16:01 <Tofe_> FYI, I have managed to get my SDIO card detected; however, it looks like pwrseq-simple is too simple for me, as I need to inialize an input gpio, an output low gpio, and an output high gpio...

16:02 <phh> Tofe_: uh input? it rather sounds like it's a pinctrl?

16:02 <phh> what do you need to do with that input pin?

16:03 <phh> the low gpio and the high gpio are different gpios or the same?

16:03 <Tofe_> phh: well I'm a bit over my head with this thing, so I might be doing things wrong :) I'm just following the downstream wifi initialization sequence

16:03 <Tofe_> they're different

16:04 <phh> maybe just share that source code?

16:04 <Tofe_> sure, one moment

16:05 <Tofe_> so for downstream, if we exclude the regulator intiialization, it's down to this: https://github.com/Tofee/shr-linux/blob/tenderloin/3.4/halium-9.0/arch/arm/mach-msm/hp/tenderloin/board-tenderloin-wifi.c#L292-L304

16:06 <Tofe_> we have 3 GPIO here: https://github.com/Tofee/shr-linux/blob/tenderloin/3.4/halium-9.0/arch/arm/mach-msm/hp/tenderloin/board-tenderloin-wifi.c#L183-L187

16:07 <Tofe_> And their definition for msm8660 was declared like this: https://github.com/Tofee/shr-linux/blob/tenderloin/3.4/halium-9.0/arch/arm/mach-msm/hp/tenderloin/board-tenderloin-wifi.c#L183-L187

16:07 <phh> you can safely ignore TENDERLOIN_GPIO_WL_HOST_WAKE and TENDERLOIN_GPIO_HOST_WAKE_WL_PIN until you have deep-suspend

16:07 <phh> just put TENDERLOIN_GPIO_WLAN_RST_N_PIN in the mmc-pwrseq

16:08 <Tofe_> I thought I tried that, but maybe not; however TENDERLOIN_GPIO_WLAN_RST_N_PIN seems to be defined as OUTL (low), isn't that an issue ?

16:09 <Tofe_> current I savagely override with a call to gpiod_direction_output(wlan_reset_gpio, 0); but it's very hacky and I'd like to find a nicer way

16:10 <phh> well, as I said before, just export the gpio through /sys, add broken-cd to the mmc node, and just try both options, setting the gpio to 0 or to 1

16:10 <phh> (the timing should be irrelevant, but you do need to toggle it at least once)

16:10 <Tofe_> I didn't succed in exporting anything through /sys

16:11 <Tofe_> I get errors like: "echo: write error: Device or resource busy"

16:11 <Tofe_> when exporting

16:12 <Tofe_> but first let me try with just TENDERLOIN_GPIO_WLAN_RST_N_PIN, if that's enough it'll simplify things a bit

16:17 <phh> i'm guessing pwrseq had already registered the gpio so you couldn't

16:17 <phh> but you can check /sys/kernel/debug/gpio for more infos

16:17 <Tofe_> ah, yes, I'll look there

16:19 <Tofe_> (also it's an old tablet, takes forever to boot, so patience is required)

17:53 <Tofe_> mmh now I've deactivated my hack, and it still works, weird

17:53 <Tofe_> I've done a power-off / power-on cycle, and the SDIO MMC card is still detected

17:54 <Tofe_> so in the end, my fix commit just looks like this https://github.com/Tofee/shr-linux/commit/9dc22fa4eded72ee715d03959a72e51b4fd596cd

17:57 <Tofe_> so right now the "full" wifi log error looks like this: https://paste.tofe.me/p/r01jQv.syslog

17:57 <Tofe_> the firmware missing seem to be "normal"

17:58 <Tofe_> -110 is ETIMEDOUT

18:03 <phh> it's possible that the -110 issue is just that it can't find any firmware

18:06 <Tofe_> possible yes, I'll investigate

18:18 <phh> 93 and 337 are TENDERLOIN_GPIO_HOST_WAKE_WL_PIN and TENDERLOIN_GPIO_HOST_WAKE_WL_PIN ?

18:18 <phh> 137*

18:20 <Tofe_> yes: #define TENDERLOIN_GPIO_WL_HOST_WAKE93 and #define TENDERLOIN_GPIO_HOST_WAKE_WL137

18:20 <Tofe_> and there's a matching table afterwards, but for my board it's the same

18:23 <Tofe_> so the firmware thing is just the ath6kl driver trying every fw-%d.bin downwards, until it finds one. I only have fw-3.bin. I guess it's for possible future upgrades. But I don't think it's related to the timeout here; I would rather thing that I messed up the mmc sdio configuration somehow

18:23 <Tofe_> think*