04:43 <beeble> lurchi_: gpio count depends on the package size. the smallest one has 4. and yes programmable serialnumber. fully custom or unique one from silicon

06:07 <rperier> beeble: yeah but you did this from uboot mainline and not from uboot rk ?

06:07 <rperier> arrfffff :/

06:27 <beeble> i don't have a uboot rk at hand and not really interested in building it. but the test would work on any

08:11 <rperier> mhhh in rk_uboot

08:12 <rperier> switching from EL3 to EL1 seems to be configurable

09:32 <rperier> well, so BL31... switches from EL3 to EL1... so no way to get this stuff working :/

09:34 <rperier> I think , I will have to ask for testing. Heiko or Klauss if you have 5 min today, could you try to replace "rockchip_rk3368_efuse_read" by "rockchip_rk3288_efuse_read" (if you have a working uboot mainline) in my patch. I should just work, I think

09:34 <rperier> I am sorry but I am completly blocked :/

09:35 <rperier> s/I/It/

09:38 <rperier> well, I will send a v2 and ask for testing on the ML, that's the best I can do

09:47 <ayufan> rperier: what are you working on?

09:48 <rperier> on efuse support for rk3368

09:48 <ayufan> ok

09:48 <ayufan> is it similar to rk3328?

09:49 <rperier> the problem is that with the vendor trusted firmware and the vendor uboot, this efuse is only accessible in EL3 so I have to use non standard ARM SMC calls , while normally this efuse should be accessible without secure mode

09:49 <ayufan> OK

09:49 <ayufan> I know that on rk3328 some region is readable, some is marked as secure: https://github.com/ayufan-rock64/linux-u-boot/commit/7a5f2a60c2f881984e8a82a6d11d723a2cff24b1

09:49 <rperier> the solution I have for this is : use uboot mainline (does not work, LPDDR3 not supported) or use upstream ATF, not working too... so... I have no solutons for now

09:50 <rperier> that's with a geekbox rk3368 btw

09:50 <rperier> yeah, the problem is when it's marked as secure, you get CPU exception when you try to access it from EL1 (normal)

09:54 <ayufan> yeah

13:51 <kloczek> hi

13:52 <kloczek> guys is it any generic way how to install linux on rk3399 firefly over PXE?

15:34 <xevious> ayufan: I'm trying to get your MAC address fixes for the RK3328 working on an A95X R2. The A95X R2 uses the gmac2phy NIC, not gmac2io. I know gmac2phy is unstable on Rock64 (caused by insufficient power, I believe) but I had it working and stable with an older build on the A95X R2. I've created a custom device tree with gmac2phy, and now the MAC address isn't changing. However, I can't use the network interface yet because I'm getting this

15:34 <xevious> error (several times) on boot: stmmac_open: Cannot attach to PHY (error: -19)

15:37 <ayufan> xevious: gmac2phy works fine for me on mainline kernel with applied fixes

15:37 <ayufan> so, it is just misconfiguration at this point :)

15:38 <ayufan> btw. I received your e-mail, give me a few days to think that throug

15:38 <ayufan> I tested rock64 gmac2phy with es9023 audio pot board

15:40 <xevious> Is the mainline kernel compatible with your MAC address fix, or does your kernel fork have changes for it?

15:41 <ayufan> My mainline kernel fork has them

15:41 <ayufan> I use now mainline uboot with fixes to solve a few issues

15:42 <ayufan> Grab my latest linux image 0.5.6 and install mainline kernel from 0.5.7 and t should work

15:42 <xevious> This system is built with your linux-kernel, linux-u-boot, and linux-build repos, with the rockchip-linux rootfs repo.

15:42 <xevious> Ok, I'll try that next. Thanks!

15:43 <xevious> ayufan: About the email - thanks! I hope you decide you want to work with us. I could definitely use your help :)

15:43 <ayufan> The biggest problem is time :)

15:43 <xevious> haha I understand that, for sure! :D

15:44 <xevious> I have to head out for the weekend, and I'll be AFK (unfortunate family stuff). I'll try linux 0.5.6 + kernel from 0.5.7 on Tuesday.

17:25 <via> paulk-gagarine: dammit, i just got it flashed, but i guess i didn't enable all the flags i needed

17:25 <via> i have 0x11

17:25 <via> i guess i need the dev flag on too -_-

17:25 <via> can i just confirm exactly what i need?

17:25 <via> i need to add 0x8?

17:50 <paulk-gagarine> via, I have 0x30 on my images

17:50 <paulk-gagarine> you probably want to have 0x10 at least

17:50 <paulk-gagarine> which is dev_boot_usb if I'm not mistaken

17:51 <via> looking at https://chromium.googlesource.com/chromiumos/platform/vboot_reference/+/master/firmware/include/gbb_header.h

17:51 <via> i'm not sure how 30 works

17:51 <via> no dev switch related stuff would be on

17:51 <via> even dev switch on didn't seem to work, trying alt-os now

17:52 <via> fcurently fllashing 0x9d

17:52 <paulk-gagarine> that's GBB_FLAG_FORCE_DEV_BOOT_USB | GBB_FLAG_DISABLE_FW_ROLLBACK_CHECK

17:53 <paulk-gagarine> if you already have dev switch on, you don't need to force it with gbb

17:53 <paulk-gagarine> (actually in my case I don't want to force it, which is why it's not there and 0x30 is enough)

17:53 <via> in fucking with it last week i might have hit the space bar to turn off dev mode

17:54 <via> it keeps telling me that the device i inserted doesn't contain chromeos

17:54 <via> i suppose i could have set up the usb stick wrong

18:00 <via> nope suddenly everything works

18:07 <via> partition table fixed, my debian install is working again

18:07 <via> thank you

18:29 <paulk-gagarine> via, neato

19:44 <phh> anyone has any news about vulkan on rk3288?

19:45 <phh> I can't find any Mali-T760 platform with vulkan driver, yet https://developer.arm.com/graphics/vulkan does claim T760 is vulkan compatible...

19:45 <phh> https://community.arm.com/graphics/f/discussions/7281/are-the-vulkan-drivers-compatible-with-mali-t764-gpu mentions that r12p0 should have vulkan driver, we're at r13...

20:03 <phh> ok, looks like r14p0 has it, though there is no public release of it

20:46 <via> anyone know if uboot support's the display on rk3288 (acer chromebook)

20:46 <kloczek> kloczek

20:54 <via> guessing not based on the very sparse 'device drivers' section in menuconfig

20:58 <via> is there *any* payload i can use for coreboot that supports some type of ui for booting?

21:06 <via> looks like depthcharge has some suppor for a basic ui but i'm struggling to find docs on it