<GitHub7>
[sinara] marmeladapk tagged Kasli/v1.1rc2 at master: https://git.io/vNPHs
<rohitksingh-demo>
sb0: Met Julius Baxter (author of mor1kx) finally just now. He tells that the mor1kx_fetch_tcm_prontoespresso module doesn't work. Its broken. And also that Cappuccino without caches *will* take atleast 3 cycles for instruction fetch. For Cappuccino he said that a new fetch module will need to be written for TCM specific purpose.
[X-Scale] has joined #m-labs
X-Scale has quit [Ping timeout: 256 seconds]
[X-Scale] is now known as X-Scale
<sb0>
_florent_, not yet, just received the kasli yesterday
<sb0>
_florent_, did you see my suggestion about jesd204 sc1?
<_florent_>
sb0: ok for drtio
<_florent_>
sb0: yes it seems fine, just need to read the documentation to see how to do it, i'll do that next week when i'll be at my office
<sb0>
_florent_, just tried drtio between sayma and kasi. nix.
<sb0>
you said there were initialization problems on the artix7 side. what are they?
<_florent_>
sb0: i think i need more information
<_florent_>
sb0: the problem was that sometime the bruteforce aligner was not able to lock
<_florent_>
sb0: i know that tx and rx fsm were always going to the end
<_florent_>
sb0: i'll do some tests next week between ultrascale and artix7 at 3gbps
kristianpaul has quit [Quit: Lost terminal]
<sb0>
_florent_, I don't have much information. just built and loaded the current artiq code, copper sfp between sayma (master) and kasli (satellite), no reaction.
<sb0>
I just changed the satellite code to use sfp2
<sb0>
clock aligner not working would cause such symptoms.
<sb0>
though it's surprising that the sayma side doesn't see anything either
<GitHub183>
[artiq] sbourdeauducq commented on issue #907: As I told you before, you need to replace things between brackets with the appropriate values for your situation, in your case kc705 instead of [board]. https://github.com/m-labs/artiq/issues/907#issuecomment-360781465
FabM has quit [Quit: ChatZilla 0.9.93 [Firefox 52.5.0/20171114221957]]
<cr1901_modern>
rjo: What's the easiest way to commit changes to openocd (Spartan 7 support so the JTAG ID is recognized)? Should I just funnel my changes to the m-labs fork and wait for you to commit them upstream?
sb0 has joined #m-labs
<rjo>
cr1901_modern: push them to the openocd review as well. and see them through yourself. i can still cherry pick them and carry them in the conda packages if that's useful.
<cr1901_modern>
rjo: Fair, and no need to cherry-pick; migen doesn't support Spartan 7 boards at present, but that will change soon (Arty S7 is working in Litex)
<rjo>
whitequark: are you working on the artiq_flash situation (windows in 3.2/3.3 and sayma amc/rtm)?
<GitHub189>
[artiq] sbourdeauducq commented on issue #906: The conda package will be built when JESD204 SC1 is done, at the latest. We don't need an issue for every single thing. And, in the present state of affairs, if you use Sayma you are generally expected to build from source and use git. https://github.com/m-labs/artiq/issues/906#issuecomment-360836963
<GitHub68>
[smoltcp] dlrobertson commented on pull request #128 e0e05f3: I'm not sure if this is correct. `header_len` returns results more similar to the older version of ICMPv4's `header_len` (can be 4 or 8). This is so that `payload` can be more useful for NDISC and MLDv2 etc that need the last 4 bytes of the header. That being said, no valid ICMPv6 message is less than 8 bytes in length. https://github.com/m-labs/smo
<GitHub183>
[smoltcp] whitequark commented on pull request #125 6a89a57: In general, we use non-prefixed names like `Type` in implementation files like `ipv6option.rs` and then re-export them under final names in `wire/mod.rs`. https://github.com/m-labs/smoltcp/pull/125#discussion_r164161137
<GitHub44>
[smoltcp] whitequark commented on pull request #125 6a89a57: I think there should be a `Repr::Unknown { length: u8, data: &'a [u8] }` option, since the options are explicitly designed such that unknown options can be still parsed to some extent. https://github.com/m-labs/smoltcp/pull/125#discussion_r164163835
<GitHub174>
[smoltcp] whitequark commented on pull request #128 e0e05f3: This will actually panic if you supply an empty buffer because `header_len` checks `msg_type`. If no valid ICMPv6 message is less than 8 bytes in length you should just use `field::UNUSED.end` (actually I prefer it to be called `field::HEADER_END`. https://github.com/m-labs/smoltcp/pull/128#discussion_r164169245
<GitHub181>
[smoltcp] batonius opened pull request #133: Replace unspecified src IP addrs with the interface IP (master...unspecified_src) https://github.com/m-labs/smoltcp/pull/133
<travis-ci>
m-labs/smoltcp#617 (master - b0995c1 : whitequark): The build passed.
<GitHub74>
[smoltcp] batonius commented on pull request #133 0ddb0e3: Since the branches in my case are equivalent, I could write a similar macro if you think that would be better. I didn't because `Ipv6Repr` and `Ipv4Repr` differ at `proto`/`next_header` so the macro would be somewhat messy. https://github.com/m-labs/smoltcp/pull/133#discussion_r164195494
<cr1901_modern>
2. I didn't even know VCD could do this
<whitequark>
it can't
<whitequark>
it's a hack.
<GitHub126>
[smoltcp] whitequark closed issue #127: smoltcp doesn't replace unspecified src addresses with local addresses for UDP and TCP sockets https://github.com/m-labs/smoltcp/issues/127
<GitHub54>
[smoltcp] whitequark closed pull request #133: Replace unspecified src IP addrs with the interface IP (master...unspecified_src) https://github.com/m-labs/smoltcp/pull/133
<GitHub153>
[artiq] dhslichter commented on issue #652: Are there still obstacles to the 3.6 upgrade? What benefits are there for the 3.6 upgrade other than playing nicer with the Anaconda default install? https://github.com/m-labs/artiq/issues/652#issuecomment-360921200
<GitHub25>
artiq/release-3 aa64b8a whitequark: runtime: build with -Cpanic=unwind....
<GitHub29>
[artiq] whitequark commented on issue #902: I've fixed the backtraces in d58393a1 (master) and aa64b8ad (release-3), can you please reproduce again with a newer firmware? You can decode the backtrace with `addr2line` if you feed the addresses and `runtime.elf` to it. https://github.com/m-labs/artiq/issues/902#issuecomment-360929122
<GitHub47>
[artiq] dhslichter commented on issue #652: @whitequark this seems like a minor convenience for the developer, but not really anything for the end user. Are there features which are compelling/useful for the end user? https://github.com/m-labs/artiq/issues/652#issuecomment-360929337
<GitHub16>
migen/master e554f07 whitequark: sayma_amc,rtm: use correct syntax for PULLUP/PULLDOWN attribute.
rohitksingh-demo has joined #m-labs
rohitksingh-demo has quit [Client Quit]
<GitHub161>
[artiq] cjbe commented on issue #908: The typical valid read region seems to be ~170 LSB on my board, so I don't think that commit (increasing the initial step from 8 LSB to 16 LSB) caused this.... https://github.com/m-labs/artiq/issues/908#issuecomment-360937141