Sunday, 2018-03-18

*** tpb has joined #timvideos00:00
*** akhil_singh has joined #timvideos00:04
*** shorne has joined #timvideos00:05
*** shorne_ has joined #timvideos00:05
akhil_singhmithro: For implementing error detector. I am thinking of something like that the error detector will use the incoming data as seeds to generate the expected patterns in the lock state. After a certain number of error-free cycles, it switches to internal seeds so that incoming erroneous bits will not disturb the error checking of the detector.00:05
akhil_singhmithro: Is this correct ?00:06
mithroakhil_singh: I think you are talking about a self synchronizing pattern detector?00:06
mithroakhil_singh: The most basic pattern should be a predefined like 010100:07
mithroshorne/shorne_: ping?00:08
akhil_singhmithro: Yeah00:09
akhil_singhmithro: Also now for pattern generation we can design a state machine which will check for idle condition and pattern generator. So if it is idle then tx_out < sync data otherwise enable pattern generator00:11
mithroakhil_singh: The pattern which is being used should be selectable from a bunch of possibilities by the user00:11
akhil_singhmithro: Okay00:12
mithroakhil_singh: A good part of your proposal should be which patterns should be available by default and why they would be useful00:13
akhil_singhmithro: What if we go from idle to pattern_generate state if (rx_freqlock & rx_patterndetect) = 1 .00:14
mithroakhil_singh: hrm?00:14
akhil_singhmithro: Will imposing this condition help us choose from a bunch of possibilities by the user.00:17
mithroakhil_singh: I'm afraid I still don't understand the question00:17
akhil_singhmithro: we can set rx_patterndetect=1 only when it is from bunch of possibilities defined by the user. In this correct ?00:19
mithroI don't get what rx_patterndetect is?00:19
cr1901_modernmithro: Is it possible for you to downgrade the Xilinx dist until the TCL problem goes away?00:20
mithrocr1901_modern: So you don't need my change?00:21
cr1901_modernmithro: I haven't built anything since you pushed your changes00:21
akhil_singhmithro: This is about Pre-define training patterns.00:22
cr1901_modernmithro: I'm asking whether you think the TCL error was _also_ due to tarball issues or potentially due to my changes.00:22
mithrocr1901_modern: The tcl problem is caused by migen using the tcl if I understand correctly...00:22
akhil_singhmithro: So that now we can transmit Pre-defined training patterns and verify for the link to establish word boundary and for the state machine to advance to the link-lock mode.00:23
cr1901_modernOkay I'll take a look then. In any case, I'm going to do my work off of new-litex starting now (new-litex-tmp is a dead end)00:23
mithroakhil_singh: Yes00:24
akhil_singhmithro: Okay.00:24
akhil_singhmithro: As you suggested that good part of your proposal should be which patterns should be available by default and why they would be useful. So i will start writing proposal and code for PRBS. Should I do write code for any other sublock now  ?00:27
mithroakhil_singh: You don't need to write the code - just describe what type / styles of patterns might be useful00:28
akhil_singhmithro: Okay00:28
akhil_singhmithro: Thanks00:28
shornemithro: hello00:29
mithroshorne: How's the gcc porting going? I saw you did some test stuff to get a feel about how to do the port?00:30
shornemithro: yes, its ok, I did a dummy architecture following a tutorial, I think that went well00:31
shornefinished that last week, so starting on the or1k port. I did some housekeeping yesterday and update binutils (which I will use for the port)00:32
shorneI decided to create a toolchain release with gdb too (
mithroshorne: Cool!00:32
tpbTitle: GDB Upstreaming and Binaries: 7.2.0 - OpenRISC (at
cr1901_modernmithro: I'd like to get these changes merged in tonight, but right now I'm afk. Be back in an hour hopefully00:32
cr1901_modernshorne: Excellent work! Seeing ppl port gcc keeps me motivated to improve00:32
mithroshorne: We build them as part of our conda packages ->
tpbTitle: conda-hdmi2usb-packages/or1k at master · timvideos/conda-hdmi2usb-packages · GitHub (at
shornemithro: but thats probably the last release, with that version of the gcc port as Ill be working on the rewrite now00:33
shornemithro: yeah, I seem to remember that00:33
shorneI create a few different toolchains00:33
shornecr1901_modern: do you have experiecen with gcc backend porting?00:34
*** akhil_singh has quit IRC00:34
cr1901_modernshorne: Dabbled in it once in 2015 for a deprecated arch. Didn't make it that far. More recently I've been reading about the RTL language00:35
cr1901_modernand well, I like it better than LLVM let's just say that :)00:35
shornecr1901_modern: I haven't messed with llvm much but I take its probably a bit easier to write the port00:38
mithrocr1901_modern: don't you let whitequark hear you say that :-P00:38
shornebut gcc is quite flexible00:38
shornemithro: I will admit I dont know what I am talking about, I never looked much at LLVM code :)00:39
cr1901_modernmithro: I'm well aware00:39
mithroshorne: I was more talking about cr1901_modern liking gcc better then llvm :-P00:39
shornehaha :)00:40
*** shorne_ has quit IRC00:41
shorneI was logged in 2 times, had trouble with my dsl connection this morning after a router firmware upgrade00:42
cr1901_modernmithro: I find s-exprs (RTL) easier to read. I am basing my opinion on aesthetics rather than any design decision concerned with what "physically existing architecture" code is generate00:42
mithroshorne: Oh - nolib is a good name for the nostdc binary :-P00:42
shornemithro: I think thats what it was called by crosstool00:42
tpbTitle: Latest cross-compilers (at
shorneoh wow, arnd has a 7.3.0 toolchain there for openrisc00:44
shornemithro: I cant find reference to 'nolib' there, I dont think I came up with it00:46
mithroIs or1k in crosstool-ng?00:46
shornethe build script they/we use is here:
tpbTitle: Git - users/segher/buildall.git/summary (at
shornemithro: I dont think crosstool-ng is related, I am just seeing that myself for the first time00:47
mithroshorne: crosstool-ng is what conda currently uses to build toolchains00:47
shornemithro: good, maybe I should look into it00:48
mithroI won't necessarily say it is _good_ - but the version they use supports building windows/linux/mac host toolchains00:48
shornemithro: I just hacked together this docker image and script to spit out toolchains
tpbTitle: GitHub - stffrdhrn/or1k-toolchain-build: OpenRISC toolchain build scripts (at
shorneits a bit slow but its repeatable so I use it00:49
mithroshorne: Yeah00:50
mithroThey use this fork ->
tpbTitle: GitHub - diorcety/crosstool-ng: crosstool-ng fork with the full support of Darwin/Linux/Windows as build and host machine (at
shornemithro: I see, for the openrisc toolchains I didnt even thing for Darwin/WIndows users00:52
shorneprobably I should :)00:52
shornemithro: do you find many people actually use the toolchains on windows? I guess Darwin/MacOS would exist00:53
mithroshorne: not yet :-P00:53
shornemithro: I need to go do some chores00:57
mithrodeclare -x HAS_ANTARES_THREE_LITTLE_FRONZIES_BADGE="true"00:57
cr1901_modern^The hell? (No I haven't left my keyboard yet)00:58
cr1901_modern"fronzie" sounds like a branded milkshake or something00:59
mithroHrm, I can't see anything about TCL path01:00
cr1901_modernDoes the TCL path issues predate new-litex?01:01
cr1901_modernWhy not restart a build of a prev successful job w/ the new Xilinx tarball01:01
*** sb0 has quit IRC01:05
cr1901_modernlike for instance:
cr1901_modernerr wait, passibly wrong build01:09
cr1901_modernmithro: I'm lost at this point: This build fails on your branch01:13
cr1901_modern This build is the merge of the build just above and master01:14
cr1901_modernNotice that one of them succeeds... why?! T_T Does the timvideos version use a different list of xilinx builds?01:14
cr1901_moderns/one of them/the upstream build/01:15
tpbTitle: AR# 62945: Vivado - When launching Vivado I see: ERROR: [Common 17-656] Repo catalog is not loaded. (at
mithrocr1901_modern: Pull requests don't run toolchain01:17
mithros/run toolchain/run using Xilinx toolchain/01:17
cr1901_modernmithro: I want to eliminate the Xilinx tarball as the source of TCL issues before diving into the code. Do you remember the earliest commit where the issue appeared?01:20
cr1901_moderns/the issue/the TCL issue/01:21
mithrocr1901_modern: Which issue are you looking into?01:22
cr1901_modernmithro: There's more than one now? The one where Vivado can't find it's TCL script01:24
cr1901_modernI just want new-litex merged. My PR is starting to block other ppls work, and porting back to migen _should_ have been a simple text substitution.01:26
mithrocr1901_modern: I think I've succeeded ->
cr1901_modernmithro: That's wonderful :D! Still worried about why it's required after the migen switchover03:43
cr1901_modern(and additionally, my local copy didn't require an env var like that)03:43
cr1901_modernI swear I tested the PR before I made it :P03:43
mithrocr1901_modern: I'm guessing you have the correct files in ~/.Xilinx03:48
cr1901_modernmithro: I do have the correct files under ~/.Xilinx03:51
cr1901_modern(8:19:35 PM) mithro: cr1901_modern: The tcl problem is caused by migen using the tcl if I understand correctly...03:57
cr1901_modernI think I see now...
tpbTitle: build: merge with 27beffe7 · enjoy-digital/[email protected] · GitHub (at
mithrocr1901_modern: Yeah
tpbTitle: build: merge with 27beffe7 · enjoy-digital/[email protected] · GitHub (at
mithrogah ->
cr1901_modern Culprit on the Migen side (I wasn't aware of this change in Migen)04:13
tpbTitle: vivado: create vivado project to enable inspection · m-labs/[email protected] · GitHub (at
cr1901_modern(That's why I was so confused about why the swap to migen would change behavior)04:13
*** shivm has joined #timvideos04:27
*** shivm has quit IRC04:38
mithroERROR:Xst:1817 - Invalid target device 'xc6slx9'.04:43
cr1901_modernSpartan 6 never existed, it was a figament of our imaginations :D! (wtaf?!)04:45
nancyHelp! The file image-gateware+bios+firmware.bin in  HDMI2USB-litex-firmware/build/opsis_video_lm32  couldnt be built!04:55
*** suhdood has joined #timvideos05:03
CarlFKnancy: paste bin the output (using again)05:16
tpbTitle: Ubuntu Pastebin (at
nancy ok sir05:24
tpbTitle: Ubuntu Pastebin (at
CarlFKnancy: sorry, I mean the output from the make command.05:44
CarlFKer, wait.. it did build: image-gateware+bios+firmware.bin05:44
CarlFKI think you mean it didn't work on my opsis.05:45
*** suhdood has quit IRC05:47
nancymake gateware  ? command05:49
nancySir , after running make gateware command i performed "make image command" which was not mentioned in repo. And after that image-gateware+bios+firmware.bin came05:51
CarlFKcr1901_modern:  can you help nancy with
tpbTitle: HDMI2USB-litex-firmware/ at master · timvideos/HDMI2USB-litex-firmware · GitHub (at
CarlFKthat doc may have gotten out of date- im not sure how to debug05:59
nancyyes sir06:00
nancyAfter make gateware i had to run make image which was not mentioned in getting-started.md06:01
cr1901_modernCarlFK, nancy: I'm going to bed, but I've been having trouble building the images recently as well. I would not be surprised if it's a latent bug.06:01
CarlFKcr1901_modern: but... the problems you are having havn't landed in the repo yet, right ?06:03
CarlFKor maybe they have, and travis.. never mind.  I think I understand enough.06:03
cr1901_modernCarlFK: Me being unable to build the images has been a thing even on the main branch if memory serves06:03
*** rohitksingh has joined #timvideos06:05
nancysir do we use Firewire ?06:11
CarlFKnancy: no06:12
CarlFKnancy: git checkout bc7feb506:17
CarlFKthat will go back about 10 commits to a version we used successfully at an event.06:18
CarlFKand then gateware command06:19
nancyyes sir06:25
*** rohitksingh has quit IRC06:33
nancyi am figuring it out sir06:35
*** rohitksingh has joined #timvideos06:43
nancynow it is not entering the (H2U P=opsis) #06:49
*** rohitksingh has quit IRC06:50
CarlFKnancy: sorry, im not sure why that would happen.07:05
CarlFKhave a look at the .sh script, it isn't too long07:06
CarlFKsadly, I need to get to bed07:06
*** tsukasa_au has quit IRC07:19
*** Ishan_Bansal_ has joined #timvideos07:20
*** jimmo_ has joined #timvideos07:20
*** thaytan_ has joined #timvideos07:22
*** jimmo has quit IRC07:22
*** tsukasa_au has joined #timvideos07:22
*** Ishan_Bansal has quit IRC07:22
*** mcarden_ has quit IRC07:22
*** thaytan has quit IRC07:22
*** xobs has quit IRC07:22
*** tsukasa_au has quit IRC07:22
*** tsukasa_au has joined #timvideos07:22
*** Ishan_Bansal_ is now known as Ishan_Bansal07:22
*** jimmo_ is now known as jimmo07:22
*** mcarden has joined #timvideos07:22
*** xobs has joined #timvideos07:23
*** xfxf_ has joined #timvideos07:30
*** tsukasa_au__ has joined #timvideos07:30
*** xfxf has quit IRC07:32
*** tsukasa_au has quit IRC07:32
*** xfxf_ is now known as xfxf07:32
*** tsukasa_au__ is now known as tsukasa_au07:32
*** shivm has joined #timvideos07:56
*** shivm_ has joined #timvideos08:18
*** shivm has quit IRC08:20
*** shivm__ has joined #timvideos08:23
*** shivm_ has quit IRC08:26
*** harshgugale has joined #timvideos08:42
*** nancy has quit IRC09:00
*** shivm__ has quit IRC09:38
*** harshgugale has quit IRC10:29
*** shivm has joined #timvideos10:58
*** sb0 has joined #timvideos12:29
*** anant3110 has joined #timvideos13:20
CarlFKme too: 502 Bad Gateway
tpbTitle: PS:One Member Site (at
*** shivm_ has joined #timvideos13:58
*** shivm has quit IRC14:01
*** shivm__ has joined #timvideos14:07
*** anant3110 has quit IRC14:41
CarlFKwhoops, wrong #chan...15:09
*** nancy has joined #timvideos15:23
*** nancy has quit IRC16:14
*** nancy has joined #timvideos16:14
nancy@CarlFk sir , kindly review
tpbTitle: Ubuntu Pastebin (at
*** shivm__ has quit IRC17:02
*** shivm_ has quit IRC17:02
*** shivm has joined #timvideos17:03
*** shivm_ has joined #timvideos17:03
CarlFKnancy: looking...17:04
*** Samidhamdhi has joined #timvideos17:26
*** Samidhamdhi has quit IRC17:32
CarlFKnancy: do this for me...   mv image-gateware+bios+firmware.bin image-gateware+bios+firmware-nancy-v1.bin17:34
CarlFKmd5sum image-gateware+bios+firmware-nancy-v1.bin17:34
CarlFKand email me that file and the md5sum17:34
*** samishamshi has joined #timvideos17:35
CarlFKthe rename is to make sure I don't end up flashing the wrong file, the md5sum is to make sure it isn't corrupted17:39
nancyok sir17:41
nancylet me mail17:42
nancyyes sir , kindly review17:44
CarlFKnancy: [email protected]17:49
CarlFKnancy: [email protected]17:49
CarlFKno need to fill everyone's inbox with the attachment17:49
nancyi am sorry sir !17:49
nancydidnt ask your id earlier17:50
nancyi am just sending it again17:50
CarlFKno problem - we dont really have a procedure for this kind of thing17:52
*** samishamshi has quit IRC18:03
nancyyes sir18:04
CarlFKnancy: same result: opsis doesn't boot.  :(18:15
CarlFKnancy: is this the same?  a1955f00ccf014c0f071073e6f890430  image-gateware+bios+firmware-nancy-v1.bin18:15
shivm_CarlFK: Hi, Do I need any specialised USB cable to flash .bin on board?18:16
shivm_Currently, I am using just a standard USB cable to flash the board.18:17
CarlFKshivm - no - normal usb cable.  (two of them.. well 1 if you know which port to use, I just hook up both18:17
shivm_Did my .bin file work?18:18
shivm_ok thanks18:18
CarlFKshivm: do you have a machine you can wipe/install a fresh OS?18:18
nancyit is the same18:19
shivm_But I can try setting up things on in one of the computers in the lab.18:20
shivm_What all do I need?18:20
CarlFKshivm  shucks.  we have a pretty good script to : make usb installer - wipe, install debian and the few things needed18:20
nancywhat might be the probelm sir ?18:21
CarlFKshivm: this is 'everything' which includes drive partitions, user names.. and the few things that are relevant to your flashing problem
tpbTitle: video-stack-deploy/scripts at master · CarlFK/video-stack-deploy · GitHub (at
shivm_CarlFK: thanks a lot!!18:23
CarlFKshivm iv you have something like a raspberry pi we could make that work18:24
shivm_Currently, I have access to Arduino UNO. I can ask for Rpi from the lab tomorrow.18:25
CarlFKnancy: really I don't know.  I have 2 guesses: 1. it isn't the right platform/target.  (but it looks like that is right)  2. the version you are trying to build is broken on our side18:26
nancybecause i deleted and set the environment again today sir18:26
CarlFKshivm: needs to be something that will run linux.  ideally something that can boot from usb.  so an old laptop from 1995 would be ok18:27
nancyi just want to bring one thing in notice, that after i followed all instruction as given in getting started.md18:28
nancyand after make gateware , make image was followed18:28
nancyand than image-gateware+bios+firmware-nancy-v1.bin file was build18:29
CarlFKmithro: and cr1901_modern and maybe shivm know more about the build process than I do18:31
CarlFKnancy: can you pastebin the output of make gateware ?18:34
*** samishamshi has joined #timvideos18:36
*** nancy has quit IRC18:36
*** nancy has joined #timvideos18:37
nancyshivam seeking your help !18:43
shivm_CarlFK: I will issue an Rpy tomorrow from the lab and then follow the link. Currently, I am in the process of writing proposal for GSOC.18:45
CarlFKshivm: you won't be able to boot the pi from usb. you are going to need to put a minimal debian install and then use ansible to set up the things you need18:47
CarlFKand you can do all of that on the pi18:47
CarlFKnancy: can you pastebin the output of make gateware18:47
shivm_Currently, I have installed Ubuntu 16.04 and Windows 10 on my laptop.18:48
CarlFKnancy: I am hoping that will help anyone see what the problem is18:48
CarlFKshivm: I use ubuntu too. I use both debian and ubuntu for testing -18:49
CarlFKshivm: mithro likes ubuntu, tumbleweed and a few others like debian.  for 'this' I think you want as simple as possible, so debian18:49
shivm_I think there is some issue with Ubuntu 16.04. Is there any need to change permissions for serial port?18:50
shivm_I can install debian on Virtual Machine and then check, right?18:51
nancyyes i am doing it18:52
CarlFKshivm: don't bother with the VM - it can't override whatever the problem is with the host os18:53
CarlFKshivm: i suspect udev rules - either one is doing something we dont want, or you need one to do soemthing and it isn't18:54
CarlFKthis is why starting with a bare system that we know doesn't do anything is a good place to start18:54
shivm_Okay I will try something.18:56
*** samishamshi has quit IRC18:56
nancysir in issue #[HDMI2USB] Get Milkymist "Video DJing" functionality on the Numato Opsis board #33, i read a document given18:59
nancyi have a query that TimVideos is developing on  Option 3: Porting Linux to run on MiSoC based gateware and porting Flickernoise to Linux19:00
CarlFKnancy: I think porting Linux has been done, but... I suspect it isn't well documented and no tests.  ...19:05
CarlFKright now I do most of the testing, and I don't even have .. a procedure19:05
CarlFKit would be wonderful if there was something that would flash a file and check for results19:06
CarlFKone problem with that is "flash a file" is not reliable.   so that would have to be accounted for somehow.19:07
nancythe code base for this is in  ?19:10
tpbTitle: GitHub - m-labs/linux-milkymist: Linux kernel with Milkymist support patches (at
CarlFKnancy: I would ignore milkyumist part19:12
CarlFKnancy: well, I am assuming there is an easier way to boot Linux on Opsis.  but maybe that is the only way.19:17
*** nancy has quit IRC19:19
*** nancy has joined #timvideos19:20
CarlFKshivm: when you plug in / power cycle the atlys, does dmesg show ttyVIZ019:22
shivm_I tried changing permissions for serial port and tried one or two scripts from internet but nothing worked.19:25
CarlFKI bet you don't have the driver installed19:25
CarlFKwhich I think is needed - for sure i is needed to operate, and maybe  you need it to flash too19:25
shivm_I were able to load gateware using PROG port but there is some issue with USB/JTAG port19:26
mithroI'll be around in about 1 hour19:27
CarlFKsudo apt install vizzini-dkms19:27
nancysir should i paste the output of make gateware at ubuntu pastebin !19:27
mithrocr1901_modern: it's merged!19:28
CarlFKnancy: yes19:34
shivm_CarlFK: I just installed vizzini-dkms19:34
shivm_Now, what command should I use to see if board is connected or not?19:35
CarlFKshivm:power cycele the atlys, dmesg, look for ttyVIZ019:35
shivm_Ok, it's there.19:35
tpbTitle: Flashing Firmware · timvideos/HDMI2USB Wiki · GitHub (at
CarlFKhdmi2usb-mode-switch -v --flash-image <your image filename>19:38
shivm_Error: Unknown flash device (ID 0x00ffffff)19:39
CarlFKhey mithro19:41
mithroI'm around now and have about 20 minutes to help out before I have to get to work on a presentation for work19:41
tpbTitle: Ubuntu Pastebin (at
nancythis is the output of make gateware command19:42
CarlFKmithro: nancy is building, she sends me a .bin, I flash it to the Opsis, and I can't connect a term (that is the only thing I hav etested..)19:42
CarlFKmithro: shivm can build, sends me bin, I flash to my Atlys, it works.  *yay*19:43
mithroThat target is nacy building for?19:43
mithroI can't tell from the pastebin...19:44
nancyI am trying to build image-gateware+bios+firmware.bin the file19:45
shivm_mithro: Hi, I am doing some research about generic debug system issue and posted some queries on this thread:!topic/timvideos/E6XkGPLOVbk. Can you please take a look at it?19:45
tpbTitle: Google Groups (at
CarlFKmithro: i hope so.  should this confirm?  strings image-gateware+bios+firmware-nancy-v1.bin | grep opsis19:46
CarlFK opsis_eeprom: NACK while writing opsis_eeprom address!19:46
mithroshivm_: I'll try to, I managed to reply to some of the emails to the list yesterday19:46
shivm_mithro: thanks a lot!!19:48
mithroshivm_: What is your email address? Since you are here in the channel, I'll try and reply to yours first19:49
mithroshivm_: Have you been talking to the opendebugsoc people at all?19:50
nancysir , what could be the bugs19:52
nancyi have set up the environment twice and did make gateware19:52
nancy.bin doesnt work Atlys19:53
mithronancy: Can you explain to me the exact steps you did to set up the environment and build the gateware? Assume I know *nothing* about how to set up the gateware19:53
nancy I installed the Xilinix ISE with license in default directory /opt/xilinix19:54
mithronancy: /opt/xilinix or /opt/Xilinx case and spelling matter...19:56
nancyand than i ran the bootsrtrap script19:56
nancycurl -fsS | bash19:56
nancyi could the output as given in getting strted.md19:56
shivm_mithro: my email address: [email protected]19:57
nancythan i went to the directory of scripts and ran sudo bash env-env.sh19:57
mithronancy: Couple of quick tips - when you are experiencing bugs the small details **really** matter19:59
nancyexport PLATFORM =opsis,  cd HDMI2USB-litex-firmware , source scripts/enter-env.sh19:59
mithronancy: so please make sure what you are saying here is *exactly* what you did -- all the way down to the white space20:00
shivm_mithro: I am still reading about adv_dbg_sys and have not talked to opendebugsoc people.20:00
mithronancy: This is why we ask people to use a pastebin, because it is much harder to accidentally say something different to what you actually did or saw20:01
mithronancy: Something like a single space at the wrong point can *drastically* affect the results you get20:03
nancyyes sir20:04
nancyi get it20:04
mithronancy: Great! So can you answer my previous questions with that in mind?20:05
nancy i am sorry sir, i thing i have commit some step wrong , while installing . i will try again from step 120:10
CarlFKnancy: what did you find?20:11
*** nancy_ has joined #timvideos20:17
*** nancy has quit IRC20:17
nancy_sir , i followed same steps but after make gateware command i ran make image command20:17
shivm_I am getting same error while loading firmware: Any idea?20:22
tpbTitle: Ubuntu Pastebin (at
CarlFKnancy_: please pastebin whatever is in your terminal buffer right now so we can see what is going on20:24
shivm_CarlFK: this issue is also listed here:
tpbTitle: Detect "Error: Unknown flash device (ID 0x00ffffff)" on Digilent Atlys · Issue #24 · timvideos/HDMI2USB-mode-switch · GitHub (at
CarlFKshivm: yay!! that is progress20:26
CarlFKshivm: there may be notes on how to fix this.. we have a few pages like this
tpbTitle: HDMI2USB-litex-firmware/ at master · timvideos/HDMI2USB-litex-firmware · GitHub (at
tpbTitle: Ubuntu Pastebin (at
mithrosorry, I was just replying to shivm_'20:30
mithrosorry, I was just replying to shivm_'s email20:30
mithroCarlFK: I know what shivm_'s issue is20:31
mithroCarlFK: You are correct it is this iss ->
tpbTitle: Detect "Error: Unknown flash device (ID 0x00ffffff)" on Digilent Atlys · Issue #24 · timvideos/HDMI2USB-mode-switch · GitHub (at
mithroDammit past Tim, why didn't you link to the special instructions for the work around :-(20:32
tpbTitle: Digilent Atlys Board: Jumpers and Video Connections · timvideos/HDMI2USB Wiki · GitHub (at
mithroJP11 is used to disable the FPGA from loading the configuration from the onboard SPI flash. This header may not be populated on your board. During normal operation this should not have a jumper on it. This jumper is only needed when loading the HDMI2USB firmware onto the board permanently for the first time. From the Digilent Atlys manual;20:36
mithroAn on-board mode jumper (JP11) selects between JTAG/USB and ROM programming modes. If JP11 is not loaded, the FPGA will automatically configure itself from the ROM. If JP11 is loaded, the FPGA will remain idle after power-on until configured from the JTAG or Serial programming port.20:36
tpbTitle: digilent_atlys [Generation of Reconfigurable Circuits from Machine Code] (at
shivm_mithro: I did that too. Are there any other configuration jumpers on the board?20:37
mithroshivm_: The steps are very fiddly and it's very important to do them in the right order20:37
shivm_mithro: My .bin image is working on the board tested by Carl.20:38
mithroshivm_: Great!20:39
CarlFKshivm: in case it isn't clear: Atlys board ships with firmware.  to update it, you need to fiddle with jumpers20:39
shivm_Ya I know.20:39
CarlFKonce you get our firmware on it, you no longer need to touch the jumpers20:39
CarlFKI forgot all about that20:40
mithroshivm_: The issue is that the Atlys ships with firmware which *leaves* the flash into QSPI mode -- when the flash is in QSPI mode openocd / hdmi2usb-mode-switch is unable to program the flash20:40
shivm_So, how can I disable QPSI mode other than using a header on JUMPER J1120:42
mithronancy_: So you ran "make gateware" to get the output in the above pastebin?20:42
mithronancy_: How did you get the image-gateware+bios+firmware-nancy-v1.bin file?20:43
nancy_make gateware  and after that make image20:43
mithronancy_: Can you give me the output of the "make image" command and the output of "git describe; git status"?20:44
mithroshivm_: The only way I know is putting a jumper on J11, powering on the board, removing the jumper from J11 while the power is still connected, then programming using mode-switch without power cycling, **then** power cycle...20:45
*** nancy has joined #timvideos20:47
mithroshivm_: The process is pretty frustrating to get right20:47
*** nancy_ has quit IRC20:47
shivm_thanks. Then I will try again later with another board.20:47
mithronancy_: Can you give me the output of the "make image" command and the output of "git describe; git status"?20:47
mithroshivm_: I'm pretty sure this will work on your current board20:48
*** shivm_ has quit IRC20:48
*** shivm has quit IRC20:48
*** Samishamshi has joined #timvideos20:48
nancy   output of make image20:49
tpbTitle: Ubuntu Pastebin (at
*** samishamshi_ has joined #timvideos20:50
mithronancy: That all looks good to me!20:51
mithroCarlFK: Can you also test the latest hdmi2usb builds from last night? They are probably broken - but if they are not then they should have the edid fix -- archive/master/v0.0.4-190-gcacbf4020:53
*** akhil_singh has joined #timvideos20:53
*** Samishamshi has quit IRC20:53
mithroCarlFK: FYI - You should **never** put the same name in the "Tag Name" column twice in the spreadsheet here ->
tpbTitle: HDMI2USB Firmware Versions - Google Sheets (at
nancyyes sir this is all i build output of git describe; git status20:55
tpbTitle: Ubuntu Pastebin (at
*** samishamshi_ has quit IRC20:55
akhil_singhmithro: Is there any memory constraint on how much we can use for bit error tester on FPGA.20:56
akhil_singhmithro: because if not then we can store all patterns we generate and do error checking.20:57
akhil_singhmithro: or if there is some memory constraint then we will need to do some form of compression for the same.20:58
mithroakhil_singh: I don't think the bit patterns will be that big?21:00
*** nancy_ has joined #timvideos21:00
mithronancy: That all looks good!21:01
*** nancy has quit IRC21:02
akhil_singhmithro: what should be the number of bits for each pattern ?21:02
mithroakhil_singh: Most patterns should be repeating right?21:02
akhil_singhI am thinking of random which will generate from 0 to 2^(n-1) but not in a particular sequence.21:03
mithroakhil_singh: I want you to think about why different patterns would be useful -- what is the pattern trying to test21:04
akhil_singhmithro: Because how will one know which patterns are important for testing. We need to check for every possible pattern that trans receiver are working fine. Right ?21:05
mithroakhil_singh: Do you know much about unit testing / test coverage?21:07
mithroakhil_singh: The key to good testing is coming up with tests which represent what you actually want to do with the code or device21:07
akhil_singhmithro: I have knowledge of Design Verification and Test of Digital VLSI Designs.21:10
akhil_singhmithro: Can you tell what we are trying to do with these high speed trans receivers so that i can think of what patterns might be useful for that case ?21:11
mithroakhil_singh: You should try and figure out what we could possibly want first, it is much quicker for me to check a list then try and enumerate them from scratch21:13
mithroakhil_singh: I would probably start by looking at what Xilinx advertises the GTPs being useful for21:15
akhil_singhmithro: Okay i will look at that and will update you what patterns might be useful.21:19
mithrointeresting looking device...
mithroGot to run - see ya!21:29
*** akhil_singh has quit IRC21:35
CarlFKnancy_: I am back, so let me know if you have a new image for me to test21:52
CarlFKnancy_: also, I'll set up an ftp server that you can send files to, that might be easier than email21:53
*** techman83 has quit IRC22:03
nancy_yes sir22:04
nancy_@CarlFK should i mail you22:04
*** techman83 has joined #timvideos22:04
*** ChanServ sets mode: +v techman8322:04
CarlFKnancy_: no - give me a min22:04
CarlFKnancy_: well, yes.  but I may have someting else soon22:04
nancy_sir mithro sir checked it and he said everything is fine !22:05
mithronancy_: I more said that everything looks good and I can't see any errors at first inspection :-P22:18
mithrobblr - Have to concentrate on this presentation....22:20
*** shorne has quit IRC22:38
*** shorne has joined #timvideos22:39
*** NIIX0Bavelardi has joined #timvideos23:26

Generated by 2.13.1 by Marius Gedminas - find it at!