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/20150826-io-chatlog | 267 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 267 insertions(+) create mode 100644 wiki/Chat_log/20150826-io-chatlog (limited to 'wiki/Chat_log/20150826-io-chatlog') diff --git a/wiki/Chat_log/20150826-io-chatlog b/wiki/Chat_log/20150826-io-chatlog new file mode 100644 index 0000000..cf64c17 --- /dev/null +++ b/wiki/Chat_log/20150826-io-chatlog @@ -0,0 +1,267 @@ +--- Log opened Wed Aug 26 10:55:12 2015 +10:55 -!- wsa_ [~wsa@p4FE25F24.dip0.t-ipconnect.de] has joined #periperi +10:55 -!- Irssi: #periperi: Total of 8 nicks [0 ops, 0 halfops, 0 voices, 8 normal] +10:55 -!- Irssi: Join to #periperi was synced in 0 secs +10:55 < geertu> Hi Wolfram +10:55 < wsa_> heyho! +10:55 < shimoda> Hi Wolfram-san! +10:55 < horms> Hi + Bye Wolfram +10:56 < wsa_> bye? +10:56 < wsa_> No IO Meeting for you Simon? +10:56 * horms goes home +10:56 < horms> oh we have a meeting? +10:57 < horms> sorry i missed that +10:57 < horms> i'll run home and log back in +10:57 < wsa_> ok +10:57 < wsa_> :9 +10:57 < wsa_> :) +10:57 < wsa_> hi geert, shimoda-san! +10:59 < shimoda> hi! +11:01 < geertu> hi all +11:01 -!- horms [~horms@reginn.isobedori.kobe.vergenet.net] has quit Ping timeout: 244 seconds +11:01 < wsa_> so, simon will come back soon +11:01 < wsa_> is morimoto-san around? +11:02 < morimoto> Hi +11:02 < geertu> His bot is alive ;-) +11:02 < wsa_> :) +11:02 < shimoda> :) +11:03 < wsa_> a meeting bot +11:03 < wsa_> i guess that's a way to make $$$ +11:03 < wsa_> ;) +11:04 < geertu> My former project manager was once convinced I was in a phone conf call and said things, while I wasn't there ;-) +11:05 < geertu> With IRC it's even easier... +11:05 -!- horms [~horms@121-80-213-238f1.hyg1.eonet.ne.jp] has joined #periperi +11:05 < geertu> welcome back +11:05 < horms> sorry about that. for some reason i hadn't added this meeting to my schedule +11:05 < wsa_> geertu: I hope you said good things while you weren't there +11:06 < horms> i said "I have a meeting" ! +11:06 < horms> + hello to someone in the elevator +11:06 < wsa_> well, you are here nonetheless +11:06 < wsa_> so fine +11:07 < wsa_> and I will make sure I get confirmations for the meeting dates from all of you explicitly ;) +11:07 < wsa_> so, let's start I'd say +11:07 < wsa_> I see two topics +11:08 < wsa_> 1) Gen3 status update +11:08 < wsa_> 2) SCIF +11:08 < wsa_> and miscellaneous +11:08 < wsa_> any other topics from you? +11:09 < horms> not from me +11:10 < wsa_> then let's start with Gen3 +11:10 < wsa_> we can add new topics as they come +11:11 < wsa_> is there an update to the timeline when which devices need to be enabled when? +11:11 < horms> i can give an etheravb update of sorts; when you are ready +11:12 < wsa_> i did not get any feedback to my I2C patches, I guess everyone was busy doing core bringup... +11:12 < shimoda> The current plan is, SPI needs by 1th Sep. but, i don't know it can be used on gen3 board +11:12 < morimoto> This is different topics, but now I could use GPIO and I2C on Gen3 board. it is not included on v7, but I can add it on v8 +11:13 < geertu> That's MSIOF? +11:13 < shimoda> geertu: yes +11:13 < geertu> The only usable point so far is SoftSW +11:13 < geertu> Haven't checked pinctrl and EXIO yet +11:13 < geertu> BTW, the pintrl part in the datasheet is "difficult" +11:13 < wsa_> morimoto-san: you mean I2C native or I2C using GPIO? +11:14 < geertu> There are no tables like on Gen2 (I want to add HSCIF pinctrl) +11:14 < geertu> Or am I missing something> +11:14 < geertu> Still, SoftSW would require soldering +11:15 < geertu> Or the CP[1-4] test points. +11:15 < shimoda> usb2 host needs by 25th August, and renesas-drivers-2015-08-25-v4.2-rc8 works correctly +11:15 < morimoto> geertu: Gen3 datasheet was exchanged. yes it is difficult to read +11:16 < wsa_> so I2C is closed case now? +11:17 < wsa_> and GPIO? (which is requirement for EtherAVB) +11:17 < morimoto> wsa_: I'm using renesas,i2c-r8a7795 for i2c. i2c driver seems working. I could chance to use gpio-i2c too +11:17 < wsa_> well, i am fine with renesas,i2c-r8a7795. that's enough for me. +11:17 < wsa_> nice +11:18 < morimoto> wsa_: please keep i2c/gpio + 1week. I need to check it on new board. (I used old and broken board) +11:18 < wsa_> ok +11:19 < shimoda> HSCIF and USB func need by 22th Sep. +11:20 < geertu> Working on HSCIF, now I have a serial link to DEBUG SERIAL-1 +11:20 < geertu> SCIF1 on those pins works +11:21 < shimoda> nice! +11:21 < geertu> shimoda: Or is MSIOF the slave-only DRIF? +11:21 < wsa_> so, the next work items would be: SPI (but we don't have proper HW setup currently) and testing I2C/GPIO with the new board (morimoto-san) +11:21 < wsa_> usb works +11:22 < morimoto> I didn't try ii2 +11:22 < morimoto> but i2 +11:22 < morimoto> i2c +11:22 < morimoto> please care ii2 +11:23 < morimoto> s/ii2/iic/g +11:23 < shimoda> geertu: just MSIOF, not DRIF +11:24 < wsa_> morimoto: what do you mean by care? +11:25 < morimoto> Gen3 has I2C/IIC. I can try I2C, but I don't try IIC +11:25 < horms> care -> please look after it +11:25 < wsa_> by the way, does magnus have a "new board" or an "old broken board" in his lab? +11:26 < morimoto> His one is OK board +11:26 < morimoto> but he is using old uboot +11:26 < wsa_> good +11:26 < morimoto> Ah, but 1 note. +11:26 < wsa_> morimoto: do you have a patch for adding all i2c to the r8a7795 dtsi or shall I make one? +11:27 < morimoto> His one has 16.000Mhz as extal clock. +11:27 < morimoto> new board has 16.666HMhz +11:27 < morimoto> First plan was 33.333MHz, but... +11:28 < morimoto> wsa_: I have it, I can send it in v8 +11:28 < wsa_> ok +11:28 < morimoto> wsa_: it is including PFC / MSTP too +11:28 < wsa_> great +11:28 < geertu> Before someone starts writing the SCIF/HSCIF dtsi patch, I have it +11:29 < morimoto> Mine has only SCIF2, can I update it to your one? or do you send it as incremental patch ? +11:29 < morimoto> in v8 +11:29 < geertu> I will send an incremental patch +11:29 < wsa_> is there some schedule when to use magnus gen3 board? +11:29 < morimoto> OK +11:30 < wsa_> i assume core group needs it a lot? +11:30 < wsa_> or first come first serve? :) +11:30 < geertu> It's mutex is kept on #periperi +11:30 < geertu> Currently I have it ;-) +11:30 < wsa_> ok +11:31 < wsa_> anything else related to gen3? +11:32 < horms> yes +11:32 < horms> the bsp team have an etheravb patchset, which i now also have +11:33 < horms> it depends on gpio which ulrich is handling; there were patches for that too which he now has +11:33 < horms> i plan to test the patchset in the near future +11:33 < horms> a blocking item is that it has some non-trivial driver changes that we would like to check for regresssions on 32bit +11:33 < horms> which requires testing +11:33 < wsa_> i see +11:33 < horms> i believe shimoda-san will handle that when he has time +11:34 < wsa_> are the changes needed for basic support or for the extended feature set? +11:34 < horms> in the short term we may be able to get 100Mbit working and on a ML +11:34 < shimoda> horms: yes, i will try that on alt + expantion board +11:34 < horms> there is some special sauce required for futher support; a documentation update is a dependency there +11:34 < shimoda> s/expantion/expansion/ +11:35 < horms> wsa_: for basic support, i believe +11:35 < horms> i'll double check that +11:35 < horms> basucically i think we are talking about crash-bugs on 64bit +11:35 < wsa_> i see +11:36 < horms> probably the best things is for me to test things on 64bit, polish things up and post an rfc +11:36 < horms> with no special sauce +11:36 < horms> so i nominate that as my near term goal +11:36 < wsa_> sounds reasonable +11:37 < horms> one question i have; i assume the main use of this to us is nfs root. but that also seems not very useful to the people in europe +11:37 < horms> which is a good portion of us +11:37 < horms> am I reading things right? +11:37 < horms> shimoda: thanks +11:38 < geertu> At least NFS would allow to import/export files in the initrd +11:38 < wsa_> I mainly use initramfs, but my userspace needs are not very high... +11:38 < horms> geertu: good point +11:38 < geertu> But from my experience, just mounting NFS and writing a small file takes a minute +11:38 < horms> anyway, i'll work on it +11:39 < geertu> Right now I'd like to copy my HSCIF1 overlay dtbo to my initrd ;-) +11:39 < geertu> I could modify the initrd +11:39 < geertu> But for exporting, that doesn't work +11:39 < horms> so you tftp things once you are in usespace; in an ideal situation +11:39 < geertu> Probably we can set up an NFS root on the management VM, too? +11:40 < horms> probably +11:40 < geertu> AH, TFTP is an alternative +11:40 < geertu> Probably faster +11:40 < horms> i believe that is magnus's preferred method for himself +11:40 < wsa_> Yup, TFTP does it for me. Also, less to setup +11:40 < horms> but i agree a more local nfs server would be useful +11:41 < geertu> And we already have tftp setup on the management vm +11:41 < horms> but i think its somewhat off topic for this meeting +11:41 < geertu> yep +11:41 < horms> i think thats all for etheravb from me +11:41 < wsa_> thanks, simon +11:42 < wsa_> so, let's move to SCIF +11:42 < geertu> I submitted a new patch series last week +11:42 < wsa_> geertu has sent an awesome new series (actually two) which make DMA work now +11:42 < wsa_> so \o/ +11:43 < horms> yay indeed +11:43 < geertu> Major change is submitting the RX DMA request from the RX interrupt, which is what many other serial drivers do +11:43 < geertu> + lots of cleanups, and reordering of the patches to make more sense +11:44 < wsa_> did that give you any more insights about how all this affects non-Rcar SCIF and how/what could be backported to the BSP? +11:44 < geertu> I have to do more testing, though, as you may receive bogus data during overrunrs +11:45 < geertu> Non-R-Car SCIF* may be an issue +11:46 < geertu> We don't have DT-based DMA on any other ARM platform +11:46 < shimoda> wsa_: I think we don't need backport to the R-Car Gen2 BSP because it's maintenance mode +11:46 < geertu> kzm9g/armadillo/ape6evm use SCIFA for console, though +11:47 < geertu> in -legacy they had some DMA support, but it was not enabled for SCIFA (and didn't work when I tried to enable it on Armadillo) +11:47 < horms> do you have access to any of those boards? +11:47 < geertu> I have kzm9g/armadillo/ape6evm +11:47 < horms> a flush! +11:47 < geertu> So PIO SCIFA keeps on working +11:49 < geertu> I could try SCIFA DMA on e.g. Armadillo if I (1) revive -legacy and (2) try to fix the legacy shdma driver for modern residue handling (untested RFC patch posted) +11:49 < geertu> But in the long term, that's insignificant +11:49 < wsa_> i agree +11:49 < geertu> SCIF DMA was originally written for SH (sh7724?) +11:50 < geertu> hut never wired up in the sh platform code, so there are no users. +11:50 < geertu> s/hut/but/ +11:50 < geertu> So for now R-Car Gen2 (M2/koelsch) is what we have to work with +11:51 < wsa_> I think we are fine as long as PIO keeps working +11:51 < geertu> And Gen3 soon +11:51 < wsa_> on the old platforms +11:51 < wsa_> yes +11:51 < wsa_> is there any gain if I test them on lager? +11:51 < geertu> I think Shimoda-san already did, at least with the previous version +11:51 < wsa_> except that more testing is always a gain ;) +11:52 < geertu> On Koelsch, I test SCIF/SCIFA/SCIFB/HSCIF on EXIO +11:52 < shimoda> geertu: yes, I tested v2 version on lager. but i don't test v3 yet. +11:53 < geertu> BTW, SCIFB receive doesn't seem to work on one of the EXIOs. It may be a bug in pinctrl (code/doc) +11:53 < wsa_> so, gen2 bsp is in maintenance mode and gen3 bsp could pick up geert's patches as is? +11:54 < wsa_> or do we still have a version gap? +11:54 < wsa_> we talked last time about it +11:54 < geertu> I think Magnus would like the BSP team to switch over once my series is "finished" +11:54 < wsa_> yes, we would all like that, I assume :) +11:55 < horms> that makes sense to me unless there is a pressing need for it in the gen3 bsp in the near term +11:56 < wsa_> sounds good, i don't see any obstacles in how to go forward... +11:56 < shimoda> wsa_: about gen3 bsp plan, it uses upstream version (renesas-driver.git) until end of next year +11:57 < wsa_> next year? wow! +11:58 < wsa_> and yay! +11:58 < wsa_> so, that's it for SCIF? +11:58 < shimoda> oops, gen3 bsp target is "LTSI-2017". so, this might mean mid of next year +11:59 < geertu> I have one more question +11:59 < shimoda> anyway, bsp team looks upstream in this year +11:59 < geertu> What's up with the IRDA MSTP clock and SCIF2? +11:59 < wsa_> mid-2016 is still good IMO +12:00 < morimoto> IrDA on Gen3 is secret feature. +12:00 < morimoto> It will be removed from datasheet +12:00 < morimoto> and, please don't ask it on ML :) +12:01 < geertu> So there's a hole in the SCIFx MSTP2 range, and SCIF2 uses a clock from MSTP3 +12:01 < morimoto> About MSTP, IrDA and SCIF2 is using same one +12:02 < morimoto> Yes +12:06 < wsa_> ok +12:07 < wsa_> misc +12:07 < wsa_> i pulled in my i2c-slave updates +12:08 < wsa_> and I am currently scoping the rcar-i2c core to understand all the repeated start/NACK/auto retransmission issues we were seeing +12:09 < horms> shimoda san sent me a Kconfig fix for our ethernet drivers which has been merged +12:09 < wsa_> shimoda-san also got his PWM driver accepted \o/ +12:10 < horms> https://git.kernel.org/cgit/linux/kernel/git/davem/net.git/commit/?id=83bc805bff89854e4c81c67633a52ce1015b1502 +12:11 < wsa_> that was my misc part +12:11 < shimoda> wsa_: about PWM, yes, the PWM driver seems accepted but i cannot find the pwm maintainer git :) +12:12 < wsa_> yes, he said he will do the final changes but it is not clear if he actually did them +12:12 < geertu> git://git.kernel.org/pub/scm/linux/kernel/git/thierry.reding/linux-pwm.git +12:12 < geertu> it's part of renesas-drivers +12:14 < shimoda> oops, i meant that i could not find the PWM driver commit in his repo +12:16 < wsa_> so, final question would be when to meet next time +12:17 < wsa_> to be honest, I don't think we need the next meeting in two weeks. what about three weeks? +12:18 < wsa_> sorry, just found one thing +12:18 < wsa_> shimoda: what is the status about USB+IPMMU issues? +12:18 < horms> so far my evening schedule is open in both 2 and 3 weeks +12:18 < geertu> shimoda: It's indeed not included yet, probably because it's meant for v4.4 +12:19 < geertu> My schedule is open in both 2 and 3 weeks, except on Wednesdays +12:20 < shimoda> wsa_: no update for now +12:20 < shimoda> geertu: thank you for the checking +12:21 < geertu> shimoda: we have to have a bit more patience, as not many patches will be accepted between now and v4.3-rc1 +12:22 < geertu> (this also applies to the Gen3 patch stack!) +12:22 < morimoto> About next meeting, I don't want 16th Sep, other days are OK +12:22 < morimoto> 16th and 23th Sep +12:23 < wsa_> geertu: well, thierry could apply the new-driver-rule. but it is his choice, of course... +12:23 < horms> btw, 21, 22, 23 are holidays in the land of the rising sun +12:23 < wsa_> i suggest Sep 15th +12:23 < wsa_> same time +12:23 < morimoto> 15th Sep is OK for me +12:24 < horms> This time its in my schedule properly :) +12:24 < shimoda> geertu: I got it +12:24 < wsa_> \o/ +12:24 < shimoda> wsa_: 15th is OK to me +12:26 < geertu> booked +12:26 < morimoto> same here +12:26 < wsa_> great +12:27 < wsa_> that's it I assume +12:27 < wsa_> thank you! +12:27 < horms> wsa_: thanks +12:27 < geertu> thx +12:27 < morimoto> Thanks +12:27 < morimoto> Bye x2 +12:27 < wsa_> have a great evening/rest of day! +12:28 < shimoda> thank you, bye! +12:28 -!- shimoda [~shimoda@relprex2.renesas.com] has quit Quit: WeeChat 0.4.2 +12:28 < geertu> bye +--- Log closed Wed Aug 26 12:30:04 2015 -- cgit v1.2.3