ChanServ changed the topic of #armlinux to: ARM kernel talk [Upstream kernel, find your vendor forums for questions about their kernels] | https://libera.irclog.whitequark.org/armlinux
russ has joined #armlinux
naoki has joined #armlinux
apritzel_ has quit [Ping timeout: 252 seconds]
Guest824 has quit [Ping timeout: 260 seconds]
nsc has joined #armlinux
nsc is now known as Guest6226
heartburn has quit [Ping timeout: 252 seconds]
heartburn has joined #armlinux
heat has quit [Ping timeout: 248 seconds]
bochenchen has quit [Remote host closed the connection]
bochenchen has joined #armlinux
guillaume_g has joined #armlinux
heartburn has quit [Ping timeout: 248 seconds]
amitk_ has joined #armlinux
amitk has quit [Ping timeout: 248 seconds]
heartburn has joined #armlinux
headless has joined #armlinux
amitk has joined #armlinux
bjoto has quit [Ping timeout: 252 seconds]
iivanov has joined #armlinux
dougg3 has quit [Ping timeout: 255 seconds]
dougg3 has joined #armlinux
amitk has quit [Ping timeout: 246 seconds]
mcoquelin has quit [Remote host closed the connection]
mcoquelin has joined #armlinux
hanetzer has quit [Ping timeout: 248 seconds]
hanetzer has joined #armlinux
mcoquelin has quit [Ping timeout: 248 seconds]
bjoto has joined #armlinux
mcoquelin has joined #armlinux
sszy has joined #armlinux
viorel has joined #armlinux
bjoto has quit [Quit: WeeChat 3.6]
prabhakarlad has joined #armlinux
rvalue has quit [Ping timeout: 246 seconds]
matthias_bgg has quit [Quit: Leaving]
rvalue has joined #armlinux
apritzel has joined #armlinux
<bjdooks>
i've got the odd patch for the tegra tx2 modules, is sending them to the tegra list and cc'ing arnd and co ok?
<arnd>
bjdooks: no need to cc me, just send it to the platform maintainers (tagr is here), cc linux-tegra cc linux-arm-kernel
<arnd>
of course you can cc me if you think I have an opinion on the patch
<tagr>
bjdooks: I don't think I've ever used wifi on a Jetson TX2, but I've got one set up, so I could run some tests if you need anything (or a second opinion/datapoint)
elastic_dog has quit [Read error: Connection reset by peer]
elastic_dog has joined #armlinux
headless has quit [Quit: Konversation terminated!]
<bjdooks>
tagr: ok, thanks, our official nvidia contact has been useless... and since we don't know how the wifi chip is wired on the tx2 we can't really narrow down is it a pmic/pinmux/etc issue.... i know the vqmmc isn't there, but that's not fixed it... we've add a gpio for the wifi-en pin, but again not fixed hte issue
<tagr>
bjdooks: I thought schematics for TX2 were published on the L4T website
<bjdooks>
the baseboard and pinouts are, the tx2 module (p3310 iirc) don't seem to be, or are hidden so well i've not found them
<tagr>
so I see there's a WIFI_EN_CVM GPIO, which is probably what you were referring to
<tagr>
VDD_1V8_AP seems to be relevant as well, but that should be enabled for eMMC (SDMMC4) already
<bjdooks>
yeah, not sure if there's any further load gating though
<bjdooks>
wifi-enable {
<bjdooks>
output-high;
<bjdooks>
gpios = <TEGRA186_MAIN_GPIO(N, 0) 0>;
<bjdooks>
gpio-hog;
<bjdooks>
line-name = "wifi-enable";
<bjdooks>
};
monstr has joined #armlinux
cbeznea has joined #armlinux
<tagr>
bjdooks: to clarify: is the problem that the BCM chip isn't detected at all, or is the chip detected but doesn't get a signal?
<bjdooks>
no detection, mmc2 is showing no card
<tagr>
bjdooks: mmc3 you mean?
<bjdooks>
on my kernel mmc2 is the one connected to the wifi chip
<tagr>
oh, okay, so mmc2 at the kernel level, yeah, that might actually be SDMMC3 in hardware
<bjdooks>
i think it is labelled as sdmmc3 in the dts
<tagr>
we don't enable SDMMC2 on P3310
<bjdooks>
[ 4.768299] mmc2: SDHCI controller on 3440000.mmc [3440000.mmc] using ADMA 64-bit
<bjdooks>
i think that's the one, and it does work on the 4.9-l4t kenrle installed on the system
<tagr>
yeah, that's the right SDMMC interface alright
<tagr>
bjdooks: can you double-check if VDD_1V8_AP is being enabled somewhere? it's needed by both SDMMC3 and the BCM
<bjdooks>
looks like sdmmc4 has it as vqmmc-supply
<tagr>
can't spot anything else from the schematics
<tagr>
I'm completely unfamiliar with WIFI/MMC modules, so not sure how this usually works
<bjdooks>
just to confirm, WIFI_EN_CVM is N0 ?
<tagr>
yeah, PN.00
<bjdooks>
also, does the wifi chip need any clock or resets to it?
* bjdooks
was also wondering if there is some other power switch we're missing
<tagr>
as far as I can tell no controllable clocks or resets
<bjdooks>
some modules reqiore a 32khz clock as well as their local crystal
* bjdooks
wonders if he can just badger nvidia until info is derived
<tagr>
do you perhaps need bus-width and non-removable properties?
<bjdooks>
I'll try non removal later.
<tagr>
according to the mmc-controller.yaml bindings, bus-width defaults to 1, but SDMMC3 is 4 DAT lines
<bjdooks>
I know it gets unhappy if you remove the vdmmc3
<bjdooks>
I'll get back to this once I've been meetings to death
<robmur01>
Yeah, if the controller isn't even attempting to detect a hard-wired module, "non-removable" (or "broken-cd") would be my next guess
<robmur01>
missing power or clocks would be the thing to look for if it's spewing timeouts or communication errors because it *does* know something's there, but it's not replying
<robmur01>
(although of course that only happens when there *is* a working card detect pin)
punit has joined #armlinux
bjoto has joined #armlinux
<bjdooks>
robmur01: yeah, tried both of those now
<bjdooks>
robmur01: going to try proding the bt wakeup just in csae
<bjdooks>
but without schematics it is very difficult to go through everything checking
cbeznea has quit [Quit: Leaving.]
<bjdooks>
tagr: iirc, it defaults to 1 to id the card and then should try and up the game
<bjdooks>
i think wifi has been broken for some time, which is weird since no-one has complained about it yet
<robmur01>
:(
jclsn has joined #armlinux
<tagr>
bjdooks: are you sure it ever worked in the first place?
<bjdooks>
Upstream? No ideam
<bjdooks>
I've not had time to go back. Everything else we either have customer or Nvidia
<robmur01>
I think that's about the point where I gave up on reasoning and just probed all the relevant pins on the module itself to see what the vendor kernel did, then worked backwards to make mainline achieve the same
<bjdooks>
The module is under a great deal of heatsink
<robmur01>
you have my sympathies :(
jclsn has quit [Client Quit]
jclsn has joined #armlinux
<bjdooks>
I did suggest tearing it to bits and that might just happen next week.
<geertu>
bjdooks: How hard Röntgen do you need to go through the heatsink?
cbeznea has joined #armlinux
<bjdooks>
It might be removed by hard impact with a office wall
<tagr>
I can try to find out if there's a way to release the schematics if that helps
<bjdooks>
tagr that would be helpful. I'll continue badgering our sales and management too
<bjdooks>
I might try and send a gpio hog patch for the WiFi gpios
torez has joined #armlinux
heat has joined #armlinux
heat has quit [Remote host closed the connection]
iivanov has quit [Quit: Leaving...]
heat has joined #armlinux
iivanov has joined #armlinux
<bjdooks>
tagr: hmm, there's a whole bluetooth_pm driver which has GPIO L6 as a "reset" to the bluetooth sectin
<bjdooks>
or maybe H5, depending on which bt of this is vald
naoki has quit [Quit: naoki]
torez has quit [Quit: torez]
torez has joined #armlinux
<tagr>
bjdooks: downstream device trees use L6 for camera
<tagr>
H5 seems related to something called "bluedroid"
<tagr>
by the way, I do get these:
<tagr>
[ 3.880956] VDDIO_SDMMC3_AP: voltage operation not allowed
monstr has quit [Remote host closed the connection]
<tagr>
downstream has it configured as scalable between 1.8 and 3.3 V
<tagr>
doesn't seem to change anything, though... still no response
<tagr>
hm... do I need to enable any options to make it attempt to probe the BCM?
bochenchen has quit [Remote host closed the connection]
headless has joined #armlinux
<bjdooks>
tagr: I did an update to change vdmmc3 to allow 1.8..3.3v
<bjdooks>
tagr: I don't think so, it should at least see a SDIO device on MMC2.
<tagr>
looks like on some boards wifi-enable can also be B6
<bjdooks>
tagr: i've got to head off soon as it is almost home time
<tagr>
no success with any of that, though
<bjdooks>
iirc there's also an external slot
* bjdooks
will try and prod it over the weekend if possible, but not sure what i can get move
<tagr>
bjdooks: alright, good luck with that
<tagr>
need to go send PRs now... I may have some more time to look at this next week
<bjdooks>
tagr: ok, thanks
<bjdooks>
tagr: i'll try and make sense of the 4.9 stuff, might be some over implicty shite going on there, there is a drivers/misc bluedroid-pm driver on 4.9
<tagr>
shouldn't that be optional, though? i.e. the chip should be able to do wifi without any of the bluetooth stuff enabled
<geertu>
tagr: Thanks for reminding me! I have to send out my PRs, too, but I was waiting for resuslts from the buildbots...
<geertu>
Looks like I haven't received any during the last 48h?
<bjdooks>
tagr: yeah, but maybe there's some sort of pm/reset/etc with that
<bjdooks>
it does also do some regulator work
viorel has quit [Quit: WeeChat 3.6]
guillaume_g has quit [Quit: Konversation terminated!]
<robmur01>
that's pretty much how the downstream Rockchip stuff worked - completely separate rfkill drivers that wiggle the power/enable/etc. pins for the module, with the WiFi driver then binding via normal SDIO autoprobe, and Bluetooth over a UART entirely from userspace.
tachoknight has joined #armlinux
<robmur01>
also IIRC on a lot of the Broadcom chips the WiFi and BT enables aren't entirely independent of each other