*** tpb has joined #timvideos | 00:00 | |
shorne | mithro: I am working on it :) I got my gdb and qemu patches done | 00:41 |
---|---|---|
shorne | Right now I am working on vivado (maybe not needed yet) but just need to get to it | 00:42 |
mithro | shorne: awesome! I'll be around in like an hour for most of the afternoon | 00:42 |
mithro | Vivado is needed for the Arty | 00:42 |
shorne | I had a read through your google doc, I think I can complete that tonight | 00:43 |
shorne | today there are a few family things to do, and all this week work | 00:43 |
mithro | ISE is needed for MimasV2 | 00:43 |
shorne | So goal, by tonight have it running in qemu | 00:43 |
mithro | It should only take a few hours to get set up unless you hit issues | 00:44 |
mithro | Qemu doesn't need ISE or Vivado | 00:45 |
shorne | right, I noticed that there was no spartan-6 support in the vivado I downloaded | 00:45 |
mithro | Yeah :-/ | 00:45 |
shorne | which is a bit unfortunate after a 14gb download | 00:46 |
mithro | The mimasv2 doesn't quite work yet | 00:46 |
shorne | Yeah, but I figured I would blink and LED or something first to understand the workflow on these boards | 00:47 |
mithro | At the moment the DDR controller needs the core to run at 83mhz and ISE has trouble routing things on the lx9 at that speed | 00:48 |
mithro | Works fine at 50mhz but then the DDR isn't working correctly it seems.. | 00:49 |
mithro | Need to chat with _florent_ about the fix for that | 00:49 |
shorne | what is the lx9? | 00:50 |
mithro | Spartan 6 LX9 - means it has about 9000 cells | 00:50 |
shorne | maybe we need multi-clock domains with some cross clock domain fifos? | 00:51 |
shorne | I see right I remember you saying that the higher clock rate doesnt route | 00:51 |
mithro | The Opsis and Atlys have LX45 (45000 cells) | 00:51 |
mithro | The Arty is an Artix-7 35T which means it has ~35,000 cells | 00:52 |
shorne | I see, those numbers make sense then | 00:54 |
shorne | 9,45,35 :) | 00:54 |
mithro | They are generous in rounding up | 00:55 |
mithro | And the Spartan 6 cells and a bit different to the Series 7 cells | 00:55 |
mithro | S/and/are/ | 00:55 |
shorne | well, I am getting started.. now family calling | 00:57 |
mithro | But yeah, generally the Xilinx parts have the cell counts in the part name | 00:57 |
mithro | No worries, have fun! | 00:57 |
shorne | right, I never noticed any logic to the numbers on the altera parts | 00:57 |
shorne | mithro: for the first part litex-firmware | 00:59 |
shorne | I find these branches | 01:00 |
shorne | remotes/origin/or1k-debug | 01:00 |
shorne | remotes/origin/or1k-fixing | 01:00 |
shorne | or1k-fixing being the latest | 01:00 |
mithro | Spartan -> Artix -> Kintex -> Virtex -> Ultrascale | 01:00 |
mithro | shorne: check my email about the Arty? | 01:01 |
mithro | On my phone, so can't look easily | 01:02 |
mithro | Will be at a computer in about 30 minutes | 01:03 |
mithro | https://github.com/mithro/HDMI2USB-litex-firmware/tree/merge-arty-soc?files=1 | 01:07 |
tpb | Title: HDMI2USB-litex-firmware/ at merge-arty-soc · mithro/HDMI2USB-litex-firmware · GitHub (at github.com) | 01:07 |
mithro | https://groups.google.com/forum/m/#!topic/linux-litex/CzGITacfEtY | 01:09 |
tpb | Title: Google Groups (at groups.google.com) | 01:09 |
*** sb0 has quit IRC | 01:12 | |
mithro | shorne: I'm at a computer now, but I assume you have probably gone to do stuff with your family? | 01:22 |
*** shorne_ has joined #timvideos | 01:33 | |
shorne_ | mithro: yes... hmm somehow my connection reset | 01:33 |
shorne_ | I need to be away | 01:33 |
mithro | shorne_: No worries, maybe I'll have stuff merged by the time you get back... | 01:34 |
*** shorne has quit IRC | 01:35 | |
*** andi-m has quit IRC | 01:57 | |
mithro | shorne_: Gah - it turns out that I had not merged a bunch of patches - doing now | 01:59 |
*** andi-m has joined #timvideos | 02:14 | |
*** andi-m has quit IRC | 02:20 | |
*** andi-m has joined #timvideos | 02:23 | |
*** miselin has quit IRC | 04:32 | |
mithro | Woot! The merge builds! | 04:42 |
mithro | CarlFK: I'm about to do a super huge merge - can you do some testing after I do the merge? | 05:00 |
CarlFK | mithro: I only have Atlys (and I am in Los Gatos, CA right now, fly home in 12 hours ) | 05:01 |
*** thaytan has quit IRC | 05:01 | |
CarlFK | tumbleweed has my Opsis. I think he is back 'here' in a few days | 05:02 |
mithro | I think he arrives like 30 minutes before I do | 05:03 |
*** thaytan has joined #timvideos | 05:04 | |
*** ChanServ sets mode: +v thaytan | 05:04 | |
*** thaytan has quit IRC | 05:39 | |
*** thaytan has joined #timvideos | 05:39 | |
*** ChanServ sets mode: +v thaytan | 05:39 | |
*** andi-m has quit IRC | 05:45 | |
*** andi-m has joined #timvideos | 05:51 | |
mithro | CarlFK: Will you have time to test when you get back home? | 05:56 |
CarlFK[m] | mithro (IRC): yes - I have an atlys and desktop sitting near each other, might even be plugged in, ready to go | 05:58 |
*** rohitksingh has joined #timvideos | 06:47 | |
shorne_ | running download-env.sh | 07:29 |
shorne_ | interesting, this really is a whole environment with python and everything | 07:30 |
shorne_ | mithro: ok, for the first issue with fxload missing | 07:31 |
mithro | shorne_: You most definitely want to update to what I pushed an hour or two ago... | 07:32 |
mithro | shorne_: BTW - I'm hoping to merge that branch in the next couple of hours | 07:32 |
shorne_ | mithro: | 07:33 |
shorne_ | +echo | 07:33 |
shorne_ | +echo "Installing fxload (tool for Cypress FX2)" | 07:33 |
shorne_ | +# conda install fxload | 07:33 |
shorne_ | check_exists fxload | 07:33 |
shorne_ | why is "conda install fxload' commented out? | 07:33 |
shorne_ | I updated to the latest | 07:34 |
shorne_ | IM going to uncomment that and try again | 07:34 |
shorne_ | I dont need cypress but might as well | 07:34 |
mithro | shorne_: I think because we don't have a conda package for fxload yet.... | 07:34 |
shorne_ | ah, Ill see, ill just comment out the check | 07:34 |
shorne_ | I am assuming I dont need it for qemu firmware build | 07:35 |
mithro | Nope | 07:35 |
mithro | shorne_: ./scripts/download-env-root.sh <-- that installs fxload via apt-get | 07:36 |
shorne_ | I see | 07:36 |
shorne_ | Ill put a note in the docs on that | 07:36 |
mithro | shorne_: Do you have edit access? | 07:37 |
shorne_ | I can comment | 07:37 |
shorne_ | ok, i need the ftdi lib/driver too ill run the root env thing... but I run Fedora :) | 07:40 |
shorne_ | let me convert this to dnf | 07:41 |
mithro | dnf? | 07:41 |
mithro | That libftdi stuff should be fixed soon -- I think I have figure out how to make openocd statically link against libftdi and libusb | 07:41 |
shorne_ | well, yum (they call it dnf in fedora now) | 07:41 |
shorne_ | the package manager | 07:42 |
shorne_ | apt-get equivalent | 07:42 |
shorne_ | hmm, the ftdi stuff doesnt match fedora 'looking for libftdi.so.1', but fedora has libftdi1.so | 07:45 |
shorne_ | let me see what it really needs | 07:45 |
mithro | Oh, I think we had instructions for that somewhere.... | 07:45 |
shorne_ | I am trying `sudo ln -s /usr/lib64/libftdi1.so /usr/lib64/libftdi.so.1` it seems to work | 07:55 |
mithro | shorne_: You could actually do a "ln -s /usr/lib64/libftdi1.so build/conda/lib/libftdi.so.1" probably | 07:57 |
shorne_ | ah, just into the local path | 08:05 |
shorne_ | right, it seems to work either way | 08:05 |
mithro | The environment tries very hard to get everyone *exactly* the same environment... | 08:06 |
shorne_ | mithro: environment seems to be loaded | 08:10 |
mithro | shorne_: Awesome! | 08:10 |
shorne_ | running build-qemu | 08:10 |
mithro | Hopefully that works - I haven't run it in the last couple of hours :-P | 08:12 |
shorne_ | yeah. I guess running on fedora I am in a bit of a no mans land? | 08:12 |
shorne_ | I should probably try to run this on docker | 08:13 |
mithro | A bit, but once you have the enviroment it shouldn't be all that dependent on the outside environment | 08:13 |
mithro | I guess qemu is possible the one case where it would be... | 08:13 |
*** anant3110 has joined #timvideos | 08:14 | |
anant3110 | Hi, I am just starting out and I love the work being done in your organisation specially the HDMI2USB project. I know basic to intermediate verilog programming, java, python and C++. Is there anyway I can contribute? | 08:19 |
mithro | anant3110: Hello! | 08:19 |
mithro | anant3110: There are plenty of things to contribute too, how much time do you have? :-P | 08:20 |
anant3110 | I can give 7-12 hours per week. | 08:21 |
mithro | anant3110: Cool, were in the world are you located? I'm guessing maybe India? | 08:22 |
anant3110 | Yes! | 08:24 |
mithro | anant3110: So, what part interests you more? You interested in doing more stuff running on hardware side of things - or more stuff running on the computer? | 08:25 |
anant3110 | I would be much more interested in the hardware part as long as I get the resources easily. My institute does have a lot of FPGA models but I will have to check if your requirements match with them. If that doesn't work out, I would be okay to do stuff running on the computer. | 08:28 |
mithro | anant3110: The boards you should look out for are "Digilent Arty" and "Digilent Atlys" | 08:32 |
anant3110 | I m pretty sure I have seen a Digilent Atlys in my lab. But I m not sure about Arty. | 08:35 |
mithro | shorne_: Hows that build going? | 08:36 |
shorne_ | mithro: sorry, had to be away | 08:38 |
mithro | anant3110: You can get the build environment setup before having hardware... | 08:38 |
shorne_ | I got an error 'ERROR: configure test passed without -Werror but failed with -Werror.' ... | 08:39 |
shorne_ | no time to loo kat right now | 08:39 |
shorne_ | on my way out again | 08:39 |
mithro | shorne_: Okay | 08:39 |
shorne_ | (that was on qemu) | 08:42 |
anant3110 | mithro: That would be cool. Can you give me a link to the code base? If not this : https://github.com/timvideos/HDMI2USB-litex-firmware | 08:45 |
tpb | Title: GitHub - timvideos/HDMI2USB-litex-firmware: A version of the HDMI2USB firmware based around LiteX tools produced by @Enjoy-Digital (based on misoc+migen created by @M-Labs) (at github.com) | 08:45 |
mithro | anant3110: Yes - that is the correct thing to start with, there should be a "getting-started.md" file in the repo | 08:46 |
mithro | shorne_: Will test here | 08:46 |
anant3110 | mithro: I ll get back to you once I m done! | 08:48 |
mithro | anant3110: I might be in bed by then, but I should be around again tomorrow | 08:49 |
anant3110 | mithro: Sure! | 08:50 |
*** anant3110 has quit IRC | 09:48 | |
*** tsglove has quit IRC | 10:28 | |
*** rohitksingh has quit IRC | 10:44 | |
*** rohitksingh has joined #timvideos | 10:48 | |
jimmo | @mithro just compiling linux now. qemu worked with gcc7 after i disabled Werror | 11:29 |
mithro | jimmo: Ahh cool | 11:29 |
jimmo | building a buildroot | 11:31 |
jimmo | @mithro do you have instructions for that? | 11:31 |
mithro | jimmo: No - I can | 11:32 |
mithro | jimmo: No, I can't remember how I created mine... | 11:32 |
mithro | I'm pretty sure it didn't work either, I ended up having one from Joel | 11:32 |
mithro | jimmo: https://drive.google.com/file/d/0B5VlNZ_Rvdw6d21LWXdHQlZuOVU/view?usp=sharing <-- shenki-openrisc-roofs.cpio | 11:42 |
tpb | Title: shenki-openrisc-rootfs.cpio - Google Drive (at drive.google.com) | 11:42 |
mithro | jimmo: That is the one I have been using... | 11:43 |
mithro | jimmo: Any luck? | 11:58 |
jimmo | i gziped that file and put it in the right place i think | 11:59 |
jimmo | now it's just compiling | 11:59 |
jimmo | just failed | 11:59 |
jimmo | or1k-elf-ld: cannot find libgcc.a: No such file or directory | 11:59 |
mithro | jimmo: try "conda install gcc-or1k-elf-newlib" ? | 12:00 |
jimmo | conda was more fun when we were on google's internet | 12:01 |
jimmo | @mithro are you going to update scripts/download-env or should i? | 12:02 |
mithro | jimmo: xxx-newlib seems to be only needed for Linux / MicroPython firmware -- not the bare metal ones | 12:03 |
jimmo | what libc does hdmi2usb use? | 12:04 |
mithro | jimmo: None! \o/ :-P | 12:04 |
jimmo | i feel like download-env should grab it anyway? | 12:04 |
mithro | jimmo: See scripts/build-micropython.sh ? | 12:05 |
jimmo | hrmm | 12:08 |
jimmo | why not just install by default | 12:08 |
jimmo | CRC failed... | 12:08 |
jimmo | https://www.irccloud.com/pastebin/oQDjfBa8/ | 12:08 |
tpb | Title: Snippet | IRCCloud (at www.irccloud.com) | 12:08 |
mithro | jimmo: Need more further back then that... | 12:09 |
mithro | But I'm guessing that you are trying to use Linux on the flash with the mimasv2? | 12:09 |
jimmo | yes | 12:11 |
jimmo | https://www.irccloud.com/pastebin/wGkqUHK4/ | 12:11 |
tpb | Title: Snippet | IRCCloud (at www.irccloud.com) | 12:12 |
mithro | The mimasv2 flash size is too small to store the firmware... | 12:12 |
mithro | Remaining space -2174436 bytes (-16 Megabits, -2.07 Megabytes) | 12:13 |
mithro | I wonder why that didn't error.... | 12:13 |
jimmo | argh. ok trying minispartan6 now | 12:14 |
jimmo | nope. unsupported SPI flash chip | 12:16 |
mithro | Use the opsis target | 12:20 |
mithro | s/target/platform/ | 12:20 |
jimmo | ah yeah, looks like no SPI flash support for the arty either? | 12:25 |
mithro | jimmo: I've yet to get it working... | 12:25 |
jimmo | cool well not a login prompt, but linux is booting | 12:28 |
mithro | jimmo: Pastebin? | 12:28 |
jimmo | https://www.irccloud.com/pastebin/jqxvUnv1/ | 12:29 |
tpb | Title: Snippet | IRCCloud (at www.irccloud.com) | 12:29 |
mithro | jimmo: Yay! That is what happens on the hardware | 12:29 |
mithro | Will without the unexpected IRQ trap... | 12:30 |
jimmo | but your doc says it boots to login ? | 12:31 |
mithro | jimmo: Yeah - something has obviously broken since then... I blame shenki :-P | 12:31 |
mithro | jimmo: Try editting third_party/litex/litex/soc/integration/soc_core.py and replace move the uart to "interrupt_map" IRQ 2 | 12:32 |
mithro | jimmo: Are you using the base target or the net target? | 12:33 |
jimmo | base | 12:33 |
shorne_ | mithro: I fixed the qemu thing it was an issue with my CFLAGS | 12:34 |
shorne_ | I deleted a path earlier today it cause a issue with qemu configure | 12:34 |
mithro | Hey shorne_! | 12:34 |
shorne_ | nothing related to litex | 12:34 |
shorne_ | but.. my time is up today unfortunately | 12:35 |
shorne_ | looks like I got hup to qemu build | 12:35 |
mithro | I think jimmo might have accidently stumbled on something interesting... | 12:35 |
shorne_ | next will be linux tomorrow if I get time | 12:35 |
jimmo | changing the irq got rid of the irq 0 error, but otherwise identical | 12:35 |
mithro | jimmo: What did you rebuild after changing it? | 12:36 |
jimmo | scripts/build-qemu.sh | 12:36 |
jimmo | (i don't think i have built anything else?) | 12:36 |
mithro | jimmo: Damn, getting there here too... | 12:38 |
mithro | I wonder if Joel broke something when rebasing | 12:39 |
shorne_ | on qemu build I am getting | 12:41 |
shorne_ | block/blkverify.c:318:29: error: ‘%s’ directive output may be truncated | 12:41 |
mithro | shorne_: Ask jimmo about that | 12:42 |
jimmo | shorne_: | 12:42 |
jimmo | shorne_: gcc --version ? | 12:42 |
shorne_ | gcc (GCC) 7.2.1 20170915 (Red Hat 7.2.1-2) | 12:43 |
shorne_ | ah, maybe too new | 12:43 |
jimmo | same as me | 12:43 |
shorne_ | and qemu to old? | 12:43 |
mithro | Yeah | 12:43 |
shorne_ | ok, then ill look into it | 12:43 |
mithro | I started rebasing and splitting each feature into their own branch | 12:43 |
jimmo | or remove -Werror from configure:4752 | 12:43 |
mithro | shorne_: For now you can just disable the -Werror | 12:43 |
shorne_ | ok, I can do that too | 12:43 |
mithro | shorne_: I was doing the rebase stuff here -> https://github.com/mithro/qemu/branches/all | 12:44 |
tpb | Title: Branches · mithro/qemu · GitHub (at github.com) | 12:44 |
shorne_ | oh... it builds | 12:49 |
shorne_ | then I guess runs | 12:49 |
shorne_ | Executing booted program at 0x40000000 | 12:49 |
shorne_ | HDMI2USB firmware booting... | 12:49 |
shorne_ | opsis_eeprom: Init...finished. | 12:49 |
shorne_ | then a bunch of other stuff | 12:49 |
mithro | shorne_: You should get all the way to the prompt | 12:49 |
mithro | shorne_: Is this lm32 or or1k CPU? | 12:49 |
shorne_ | my env was or1k | 12:50 |
mithro | okay cool | 12:50 |
shorne_ | and yes all the way to the H2U propt | 12:50 |
mithro | shorne_: great! | 12:51 |
shorne_ | mithro: cool, on splitting features out to branches, I was splitting all of the openrisc backlock out to feature branches | 12:52 |
shorne_ | it made the work of upstreaming a bit more pragmatic | 12:52 |
shorne_ | I tried to document what I was doing here https://github.com/stffrdhrn/linux/wiki/commit-batches | 12:53 |
tpb | Title: commit batches · stffrdhrn/linux Wiki · GitHub (at github.com) | 12:53 |
shorne_ | but... sorry, need to go good night, finally :) | 12:54 |
mithro | shorne_: Thanks! | 12:55 |
mithro | jimmo: So, you going to try get hardware Linux? | 12:57 |
jimmo | @mithro - that's only possible on the arty? | 13:08 |
jimmo | i'll have to install vivado then | 13:08 |
mithro | jimmo: Or the Opsis - which you have two of | 13:08 |
*** CarlFK has quit IRC | 13:13 | |
jimmo | @mithro i'll leave vivado downloading and opsis gateware building | 13:20 |
mithro | Opsis gateware should only take a couple of minutes | 13:20 |
jimmo | @mithro the opsis has no programmer firmware right? because the fx2 gets its firmware as part of the programming process? | 13:37 |
mithro | jimmo: More complicated than that - but you can think of it like that for now | 13:38 |
jimmo | @mithro so i don't need to do anything on those two boards - i should just be able to gateware-load or gateware-flash? | 13:39 |
mithro | Yeah | 13:40 |
jimmo | cool ok i will try tomorrow | 13:40 |
mithro | How did it get to almost 1:00am :-( | 13:42 |
*** CarlFK has joined #timvideos | 14:20 | |
*** ChanServ sets mode: +v CarlFK | 14:20 | |
*** CarlFK has quit IRC | 14:48 | |
*** rohitksingh has quit IRC | 17:11 | |
*** sb0 has joined #timvideos | 17:34 |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!