06:58 <pivi> Hello all, I am currently implementing the equivalent of Linux commit c7666240ec76 ("drivers: mmc: sdhci_am654: Add the quirk to set TESTCD bit") in U-Boot. When using the SDHCI CD as GPIO this is required.

06:59 <pivi> From what I can tell current U-Boot master, because of this, should have a non working SD interface also on the beagleplay.

07:00 <pivi> Can anybody confirm? I do not have a beagleplay at hand to confirm this

07:03 <javierm_> pivi: what u-boot version should have that problem? I'm using U-Boot 2024.01-rc4-00036-g65eed687729c and the uSD works correctly

07:05 <pivi> javierm_: that version should be affected. I would say there is something missing in my reasoning

07:06 <pivi> from what I see on our hardware, when using the CD as GPIO (likewise beagleplay is doing), "ti,fails-without-test-cd" property needs to be correctly handled. But this is not done in U-Boot

07:07 <pivi> but again, if it works for you, I am missing something

07:07 <pivi> wait

07:07 <pivi> no, this version is ok

07:08 <pivi> v2024.07 is affected

07:08 <pivi> v2024.04 I have not checked, it depends when the upstream DTS was merged into U-Boot, but for sure your old version is fine

07:09 <pivi> (2024.01 is so old that my mind filtered out the .01 ;-)

07:20 <javierm_> pivi: I can try updating my u-boot later then

07:39 <pivi> javierm_: if you try it let me know, thanks!

07:52 <pivi> vigneshr: hello! I was just discussing one of your old patches, maybe you can have a look at the logs ;-) https://libera.irclog.whitequark.org/linux-ti/2024-06-28

08:00 <javierm_> pivi: it failed to boot with U-Boot SPL 2024.07-rc3-00013-g46ff00bea5dd (Jun 28 2024 - 09:41:31 +0200)

08:00 <javierm_> but last message I see is "[ 0.000001] sched_clock: 58 bits at 200MHz, resolution 5ns, wraps every 4398046511102ns"

08:00 <javierm_> this is with kernel 6.9.5-200.fc40.aarch64

08:03 <javierm_> pivi: it doesn't sound like your issue right ?

08:33 <pivi> javierm_: nope, my expecation is that the SD card is not working at all in U-Boot, way before loading the Linux kernel

08:34 <pivi> javierm_: unless you have a fallback to boot to eMMC and you did not notice the error

08:34 <javierm_> pivi: I don't, I've wiped the eMMC completely

08:35 <pivi> that's weird. something I am not understanding.

08:35 <pivi> thanks for testing

08:35 <pivi> you just discovered some other unrelated issue, BTW

08:35 <pivi> in case you were getting bored ;-)

08:39 <javierm_> so yeah, u-boot -> grub works correctly

08:39 <javierm_> pivi: https://paste.centos.org/view/raw/ef009318 my boot log

08:39 <javierm_> pivi: I always have more than I can take on my plate but yes, it seems is an unrelated issue indeed

08:39 <javierm_> you broke my boot :P

08:40 <javierm_> just joking, I'll try to dig on this

08:54 <vigneshr> javierm_: https://lore.kernel.org/u-boot/20240619-play-fdt-fixup-v1-1-18e2a728ca45@ti.com/ should help in non grub case

08:55 <vigneshr> pivi: I am not sure how U-Boot works/handles TESTCD. let me check internally

09:02 <javierm_> vigneshr: wooo, that seems to fix it!

09:13 <javierm_> vigneshr: confirmed that boots with that patch, even with the u-boot provided FDT

09:13 <javierm_> pivi: so I was not able to reproduce your issue

09:13 <vigneshr> javierm_: good to know, thanks!

09:14 <javierm_> can someone remind me from where I could fetch the ti-connectivity/wl18xx-conf.bin firmware for the BeaglePlay ?

09:15 <vigneshr> pivi: I know that mmc2 instance definitely fails if CD isnt connected (found this while testing differen wifi cards). I am not sure how SDCD line is internally tied off for mmc0/1 which maybe helping here...

09:15 <javierm_> I remember that it was mentioned before that was optional? But WiFi isn't working for me

09:16 <javierm_> vigneshr: that could be why the WiFi SDIO isn't working then? Due the CD not connected and the "ti,fails-without-test-cd" quirk not present in the DT ?

09:17 <vigneshr> javierm_: if you dont have "ti,fails-without-test-cd" then yes MMC2 would not enumerate cards correctly... but i see k3-am625-beagleplay.dts has this set

09:18 <javierm_> vigneshr: ah, another issue then. What about the FW, it's optional as was told or required ?

09:18 <javierm_> [ 44.080753] wl18xx_driver wl18xx.6.auto: Direct firmware load for ti-connectivity/wl18xx-conf.bin failed with error -2

09:25 <vigneshr> javierm_: Sorry, havent used wl8xx much, all firmwares should ideally be in linux-firmware and see I dont see above, I am thinking thats optional.. Will let you know if thats not the case

09:26 <javierm_> vigneshr: sure, no worries. And thanks again for solving the boot issue!

09:38 <pivi> vigneshr: the concern here is in U-Boot, where the "ti,fails-without-test-cd" is not handled by the driver ... I am surprised this is working fine with beagleplay ...

11:38 <javierm_> TIL about bootflow and the standard boot

11:38 <javierm_> it is indeed much nicer than the complex distro boot commands and environment variables