apritzel has quit [Remote host closed the connection]
apritzel has joined #linux-sunxi
<warpme>
apritzel: i remember you mention you compared a523 docs for gmac and they seemed very similar, if not identical, to the previous one (A64) - but simple using A64 compatible on a523 not works. I looked on orange 4a bsp (t527 sbc) dts and also on h728 vendor dts and i see both boards (for ge) are using sunxi-gmac-200. gmac-200 hw seems to have dedicated driver. Most recent ver. (0.3.0) i found is in avaota bsp code
<apritzel>
warpme: please remember to wear gloves and protective goggles ;-) Most BSP code I have seen is really awful, and especially the abstraction, driver design and code reuse is typically really bad
<apritzel>
but it might be good enough to take a peek and see how the "g200" differs from the previous ones
<apritzel>
so thanks for the pointer, I might have a look later
<warpme>
btw: for h728 x96 pro+ box with MAE0621 phy - it looks like 6.12 kernel has already mainlined maxio driver for this chip :-)
<apritzel>
oh really? what's the driver name? I couldn't find any quickly when grep'ing for either maxio or parts of the chip name
<apritzel>
warpme: also, do you know which "gmac-xxx" number the others SoCs refer to, for instance the H616 or A64? I didn't find any BSP DT files in that repo ...
<apritzel>
oh, and btw: when seeing that Syterkit couldn't identify the H728 chip, I had the idea of supporting both the Avaota and the X96QPro+ boards in one Syterkit build
<apritzel>
so now there is one binary that boots on both: if it sees an H728 chip (0xff00), it uses the DDR3 DRAM parameters and the X96QPro+ PMIC settings, and the Avaota ones otherwise
<apritzel>
still hoping for someone to come around with proper mainline DRAM code for the A523/T527 ... jernej, MasterR3C0RD: any updates on that?
<warpme>
apritzel: well - any reference to sunxi-gmac-xxx i started see when a523/t527 bsp code appears in the Internet. Indeed i'm curious: is sunxi-gmac-xxx an approach to: a)have universal gmac driver covering old hw (v110) and new hw (v2xx) or rather b) aw has multiple new hw variants and v200 is first we see in public. I suspect a) might be more probable as v110 might be for something older than a523 (e.g. old hw like a64?)
<apritzel>
it might be all smoke and mirrors, who knows. All SoCs supported in mainline have a TX delay limit of 7, which only matches the GMAC-200 in that code
<apritzel>
warpme: have you actually tried the mainline driver? IIRC from the vendor DT, then the X96QPro+ uses the second GMAC, so pinmux and resources must account for that
warpme has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
<warpme>
"just wanted to give you a heads up" - mllion thx for this!
<apritzel>
there is indeed only an RST_BUS_EMAC1, but no CLK_BUS_EMAC1
<apritzel>
I thought I left those issues in as a motivation for reviewers ;-)
<warpme>
:-))))))
<apritzel>
warpme: do you have a pointer to that Maxio PHY driver? So far I just see that one patch in your github, which is not easily mainline-able
gnarface has quit [Quit: Leaving]
gnarface has joined #linux-sunxi
<warpme>
apritzel : omg: it looks like i was looking on wrong dir for maxio (instead of mainline i was looking on mine sources dir). sorry - maxio in mainline is false positive argh....
<apritzel>
ah, ok, that explains it ;-)
<warpme>
yeah - to much kernel for today.....
warpme has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
<anarsoul>
fraolt: any luck with testing my patches?