path: root/wiki/Chat_log/20210520-core-chatlog

09:33 < geertu> Welcome to today's Core Group Chat Meeting!
09:33 < geertu> Agenda:
09:33 < geertu> 1. Status Updates
09:33 < geertu> 2. Discussion Topics
09:33 < geertu> Topic 1. Status updates
09:33 < geertu> A) What have we done since last time:
09:33 < geertu> Marek updated ATF patches, fixed MISRA C issues in ATF, and upported
09:33 < geertu> R-Car V3U U-Boot support.
09:33 < geertu> Shimoda-san reported about OSAL, asked the hardware team about R-Car D3
09:33 < geertu> PFC documentation bugs, got information about R-Car Gen3e, and inspected
09:33 < geertu> the R-Car V4H spec.
09:33 < geertu> Geert did periject bsp-41x triage, upported BSP opp-table patches,
09:33 < geertu> converted remaning core DT bindings to json-schema, added pin bias
09:33 < geertu> support for R-Car SoCs that still lacked them, and added missing GPIO
09:33 < geertu> keys pull-ups for R-Car Gen2 boards.
09:34 < geertu> B) What we plan to do till next time:
09:34 < geertu> Marek plans to work on OpenOCD support and U-Boot upstreaming for R-Car
09:34 < geertu> V3U.
09:34 < geertu> Shimoda-san plans to follow up R-Car V3U IPMMU work and R-Car D3 PFC
09:34 < geertu> issues, and continue to inspect the R-Car V4H spec.
09:34 < geertu> Geert plans to review the RZ/G2L patches, and post more pinctrl
09:34 < geertu> validation code.
09:34 < geertu> C) Problems we have currently:
09:34 < geertu> Geert wants to kill "renesas,shdma-mux", as (un)used on R-Mobile APE6.
09:34 < geertu> ---EOT---
09:34 < geertu> Anything I missed?
09:36 < geertu> If anyone with hardware access could test the GPIO key pull-ups, that would be great
09:37 < wsa> geertu: just testing if they work?
09:37 < geertu> wsa: yes, regression testing, and optionally measuring the pin voltage with bis enabled and disabled
09:38 < geertu> s/bis/bias/
09:38 < geertu> Topic 2. Discussion Topics
09:39 < geertu> Are there plans to make R-Car Gen3 available to us, e.g. with remote access?
09:39 < wsa> I can surely test the keys on Lager. I'll try to check the bias for H2/E2, too...
09:39 < geertu> wsa: thx!
09:40 < wsa> "Gen3e" probably?
09:40 < geertu> oops, yes
09:43 < wsa> shimoday: are there already plans for "Gen3e" availability
09:43 < wsa> ?
09:44 -!- damm [~damm@68.183.84.28] has quit [Ping timeout: 245 seconds]
09:44 -!- damm [~damm@68.183.84.28] has joined #periperi
09:44 < shimoday> wsa: i'd like to support Gen3e in this year somehow
09:45 < wsa> geertu: I sent a patch to periperi moving optee-patches to non-target. are you okay with it?
09:45 < shimoday> we can use "Gen3" boards as "Gen3e" even if 2G mode
09:45 < wsa> i think optee is core, so i didn't want to push it right away
09:45 < geertu> wsa: I'm happy to get rid of any optee patches ;-)
09:45 < shimoday> because the SoCs are completely the same
09:45 < geertu> marex: Are the BSP optee patches any good?
09:46 < geertu> shimoday: sure, but one day we'll have to create DTS files for them, right?
09:46 < marex> geertu: which ones ?
09:47 < wsa> I found a message saying that the drives quite diverged, and it seems the patches are fixes to local code...
09:47 < wsa> Quoting bsp392_tee_driver.yaml:
09:47 < wsa> "It's difficult to refactor this driver for upstream. So, mark Abandoned"
09:47 < shimoday> geertu: yes, we have to create DTS files somehow
09:47 < geertu> shimoday: For Gen3e I agree, but won't we overclock our non-2G parts?
09:47 < marex> geertu: upstream optee works without extra patches
09:47 < wsa> marex: check "[PATCH periject] linux: bsp-41x: non-target: move optee changes" from periperi-ml
09:49 < shimoday> geertu: i heard almost all Gen3 chips can overclock to 2G. If we are unlucky, the SoC just cannot work correctly without broken, I heard
09:49 < geertu> shimoday: Ok (so the fans do serve a good purpose ;-)
09:50 -!- damm [~damm@68.183.84.28] has quit [Ping timeout: 252 seconds]
09:51 -!- damm [~damm@68.183.84.28] has joined #periperi
09:51 < shimoday> geertu: yes :)
09:51 < wsa> time to remount the fans then :)
09:53 < marex> wsa: seems like a whole lot of it is related to the SMC "update" call, i.e. renesas has custom SMC service where you put in bootloader update and OpTee has enough permissions to write it into RPC HF
09:53 < marex> wsa: that's what my initial look tells me anyway
09:54 < geertu> marex: thx for checking! Hence non-target it'll be?
09:54 < marex> geertu: non-target ?
09:55 < wsa> non-target == no upporting
09:55 < marex> wsa: for this custom SMC call, no, no upporting
09:56 < wsa> good, thanks for looking
09:56 < wsa> into it
09:57 < marex> sure ...
09:57 < geertu> Anything else to discuss?
09:57 < marex> which however brings me to another question
09:57 < marex> yes
09:57 < marex> ATF upstream is starting to turn increasingly ... uh ... not great
09:58 < marex> so I was thinking, it might make sense to stick to the U-Boot PSCI implementation for upstream V3U port
09:58 < marex> (i.e. avoid ATF there altogether)
09:58 < marex> I hope that would be OK with you ?
09:59 < geertu> marex: Does it support SMP, suspend, ...?
09:59 < marex> geertu: whatever is missing should be added ; I still don't have the ATF sources for V3U , so I cannot tell
10:00 < marex> geertu: but the idea would be to have minimal PSCI support to toggle CPU cores on/off as mandated by the arm64 spec, and not any extra stuff, let Linux do its job
10:01 < geertu> marex: fine for me, if upstream arm64 won't complain
10:01 < marex> geertu: it won't, since it is compliant to the spec
10:01 < marex> geertu: also, they didn't complain when intel used it on their aarch64, so why should they complain now
10:02 < geertu> ok
10:03 < geertu> Anything else to discuss?
10:03 < geertu> Perhaps #periperi and freenode?
10:03 < geertu> Or too early for that?
10:03 < wsa> next meeting on June, 17th?
10:03 < wsa> is something wrong with Freenode?
10:04 < pinchartl> geertu: I think it's not too early to discuss a possible move, but too early for a conclusion. we need the dust to settle first, there's too much ongoing drama
10:04 < geertu> https://lwn.net/Articles/856543/
10:05 < pinchartl> what would be the options ?
10:06 < geertu> perichat ;-)
10:06 < geertu> From the knights of NIH...
10:06 < pinchartl> - another public IRC server (libera.chat ?) - this would likely suffer from similar privacy concerns
10:06 < kbingham> NIH
10:06 < pinchartl> - a private IRC server ?
10:06 < pinchartl> - matrix ?
10:06 < damm> i can setup a local IRC server in a VM if needed
10:07 < damm> or we can put it in the cloud to improve latency
10:07 < neg> In case it's needed I registered #periperi at libera.chat
10:08 < marex> neg: it is real nice
10:09 < geertu> OK, let's wait and see
10:09 < neg> damm: I love the zero maintenance of the could :-)
10:09 < geertu> June 17th is fine for me
10:09 < geertu> neg: cloud?
10:09 < shimoday> renesas has MS Teams so if we all have Microsoft account I could invite you :) This is a voice meeting though.
10:10 < geertu> shimoday: can;t you access Teams in a web browser, without account?
10:10 < pinchartl> noooooooo not teams :-)
10:10 < shimoday> geertu: I'm not sure about the detail. perhaps, Magnus-san knows
10:11 < shimoday> pinchartl: :)
10:11 < pinchartl> would everybody generally prefer to keep using IRC, or move to matrix ?
10:11 < geertu> pinchartl: Google Hangouts? ;-)
10:11 < pinchartl> (we could also consider a voice discussion instead of IRC discussion)
10:12 < wsa> wen can meet in one of the C64 BBSes available via telnet ;)
10:12 < damm> maybe geert has something compatible with 68k as well? =)
10:12 < wsa> seriously, I agree with Laurent, let's wait a little until the dust settles
10:13 < neg> Why not meet in person each month in Tokyo?
10:13 < wsa> I'd personally like to stay with IRC, either Freenode or LibraChat
10:13 < geertu> OK, let's wait and see
10:13 < marex> wsa: yes
10:13 < wsa> shimoday: moriperi: damm: 17th of June is good for you?
10:14 < geertu> neg: plus two weeks in quarantaine before, and after
10:14 < wsa> pinchartl: ^?
10:14 < shimoday> wsa: yes, 17th of June is good to me
10:14 < moriperi> wsa: yes, it is OK for me. thanks
10:14 < geertu> 17th it'll be
10:14 < geertu> Thanks for joining, and have a nice continued day!