path: root/wiki/Chat_log/20180322-io-chatlog

09:01 < wsa_> hiya!
09:02 -!- dammsan [~dammsan@s214090.ppp.asahi-net.or.jp] has joined #periperi
09:03 < jmondi> and now I have received the emails, but still can't send any
09:03 < wsa_> Greetings from snowy Berlin...
09:04 < wsa_> So, Vaishali is not here yet. Will she come, is that known?
09:05 < morimoto> Ohayo
09:06 < wsa_> well, let's start then, maybe she'll come later
09:06 < wsa_> welcome to the IO meeting!
09:07 < wsa_> is everyone fine? it is so quiet here...
09:07 < morimoto> I'm fine :)
09:08 < jmondi> email issues apart, I'm fine as well :p
09:08 < neg> I'm fine, just waking up and out of coffey :-)
09:08 < dammsan> i assume vaishali will join, but i dont know if she got the meeting announcement?
09:08 < wsa_> is she on the list?
09:08 < horms> I can check
09:09 < dammsan> thanks guys
09:09 < shimoda> hi, sorry i talked with customer team guy now
09:09 < horms> no, i don't think she is
09:09 < horms> let me know an address to add if appropriate
09:09 < wsa_> I think she should be...
09:10 < wsa_> So, let's start with the status updates:
09:10 < wsa_> A)
09:10 < wsa_> Geert removed the requirement to need DT aliases for SCIF. Niklas fixed a bad comment in RAVB and reviewed the thermal patches in the BSP. Shimoda-san upstreamed M3-N bindings for Gen3-USB2-PHY and PWM, and also added PM to PWM. Wolfram finished the QEMU-I2C-passthrough research, fixed some issues with QEMU EEPROM emulation, started working into the new upporting list, subitted a talk proposal for AGL, and 
09:10 < wsa_> did various reviews discussions, mainly about SDHI and Watchdog.
09:10 < wsa_> B)
09:10 < wsa_> Geert wants to enable MSIOF on M3-N. Niklas wants to upport thermal patches once he gets feedback from BSP team. Shimoda-san wants to research USB role swap issue nad add GPIO support to Gen3-USB2-PHY. Wolfram wants to do more SDHI work (tuning & fast modes), write feedback to the HW team for Gen4, plan new tasks (upporting & additional), and hopes to find some time to finally talk about reprogrammable I2C 
09:10 < wsa_> addresses.
09:10 < wsa_> C)
09:10 < wsa_> None
09:11 < wsa_> The summary is missing Simon's and Uli's report which came a bit too close to the deadline
09:11 < uli___> no i/o updates from me
09:12 < wsa_> horms: uli___ : please send them a tad earlier, so I have a chance ;)
09:12 < horms> wsa_: sorry about that
09:12 < horms> It completely slipped my mind
09:12  * geertu hasn't received Simon's and Uli's report yet, blames Gandi
09:12  * morimoto we need to care that non PeriPeri member are joined in periperi-ML. That is the reason why didn't post datasheet download announce mail to it this time...
09:12 < horms> morimoto: I can send you the list of people on the list
09:12 < morimoto> Thanks, but It is OK
09:12 < horms> The only person I think could be removed is Khiem-san
09:13 < dammsan> horms: morimoto-san: can you consider if vaishali should be added or not?
09:13 < dammsan> i think it would make sense
09:13 < wsa_> I agree
09:13 < morimoto> If member who has contact to Jinso / Renesas, he/she is OK
09:13 < morimoto> I mean NDA
09:13 < dammsan> morimoto: NDA download will handled separately anyway, right?
09:14 < wsa_> So, Kaneko sent the D3 thermal update and will keep at it
09:14 < geertu> dammsan: "morimoto: ... That is the reason why didn't post datasheet download announce mail"
09:14 < wsa_> And Simon finished his QEMU task about networking and wonders about HS400
09:14 < morimoto> dammsan: Yeah, so it is no problem now
09:15 < morimoto> Previous time, I posted download announce to periperi-ML, but maybe it was wrong from NDA point of view
09:15 < Marex> ( geertu forgot to mention https://www.denx.de/wiki/view/U-Boot/UbootStat_2018_03 in my report, not that I'd like to boast ;-) )
09:15 < wsa_> dammsan: about QEMU tasks: are you satisfied with the outcomes yet? Is it priority to continue with them?
09:16 < wsa_> we have this other priority now to reduce the delta between BSP and upstream, as I see it
09:16 < dammsan> wsa_: not yet checked yours - sorry will do tomorrow
09:16 < neg> Marex: nice stats, congrats :-)
09:16 < wsa_> but that doesn't necesserily mean that QEMU is out of sight, or?
09:17 < dammsan> wsa_: i think we should keep at it with virtualization
09:17 < morimoto> Marex: yes, congrats
09:18 < wsa_> horms: about HS400, can you give your summary of the current state?
09:18 < horms> I am aware of a few issues.
09:19 < horms> 1) works on my HS3 ES1.0 but not yours? Probably can be fixed by correctly handling clocks for ES1.0 as the current setup seems to be that of ES2.0.
09:19 < horms> 2) follow-up work required for M3-N
09:19 < wsa_> dammsan: ok, fine. I am still not 100% sure of all the use cases, so we might want to talk about it somewhen to form the next add. tasks?
09:19 < horms> I feel we need a plan to solve 1) first
09:19 < dammsan> wsa_: yeah i think so
09:20 < horms> what is your view of things?
09:20 < wsa_> horms: wasn't it agreed on periperi to drop HS400 for H3 ES1.0?
09:20 < wsa_> let me check...
09:20 < horms> ok, somehow I missed that
09:20 < horms> perhaps we can have a follow-up chat to clarrify
09:20 < horms> basically i'm not clear on the situation and way forwards
09:21 < wsa_> horms: Message-ID: <TY1PR06MB09920F022C0459489E7B34FBD8DA0@TY1PR06MB0992.apcprd06.prod.outlook.com>
09:21 < wsa_> horms: yes, let's do that
09:22 < dammsan> may i chime in
09:22 < wsa_> sure
09:22 < dammsan> just wanted to say that older ES versions may not require the most efficient performance really
09:22 < dammsan> so tracking latest ES version and optimizing those might be a good thing to focus on
09:23 < horms> wsa_: thanks, I see that message.
09:23 < dammsan> i highly doubt that any customer would cry if they had PIO-only for older ES versions
09:23 < Marex> neg: morimoto: thanks ;-)
09:23 < horms> dammsan: thanks, understood.
09:23 < dammsan> thanks guys
09:23 < horms> It seems the proposal from Shimoda-san is not to support HS400 on H3 ES1.x
09:23 < horms> is that acceptable to you?
09:23 < dammsan> for sure
09:23 < horms> Ok, thanks
09:24 < wsa_> I'd favor that, too
09:24 < horms> In that case I think we more or less have a way forwards. Just have to sort out the technical details.
09:24 < horms> There was one more issue relating to where tuning hooks in
09:24 < geertu> As long as we get access to the newer hardware, that's fine (for H3 ES2.0 we have, for H2/M2-W we haven't)
09:24 < horms> I did try the suggestion to move it to the usual place, which failed
09:24 < horms> I think I should try again
09:24 < wsa_> let's talk about this later in #periperi-io after the IO meeting ended (so other ppl can join, too if they want)
09:25 < horms> sure
09:25 < wsa_> thanks!
09:26 < wsa_> so, there have been requests for smaller tasks for Kaneko-san and Vaishali
09:26 < wsa_> I wonder if M3-M enablement would be suitable?
09:26 < wsa_> M3-N
09:26 < wsa_> most (if not all) IO devices are identical to M3-W 
09:27 < wsa_> so, I'd think this is a good chance to learn about IP cores from datasheets etc...
09:27 < shimoda> wsa_: enablement means adding bindings and/or adding dts?
09:27 < wsa_> yes
09:27 < wsa_> exactly
09:27 < geertu> enablement means ... + testing
09:28 < wsa_> also true
09:28 < shimoda> wsa_: geertu: i got it. thanks
09:28 < wsa_> which can be done even by remote for quite some IO devices
09:28 < wsa_> but they'd need access to the datasheets
09:29 < wsa_> is it possible? does it make sense?
09:29 < wsa_> it would reduce the periupport list also quite a lot, I'd think :)
09:30 < wsa_> Marex: BTW is there Gen3 WDT support in U-Boot?
09:30 < dammsan> wsa_: there is no NDA in place for Vaishali yet
09:30 < dammsan> wsa_: but remote access is possible
09:31 < geertu> Without NDA, she can access RZ/G1, RZ/A1, and RZ/N1 datasheets
09:31 < wsa_> hmm, well it would be a bit ... difficult ... to add DTS magic number without the datasheet
09:31 < wsa_> horms: what about Kaneko-san?
09:32 < dammsan> geertu: yeah so older devices might be more suitable
09:32 < horms> I have a confidentiality agreement with Kaneko-san
09:32 < wsa_> dammsan: OK
09:32 < horms> which last time we discussed this - a long time ago - was thought to be sufficient
09:33 < wsa_> do you think such tasks are suitable for him?
09:33 < horms> Yes, I think it could work out
09:34 < geertu> horms: Does Kaneko-san have hardware access?
09:35 < horms> no, i assume some sort of remote access could be arranged
09:35 < wsa_> horms: Good news, thanks
09:35 < horms> perferably in Tokyo
09:35 < wsa_> dammsan: do you have some RZ/... boards in your farm?
09:35 < dammsan> i have a genmai boars setup
09:36 < dammsan> and another one in a box
09:36 < dammsan> waiting to get a boot loader
09:36 < wsa_> good, i might have a task for vaishali then where she could use the genmai
09:36 < dammsan> good!
09:36 < dammsan> please contact her
09:37 < wsa_> I will
09:37 < geertu> dammsan: "another one in a box" == rskrza1?
09:37 < dammsan> yep i think so
09:37 < dammsan> and a peach as well
09:37 < wsa_> good
09:38 < wsa_> so, there will be HS400 discussion right after this meeting in #periperi-io and I will contact a few people in private chat about other stuff, too
09:38 < wsa_> that was it from my side
09:38 < wsa_> anything from your side?
09:40 < wsa_> well, then
09:40 < wsa_> I think it is time for core now
09:40 < wsa_> Thanks for this meeting!
09:40 < wsa_> Have fun, Geert!

==== split out discussion about HS400 follows

09:43 < wsa_> hiya
09:46 < wsa_> so, about HS400
09:46 -!- shimoda [~shimoda@relprex2.renesas.com] has joined #periperi-io
09:46 < wsa_> I think we all agreed now to drop HS400 support for ES1.x
09:46 < wsa_> which makes a lot of sense to me
09:47 < wsa_> horms said he would try using the core hook for HS400 preparation again
09:47 < wsa_> that would be really good if it worked out
09:48 <@horms> Ok, I will look into that
09:49 <@horms> Regarding H3 ES1.x, do we have an idea of how to exlude HS400 support for those SoCs?
09:49 < wsa_> But there is still the issue that M3-N changes the HS400 game
09:49 < wsa_> (as well as all future ES of H3 and M3-W)
09:50 < wsa_> It needs a different tuning clock (200 instead of 400MHz) and has 8 taps instead of 4
09:50 < wsa_> (if I recall correctly)
09:50 < geertu> @horms: soc_device_match() and a maximum speed field in of_device_id.data?
09:51 <@horms> ok, could be as easy as that
09:51 <@horms> I'll investigate
09:51 < wsa_> soc_device_match() and disable the capability?
09:51 < wsa_> HS400 capability
09:51 <@horms> Right, I'll look into a good way to use soc_device_match()
09:53 < wsa_> I tend to think we should upstream HS400-M3N-support incrementally
09:53 < wsa_> so we can upstream this series now
09:53 < wsa_> but I am open to other thoughts
10:00 < wsa_> however, we should be prepared for that
10:01 <@horms> wsa_: yes, I think so too
10:01 <@horms> support is present in the BSP
10:01 <@horms> so it doesn't look like a major headache
10:01 <@horms> but it needs something to be based on
10:01 <@horms> and we may as well get the base right first
10:04 < wsa_> the SD clk divider patch belongs to that base, too
10:05 < wsa_> There is still an open issue:
10:05 < wsa_> Message-ID: <20180227215352.gztj6ezjm7wtbm2z@katana>
10:07 < wsa_> shimoda: can you help here?
10:07 < wsa_> I will check the latest BSP code...
10:12 < wsa_> still the same
10:14 <@horms> wsa_: regarding SD clk patch, ack. I think it is fine as-is given that we no longer want to support H3 ES1.x. Do you agree?
10:15 < wsa_> Well, the code matches the data sheet
10:16 < wsa_> However, AFAIU, M3-N tunes with a clock of 100MHz
10:16 < wsa_> I would like to have this confirmed by the HW engineers
10:17 < wsa_> or to be proven wrong :)
10:17 < wsa_> but this M3-N which we agreed to be incremental
10:17 <@horms> Ok, I see
10:17 <@horms> Maybe we can confirm this using the BSP?
10:18 < wsa_> Not sure
10:18 < wsa_> I need to dive in more
10:19 < wsa_> Otherwise, it could just work... but at the speed of 100MHz
10:19 <@horms> Ok, do you think you could look into that?
10:20 < wsa_> This is mainly about asking HW engineers
10:24 <@horms> Ok, but perhaps we can examine the BSP to get some insight? Or do you think the BSP may be wrong in this regard?
10:27 < wsa_> It matches the datasheet
10:27 < wsa_> I don't know if the datasheet is correct
10:27 < wsa_> That's what I want to ask the HW guys :)
10:27 <@horms> Ok, got it
10:31 < shimoda> wsa_: sorry missed your comment, I should check the email of 20180227...?
10:31 <@horms> I will take a break now, should be back soon
10:35 <@horms> back
10:39 < wsa_> shimoda: I think I will rewrite my question in a seperate mail
10:39 < wsa_> shimoda: so you can forward it to HW engineers more easily
10:44 < shimoda> wsa_: i got it. Thanks!