Core-chat-meeting-2017-11-09 09:38 < geertu> Welcome to today's core group meeting! 09:38 < geertu> Agenda: 09:38 < geertu> 1. Status updates 09:38 < geertu> 2. Discussion Topics 09:38 < geertu> Topic 1. Status updates 09:39 < geertu> I'll just replay (a sometimes shortened version of) the update sent by email 09:39 < geertu> A) What have I done since last time 09:39 < geertu> Geert: 09:39 < geertu> - Sent 2nd pull requests for clk-renesas and sh-pfc, 09:39 < geertu> - Attendend ELCE and DT Workshop (Thursday), 09:39 < geertu> - Sent patches to enable SMP on r8a77970/eagle, 09:39 < geertu> - Sent patches for initial support for Salvator-XS with R-Car M3-W. 09:39 < geertu> - Looked a bit into initial support for Salvator-X with R-Car M3-N. 09:39 < geertu> - Fixing genpd for wake-up devices: 09:39 < geertu> - Sent active_wakeup cleanups, 09:39 < geertu> - Working on part two (interrupt routing), 09:39 < geertu> Jacopo: 09:39 < geertu> - Attended ELCE and multimedia meeting on 26/27 November 09:39 < geertu> Marek: 09:39 < geertu> - UBoot: Fixed eMMC data line voltage setting bug in U-Boot DT 09:39 < geertu> - UBoot: Final testing in preparation for the 2017.11 release 09:39 < geertu> Morimoto: 09:39 < geertu> - Posted DMAC TCR/TCRB patch 09:39 < geertu> Niklas: 09:39 < geertu> - [PATCH] net: ethtool: remove error check for legacy setting transceiver 09:39 < geertu> type 09:39 < geertu> - Received a V3M. 09:40 < geertu> Shimoda: 09:40 < geertu> - Prepared test environment for a new feature of M3-N's IPMMU (stage-2 09:40 < geertu> format). 09:40 < geertu> - Handle our 3rd party (EPAM) requests for Xen. 09:40 < geertu> - Handle our kernel test team for testing v4.14. 09:40 < geertu> - Merged the following patch(es): 09:40 < geertu> - arm64: dts: renesas: salvator-common: add dr_mode property for USB2.0 09:40 < geertu> channel 0 09:40 < geertu> Simon: 09:40 < geertu> - Posted, changes requested: 09:40 < geertu> - Gen3 SoC IPMMU DTS patches 09:40 < geertu> - Posted, now under review: 09:40 < geertu> - [PATCH 0/2] iommu/ipmmu-vmsa: r8a779(70|95) support 09:40 < geertu> - Posted, now accepted: 09:40 < geertu> [PATCH] ARM: dts: koelsch: Move cec_clock to root node 09:40 < geertu> [pause to let people catch up] 09:41 < geertu> Any persone I missed? 09:43 < geertu> B) What I plan to do till next time 09:43 < geertu> Geert: 09:43 < geertu> - Finalize genpd for wake-up devices, 09:43 < geertu> - Initial support for Salvator-X with R-Car M3-N? 09:43 < geertu> Marek (U-Boot): 09:43 < geertu> - UBoot: Start converting Gen2 to DM/DT 09:43 < geertu> Morimoto: 09:43 < geertu> - Solve DMAC TCR/TCRB issue. 09:44 < geertu> Niklas: 09:44 < geertu> - Hook up V3M and make sure I can boot it. 09:44 < geertu> - Start to look at 'Add support for more than 32-bits IOVA space' 09:44 < geertu> Shimoda: 09:44 < geertu> - Add usb3.0 phy device node for r8a779[56].dtsi. 09:44 < geertu> Simon: 09:44 < geertu> - Address review of CPUFreq/Z,Z2 Clock patches 09:44 < geertu> - Address review of Gen3 SoC IPMMU DTS patches 09:44 < geertu> - Upport lots of BSP pinctrl patches 09:45 < geertu> My comment here is that several of the patches listed are alrwady upstream ;-) 09:45 < geertu> s/alrwady/already/ 09:45 < horms> geertu: ok, I thought I had pruned those. I'll prune again. 09:45 < horms> Unless kaneko-san posts them before I get that far 09:48 < geertu> C) Problems I have currently 09:48 < geertu> Morimoto: 09:48 < geertu> - I still don't understand DMAC TCR/TCRB issue on serial. 09:48 < geertu> I think problem is in rx_timer_fn(). 09:48 < geertu> According to Geert, "With your patch, the serial driver doesn't know 09:48 < geertu> about them (residue == 0), and the console doesn't see them." What does 09:48 < geertu> this mean? 09:48 < geertu> BTW, about commit e7327c09def48ccf ("serial: sh-sci: Pause DMA engine and 09:48 < geertu> get DMA status again"), I guess the reason why "transaction completes 09:48 < geertu> _after_ DMA stopped" is that it used TCR instead of TCRB? 09:49 < geertu> Perhaps I wasn't very clear in my initial bug report, but serial console input just doesn't work when SCIF DMA is enabled. 09:50 < morimoto> geertu: thank you for your report, we will check it again. 09:51 < wsa_> gotta go, cya guys! 09:51 -!- wsa_ [~wsa@p54B33A46.dip0.t-ipconnect.de] has quit [Quit: ...] 09:51 < horms> ciao 09:51 < geertu> horms: Goodbye! 09:51 < geertu> Oops, wsa_: Goodbye! 09:52 < geertu> morimoto: Checking the direction again may help. 09:52 < morimoto> Yeah, it seems we can guarantee data if we can check DE bit 09:53 < morimoto> geertu: can you agree ? 09:53 < geertu> morimoto: But DE is not the direction bit? 09:54 < morimoto> Yes. using TCR vs TCRB is checking direction 09:54 < morimoto> and guarantee is checking DE bit 09:54 < geertu> With "guarantee", you mean to be sure the DMA has completed? 09:54 < morimoto> yes 09:55 <@pinchartl> geertu: as far as I understand, the DE bit will be cleared the the hardware when the DMA engine finishes draining its buffers 09:55 < geertu> ok 09:55 < morimoto> pinchartl: yes 09:55 <@pinchartl> I'd like to have a clear explanation as to why SCIF doesn't work with the current patch though 09:56 <@pinchartl> I don't think we should merge anything based on guesswork 09:56 < morimoto> Current SCIF is maybe, DMAC read data from device 09:56 < morimoto> then, TCR was updated 09:56 < morimoto> but, DMAC will buffer it untile transferable 09:57 < morimoto> thus, this timing, TCRB is not updated 09:57 < morimoto> Because of this time gap, TCRB indicate 0x20 09:57 < morimoto> I guess 09:58 < geertu> OK, it's worth a try. 09:58 <@pinchartl> that's the thing, I would feel better with a "I'm sure" than a "I guess" :-) 09:58 < morimoto> :) 09:58 < morimoto> So, geertu: pinchartl: can you agree using direction and DE bit check ? 09:59 < morimoto> using direction = TCR vs TCRB 09:59 < geertu> Yes, it's worth a try. It needs to be tested, though. 10:00 < morimoto> Thanks. 10:00 < morimoto> How about pinchartl ? 10:01 < morimoto> ... 10:01 < morimoto> BTW, on commit e7327c09def48ccfd204025726f11b57a19a9c24, it seems rcar dmac driver doesn't support dmaengine_pause() ? 10:02 < geertu> That's correct, AFAIK. 10:02 <@pinchartl> morimoto: I'm not sure yet, I need to analyze the problem in details. knowing how it works when checking the DE bit would help me understanding 10:03 < morimoto> pinchartl: OK, so I will create patch, and add [RFC]. can you check it ? 10:04 <@pinchartl> morimoto: can you run the patch and include the results ? :-) 10:04 < geertu> morimoto: Please check it doesn't break the serial console before posting ;-) 10:04 < morimoto> geertu: Hehe OK :) 10:04 < morimoto> pinchartl: OK, will try 10:04 < horms> geertu: is "[PATCH 2/3] soc: renesas: Identify R-Car M3-W ES1.1" good to merge? I think so after our discussion with shimida-san on periperi ML. 10:05 < geertu> Case closed 9for now)? 10:05 < morimoto> Sound and SCIF are only user of residue ? 10:05 < geertu> horms: I think it is. 10:05 < horms> thanks, I'll add it the next time I go through my queue 10:05 < geertu> Do you want me to update the patch description with future ES info? Or is that considered too confidential for now? 10:05 < horms> Lets keep it to ourselves 10:06 < geertu> OK, fine for me. 10:06 < geertu> s/9now/(now/ 10:06 < horms> I'll just mention that we had some discussion and it seems likely to work for current and future SoCs 10:06 < geertu> If the hardware people change their mind, we need to change the tests anyway 10:06 < horms> About "[PATCH v4 0/3] iommu/ipmmu-vmsa: r8a7796 support V4". I suppose this is a question for dammsan. But should I rebase/repost? 10:06 < horms> geertu: yes, exactly 10:07 < geertu> horms: dammsan said repost in a recent email 10:07 < horms> The important bit is that its unlikely there will be a clash 10:07 < geertu> Indeed. 10:07 < geertu> Any other topics for 10:07 < geertu> Topic 2. Discussion Topics 10:07 < geertu> ? 10:07 < horms> ok. sorry for missing that. I'll do so 10:07 < morimoto> dammsan is here, please wait 10:07 < horms> I will leave very soon, btw 10:08 < dammsan> horms: i think it is fine to wait with the repost 10:08 < dammsan> but feel free to post if it helps your integration work 10:09 < horms> dammsan: what is the motivation for wating. It effects my integration work in the sense that it adds new bindings. and is a dependency for bindings for other SoCs. 10:09 < dammsan> my gut feeling is that it is good to hold off addting code to avoid potential issueeeees 10:10 < horms> ok, perhaps just repost the bindings? 10:10 < dammsan> good idea 10:10 < horms> ok, will do 10:10 < dammsan> together with other ones too if you like 10:10 < dammsan> we have to revisit non-H3 support in the driver anyway 10:10 < horms> I'll repost all the outstanding bindings but not the driver code. Is that what you have in mind? 10:10 < dammsan> sure 10:10 < dammsan> thanks 10:10 < dammsan> sounds ogod 10:11 < horms> One last question, I take it that you intentionaly did not post a fallback binding 10:11 < dammsan> yeah, the IP varies with soc 10:11 < horms> thanks, got it 10:11 < dammsan> thanks 10:12 < horms> I need to go now, sorry for any inconvenience that causes 10:12 < geertu> I think we're finished with core anyway 10:12 < geertu> horms: Bye! 10:12 < horms> perfect 10:12 < horms> bye 10:12 -!- horms [~horms@61.40.109.130] has quit [Quit: Leaving] 10:13 <@pinchartl> one question for core 10:13 <@pinchartl> (and actually for I/O too, I should have asked earlier) 10:13 <@pinchartl> what's the status of V3M support ? 10:15 < geertu> pinchartl: Boots using serial console, NFS root should work 10:15 <@pinchartl> and while waiting for an answer to that question, let's see who's ready for the multimedia meeting 10:15 <@pinchartl> hi jmondi2 10:15 <@pinchartl> hi kbingham[m] 10:15 <@pinchartl> hi dammsan 10:15 [Users #periperi] 10:15 [@pinchartl] [ jmondi2 ] [ Marex ] [ mturquette] 10:15 [ dammsan ] [ kbingham ] [ marex-cloud] [ neg ] 10:15 [ geertu ] [ kbingham[m]] [ morimoto ] [ shimoda ] 10:15 -!- Irssi: #periperi: Total of 12 nicks [1 ops, 0 halfops, 0 voices, 11 normal] 10:15 <@pinchartl> hi morimoto 10:15 < kbingham[m]> I'm here 10:15 < kbingham[m]> Morning 10:15 <@pinchartl> hi ul[tab] 10:15 <@pinchartl> no Ulrich today 10:15 <@pinchartl> hi neg 10:15 < jmondi2> I'm here too! 10:15 < neg> pinchartl: I'm not sure about the status, got a V3M yesterday and plan to try and take it for a first spin 10:16 < kbingham[m]> geertu: . Does i2c work? 10:16 <@pinchartl> geertu: any idea whether I2C is working ? how about PFC, is that still unmerged ? 10:16 < geertu> pinchartl: Sergey is working on upstreaming PFC 10:17 < geertu> Unfortunately he discovered recently he doesn't have the xls files describing pin mappings 10:17 <@pinchartl> -_-' 10:17 <@pinchartl> that's a bit of a blocker 10:18 < geertu> I told him to post the initial version anyway. We typically don't bother reviewing the register definitions, and we can review pin groups 10:19 <@pinchartl> thanks 10:19 <@pinchartl> so let's start with the multimedia meeting 10:19 <@pinchartl> welcome everybody 10:19 <@pinchartl> (I expect Magnus and Morimoto-san to come back soon) 10:19 <@pinchartl> first topic, status update 10:20 < geertu> Let's finish the core meeting: Thanks for joining, and have a nice continued day!