On 10:34-20231129, Tony Lindgren wrote: > * Nishanth Menon <nm@xxxxxx> [231123 07:10]: > > On 14:24-20231121, Tony Lindgren wrote: > > > From: Vignesh Raghavendra <vigneshr@xxxxxx> > > > > > > WL1837 WLAN card is present on the original AM625 SK board. It > > > communicates with the SoC using 4 bit SDIO through the second instance of > > > MMCSD. > > > > > > Starting with SK-AM62B, there is a M.2 WLAN device connector instead of > > > > We support AM62B-SK. > > Is that with k3-am62-lp-sk.dts though? Wondering if they should have separate > dts files for the M.2. It has been the same dts. Separating the wlan out of the dts sounds correct. > > > > the integrated WL1837 WLAN. The M.2 connector should be handled separately > > > in the k3-am62a.dtsi and k3-am62b.dtsi files as needed. > > > > Should this rather be an overlay instead of integrated dts fixup? M2 > > connector allows for various options including the newer 33xx family[2]. > > Not sure if an overlay makes sense for an integrated device.. It sure makes > sense for plug in boards though. overlay works for both variant of devices (please see additional comment below). > > > It makes sense for the regulator etc to be on the main dts file, but I > > am not convinced about it being integrated as part of the dts. > > Yeah if AM62B-SK is supported with the same dts. My vote would be for > separate dts files for the integrated variant to keep things simple :) we can handle it via dtb itself dts can remain common, the m2 is a dtso, we generate a dtbo and the original will be an additional dtb? I think this can work out without dts/dtsi duplication. arch/arm64/boot/dts/ti/Makefile has previous examples of the form. > > > Should we use mmc-pwrseq ? > > Yes probably, I think there was some issue earlier with that but sounds like > it's been solved already. > > > Looks like we have run into an issue in BeaglePlay with wlan_en > > being always on for loading firmware. so is there a need to keep the > > wlan on while suspended? > > If the wlan irq was on the first gpio bank, wake-up path would work for > the wlan. But I think it's wired to a gpio bank with no wake-up capability, > and the padconf wake related patches are still pending. So no need to keep > the wlan enabled so far AFAIK. I believe the issue was around the wlan firmware load itself, but i think the mmc-pwrseq will probably help resolve it. -- Regards, Nishanth Menon Key (0xDDB5849D1736249D) / Fingerprint: F8A2 8693 54EB 8232 17A3 1A34 DDB5 849D 1736 249D
