From 55e3b2f45880faaf06f3c660ca9e8a6d9aa14bce Mon Sep 17 00:00:00 2001 From: Kuninori Morimoto Date: Mon, 9 Dec 2019 15:29:52 +0900 Subject: wiki: Porting wiki: Porting Chat Log Signed-off-by: Kuninori Morimoto --- wiki/Chat_log/20160519-io-chatlog | 207 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 207 insertions(+) create mode 100644 wiki/Chat_log/20160519-io-chatlog (limited to 'wiki/Chat_log/20160519-io-chatlog') diff --git a/wiki/Chat_log/20160519-io-chatlog b/wiki/Chat_log/20160519-io-chatlog new file mode 100644 index 0000000..4e181a4 --- /dev/null +++ b/wiki/Chat_log/20160519-io-chatlog @@ -0,0 +1,207 @@ +--- Log opened Thu May 19 09:59:57 2016 +09:59 -!- wsa_ [~wsa@dslb-178-008-079-154.178.008.pools.vodafone-ip.de] has joined #periperi-io +09:59 -!- Irssi: #periperi-io: Total of 5 nicks [1 ops, 0 halfops, 0 voices, 4 normal] +09:59 -!- Irssi: Join to #periperi-io was synced in 1 secs +10:00 < wsa_> hiya +10:00 < shimoda> hi +10:00 <@neg> hello +10:00 < wsa_> nice to see you all despite having forgotten the reminder mail +10:00 < wsa_> :) +10:01 < horms> hi +10:01 < geertu> hi +10:02 < wsa_> so, let's start the meeting +10:02 < wsa_> i'd like to start with status updates per person +10:02 < wsa_> driven by an honest sort -R (random) +10:03 < wsa_> ;) +10:03 < wsa_> shimoda: you are first +10:03 < wsa_> what is new on your side? +10:03 < shimoda> wsa_: yes +10:04 < shimoda> about USBPHY,v4.7,public,shimoda,add extcon support, it is merged +10:04 < wsa_> great! +10:05 < shimoda> into linux-phy , next branch +10:05 < shimoda> so, perhaps it is merged in v4.7 linus branch +10:06 < wsa_> let's hope so +10:06 < shimoda> about other 2 tasks (usb3-host and USBPHY OTG support), no update +10:07 < wsa_> ok +10:07 < shimoda> and now, I'm investigate IPMMU + USB2.0 issue +10:07 < shimoda> on Gen3 +10:07 < shimoda> other devices (xhci and ohci) seem work correctly, but ehci doesn't work... +10:08 < wsa_> is this a new task? +10:08 < shimoda> yes, it is a new task +10:08 < wsa_> USB2-Host,v4.8,plan,shimoda,IPMMU issues on Gen +10:09 < wsa_> USB2-Host,v4.8,plan,shimoda,IPMMU issues on Gen3 +10:09 < shimoda> yes, thanks! +10:09 < shimoda> about i/o, these are my current status +10:10 < wsa_> thank you! +10:10 < wsa_> I'm next +10:10 < wsa_> about the SDIO task: +10:10 < wsa_> somehow done since I got the KS7010 card working here +10:11 < wsa_> however, without the card this cannot be reproduced in Japan :( +10:11 < shimoda> great! +10:11 < wsa_> Magnus has set up 2 other cards remotely for me +10:11 < wsa_> i will check them +10:12 < shimoda> yes... but to avoid export paper work, i don't want to get the card from you :) +10:12 < shimoda> sorry. I forgot to reply on your email +10:12 < wsa_> however, if there are WLAN driver issues, there is no time to debug those... +10:13 < wsa_> OK. I will bring one in July and then we can talk what to do with it +10:13 < wsa_> skipping paperwork +10:13 < shimoda> I also tried to a sdio card that Magnus-san has (mwifiex / 88W8887), it doesn't work correctly. +10:14 < wsa_> there might be a follow-up task SDIO+UHS, but we need to discuss that +10:14 < shimoda> JFYI, the card also doesn't work correctly on gen2 bsp :) +10:14 < wsa_> and we should have one WIFI card specified for that, surely +10:14 < wsa_> shimoda: thanks, that is interesting +10:14 < wsa_> BTW the KS7010 cards do NOT work on Gen2 :( +10:15 < wsa_> looks like interrupt delivery problems +10:15 < horms> i think the gen2 bsp or at least the ltsi it is based on was tested with the sdio card that I have at some point. But it was a long time ago so I don't recall the details. +10:16 < wsa_> i think we should play some SDIO card game in Japan +10:16 < horms> ok +10:16 < horms> i'm happy to loan my card to the cause +10:16 < horms> Jinso have or at least had the same card +10:16 < wsa_> thanks +10:16 < wsa_> that sounds likely ;) +10:16 < horms> I'm pretty sure I arranged for them to order some +10:17 < horms> back in the day +10:17 < horms> ... +10:17 < wsa_> they are hard to get today +10:17 < horms> probably +10:18 < wsa_> so, despite the KS7010 working fine on Gen3, I guess SDIO topics will be around for some more time +10:18 < wsa_> my other task is watchdog pretimeout support +10:18 < wsa_> I picked up the old patches +10:18 < wsa_> and am currently simplifying them +10:19 < horms> n.b. possibly one of the cards magnus has made available to you is the same as the one I have +10:19 < wsa_> watchdog core has significantly changed, so they don't work as is +10:19 < wsa_> horms: i think so +10:20 < wsa_> horms: but for debugging, a real card may be the solution +10:20 < horms> i can post mine if the need arises, just let me know :) +10:20 < wsa_> so i remove the IMO over-engineered parts from the pretimeout patches +10:21 < wsa_> i hope it makes them easier to be applied upstream +10:21 < wsa_> horms: thanks +10:22 < wsa_> other than that: basic SDHI maintenance +10:22 < wsa_> not much more +10:22 < wsa_> lots of time spending on designing additional tasks and doing additional reviews +10:22 < wsa_> that's from my side +10:23 < wsa_> geertu: your turn now +10:23 < horms> wsa_: I send you an email about task:i2c did you see it? If its pending thats fine. +10:23 < geertu> SCIF,v4.7,public,geert,Extend subsystem with GPIO-based software handling and hardware flow control +10:23 < geertu> I've sent v2 +10:23 < geertu> which received some acks +10:23 < geertu> and a question, which I refuted. +10:23 < geertu> Will resend with the acks after the merge window has closed. +10:24 < geertu> SCIF,v4.8,plan,geert,SCIF FIFO flushing +10:24 < geertu> I'm currently working on that one +10:24 < wsa_> any new insights? +10:25 < wsa_> or still data collection phase? +10:25 < geertu> Data collection phase ;-) +10:26 < geertu> That's it from my side +10:27 < wsa_> ok, thanks +10:27 < wsa_> horms: your updates? +10:28 < horms> sure +10:28 < horms> I have been working on adding sdr104 support to sdhi +10:28 < horms> I posted an initial patch set, based on what is present in the BSP +10:28 < horms> It recieved review from many of the people present here: thanks for that +10:29 < horms> I've significantly reworked things accordingly +10:29 < horms> And things look better +10:29 < wsa_> cool +10:29 < horms> Currently I'm seeing some timeouts which may be related somehow +10:29 < horms> I'm also unsure of how exactly to handle that different sdhi blocks present support different speeds +10:30 < horms> Probably that can bet discussed after i post v2 +10:30 < horms> Lastly, I'm not entirely convinced that the tuning actually does anything with regards to improving throughput +10:30 < horms> this may be because its not being performed correctly. +10:30 < wsa_> did you get more information about this undocumented SCC clock? +10:30 < horms> e.g. the BSP implementation does not seem to reflect the spec with respect to selecting a tap +10:31 < horms> no, that slipped my mind +10:31 < horms> In short, I'm working towards v2 but there are a few loose ends +10:32 < wsa_> do you have time left for this task after v2? +10:32 < horms> I expect to have time for v3 :) +10:32 < wsa_> great +10:32 < horms> But perhaps not the DMA investigation you were hoping for +10:32 < wsa_> i anticipated that +10:32 < wsa_> let's make sure to get the prototype proper +10:33 < horms> I anticipate v3 may spill over into next month, techincally after the deadline +10:33 < wsa_> the issue about selecting tap sounds interesting +10:33 < horms> hopefully that won't be a problem +10:33 < horms> Right +10:33 < horms> so if you look at the SCC document (which you have iirc) +10:33 < horms> it says to select the tap when 3 successive tests succede +10:33 < horms> not so complex +10:34 < horms> I implemented that +10:34 < wsa_> well, I assume V2 will qualify as prototype for the report +10:34 < horms> the BSP does something else +10:34 < horms> wsa_: yes, that is my assumption too +10:35 < wsa_> and besides that I still think we have DMA throughput problems :) +10:35 < horms> I believe there will be several unsolved issues :) +10:35 < wsa_> SDR50 doesn't need tuning and is already not as fast as I'd like +10:35 < horms> Or loose ends if you prefer +10:35 < horms> Fyi I think tuning has to not be done on the sdhi for SDR50 +10:36 < wsa_> so I don't expect SDR104 to have huge gains because of that +10:36 < horms> right +10:36 < wsa_> but SDR104 was a priority for this month +10:36 < horms> as i mentioned to you elsewhere I can get 40Mbit/s +10:36 < horms> or slightly less +10:36 < horms> I tried a few different cards +10:36 < horms> their speeds vary +10:36 < horms> but that is the best one, the sandisk one +10:37 < wsa_> same here +10:37 < horms> (I boght some more cards after the last time we spoke about this) +10:37 < wsa_> should complain to samsung somewhen +10:37 < horms> I can +10:37 < horms> but perhaps there are better things to spend time on +10:37 < wsa_> that may be true +10:37 < horms> the most expensive card performed the worst, btw +10:37 < wsa_> ;) +10:38 < wsa_> ok, looking forward to v2 +10:38 < horms> thats probably all from my side +10:38 < horms> I hoped to have it out for this meeting +10:38 < horms> but no such luck +10:39 < wsa_> horms: can we chat a little after this meeting about the next additional tasks? +10:39 < horms> sure +10:39 < wsa_> thanks +10:40 < wsa_> so, neg is next +10:40 <@neg> Gen3 I2C DMA support is public +10:40 < horms> neg: is that the patch I applied? +10:40 <@neg> yes thats the DT part of it +10:41 < horms> ok +10:41 <@neg> and I think wsa_ have picked up the driver patch +10:41 < wsa_> yes +10:41 <@neg> and I just sentout a incremental patch adressing Arnds comment on the driver +10:41 < wsa_> it is marked "merged" in the todo file :) +10:42 <@neg> so I hope that is it for that task :-) +10:42 < wsa_> thanks for the work! +10:42 <@neg> and that is all for me I have no other i/o task +10:42 < wsa_> any time left for IO base task? +10:43 < wsa_> not that i have something now +10:43 <@neg> hum in base no, rest of time is for core work. But if you have something small need doing I can try to squeeze it in +10:44 < wsa_> ok +10:44 < wsa_> yeah, i asked because of distribution of small things (if there is one) +10:44 < wsa_> wanted to know where we are +10:45 < wsa_> but since we fixed the r8a7740 SDHI regression, I don't see anything left there +10:45 <@neg> ofc, if you think of something let me know and I see what I can do +10:45 < wsa_> please speak up if i am wrong +10:46 < wsa_> so, neg is basically done with IO for Q2 and will (probably) have some more days in Q3 +10:47 < wsa_> so, ulrich is not here and will do the CAN driver review once he is back +10:47 < wsa_> morimoto-san is also not here but doesn't have any tasks assigned anyhow +10:48 < wsa_> so, in general i think we are on track +10:48 < wsa_> thanks guys! +10:48 < wsa_> it would be nice to have more IO time for geert, though +10:48 < shimoda> thank you! +10:50 < wsa_> ok, anything more to discuss? +10:50 <@neg> If task list is done I have a IO question +10:50 < wsa_> sure +10:51 <@neg> during boot of Salvator-X it is deleayed for ~20 sec and then i get this 'xhci-hcd ee000000.usb: can't setup: -110' +10:51 <@neg> i use the arm64_defconfig and was just wondering if this is known +10:51 < shimoda> oh, sorry it's xhci driver problem +10:52 < wsa_> there are mails about it on renesas-soc +10:52 <@neg> ok thanks +10:53 < wsa_> in short: a Kconfig dependecy needs to be set up +10:53 < shimoda> in v4.6, it should be fixed by f879fc32aa0c96fbac261b3d857a1239d554ad01 +10:54 <@neg> shimoda: great, thanks +10:55 < shimoda> neg: you're welcome! +11:04 < wsa_> ok, if no more questions... +11:04 < wsa_> ...then thanks for this meeting! +11:05 < shimoda> thank you! bye! +11:05 <@neg> thanks all +11:05 -!- shimoda [~shimoda@relprex3.renesas.com] has quit Quit: WeeChat 0.4.2 +11:06 < geertu> Data collection phase ;-)bye +11:06 < wsa_> good luck geertu ! +11:06 < geertu> Seems I have a problem with my IRC FIFO, too ;-) +--- Log closed Thu May 19 11:16:26 2016 -- cgit v1.2.3