sb0 changed the topic of #m-labs to: ARTIQ, Migen, MiSoC, Mixxeo & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs
futarisIRCcloud has joined #m-labs
<GitHub82> [smoltcp] whitequark pushed 1 new commit to master: https://github.com/m-labs/smoltcp/commit/1d2ec8c13562b602fa9b121b9b2cc80a41cc6520
<GitHub82> smoltcp/master 1d2ec8c whitequark: Remove length check that is redundant after 181083f1.
<travis-ci> m-labs/smoltcp#667 (master - 1d2ec8c : whitequark): The build was broken.
<GitHub16> [artiq] sbourdeauducq closed issue #914: misoc asyncserial https://github.com/m-labs/artiq/issues/914
<GitHub147> [artiq] sbourdeauducq commented on issue #914: Yes, install asyncserial manually if you want to install misoc this way. https://github.com/m-labs/artiq/issues/914#issuecomment-362123832
<GitHub194> [artiq] sbourdeauducq commented on issue #911: why not use conda? https://github.com/m-labs/artiq/issues/911#issuecomment-362123947
<GitHub162> [smoltcp] whitequark pushed 1 new commit to master: https://github.com/m-labs/smoltcp/commit/d2b8f526f3fd5e9bd439c7baf16aca1ffe41d617
<GitHub162> smoltcp/master d2b8f52 whitequark: Remove internally inconsistent test that was left in 1d2ec8c.
<GitHub107> [artiq] whitequark commented on issue #914: @sbourdeauducq So why wouldn't we publish asyncserial on pypi?.. https://github.com/m-labs/artiq/issues/914#issuecomment-362124042
<GitHub117> [smoltcp] dlrobertson opened pull request #141: [WIP] Use the time types instead of a u64 (master...add_timestamp) https://github.com/m-labs/smoltcp/pull/141
<GitHub55> [artiq] sbourdeauducq commented on issue #407: Okay, so this is most likely the AV and #733 should address this. https://github.com/m-labs/artiq/issues/407#issuecomment-362133843
<GitHub131> [artiq] sbourdeauducq closed issue #407: Long dead time between kernels on Windows https://github.com/m-labs/artiq/issues/407
<GitHub72> [artiq] sbourdeauducq commented on issue #854: @marmeladapk @gkasprow Any progress? This is currently one of the most important issues with Sayma. https://github.com/m-labs/artiq/issues/854#issuecomment-362134978
<GitHub12> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: It is ambiguous what `length` is. Is length the length of the buffer, the length field, or the result of `Packet::data_len`? I think it should be the length field, but I could be swayed otherwise. https://github.com/m-labs/smoltcp/pull/139#discussion_r165238242
<GitHub135> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: Misused if the input to `OPTIONS` is the result of `self.data_len`. Contradicts the use of `OPTIONS` below. https://github.com/m-labs/smoltcp/pull/139#discussion_r165239399
<GitHub108> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: `Packet` just to be in line with other implementations in `wire`. https://github.com/m-labs/smoltcp/pull/139#discussion_r165242598
<GitHub159> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: Input here is the result of `self.data_len`. Contradicts the above use of `OPTIONS`. https://github.com/m-labs/smoltcp/pull/139#discussion_r165239506
<GitHub107> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: Same https://github.com/m-labs/smoltcp/pull/139#discussion_r165239952
<GitHub95> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: We can't make the assumption that we're the last header in the buffer. https://github.com/m-labs/smoltcp/pull/139#discussion_r165238505
<GitHub78> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: `Packet as Ipv6HopByHopPacket` https://github.com/m-labs/smoltcp/pull/139#discussion_r165242545
<GitHub58> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: `len < field::LENGTH`. Line 77 could result in an out of bounds access. https://github.com/m-labs/smoltcp/pull/139#discussion_r165236807
<GitHub19> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: Same https://github.com/m-labs/smoltcp/pull/139#discussion_r165240348
<GitHub88> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: The data contained in the `Hop-by-Hop` extension header is `Ipv6Option`s. It isn't like an IPv6 packet where the data could be anything. I think either there should be no `data` member, only `length` and `next_header` which would be `emit`ed to a buffer then the IPv6 Options get `emit`ed to the `Hop-by-Hop` `Packet` `data_mut` buffer. https://github
<GitHub91> [smoltcp] hjr3 commented on pull request #139 2dbdeb4: Ah, I changed this based on feedback from the last PR. It should be using `hdr_ext_len()` function for the number of bytes. Concerning that my tests passed. I will fix this and make the variable in `field::OPTIONS()` more explicit. https://github.com/m-labs/smoltcp/pull/139#discussion_r165245496
<sb0> whitequark, ping
<whitequark> sb0: pong
<sb0> did you get my emails?
<whitequark> yes
<whitequark> those issues will take weeks if not months, I think.
<whitequark> the device-assisted scheduling is especially bad
<sb0> can you send me individual estimates for each issue?
<sb0> and it's not something we have to develop right now, it's more long-term planning
<GitHub155> [smoltcp] hjr3 commented on pull request #139 2dbdeb4: I think it has to be `len <= field::LENGTH` because the header _must_ be at least 8 octets long per https://tools.ietf.org/html/rfc8200#section-4.3 https://github.com/m-labs/smoltcp/pull/139#discussion_r165247309
<GitHub55> [smoltcp] hjr3 commented on pull request #139 2dbdeb4: Makes sense. I will check that it is at least 8 octets long and otherwise return Malformed. https://github.com/m-labs/smoltcp/pull/139#discussion_r165247369
<whitequark> okay
<whitequark> I'll take a closer look at them then
<GitHub35> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: Good point, so why not add `MIN_HEADER_SIZE = 8`, then change this to `len < field::MIN_HEADER_SIZE` https://github.com/m-labs/smoltcp/pull/139#discussion_r165247889
<sb0> whitequark, what about the rtm loading and the camera driver?
<whitequark> sb0: no progress yet, I've been feeling sick over the last few days
rohitksingh_work has joined #m-labs
rohitksingh_wor1 has joined #m-labs
rohitksingh_work has quit [Ping timeout: 240 seconds]
<kraza> I'm sorry. I had to go to bed.
kraza has quit [Quit: Page closed]
kraza has joined #m-labs
rohitksingh-demo has joined #m-labs
rohitksingh-demo has quit [Quit: Leaving.]
<sb0> kraza, hi
<kraza> hello
<kraza> I'm sorry for disappearing last night. Had to go to bed, basically.
ohsix has quit [Ping timeout: 240 seconds]
bb-m-labs has quit [Ping timeout: 256 seconds]
bb-m-labs has joined #m-labs
<GitHub31> [artiq] hartytp commented on issue #854: Agreed, this issue has been outstanding far too long now. @gkasprow let's try to really focus on this until it's properly resolved. https://github.com/m-labs/artiq/issues/854#issuecomment-362204587
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
rohitksingh-demo has joined #m-labs
<sb0> _florent_, if the data is corrupted, could there be some issue with the gtp equalizer parameters? are those dependent on line rate?
<sb0> and clock synchronizer failure shuo
<sb0> ld be straightfoward to reproduce outside of artiq ...
<sb0> note that the kasli has the si5324 reset line disconnected from the fpga, so you can load a first design that sets up the si5324, and then your gtp test designs
rohitksingh-demo has quit [Ping timeout: 276 seconds]
<GitHub159> [artiq] sbourdeauducq commented on issue #914: If we put packages on pypi, we have to ensure that they are maintained and properly versioned there as well, otherwise there will be incorrect versions installed silently, and corresponding support issues. And pypi/setuptools don't have good ways of dealing with version dependencies, unlike e.g. cargo. So I think it is better to only support conda for packaging, which i
<sb0> _florent_, how do you program the si5324 for your tests?
<_florent_> sb0: i'm not using the si5324 in my design
<rjo> sb0: we should expand the "Users and contributors" list on the webpage by PTB, Uni Hannover, HU Berlin, BMBF, SYSU.
<rjo> sb0: and that page should mention the Kasli and EEM components.
rohitksingh-demo has joined #m-labs
rohitksingh-demo has quit [Client Quit]
<sb0> _florent_, so how are you getting a 150MHz clock?
<sb0> rjo, yes, I have a website redesign that I haven't pushed yet...
<sb0> (nor completed)
<rjo> sb0: ack. are you planning to go away from raw html?
<sb0> _florent_, would that help you debug remotely? https://hastebin.com/votesejeji.py
<rjo> IMHO it was a great idea to route the I2C bus to one FTDI port.really good for development.
<sb0> _florent_, I used that to dispatch bugs in the ethernet core gtp garbage (with the board in Berlin, I didn't have the Kasli back then), and that was pretty smooth. having the board locally would not have helped much.
<sb0> _florent_, oh and yes, you can use rjo's script to reconfigure the si5324 over USB
kuldeep_ has joined #m-labs
<rjo> sb0: could you patch the capacitors on the XADC VREFP so that we can compare temperatures with an without heat sink?
<rjo> sb0: on kasli
kuldeep has quit [Ping timeout: 276 seconds]
<sb0> rjo, will do
<rjo> sb0: they are 0201 but that's actually one of the easiest reworks i have done in a long time.
<rjo> sb0: thanks
<sb0> I have no problem with soldering 0201 but I don't have any in stock atm
<sb0> rjo, oh, it's just shorting C49, correct?
<sb0> rjo, re. website, probably a horrible NIH python-based hack. I haven't been convinced by the few frameworks I looked at
<sb0> found them too inflexible and it would take me more time to get them to do what I want than hack together a site generator script from scratch...
<sb0> jekyll is good if you're running a blog, but it's annoying for anything else
ohsix has joined #m-labs
<rjo> sb0: yes. C49.
<rjo> sb0: pandoc might be for you then. it doesn't force you into the blog corner.
X-Scale has quit [Read error: Connection reset by peer]
<rjo> whitequark: i checked that lit and checkoutput are the same as before. something else changed.
rohitksingh_wor1 has quit [Read error: Connection reset by peer]
rohitksingh-demo has joined #m-labs
Rex0r has joined #m-labs
X-Scale has joined #m-labs
<sb0> _florent_, can you work on jesd sc1 until your kasli arrives?
<_florent_> sb0: yes i started looking at the documentation, you had some plans about reading dac informations at startup for delay characterization, can you remember me what as it?
<rjo> sb0: are you sending one to him or is this from technosystem/greg?
<sb0> rjo, Tom is lending one
<sb0> _florent_, have you looked at the irc log?
<_florent_> sb0: yes, i'll have another look
<GitHub11> [smoltcp] hjr3 commented on pull request #139 2dbdeb4: This makes sense. I assume `buffer_len` should still return the total length of the header (including options data)? https://github.com/m-labs/smoltcp/pull/139#discussion_r165363965
<sb0> _florent_, so the idea is, if s/h is not met, then the dac should sometimes report synchronization slips
<sb0> at startup the firmware should scan through the phases and print the results
<sb0> then we put in the source a value that is in the middle of two areas where there are slips.
<_florent_> "scan through the phases": ie configure the delay on the hmc830 ouptut?
<GitHub114> [smoltcp] dlrobertson commented on pull request #139 2dbdeb4: Yes, that is correct. The `buffer_len` should be the minimum size buffer needed to hold the entire header. https://github.com/m-labs/smoltcp/pull/139#discussion_r165366416
<GitHub177> [smoltcp] hjr3 commented on pull request #139 92c168a: fixed https://github.com/m-labs/smoltcp/pull/139#discussion_r165369869
<GitHub88> [smoltcp] hjr3 commented on pull request #139 92c168a: i changed `OPTIONS` to use the value of `field::LENGTH`. i refactored the code a bit and was able to remove some complexity too. https://github.com/m-labs/smoltcp/pull/139#discussion_r165370179
<GitHub47> [smoltcp] hjr3 commented on pull request #139 92c168a: fixed https://github.com/m-labs/smoltcp/pull/139#discussion_r165370233
<GitHub159> [smoltcp] hjr3 commented on pull request #139 92c168a: fixed https://github.com/m-labs/smoltcp/pull/139#discussion_r165370341
<GitHub150> [smoltcp] hjr3 commented on pull request #139 92c168a: fixed https://github.com/m-labs/smoltcp/pull/139#discussion_r165370367
<GitHub19> [smoltcp] hjr3 commented on pull request #139 92c168a: fixed https://github.com/m-labs/smoltcp/pull/139#discussion_r165370390
<GitHub160> [smoltcp] hjr3 commented on pull request #139 92c168a: fixed https://github.com/m-labs/smoltcp/pull/139#discussion_r165370498
<GitHub128> [smoltcp] hjr3 commented on pull request #139 92c168a: fixed https://github.com/m-labs/smoltcp/pull/139#discussion_r165370524
<GitHub55> [smoltcp] hjr3 commented on pull request #139 92c168a: got it. data has been removed from the Repr https://github.com/m-labs/smoltcp/pull/139#discussion_r165370811
<sb0> _florent_, hmc7043
<sb0> but yes
<felix_> sb0: at least for the sdi stuff i'm currently working on the cdr parameters of the gtp receiver are different for the different line rates
<_florent_> sb0: yes sorry, i was meaning the hmc7043
<_florent_> sb0, felix: yes there are different cdr parameters (https://github.com/enjoy-digital/transceiver_test/blob/master/transceiver/gtp_7series.py#L166), but i think i'm using the right ones (at least are working with my other artix7 board)
<sb0> _florent_, those things can be fragile and working intermittently. are you using the value generated by the wizard for that line rate?
<felix_> is that parameter documented somewhere? i only saw some magic values being used and no real explaination of which bits do what
<sb0> no it's not. there are many things that are not documented in the xilinx transceivers ...
<_florent_> sb0: yes it should be the values from the wizard
<_florent_> sb0: but i'll verify that
<_florent_> sb0: and the values can change from one vivado version to another...
hozer has joined #m-labs
dlrobertson has joined #m-labs
rohitksingh-demo has quit [Quit: Leaving.]
awygle has joined #m-labs
mumptai has joined #m-labs
cr1901_modern has quit [Ping timeout: 240 seconds]
forrestv has quit [Excess Flood]
cr1901_modern has joined #m-labs
forrestv has joined #m-labs
cr1901_modern has quit [Ping timeout: 248 seconds]
<GitHub120> [smoltcp] dlrobertson commented on pull request #139 d006b48: Couldn't this just be `field::OPTIONS(self.length).end + 2`? https://github.com/m-labs/smoltcp/pull/139#discussion_r165474290
cr1901_modern has joined #m-labs
dlrobertson has quit [Ping timeout: 276 seconds]
dlrobertson has joined #m-labs
<GitHub175> [artiq] gkasprow commented on issue #854: first of all, I get valid link on my board with ARTIQ setup. Have you @sbourdeauducq tried on another board if link LED is on? https://github.com/m-labs/artiq/issues/854#issuecomment-362396204
<GitHub107> [artiq] hartytp commented on issue #854: Just throwing it out there, but is there a chance that SB's board is quite badly damaged? Could explain why he's seeing things like the 1V8 issue that aren't showing up on other boards. https://github.com/m-labs/artiq/issues/854#issuecomment-362397171
dlrobertson has quit [Read error: Connection reset by peer]
<GitHub148> [artiq] whitequark commented on issue #914: We already have pythonparser on pypi. https://github.com/m-labs/artiq/issues/914#issuecomment-362405446
<whitequark> sb0: re: website, I use nanoc, it's not blog-specific either
dlrobertson has joined #m-labs
mumptai has quit [Remote host closed the connection]
futarisIRCcloud has joined #m-labs
cr1901_modern has quit [Read error: Connection reset by peer]
cr1901_modern has joined #m-labs