09:17 <qschulz> CounterPillow: if it's not there, nobody cares about warnings and they don't get fixed :)

10:09 <CounterPillow> more like if it's there distributors will patch it out because it'll always be broken because upstream only tests with ancient toolchains and their patch submission workflow is horrible and also I'm not going to fix warnings in code that I don't know when I'm just trying to compile the damn thing

10:51 <dsimic> qschulz: I had a look at your overlay patch series, but I still need to re-read everything carefully, including the discussion in v2

10:52 <dsimic> I'll do that for sure, alas I've got some "IRL stuff" to take care of today

10:52 <qschulz> dsimic: AFAIU, the plan is to merge this early next week :)

10:52 <dsimic> let's also ping mmind00 and Kwiboo about this ^^^

10:52 <qschulz> well, mmind00 would be the one merging them so I'm sure he knows :)

10:53 <dsimic> there are still a few days before the next week, so it should be fine from the timing standpoint :)

10:54 <qschulz> dsimic: there's also FOSDEM this week-end, so that may make some people's week-end very busy :)

10:54 <dsimic> sure... I referred to myself :)

10:56 <dsimic> oh, BTW, as a nitpick for the patch descriptions... using "in order to..." is generally discouraged, and saying simply "to ..." instead is preferred

10:57 <qschulz> dsimic: discouraged by whom?

10:57 <dsimic> by the English style police, so to speak :)

10:59 <dsimic> to clarify, I spent about a couple of years contributing to English Wikipedia, which made me quite familiar with various style rules and guidelines

11:01 <qschulz> dsimic: what one website decides would be the best doesn't necessarily means it's what's preferred in the language (also American English/British English, etc...) :)

11:01 <dsimic> indeed, there are many differences between the language variants

11:02 <dsimic> however, to my knowledge, what I suggested above applies to (all) major regional variants, FWIW

11:02 <qschulz> dsimic: I also tend to be conservative when advice comes from non-native speakers which my brain believes is the case without knowing you :)

11:02 <qschulz> If you had some link(s) about this, it would have been nice to send with it, but a bit of googling seems to go in your direction

11:02 <qschulz> so thanks!

11:02 <dsimic> yes, I'm not a native speaker

11:03 <qschulz> not sure my muscle memory is going to forget it soon enough but being aware of it will maybe help in the long run!

11:03 <dsimic> but I've spent many years trying to reach near-native level, so I even dare to provide suggestions :)

11:04 <qschulz> hehe

11:04 <qschulz> (i'm not complaining about the suggestion, i'm "complaining" about the source-less suggestion :) )

11:04 <qschulz> I do actually appreciate being corrected

11:05 <dsimic> ah, I spent a few months learning the Wikipedia's Manual of Style (MoS), which is mainly derived from the Chicago Manual of Style, so the best I can provide, when it comes to references/links, is a link to the MoS :)

11:06 <dsimic> because I no longer remember where was something mentioned in the MoS, which is also quite large and split between different pages

11:07 <dsimic> I'm glad that you're fine with the corrections... being corrected is one of the ways to learn, and I used to be corrected a lot on English Wikipedia in my early days of contributing there

11:07 <dsimic> which I highly appreciated, of course

11:10 <dsimic> see also https://en.wikipedia.org/wiki/Wikipedia:Manual_of_Style

11:10 <dsimic> and https://en.wikipedia.org/wiki/Wikipedia:Manual_of_Style/Contents

11:15 <qschulz> dsimic: hihi, there's one "in order to" in the Manual Of Style

11:15 <dsimic> dang! :D

11:16 <dsimic> oh, and it's "Manual of Style", with the lowercase "o", BTW

11:16 <qschulz> like System of a Down, noted

11:17 <dsimic> capitalizing minor words in titles is a common "modern" mistake, which perhaps originates from the careless use of the PHP's ucwords() function on the web :)

11:18 <qschulz> dsimic: everything's PHP's fault

11:18 <qschulz> or JavaScript nowadays

11:18 <dsimic> yeah, PHP did a lot of good stuff, but also degraded some other stuff

11:27 <beeble> qschulz: https://paste.debian.net/1347696/

12:06 <dsimic> beeble: ha, that's a good example

12:07 <dsimic> saying "to use Device Tree overlays, specify fdtoverlay_addr_r" would be better

12:07 <dsimic> also note lowercase "o"

12:08 <dsimic> or maybe "fdtoverlay_addr_r must be specified to use Device Tree overlays"

12:08 <dsimic> passive voice is good there, because it puts emphasis on fdtoverlay_addr_r

12:09 <qschulz> It's actually Devicetree Overlay

12:09 <qschulz> https://www.kernel.org/doc/html/latest/devicetree/overlay-notes.html

12:09 <dsimic> hmm

12:09 <qschulz> ah, maybe a case of capitalization again

12:09 <qschulz> later it's said Devicetree's overlay

12:09 <dsimic> I don't think it should be capitalized

12:09 <qschulz> this DT stuff is a bit of a mess

12:10 <qschulz> U-Boot uses different terms too

12:10 <qschulz> FDT and FDTO for example

12:10 <qschulz> (Flattened ...)

12:10 <dsimic> yeah, it could use some copyediting :)

12:10 <qschulz> some peeople use DTBO or DTO

12:10 <qschulz> in the kernel

12:10 <qschulz> Devicetree/Device Tree, I never know

12:11 <dsimic> it should be "Device Tree"

12:11 <dsimic> because if it's "Devicetree", then "DT" isn't an obvious abbreviation

12:12 <dsimic> using "DTBO" technically isn't correct, because it's actually "dtbo"

12:12 <dsimic> i.e. it isn't an abbreviation, but a file "extension"

12:12 <dsimic> s/file/filename/

12:12 <qschulz> DTO would though

12:13 <qschulz> DTO is a dtso I guess?

12:13 <robmur01> dsimic: style manuals are for journalism and formal publications, they are not absolute rules to be applied to less formal, conversational, writing, as commit messages often are ;)

12:13 <dsimic> yes, DTO is a rather fine abbreviation, although a bit hard to grasp, frankly :)

12:13 <dsimic> robmur01: true, but why not make commit messages better? they're our publication :)

12:14 <CounterPillow> focus on real work instead

12:14 <dsimic> real work and suboptimal commit messages don't go together, IMHO

12:16 <robmur01> I would certainly use "X and Y in order to Z" over "X and Y to Z" to emphasize an implied dependency unless it was blindingly obvious that X and Y are the only possible things leading to Z

12:16 <dsimic> qschulz: I'd say that DTO should mean Device Tree overlay, actually

12:17 <dsimic> robmur01: indeed, it isn't that "in order to" is prohibited, it actually fits very well in some places

12:20 <dsimic> qschulz: FWIW, I'd also say that "DT" is an umbrella quasi-term that may refer to the concept, the dts files or dtb files, while "DTO" is another umbrella "term" that may refers to the concept, dtso and dtbo files

12:20 <dsimic> s/may refers/may refer/

17:21 <CounterPillow> Kwiboo: out of morbid curiosity, how'd you figure out the writel(0x3ffff800, 0x3ff803b0) rk3576 thing?

18:05 <chewitt> some people just think in code :)

21:13 <Kwiboo> CounterPillow: https://github.com/rockchip-linux/rkbin/commit/9b6f7661267cee51210e7850f83bbb695841594a hint at an issue in BootROM "BootROM load ddr.bin fail in SD card", the 3ff80000-3fffffff range is sram, a quick peek and it looked like only uart0 regs and sram was modified

21:15 <Kwiboo> 0x3ff803b0 is probably in the BootROM stack, TPL start at 0x3ff81000, and 0x3ffff800 is possible an end address or something, so issue is probably how many 4k blocks is read into sram before executing

21:16 <CounterPillow> interesting, thank you

21:17 <Kwiboo> applying the workaround seem to work for both sd-card and emmc, however emmc is not affected only sd-card boot, only have an armsom sige5 so have not been able to test with spi flash boot

21:19 <CounterPillow> just received an armsom sige5 today as well, actually used your u-boot wip branch for it which worked beautifully, still fighting with some networking gremlins though. Ultimate goal is to just have a small mainline-based u-boot with network support on the eMMC that automatically fetches a kernel and initramfs over HTTP or whatever to make development a lot easier

21:24 <Kwiboo> cool and great, I may not have tested ethernet enough, only synced with the code in mainline Linux compared to mmind00's series, mostly tested my board booted into U-Boot CLI

21:27 <Kwiboo> also suggest you avoid using the BOOT1_PARAM WORD_2=0x100 from https://github.com/rockchip-linux/rkbin/commit/cbbc6868221fb416d4f0d86a10e493dbbbc1f117 they are used to enable the 1GHz hp timer, possible done in BootROM or params is passed to TPL

21:28 <Kwiboo> using the 1ghz rate for hp timer seem to require changes to u-boot timer handling

21:36 <CounterPillow> kk, thanks for the heads up

21:39 <CounterPillow> I did manage to get the PHY to respond with some additional config fiddling and then do pings, but u-boot's wget wasn't too happy. I'm about 50:50 on whether that's due to firewalld being obtuse nonsense or whether the rx timeout I got in u-boot was due to something more sinister

21:43 <Kwiboo> one issue could be that u-boot will always set phy delay enabled, but use 0 as delay value with it should be disabled, should possible be changed both in linux and u-boot

21:45 <Kwiboo> without having access to a TRM it is hard to verify what is being configured in hw :-)

21:46 <CounterPillow> I'll check on Monday whether I do have access to the TRM and can look at whether the GMAC is included in that