diff options
Diffstat (limited to 'wiki')
-rw-r--r-- | wiki/Chat_log/20200319-core-chatlog | 106 |
1 files changed, 106 insertions, 0 deletions
diff --git a/wiki/Chat_log/20200319-core-chatlog b/wiki/Chat_log/20200319-core-chatlog new file mode 100644 index 0000000..701ead2 --- /dev/null +++ b/wiki/Chat_log/20200319-core-chatlog @@ -0,0 +1,106 @@ +09:22 < geertu> Welcome to today's Core Group Chat Meeting! +09:22 < geertu> Agenda: +09:22 < geertu> 1. Status Updates +09:22 < geertu> 2. Discussion Topics +09:23 < geertu> Topic 1. Status updates +09:23 < geertu> A) What have we done since last time: +09:23 < geertu> Marek worked on OpenOCD (patch resync), U-Boot (SDHI performance, +09:23 < geertu> Blanche and SMC911x DM conversion), and Linux (PCIe suspend). +09:23 < geertu> Shimoda-san submmited rcar-usb2-clock-sel patches, and discussed the +09:23 < geertu> plan for the R-Car V3U bootloader. +09:23 < geertu> Ulrich sent v2 of the ignore-unused clock series, and reviewed patches. +09:23 < geertu> Geert converted the CPG/MSSR DT bindings to json-schema, upported +09:23 < geertu> CPUIdle for R-Car M3-N and E3, sent a second batch of pull requests for +09:23 < geertu> v5.7, and reworked the GPIO Aggregator. +09:24 < geertu> B) What we plan to do till next time: +09:24 < geertu> Marek plans to continue SMC911x work in U-Boot for Blanche. +09:24 < geertu> Shimoda-san plans to convert DT binding docs to json-schema, and will +09:24 < geertu> continue internal discussions about BL31 code for R-Car V3U. +09:24 < geertu> Geert will post v6 of the GPIO Aggregator, and hopes to resume QEMU GPIO +09:24 < geertu> virtualization. +09:25 < geertu> C) Problems we have currently: +09:25 < geertu> Shimoda-san regrets not being able to share the R-Car V3U Hardware +09:25 < geertu> User's manual. +09:25 < geertu> Covid-19 has reached EuroPeri space. +09:25 < geertu> ---EOL--- +09:25 < geertu> Anything I missed? +09:26 < geertu> shimoda22: What is ICUMXA? +09:27 < shimoda22> This is a CPU core which R-Car V3H has (section 68A) +09:27 < shimoda22> IICU, this is RH850 CPU core so that we don't have any free compiler.. +09:29 < geertu> Thx, forgot to search in the datasheet ;-) Google didn't really know +09:29 < geertu> RH850 is different from V850? +09:31 < shimoda22> perhaps same. V850 is NEC and RH850 is Renesas +09:32 < uli_> there's a surprisingly long wikipedia article on that: https://en.wikipedia.org/wiki/V850, including a section on "ancient prom writers" :) +09:32 < geertu> yep +09:32 < geertu> and gcc has v850 support +09:33 < Marex> Variant(s) V850 Family, +09:33 < Marex> RH850 Family +09:33 < Marex> well ... +09:34 < shimoda22> Oh, I didn't know gcc supports RH850. +09:35 < geertu> perhaps not the latest variant +09:35 < geertu> And it might have been removed in recent gcc versions +09:35 < wsa> Marex: I don't see this patch in next "[RFC] PCI: rcar: Fix incorrect programming of OB windows" although you reviewed it. Could you give a ping? +09:36 < geertu> At least gcc-9.1.0 still has v850-specific options +09:36 < geertu> Now, is that of any help? ;-) +09:39 < shimoda22> I don't know :) especially, the hardware manual doesn't say ICUMXA is related to RH850. +09:40 < Marex> wsa: sure, although I suspect that if Lorenzo took forever to review patches before this current mess, he would just ignore patches by now +09:41 < shimoda22> about pci, i reviewed other patches about pcie endpoint support. +09:41 < wsa> Marex: I see. Well, might be worth pointing out that the patch is not RFC anymore +09:42 < wsa> shimoda22: yes, great! I just updated this task in periject :) +09:42 < Marex> wsa: what surprises me more is that the original author is literally sitting with Lorenzo in the same office +09:42 < Marex> oh well +09:43 < Marex> wsa: I guess that unit of ARM is a mess +09:43 < wsa> well, then pinging Andrew might make more sense? ;) +09:44 < Marex> wsa: um, that's the author of the patch, you already asked me to ping him again +09:44 < kbingham> wsa, Is that Andrew Murray? +09:44 < kbingham> wsa, Andrew Murray left ARM. +09:44 < wsa> Marex: Initially, I was aiming for the maintainer, but if the author sits next to him, the latter might make more sense... +09:45 < wsa> ... that is, unless he left the company +09:45 < wsa> kbingham: where to? +09:45 < Marex> kbingham: didn't he join like a few months ago ? +09:46 < Marex> morimoto: btw thank you for the hint to buy instant ramen, I bought a large-ish crate +09:46 < morimoto> Hehe :) no worry +09:46 < kbingham> wsa, He took some external contract or something. +09:47 < kbingham> wsa, I can hire him if you (or rather Renesas) want me to get him to work directly on Renesas :-) +09:48 < wsa> kbingham: does he have any experience in I2C dynamic address assignment? :D +09:48 < kbingham> Marex, Well techincally he 're-joined' ... I don't know why he didn't want to stay there. Maybe he preferred being independent. +09:49 < kbingham> wsa, I think he's more experienced in PCI topics :) But I'm sure he's done I2C work too :) +09:49 < geertu> Topic 2. Discussion Topics +09:50 < geertu> Any other topics to discuss? +09:50 < wsa> kbingham: it is basically the same anyhow +09:50 < kbingham> wsa, :-D +09:50 < wsa> uli_: any idea why my testing of ignore-unused clocks fail? +09:51 < geertu> wsa: Because the clock was not enabled in the first place at all? +09:51 < wsa> geertu: I started it in U-Boot +09:51 < wsa> If I don't hack the code to bail out and forbid the kernel to ping the watchdog, it will reboot +09:52 < geertu> wsa: OK. +09:53 < geertu> Looking at drivers/clk/clk.c, CLK_IGNORE_UNUSED just prevents disabling the clock. +09:53 < wsa> that's what we want +09:53 < wsa> IIUC +09:53 < geertu> But only in the clk_disable_unused late_initcall +09:54 < geertu> it can still be disabled elsewhere +09:54 < geertu> I.e. manually, or through Runtime PM? +09:54 < geertu> WDT probe fails -> detached from PM Domain -> clock disabled? +09:54 < uli_> what i have checked is whether it will be temporarily disabled when the probing is deferred +09:54 < uli_> and that works +09:54 < wsa> yup +09:54 < wsa> uli_: ok +09:55 < geertu> works == not temporarily disabled? +09:55 < uli_> yes +09:57 < geertu> Looks like cpg_mstp_clock_endisable() needs some debug code to track what happens to the WDT clock +09:57 < patersonc> wsa: geertu: Fabrizio left us in December. But in a twist of fate he's actually going to be re-joining us b/April. Huzzah! +09:58 < wsa> uli_: can you do that? +09:58 < uli_> sure +09:58 < patersonc> But this time he'll have a @renesas.com email address instead of a @bp.renesas.com address :) +09:58 < wsa> uli_: thanks! +09:58 < wsa> patersonc: leaving and rejoining is a thing these days? :) +09:59 < patersonc> wsa: It's a long complicated story! +09:59 < geertu> patersonc: Great! Looking forward to seeing a patch for .mailmap +09:59 < geertu> wsa: I'll rejoin public life in 5 weeks or so (or not) +10:00 < wsa> patersonc: anyway, glad he will be back! +10:00 < wsa> geertu: isn't it more like public life will rejoin us? +10:00 < patersonc> Same here +10:01 < geertu> Let's finish the Core meetinng with this great news. Unless someone has something to add? +10:02 < patersonc> Sorry to interrupt! +10:02 < geertu> patersonc: np, we like positive news +10:03 < geertu> Thanks for joining, and have a nice continued day! |