diff options
Diffstat (limited to 'wiki/Chat_log/20160411-io-chatlog')
| -rw-r--r-- | wiki/Chat_log/20160411-io-chatlog | 195 |
1 files changed, 195 insertions, 0 deletions
diff --git a/wiki/Chat_log/20160411-io-chatlog b/wiki/Chat_log/20160411-io-chatlog new file mode 100644 index 0000000..cea2140 --- /dev/null +++ b/wiki/Chat_log/20160411-io-chatlog @@ -0,0 +1,195 @@ +--- Log opened Mon Apr 11 09:55:04 2016 +09:55 -!- wsa_ [~wsa@p4FE259C0.dip0.t-ipconnect.de] has joined #periperi-io +09:55 -!- Irssi: #periperi-io: Total of 6 nicks [1 ops, 0 halfops, 0 voices, 5 normal] +09:55 -!- Irssi: Join to #periperi-io was synced in 0 secs +09:55 < wsa_> hiya +09:55 < wsa_> wow +09:55 < horms> Hi wsa_ +09:56 -!- shimoda [~shimoda@relprex1.renesas.com] has joined #periperi-io +09:56 < wsa_> hi guys +09:56 < shimoda> hi +09:56 < morimoto> hi +09:56 < neg> morning +09:56 < uli___> hi +09:56 < wsa_> nice to see you again! +09:57 < wsa_> so, ELC was good? +09:57 < neg> yes it was nice +09:58 < neg> good for me who are new to meet people in the flesh +09:58 < morimoto> yes I think so +09:58 < wsa_> neg: yes, hallway track is usually my fav, too :) +09:59 < wsa_> okay, then, let's get started +10:00 < wsa_> I didn't notice updates for the todo file, please tell me if I missed something +10:00 < wsa_> from my side, SDHI support for UHS-I/SDR50 was merged +10:00 < wsa_> and WDT support should show up in linux-next any day +10:01 < wsa_> SDHI probably needs DMA updates to achieve full speed +10:01 <@geertu> wsa: Have you tried Günter's restart series? +10:01 < shimoda> oh, it's nice! I (or Jinso) will test the UHS-I/SDR50 :) +10:01 < wsa_> I guess this is why we have this in the todo file: +10:01 < wsa_> SDHI,?,noplan,?,Confirm & check + tune up DMA transport speed on Gen2 +10:02 < wsa_> geertu: not yet, but will surely do today or tomorrow. It looks great to me. +10:03 < wsa_> shimoda: I get transfer rates of about 35MB/s while the card should do 90MB/s. It does 75MB/s on a Tegra-board. +10:03 <@geertu> wsa: Then the remaining question is when (if ever) to make the RWDT restart priority lower than the PSCI priority ;-) +10:04 < wsa_> geertu: yes +10:04 < wsa_> and how to activate the SWDT, because I still don't know how/where/when to change a register in secure-mode +10:05 < wsa_> oh, how impolite of me. I missed something! +10:05 < shimoda> wsa_: i see. I guess PIO is impossible to be 90MB/s on Gen3. +10:05 < wsa_> let's welcome Niklas to our group who is contracted to do IO work in Q2. +10:05 < wsa_> I didn't know until today :) +10:05 <@geertu> neg: Welcome to I/O! +10:06 < neg> thank you +10:06 < wsa_> uli___: are you also contracted for IO work or are you just interested? :) +10:06 < uli___> i'm in +10:06 < uli___> contractually, that is +10:07 < wsa_> shimoda: I agree. I first want to add eMMC support and then port the Gen3 DMA driver. +10:07 < wsa_> uli___: then welcome, too! +10:07 * geertu is not contracted for I/O work +10:07 * horms neither +10:07 < wsa_> geertu: then, bye ;))))) +10:08 < wsa_> okay, we need to talk about this contracting soon +10:08 < wsa_> in a second +10:08 <@geertu> My I/O status: Not much happened on the SCIF front due to SYSC and Easter +10:08 <@geertu> Will resume SCIF this week +10:09 < horms> geertu: quick question. am I right in thinking there is no QSPI present in Gen 3? +10:09 < shimoda> wsa_: add eMMC support first and then DMA support are good to me +10:09 < wsa_> the not surprising thing about SDHI: it sometimes really needs refactoring before adding something new to it. +10:09 < wsa_> shimoda: good +10:09 <@geertu> horms: There is. But It's for secure mode only. +10:09 < wsa_> then we will see how Gen3 DMA performs and where the bottlenecks are +10:10 < horms> geertu: ok. is it called something else in the data sheet? or absent? +10:11 < horms> wsa_: the sdhi driver has a long and cheqquered history +10:11 <@geertu> horms: It's called RPC +10:11 < horms> geertu: ok, i saw that one. thanks. +10:11 < wsa_> horms: i noticed :D +10:12 < wsa_> but i found an old TMIO datasheet which was not detailed but helped in understanding the "history" +10:12 < wsa_> and the different names for bits +10:13 < wsa_> okay, if there are no more updates to the todo, then let's talk about the situation with the contracts +10:14 < wsa_> so, i learned a lot of things today who is contracted and who is not +10:15 < wsa_> and, to be honest, i am a bit surprised to find ulrich and niklas here because they did great work for the multimedia group which is not short of tasks, is it? +10:16 < neg> my situation is that I got 5 days each for i/o, core, and multimedia (15 in total) during Q2 and ontop of that there will be additional contracts mainly for multimedia during Q2 as it looks now +10:16 < uli___> same here +10:17 < wsa_> ah, okay +10:17 < wsa_> geertu: and you are officially not contracted for IO? +10:18 < wsa_> horms: and you? +10:18 <@geertu> wsa: Indeed. Just Core Group Lead and Development +10:18 < wsa_> geertu: similar here, just with IO group, by the way +10:19 < horms> presently i am primarily working on maintainance and secondarily with Kaneko-san on upporting +10:20 < wsa_> horms: doing ethernet as a side-project hobby ;) +10:20 < horms> that has been an ongoing side project but of late its mainly been handled by Kaneko-san via the upporting work. +10:20 < wsa_> okay +10:21 < wsa_> so we need to find tasks for Uli and Niklas +10:21 < horms> iirc the main outstanding issue there is reducing dma descriptor usage for the aligned case +10:21 < horms> also Sergei sort of stepped up at some point, which imho is better than bad +10:22 < wsa_> technically, i think it would be nuts to move SCIF tasks away from Geert +10:22 < wsa_> horms: nice to hear that +10:23 < wsa_> at the the baud generator stuff is public and WIP anyhow +10:23 < wsa_> geertu: is this FIFO issue something which can be done by someone else +10:23 < wsa_> i mean reasonably +10:24 <@geertu> wsa: I assume there will be an "additional contract" in the future +10:24 <@geertu> so I don't expect to retire from I/O ;-) +10:24 < wsa_> \o/ +10:24 <@geertu> BTW, what do you mean by "WIP"? +10:24 < wsa_> ah, not baud generator +10:24 < wsa_> flow control pins +10:25 <@geertu> because BRG is in +10:25 <@geertu> There are some flushing patches in the BSP. and in Hamza's git tree +10:26 < wsa_> so, tasks I see for uli___ and neg I see from our todo-list: +10:26 < wsa_> Thermal,2016-06-30,plan,morimoto,Upstreaming For H3 +10:26 < wsa_> I2C,2016-06-30,plan,wolfram,Gen3 I2C DMA support +10:26 < wsa_> SPI,2016-05-31,plan,geert,Implement initial SPI slave prototype support for R-Car Gen2 +10:26 < wsa_> which all need reassignment +10:27 < wsa_> looks OK? did i miss something? +10:28 < morimoto> wsa_: RVC Khiem want to do Thermal for Gen3 +10:28 < morimoto> s/Khiem/Khiem-san/ +10:28 < wsa_> I see +10:28 < neg> I can start by looking at I2C DMA since I poke around with DMA anyhow +10:28 < horms> Not wanting to go around in circles, but is there any application of the SPI task to Gen 3? If not, I'm curious to know its motivation. +10:29 < wsa_> IIRC think it was a customer request? shimoda-san, do you recall more? +10:29 <@geertu> The other SPI on Gen3 is MSIOF, which has hardware issues that are supposed to be fixed in ES2, or in M3-W +10:30 < horms> Ok, so it might be used in the context of MSIOF on Gen 3? +10:30 <@geertu> And DRIF, which is sort-of half-duplex SPI slave +10:30 < wsa_> could we use Gen2 for the prototype? +10:30 <@geertu> That's what it says: "... for R-Car Gen2" +10:30 < wsa_> morimoto: can i add khiem-san to the todo file as "khiem"? +10:30 < horms> wsa_: no argument about prototyping on Gen 2 +10:30 < horms> I was just wondering what the end game was +10:30 < wsa_> geertu: got me there :D +10:31 < horms> neg: my notes tell me SYS-DMAC Supports I2C 0 to 6 and IIC-DVFS. v0.51e 17.1.1 +10:31 < morimoto> wsa_: that is good idea for me :) +10:31 < horms> neg: in case that is of any help +10:32 < neg> horms: thanks +10:33 < wsa_> uli___: how does implementing SPI slave support sound to you? that's a new thing for linux... +10:33 < horms> neg: i think the implication is that you can do dma to/from 40 bit addresses using the SYS-DMAC. I don't have notes of any other dma capabilities for I2C +10:33 < wsa_> geertu: would you be ok to reassign the task? +10:34 < uli___> i have no clue about spi, so i might as well as well have a look at it... +10:34 <@geertu> wsa: SPI? I think it may be a bit premature +10:35 < wsa_> geertu: what do you mean? +10:36 <@geertu> wsa: There will be additional contracts, right? +10:36 <@geertu> I already spent lots of investigation about SPI slave +10:36 < wsa_> shimoda: how is USB going? is there a place where you could need help? +10:38 < neg> horms: nice, the manual even have flow-charts for DMA operation :) +10:38 < horms> :) +10:39 < wsa_> geertu: i see +10:39 < shimoda> i am doing USB3-host, USB-PHY tasks. But, I forgot the "USB2-Func,?,noplan,shimoda,IPMMU issues" :) +10:42 < shimoda> and i have a problem about usb2 host with IPMMU (that is not listed the todo list yet) +10:42 < wsa_> shimoda: so, could the IPMMU tasks be done by someone else? +10:42 < wsa_> uli___: how does USB & IPMMU sound? +10:42 < shimoda> and I sent a question about it to Magnus-san +10:42 < shimoda> wsa_: it is nice idea +10:43 < uli___> i can look at that +10:43 < wsa_> \o/ +10:44 < shimoda> uli___: thanks! +10:45 < shimoda> Oh, i have a note about IPMMU of Gen3. +10:45 < wsa_> so, I'll assign usb2-func & IPMMU task to uli +10:45 < wsa_> and wait for shimoda-san if i should add a task about usb2-host & IPMMU +10:46 < shimoda> we need a new firmware which will be released on this month if we use Gen3 +10:46 < horms> v2.8.0? +10:47 < shimoda> horms: yes +10:48 < shimoda> the default setting of IPMMU is not suitable for linux and the register can be modified on secure-mode :) +10:48 < wsa_> I want a u-boot command to change registers in secure-mode :) +10:49 < wsa_> shimoda: can i set the usb3-host task to planned for v4.8? +10:49 < shimoda> wsa_: about usb2-host & IPMMU, I would like to get Magnus-san's reply +10:50 < wsa_> yes, i agree +10:51 < wsa_> i meant this one: +10:51 < wsa_> USB3-Host,?,noplan,shimoda,suspend problems +10:51 < wsa_> since you said you currently work on usb3-host tasks +10:51 < wsa_> or is it something else? +10:51 < shimoda> I see, i'm working the task now because Gen3 also has such an issue :) +10:52 < shimoda> I would like to ask HW team why this issue happen +10:53 < shimoda> so now i'm trying to get register dump and then pass it to hw team +10:54 < wsa_> i see +10:54 < shimoda> about workabout is simple, we just add XHCI_SLOW_SUSPEND quirk on Gen3, but we need a reason for upstreaming +10:55 < shimoda> s/workabout/workaround/ +10:57 < wsa_> ok +10:57 < horms> shimoda: we need to know which hw is effected and why? +10:58 < wsa_> uli___: I set the the milestone to v4.8. let me know if this is realistic after you get an overview of the problem +10:58 < shimoda> horms: yes, i will try to get such information +10:59 < horms> shimoda: thanks. fwiw that makes sense to me +10:59 < uli___> will do +10:59 < wsa_> shimoda: can you mail uli a description of the problem? +10:59 < wsa_> neg: I will give some introduction to the I2C DMA topic as well +11:00 < wsa_> OK, I think we are done then? +11:00 < shimoda> wsa_: sure +11:00 < neg> wsa_: thanks +11:01 < uli___> shimoda: thank you +11:01 < wsa_> we included three new members today! +11:01 < wsa_> the new contracts really shuffle things around ;) +11:02 < wsa_> last chance for news from your side +11:02 < horms> nothing from my side +11:03 < wsa_> Then, thank you all! It seems we found suitable tasks for everyone in the end. I hope everyone is happy. +11:03 < shimoda> nothing from me +11:04 < wsa_> then, see you next time :) +11:05 < uli___> see you +11:05 < neg> thanks all, see you all next time +11:05 < horms> likewise +11:05 < morimoto> wsa_: can you add this meeting log to Redmine ? +11:05 <@geertu> Thx, CU +11:06 < wsa_> morimoto: sure thing +11:06 < morimoto> Thanks. and bye +11:06 -!- horms [~horms@reginn.isobedori.kobe.vergenet.net] has quit Quit: Leaving +11:06 < shimoda> thank you, bye! +11:06 -!- morimoto [~user@relprex2.renesas.com] has left #periperi-io ["ERC Version 5.3 (IRC client for Emacs)"] +11:06 -!- shimoda [~shimoda@relprex1.renesas.com] has quit Quit: WeeChat 0.4.2 +11:12 -!- neg [~neg@unaffiliated/neg] has left #periperi-io [] +11:15 -!- geertu [~geert@d54C36A7B.access.telenet.be] has left #periperi-io [] +11:20 -!- Irssi: #periperi-io: Total of 2 nicks [0 ops, 0 halfops, 0 voices, 2 normal] +--- Log closed Mon Apr 11 11:36:19 2016 |