Monday, 2017-07-17

*** tpb has joined #timvideos00:00
xfxfjea: pong00:13
mithrocr1901_modern: ping?00:28
cr1901_modernmithro: pong00:29
mithrocr1901_modern: So, IIRC you were looking at writing to the spiflash?00:36
cr1901_modernmithro: Correct, reading and writing spiflash, and adding xmodem/sfl loading00:37
mithrocr1901_modern: Where did you get to on that?00:37
cr1901_modernwriting xmodem/sfl impl, spiflash reading and writing works in 1x mode. I didn't get it working in 4x mode, and I need a new power supply for my Opsis >>00:38
cr1901_modernCould you recommend an Amazon link for a power supply? I don't have the CPU supply I used anymore (long story)00:38
cr1901_modernCurrently writing* xmodem/sfl impl.00:39
mithrocr1901_modern: I sure document somewhere the connector size on the Opsis00:39
mithrocr1901_modern: Something like this should work ->
mithroOh wait, yay past Mithro ->
mithroInner Diameter: 2.0mm00:40
mithroOuter Diameter: 5.5mm00:40
mithroI think that one I linked should work00:41
mithroWe should figure out a good power supply and put an amazon affiliate link to them on that page00:41
* cr1901_modern agrees lol00:41
cr1901_modernYou also wanted me to implement loading on QEMU00:42
mithrocr1901_modern: Loading on QEMU?00:42
cr1901_modernspiflash emulation* on QEMU so you can load firmware over a virtual serial port00:43
mithrocr1901_modern: We have spiflash emulation in QEMU already (both reading and writing)00:43
cr1901_modernOh, then it was just making sure loading firmware works00:44
cr1901_modernthrough a virtual serial port*00:44
cr1901_modernusing QEMU**00:45
jeaxfxf: i'm going to prefix this by saying i expect the answer to be no, but just want to check anyway00:51
jeaxfxf: is there any possibility of getting 2 LA opsis sent to brisbane this week for me to use at wordcamp (an LA event), and then i send them back next monday after the event?00:52
jeai know it is close to pycon, so that makes it more unlikely, but just thought i would ask00:53
mithrocr1901_modern: I think it already does - haven't bothered testing it yet00:55
mithrojea: I have 2 * Opsis that I could lend you00:55
jeamithro: that would be awsome if I could borrow them00:56
cr1901_modernmithro: Could you please test for me? Running QEMU at the moment is not possible (on Windoze)00:57
mithrojea: I'm not sure the best way to ship you them?00:57
mithrojea: They also come with the implication that you'll do some more Opsis hacking :-P00:57
jeahaha yes, more opsis hacking is on the cards. i have a dev environment properly setup as of the weekend, so it will be much simpler to do now :D00:58
mithrojea: When do you need them by?00:58
jeaI'm not too sure on the best shipping. the event is this weekend (late notice sorry - was working on other options but they seem to have gone out the window)00:59
mithrojea: Okay, I'll try and pop out at lunch time and ship them by overnight express00:59
mithrojea: Can you give me an address + mobile number to ship too?00:59
mithrojea: I don't have power supplies but do have redmere cables01:00
jeayep, i can do all that. power supplies should be fine to get, and I now have 8 of the new monoprice active hdmi cables (i think they replaced redmere), so I think that should be covered01:02
mithrojea: I keep redmere cables in the Opsis boxes :-)01:02
jeathese ones
jeaah, good idea!01:02
jeawill I email you the address?01:03
mithrojea: please do01:05
mithrojea: Remind me about 1-2pm01:05
cr1901_modernmithro: This is a stupid question, but what is SSI flash?
tpbTitle: litex: Working spiflash emulation. · timvideos/[email protected] · GitHub (at
jeamithro: ok, will do :)01:05
cr1901_modernWhat does SSI stand for*?01:05
mithroSSI is what qemu calls SPI01:05
cr1901_modernOkay, so the idea is that you just compile the firmware, and qemu provides emulation of litex peripherals?01:06
mithrocr1901_modern: /* QEMU Synchronous Serial Interface support.  */01:06
mithroAcording to
tpbTitle: qemu/ssi.h at master · openvstorage/qemu · GitHub (at
mithrocr1901_modern: Yes01:06
mithrocr1901_modern: In theory qemu the way we use it should actually run under cygwin....01:07
cr1901_modernQEMU Needs KVM IIC01:07
mithrocr1901_modern: Its actually the other way around KVM needs QEMU01:12
mithrocr1901_modern: KVM uses QEMU for emulation of hardware peripherals01:12
mithrocr1901_modern: Seems to be some older info at and some more modern info at
tpbTitle: Qemu On Windows (at
cr1901_modernmithro: My mistake I will take a look01:13
cr1901_modernmithro: Some Litex changes recently (?) broke the build (_florent_ refactored the cores), so I need to investigate those first.01:19
cr1901_modernrefactored the python module paths*01:19
mithrocr1901_modern: Okay01:19
cr1901_modernmithro: I have an old XMODEM implementation in C89. Would you be open to me trying to use that (as an external dep)? If not, I can rip out the functionality from it. And lastly, have you looked into ekermit?01:22
cr1901_modernI despise the documentation but it is a smarter/more efficient protocol than XMODEM01:22
mithrocr1901_modern: I think I linked to xobs xmodem implementation? That seems like a better thing to steal from?01:23
mithrocr1901_modern: The point of xmodem is not because it is a good protocol but because everything under the sun already supports it01:23
cr1901_modernmithro: Well, would xobs object to me stealing it? That's my only concern.01:26
mithrocr1901_modern: No - it's FOSS so you can steal from it :-)01:26
cr1901_modernFair lol01:26
cr1901_modernmithro: More bikeshed painting: if implementing the full spec of xmodem, the recv loop is kinda ugly. I think xob's only implements the 1K variant, where my impl impls the whole spec (supposed to fall back to 128byte if 1K fails). Of course, xob's version is nicer to read :).01:28
mithroFeel free to improve xobs version and send him an update :-P01:29
cr1901_modernThis was me asking: You want me to impl the whole spec?01:29
mithrocr1901_modern: That is unclear to me01:30
mithrocr1901_modern: Well, you are only implementing receiving....01:30
mithrocr1901_modern: Support for the different sizes would be nice I think?01:30
cr1901_modernxmodem is receiver driven protocol; it sends out a byte requesting 1024-byte packets three times. If it doesn't receive a response back within 3 times, it falls back to sending out a request for 128-byte packets w/ CRC. If the receiver receives a byte from the transmitter indicating "I don't support either of 1024-byte or 128-byte w/ CRC", the receiver falls back to 128 byte with 8-bit checksum.01:32
cr1901_modernmithro: Okay, that answers my q01:32
jeamithro: reminder as requested03:41
mithrojea: Try again in an hour :-P03:41
mithropaddatrapper: I think I've reviewed everything you had outstanding?04:09
mithrojea: Okay, heading over to the post office now04:15
jeamithro: thanks :)04:15
mithroyou need 2 right?04:16
mithrojea: ^04:17
jea2 assuming atlys works tonight, otherwise 304:17
jeai'm hoping atlys will work04:17
*** rk[ghost] has quit IRC04:48
vysmithro: the atlys has separate flashing/console USB connections, correct?04:52
*** rk[ghost] has joined #timvideos04:55
mithrojea: you have email04:58
mithrovys: not sure what you mean?04:58
mithrovys: The atlys does have a hardware UART on it04:58
mithrovys: That is only used for the control serial port console05:01
vysthe original that you gave jea has two uUSB jacks on it, one with UART acessible via the /dev/ttyACM0 file, and another one that reports a different VID/PID and enumeration doesn no ttyACM005:01
vysdoesn't create ttyACM0*05:02
mithrovys: Eventually I hope that we don't have to use that any more05:02
mithrovys: The hardware UART on the second jack doesn't work with the ACM driver you need the exart driver05:02
vysthat answers my next question05:02
mithroexart / vizzini module - # Get the vizzini module, only needed for the Atlys board #apt-get install -y vizzini-dkms05:03
vysjea: so perhaps the atlys works fine05:04
vysI wrote a thing to use instead of flterm05:04
tpbTitle: Connect to HDMI2USB serial using bash and cat · GitHub (at
vysbut I wasn't getting anything out of /dev/ttyACM, so I'll try that when I have an opportunity to do so05:05
mithrovys: I doubt that will work05:05
vysthanks mithro05:05
mithrovys: The UARTs are kind of grumpy / borked05:05
vysthis was running flawlessly on the opsis last night05:06
mithrovys: You should use hdmi2usbd ->
tpbTitle: GitHub - timvideos/HDMI2USB-controlproxy-daemon: Simple serial to network proxy daemon suitable for use with HDMI2USB devices. (at
mithrovys: There is no flow control with the above, so if your computer happens to be faster or the opsis happens to be a bit slower you'll end up with weird behaviour05:07
mithrovys: Very dependent on the computer05:07
vysI'll believe that05:07
vysI won't advertise it then :)05:07
vysI was using it on my laptop with busted USB that only talks to opsis :)05:08
mithrovys: I'm hoping that as paddatrapper gets more familiar with the fx2 firmware I can persuade him into doing work like making the USB-CDC ACM interface on the FX2 be connected to the FPGA on the Atlys via a bit banging serial port (on the Opsis there is hardware serial port that can be used).05:11
tpbTitle: udev rules should blacklist the exart / vizzini hardware UART on the Atlys from the USB-CDC ACM driver (and maybe warn about the condition) · Issue #58 · timvideos/HDMI2USB-mode-switch · GitHub (at
tpbTitle: GitHub - mithro/exar-uart-driver: Linux kernel driver for the Exar xr21v141x "vizzini" UART (at
*** shenki has joined #timvideos05:17
*** ChanServ sets mode: +v shenki05:17
* mithro beats shenki with a large trout05:28
mithroshenki: If you had gotten that driver into the kernel back in Nov 15, 2015 when we wrote that code, it would probably be in LTS distros by now!05:29
mithroshenki: But you did not....05:35
shenkinot google's lts tho. they still run 14.0405:35
mithroso :-(05:36
shenkimithro: ok. did you find the patch that the debian dude wrote for is?05:36
shenkimithro: ok. did you find the patch that the debian dude wrote for us?05:36
mithroshenki: I can find him, why did we want to find him?05:36
shenkinot the person, just the patch05:37
shenkii will reivew it and send it upstream05:37
shenkithe blocker is i started running my laptop with secure boot turned on a few years ago05:38
shenkiwhich means i can't test it on the laptop05:38
shenkii do have random linux boards to test it on, so i'll use a raspi or something to test it05:38
tpbTitle: udev rules should blacklist the exart / vizzini hardware UART on the Atlys from the USB-CDC ACM driver (and maybe warn about the condition) · Issue #58 · timvideos/HDMI2USB-mode-switch · GitHub (at
mithro@shenki - My email says to look here ->
tpbTitle: GitHub - shenki/linux at vizzini (at
mithro@shenki - IIRC We got to the driver working but wanted to refactor things slightly. I'm pretty sure we wanted to move this hack -> into the probe function because the vizzini module has similar functionality at
tpbTitle: linux/usb-serial.c at vizzini · shenki/linux · GitHub (at
shenkithat's not the one that the debian guy wrote05:43
tpbTitle: usb-serial: Add vizzini driver for Exar USB 2 serial adapters · shenki/[email protected] · GitHub (at
tpbTitle: Bug#804350: ITP: vizzini -- Kernel driver for Exar XR21V1414 USB UART (at
mithroshenki: bwhacks is Ben Hutchings the debian guy05:44
mithroshenki: You added a patch on top of his patch in that branch05:44
mithroBen Hutchings the debian guy doesn't have an Atlys board05:45
shenkiso we need to:05:45
shenkiapply it to something modern, like 4.13-rc105:46
shenkitest it with the hardware05:46
shenkisend it to lkml05:46
mithroIf you apply it to something modern I'm sure I can find you a hardware tester05:47
mithro<cough> CarlFK <cough> vys <cough>05:47
vysI do in fact have hardware05:48
mithroshenki: It should be pretty simple rebase, I doubt the usb-serial driver gets much love :-P05:48
mithroshenki: Help get your commit count up ;-)05:48
shenkii don't get to commit it - that's gregkh's tree :)05:49
shenkivys: do you know how to build + boot your own kernels?05:49
vysI do it at work for our hardware, but there is a gap between that and my laptop05:50
vysI can read things and ask things though05:51
shenkiheh. you could test it on your hardware :)05:51
shenkiyeah, i can guide you through it05:51
mithroshenki: I'm not sure how you get grub to find your own custom kernel these days, do you just put something in /boot ?05:53
shenkieasiest way to do it is to 'sudo make install'05:54
shenkithat will call your distro's helper scripts to generate an initramfs and boot entires05:54
shenkiotherwise we can try to build an out of tree module against the system's headers05:55
jeamithro: thank you very much for that! it will help immensely06:05
shenkivys: let me know when you're ready to do some testing07:47
shenkivys: what distro do you have on your laptop?07:47
shenkivys: grab that, untar, make && make install07:54
shenkithen plug in your device07:54
CarlFKmithro: huh what?07:55
shenkior anyone who is lurking and has a Atlys board in arms reach07:56
paddatrappermithro: thanks for the reviews, I'll address them this morning08:10
Ishan_Bansalshenki : ping08:29
shenkiIshan_Bansal: hello08:29
Ishan_Bansalshenki : Hi, I have added the RLEmain and the entrophycoder in the same branch as the rle-core.08:30
Ishan_BansalIs it ok, or should I create a new one and add them in those ?08:31
Ishan_Bansalfor the present they are in the rle-core itself.08:31
shenkiIshan_Bansal: that's fine for now08:50
shenkiIshan_Bansal: the URL on your blog post is malformed08:50
CarlFKshenki: I have an atlys hooked up to a box you can ssh to08:52
Ishan_Bansalshenki: Let me see.09:08
shenkiCarlFK: cool! want to email me the details?09:13
shenkiCarlFK: does it have the usb cable plugged in to the exar device?09:13
shenki(i think that's the usb port on the side, instead of the one where you get the uvc device from)09:13
Ishan_Bansalshenki : I have updated the blog.09:18
CarlFKshenki: both of them are plugged in09:21
shenkiCarlFK: cool09:21
shenkiCarlFK: can you unplug and re-plug the exar usb cable?09:26
CarlFKcan I do both?09:27
CarlFKim not sure which is wich09:27
shenkihrm. something is broken09:28
shenkiCarlFK: thanks! will do some debugging09:28
shenkimithro: did you patch the driver you were using to make it show up as /dev/ttyXRUSB0?09:30
CarlFKshenki: there may be a udev rule on that does that09:31
CarlFKshenki: im heading back to bed for 4 hours - so if you need it touched.. sorry :p09:45
vysI don't have an atlys/opsis11:05
vysthey are all at jea's house11:05
vys17:47 <+shenki> vys: what distro do you have on your laptop?11:05
vysI'm using fedora11:06
vyshe might have a box I can ssh into11:07
jeai will have a box soon. need to make it work again11:12
CarlFKim up13:01
xfxfjea: sorry missed your message earlier - yes more than happy to send you any of the LA opsis's, but you need it by this weekend, right?13:23
xfxfmy criteria is i need them tested on a recent firmware - you're doing that for me by using them.  bringing them back when you come to pycon is fine13:23
mithroxfxf: I've already sent him some of mine14:02
mithropaddatrapper: Lets keep fx2lib board agnostic and HDMI2USB-fx2-firmware board specific15:38
CarlFKmithro: don't forget your reminder to make symlinks
tpbTitle: HDMI2USB-firmware-prebuilt links to unstable testing, stable · Issue #328 · timvideos/HDMI2USB-litex-firmware · GitHub (at
*** MatrixTraveler[m has quit IRC16:40
*** jea[m] has quit IRC17:11
*** jea[m] has joined #timvideos17:20
paddatrappermithro: ok cool17:28
*** rohitksingh has joined #timvideos17:53
*** rohitksingh has quit IRC18:22
jeaxfxf: thanks for that. mithro sent me some (which just arrived, thanks tim!), so that should cover me now23:22
mithrojea: Oh, great!23:48
jeamithro: there are lots of connectors in the box. i think they are right angle things for the tofe interface?23:49
mithrojea: They are for the Numato to Digilent PMOD converter23:51
mithroI just scored !23:52
mithroshenki: Did you get the Atlys going in the end?23:52
jeamithro: oh ok. not sure i realised that was a thing23:54
mithrojea: You have a MimasV2 right?23:54
mithroWell you or vys?23:54
jeaI don't think the atlys stuff has been tested yet. i need to turn mine on to test with. though maybe he did it with CarlFK's atlys?23:55
mithrojea: I think he found his own Atlys23:55
jeaI don't yet. Was going to buy one, but just haven't done it yet23:55
mithrojea: When is wordcamp?23:55
jeaah, that would make it easier!23:55
jeawordcamp is this friday23:55
jeasorry, saturday23:55
jeatoo much stuff in my head23:56
*** futarisIRCcloud has joined #timvideos23:56
mithrofutarisIRCcloud: Wordcamp is on Saturday in Brisbane if I understand jea correctly23:57
mithrofutarisIRCcloud: jea will be using Opsis boards to record the event23:57
jeayes, wordcamp is on at QUT on sat and sun23:57
mithrojea any chance if futarisIRCcloud is available you could have him come along and meet you?23:58
mithrojea, futarisIRCcloud is the guy who has been doing a lot of the Linux on LiteX stuff -- I send him a MimasV223:58
jeayes, he is more than welcome to come along :) i'll probably be busy at the start of sat, but by lunchtime I should be fine23:59
jealinux on litex would be awesome23:59

Generated by 2.13.1 by Marius Gedminas - find it at!