization tables` that are (among other things) used to initialize the interrupt controller.
16:38
vup
Currently we have basic c / c++ code working and even the current remote firmware somewhat works, but we are still unable to get interrupts to work.
16:39
vup
I am currently in the process of writing up / collecting the stuff I have done and uploading it to github, so when I have done that maybe others are interested in taking a look at that aswell.
16:39
se6ast1an
great, whats the current state with the usb3 module now?
16:40
vup
Finally we did some general work on the fw2.x, like adding the bitstreams to the github repository, fixing a but in the control daemon that broke the beta configuration and some minor general improvements / bugfixes.
16:41
vup
se6ast1an: the jtag debugging should work now, but is untested I think
16:41
vup
the receiver side (that receives the pixel data from the zynq) is fully implemented, but also untested
16:41
vup
so there are probably some bugs to be fixed there
16:42
se6ast1an
your availability is reduced now drastically IIRC but would anuejn be able to continue with the usb3 stuff?
16:42
vup
we also had some problems with a recently added feature of nmigen fifo's, that was / is broken in basically every way possible
16:42
se6ast1an
while its "hot" :)
16:43
vup
se6ast1an: Yes anu3jn should be able to continue with the usb3 stuff.
16:43
vup
And me aswell, if I find the time for it.
16:43
se6ast1an
great
16:44
se6ast1an
regarding the pic32 gcc I have to admit I have some concern about the general usefulness and "pain" that this solves but that should be discussed later/seperately
16:44
vup
Bertl: one question for you, how is the remote remote currently hooked up to the zbox?
16:44
vup
se6ast1an: sure
16:45
Bertl
the remote in our remote setup is hooked up via pickit/ICSP and serial
16:45
vup
which version of pickit?
16:45
Bertl
it is a chinese pickit2 clone
16:45
vup
is the one se6ast1an had also a chinese clone?
16:45
vup
ah on the bus back to germany we also did a bit of routing on the micro r3 mainboard, but nothing too exciting, just some further ddr3 routing.
16:46
Bertl
very likely
16:46
vup
hmm
16:46
Bertl
what's the problem/issue/idea?
16:46
vup
well I tried to use this: https://github.com/sergev/ejtagproxy
16:46
vup
to debug the remote using gdb
16:47
se6ast1an
vup you wanted to do a audio chat session with Bertl using nextcloud talk to discuss the micro r3 schematics IIRC, maybe you still find the time to do that soon
16:47
vup
however I was unable to step more than a few instructions, before something broke
16:47
Bertl
doesn't list mz devices
16:48
anuejn
I will continue to play with the usb3 module :)
16:48
vup
Bertl: also true, but considering it worked somewhat and I was able to step a single instruction, I think it would be interesting, if a other jtag / icsp adapter would work better
16:49
Bertl
well, no problem, we can try that
16:49
se6ast1an
<anuejn> I will continue to play with the usb3 module :) <- most excellent! please let me know if anything can be supplied/done/ from our side currently
16:49
Bertl
note that the pickit2 is quite solid while the pickit3 is rather problematic for open source interfaces
16:50
se6ast1an
anuejn/vup did you make any progress with the wifi usb stick kernel drivers?
16:50
vup
se6ast1an: well I don't know if I should call this progress or reverse progress :)
16:51
vup
first of all, we enabled a staging driver for one of the wifi chipsets, however for our current kernel version, that driver doesn't seem to properly support ap mode
16:52
vup
furthermore, we ran into a bunch of problems on the micro with the working usb wifi sticks, that made linux lock up / freeze when interacting with the wifi stick
16:52
vup
But I started working on updating the linux kernel, so that might help with that
16:52
vup
Bertl: I see, it also mentions you need the pickit3 with scripting firmware (not sure what that means, maybe you now more...)
16:53
Bertl
the 'scripting firmware' is basically a firmware which makes the pickit3 behave like the pickit2
16:53
se6ast1an
<vup> se6ast1an: well I don't know if I should call this progress or reverse progress :) <- I see :)
16:54
Bertl
it was released by microchip after many folks complained that the pickit3 sucks and that essential features of the pickit2 were missing
16:55
Bertl
but it was removed again and is not maintained by microchip
16:55
vup
I see
16:55
Bertl
but as I said, we can test both, original pickit2 as well as pickit3 (with scripting firmware)
16:56
vup
sounds good
16:56
Bertl
BAndiT1983 is conveniently 'away' so we won't get a report there now ...
16:57
Bertl
which leaves me with a short report on what happened during the last week
16:57
Bertl
as mentioned before, vup and anuejn visited us in Vienna and hopefully had a great time here
16:58
Bertl
after the success with the USB3 plugins and reasonably high data rates of 360+ MB/s
16:58
Bertl
we decided to adapt the Axiom Beta gateware to support the plugin module
16:59
Bertl
the modified version removes the HDMI encoding but keeps the scan mechanism (well, at least a similar one) to control data output
17:00
Bertl
basically data from memory is retrieved and sent via LVDS channels to the plugin which in turn transfers this data to the receiving side
17:00
Bertl
the new gateware was tested in a low speed configuration (25MHz) and verified by sampling with a logic analyzer and decoding the output
17:01
Bertl
once the gateware for the plugin modules will be ready to decode the data, we should be able to transfer large chunks of memory via USB3
17:02
Bertl
I also took the opportunity to dig (once again) into the gateware and besides some clean up and testing, I created a version which removes the RCN unit as well as the linearization LUTs (which are not required for raw data transfers via USB3)
17:03
Bertl
and I'm currently investigating our options regarding higher transfer rates from the sensor
17:03
Bertl
finally I got SPI working on the new firmware 2.x (which I will test and use shortly) so that userspace can interface with SPI devices over PL
17:04
Bertl
some more gateware testing is planned but we'll see how much time I find for that ...
17:05
Bertl
that's it from my side!
17:05
se6ast1an
great!
17:05
se6ast1an
many thanks, also for leading the meeting today!
17:05
Bertl
you're welcome!
17:05
se6ast1an
and everyone else who participated
17:05
Bertl
as a note to our GSoC students, as final evaluations are coming up, we will have a project discussion today and tomorrow!
17:06
Bertl
(so please stick around a little longer :)
17:36
BAndiT1983|away
changed nick to: BAndiT1983
17:57
BAndiT1983
changed nick to: BAndiT1983|away
18:28
RexOrCine
left the channel
18:46
RexOrCine
joined the channel
18:57
BAndiT1983|away
changed nick to: BAndiT1983
19:03
BAndiT1983
changed nick to: BAndiT1983|away
19:15
RexOrCine
left the channel
19:26
RexOrCine
joined the channel
19:55
RexOrCine
left the channel
19:55
RexOrCine
joined the channel
20:35
Bertl
off for now ... bbl
20:35
Bertl
changed nick to: Bertl_oO
20:48
BAndiT1983|away
changed nick to: BAndiT1983
21:37
BAndiT1983
changed nick to: BAndiT1983|away
22:21
mumptai
left the channel
22:37
illwieckz
left the channel