path: root/wiki/Chat_log/20181108-io-chatlog

09:02 < wsa> again, welcome everyone. IO meeting time now
09:02 < wsa> status updates:
09:02 < wsa> Status updates
09:02 < wsa> ==============
09:02 < wsa> A - what have I done since last time
09:02 < wsa> ------------------------------------
09:02 < wsa> Kaneko-san
09:02 < wsa> : upported I2C, SCIF, HSICF enablement for E3
09:02 < wsa> Marek
09:02 < wsa> : upported SCIF2 pinmux patches, SDHI patches for E3 and submitted V2 for
09:02 < wsa>   the PCIe L1 ATF workaround
09:02 < wsa> Niklas
09:02 < wsa> : (did various SDHI patches to enable HS400 but missed to send in his report)
09:02 < wsa> Shimoda-san
09:02 < wsa> : fixed remove function of SCIF driver and investigated an issue for USB3.0
09:02 < wsa>   peripheral with E3
09:02 < wsa> Ulrich
09:02 < wsa> : reviewed two SCIF patches
09:02 < wsa> Wolfram
09:02 < wsa> : organied and participted in SDHI hackfest, imported periupport priorities to
09:02 < wsa>   latest ticket file, picked up TDSEL and WDT patches again, reviewed lots of
09:02 < wsa>   SDHI patches and some general PM rework, updated periupport, and started to
09:02 < wsa>   pick up I2C core PM rework again
09:02 < wsa> B - what I want to do until next time
09:02 < wsa> -------------------------------------
09:02 < wsa> Geert
09:02 < wsa> : wants to resubmit fixes for fallback to PIO in the sh-sci driver
09:02 < wsa> Kaneko-san
09:02 < wsa> : wants to upport HSCIF, PWM, and PCIE for E3
09:02 < wsa> Marek
09:02 < wsa> : wants to continue on PCI, DVFS I2C and PMIC upporting for E3 and PCA953x
09:02 < wsa>   suspend/resume support
09:02 < wsa> Shimoda-san
09:02 < wsa> : wants to fix the USB3.0 issue for E3 and investigate a USB related reset
09:02 < wsa>   issue for Gen2
09:02 < wsa> Simon
09:02 < wsa> : wants to investigate RAVB no-link patches in BSP 3.7.2
09:02 < wsa> Ulrich
09:02 < wsa> : wants to continue reviewing patches
09:02 < wsa> Wolfram
09:02 < wsa> : wants to continue I2C core PM rework and upport BSP patches
09:02 < Marex> wsa: clock
09:03 < wsa> C - problems I currently have
09:03 < wsa> -----------------------------
09:03 < wsa> Shimoda-san:
09:03 < wsa> asks for documentation introducing to virtualization
09:03 < wsa> Simon:
09:03 < wsa> has D3 thermal support tested but patches stuck on upstream mailing list
09:03 < wsa> my questions:
09:04 < wsa> horms: I assume the IIC issue with ICSTART are resolved now?
09:04 < wsa> horms: any plan when you can start working on the RAVB no-link patches?
09:05 < wsa> Marex: did you coordinate your E3 upporting work with someone? I mean it is great to see progress there, but there is work assigned to Kaneko-san already and I'd like to avoid duplication of work...
09:06 < Marex> wsa: I checked with geertu, he said it was OK
09:06 < Marex> wsa: I was concerned about Kaneko-san, yes
09:07 < Marex> wsa: is the PCI/I2C/PMIC task assigned ?
09:08 < wsa> Marex: Good, thanks! I will check with geertu then how to handle this in the future...
09:08 < wsa> PCIE and IIC is, PMIC not
09:08 < geertu> Marex: i2c was not mentioned in last's IO meeting report, PMIC was for core
09:09 < wsa> It probably makes sense to combine IIC and PMIC
09:09 < wsa> to you
09:09 < geertu> I've been running with iic and pmic enabled on e3
09:09 < wsa> It is a grey area, I think, because PFC is core and I2C itself is IO
09:09 < geertu> iic works
09:10 < geertu> pmic is blocked by Ebisu firmware not supporting s2ram on the 4D variant yet
09:10 < geertu> Else I would have had submitted it myself the week I got the board.
09:10 < wsa> so, not much harm done (I think), but a sign to install some communication between geert and me here
09:12 < wsa> I also think it would make sense to assign PCIE to Marex but we also need work for Kaneko-san
09:12 < wsa> I will check this when I update the lists after the meeting
09:12 < wsa> horms: is this all OK with you?
09:12 < Marex> wsa: OK
09:12 < geertu> Can Kaneko-san test his work on some board now?
09:12 < geertu> Apparently his last PFC patch was not even compile-tested
09:12 < Marex> wsa: I tested the PCIe on Ebisu 4D already
09:13 < Marex> wsa: so I can roll out the patch right away
09:13 < wsa> neg: you there?
09:13 < wsa> Marex: heh
09:15 < wsa> I also wonder about Jinso sending thermal patches...
09:16 < wsa> why exactly this driver? :)
09:16 < wsa> damm: do you know about this?
09:16 < damm> nope sorry
09:17 < wsa> pity neg seems not here, too
09:17 < wsa> neither he sent a report, I hope he is well
09:17 < Marex> wsa: probably on his way to Canada ?
09:17 < Marex> wsa: so anyway, I'll wait with the PCIe patches ?
09:18 < geertu> He said his mail and git trees will be cut off for a week by its hosting provider, perhaps that has already started, and caused more damage than expected?
09:18 < wsa> Marex: yes, but probably just until this evening
09:18 < wsa> just until I updated all my lists
09:18 < Marex> wsa: got it, will do
09:19 < wsa> Marex: thanks
09:20 < wsa> so, the updated periupport list with imported priorities...
09:20 < wsa> morimoto: are you okay with this (in general)?
09:20 < wsa> geertu: did you have a chance to look at it?
09:20  * morimoto talking with our Boss now
09:21 < wsa> I would be nice to merge it rather soonish, I'd think. pinchartl and my updates are already depending on it...
09:22 < wsa> pinchartl: thanks for the prompt review!
09:22 < geertu> wsa: I only had a brief look.
09:23 < wsa> Marex: did the HS400 clock fix from this night also fix the HS400 issue in Linux then?
09:23 < geertu> I did do some auto-updates and fixed obsolete linux-next references
09:23 < wsa> geertu: before or after the conversion?
09:24 < Marex> wsa: no, I'm still seeing that and I'm not sure why
09:24 < Marex> wsa: I have a hypothesis, but I need to test it
09:24 < geertu> wsa: on master
09:25 < wsa> Marex: which hypothesis?
09:25 < morimoto> wsa: basically I have no objection (= for periupport)
09:25 < Marex> wsa: that it has to do somehow with the board physical properties, possibly temperature
09:25 < wsa> Marex: there are HS400 related patches in the latest BSP dealing with temperature
09:26 < Marex> wsa: I have very little samples to support the hypothesis, but the HS400 worked in Linux if I cold started the board after some idle period
09:26 < Marex> wsa: on the next reboot or power cycle, it failed
09:26 < Marex> wsa: even if I waited a bit to discharge the caps
09:27 < Marex> thus, I suspect temperature, but like I said, I have very little samples to support this
09:27 < Marex> I need to test more
09:27 < wsa>
09:27 < wsa>     Since Gen3 SDHI has a internal DS signal AC-spec violation in HS400 mode,
09:27 < wsa>     CRC-error may occur in read command. It is only HS400 mode.
09:27 < wsa>     This phoenomenon occurs at low/High temperature.
09:27  * wsa wonders about Marex room temperature ;)
09:29 < wsa> geertu: do you want to have time for another look at the periupport changes?
09:30 < Marex> wsa: I saw that patch
09:30 < Marex> wsa: I rather wonder if this has to do with the CPU heating up the board, or itself
09:30 < wsa> Marex: sure, was j/k
09:30 < geertu> wsa: I'll have a closer look
09:31 < geertu> (using the pre-commit hook or make patch ;-)
09:31 < Marex> wsa: I suspect that violation is also present on older Gen3s and that results in the 4TAP 400MHz SDnH limitation ?
09:31 < wsa> geertu: ok, let me know what you think by mail then. Thanks!
09:32 < Marex> wsa: but then I wonder if on 4tap SoCs, the performance in the HS400 mode is worse than on 8tap SoC ...
09:33 < wsa> Marex: my gut feeling is there is some bigger thing going wrong, but it's all speculation
09:33 < wsa> and not terribly important why exactly there are 4 taps
09:33 < Marex> wsa: the thing is, when I configured the clock correctly for the 8tap/HS400, the raw read performance almost doubled ...
09:34 < Marex> wsa: so I wonder if on 4tap SoCs, the read performance in HS400 would be lower than on 8tap SoC
09:34 < wsa> Marex: I expect the 4tap performance to be worse than 8tap
09:34 < wsa> but as long as it is > HS200...
09:35 < Marex> wsa: it was exactly equal for me
09:35 < Marex> wsa: but I need to revisit the 4tap case
09:35 < wsa> AFAICS there is not much we can do, and the focus should be the latest ES version anyhow
09:36 < wsa> neg did a chart
09:36 < Marex> wsa: right, the scratchpad
09:36 < geertu> Who has access to H3 ES3.0?
09:37 < Marex> geertu: I had it in my hands once
09:37 < wsa> neg had a slight increase on H3 ES2.0. that matches my experiences
09:37  * wsa notes to update the chart with his results
09:38 < wsa> okay, are there questions from your side
09:38 < wsa> ?
09:39 < wsa> sadly, no responses from horms about questions to him. Simon, if you read this later, please respond by mail to periperi
09:39 < Marex> wsa: on H3, I had 170MB/s in both HS200 and HS400, so no increase for me :(
09:40 < Marex> wsa: but I suspect 200 MB/s might be the limit for 4tap configs
09:40 < wsa> let's update the table and continue the discussion by mail, too
09:40 < Marex> wsa: sounds good
09:40 < wsa> Marex: thanks!
09:41 < wsa> let's get to the core now :)
09:41 < wsa> geertu: have fun!