[Librem-5-dev] queries with dev kit environment set up

[Guido Günther]

Hi,
On Wed, May 29, 2019 at 11:36:29AM +0930, Dan Kortschak wrote:
> Thanks. Since the board was fresh out of the box due to a later deliver
> as stated above I had assumed that I would be working from a common
> starting point and could just follow the provided documentation.
> 
> It's unclear to me from the outset exactly what information would be
> useful, so this is why I asked for suggestions about how to debug it.
> 
> $ uname -a
> Linux pureos 4.18.11-g418cc45f #1 SMP PREEMPT Wed May 22 06:24:21 PDT
> 2019 aarch64 GNU/Linux
> 
> 
> * HDMI
> 
> $ cat /sys/firmware/devicetree/base/hdmi at 32c00000/status
> disabled
> 
> but the link from /boot/dtb looks like it is pointing at the right
> thing
> 
> $ ls -l /boot/dtb
> lrwxrwxrwx 1 root root 40 May 24 21:59 /boot/dtb -> dtbs/4.18.11-
> g418cc45f/./librem5-evk.dtb

Is that the location you copied the HDMI device tree to? Can you try
copying the HDMI DT to /etc/flashkernel/dtbs and run flash-kernel and
see if that helps with HDMI?

> 
> which exists
> 
> $ ls -l dtbs/4.18.11-g418cc45f/./librem5-evk.dtb
> -rw-r--r-- 1 root root 42276 May 24 21:59 dtbs/4.18.11-
> g418cc45f/./librem5-evk.dtb
> 
> 
> * WIFI
> 
> No, I did not remove the blacklist line as I didn't realise it was
> blacklisted. I have now and that resolves that issue. I will send an MR
> to add that instruction or at least note that it is blacklisted.

I've filed https://source.puri.sm/Librem5/developer.puri.sm/issues/122
for this.

> 
> 
> 
> If the documentation that's at the set up guide page is not correct, is
> there documentation that I can follow that is?

It's not that the docs aren't correct in general - they're a great
resource to get started. It's just the DTB part that needs updating:

Check https://source.puri.sm/Librem5/developer.puri.sm/issues/121

Cheers,
 -- Guido

> 
> 
> cheers
> Dan
> 
> 
> On Tue, 2019-05-28 at 13:14 +0200, Guido Günther wrote:
> > Hi,
> > On Tue, May 28, 2019 at 08:13:45PM +0930, Dan Kortschak via Librem-5-
> > dev wrote:
> > > Can anyone suggest how I can debug these two issues?
> > 
> > In general providing the version (uname -a), related logs (dmesg,
> > journalctl -b0), information from/sys/firmware/devicetree and
> > e.g. lsmod helps.
> > 
> > > thanks
> > > Dan
> > > 
> > > On Sun, 2019-05-26 at 08:45 +0930, Dan Kortschak wrote:
> > > > I think this is the last couple of questions from me on this
> > > > page.
> > > > 
> > > > I have successfully reflashed the eMMC and that has all worked.
> > > > Setting
> > > > up the HDMI does not appear to work though.
> > > > 
> > > >  * I have performed the change at [1] but this does not bring up
> > > > the
> > > >    HDMI connection and the on-board display remains on.
> > 
> > uboot is likely not picking up the right DTB then. Check whether
> > /boot/dtb symlink points to the right hdmi dtb.
> > Also this should print okay when you're using the right device tree:
> > 
> > $ cat /sys/firmware/devicetree/base/hdmi at 32c00000/status
> > 
> > I think the docs at
> > 
> > 
> https://developer.puri.sm/Librem5/Development_Environment/Boards/imx8.html#hdmi
> > 
> > don't match what needs to be done since the image is using uboot's
> > distro boot and flash-kernel. When using flash-kernel "custom" dtbs
> > should go to /etc/flash-kernel/dtbs (untested):
> > 
> >    cp <hdmidtb> /etc/flash-kernel/dtbs/
> >    flash-kernel
> > 
> > > >  * The second issue is how to get rsi_sdio to load at boot. I
> > > > have
> > > >    added rsi_sdio to /etc/modules, but it remains unloaded,
> > > > leaving
> > > > wi-
> > > >    fi off without manual intervention. The wi-fi switch is
> > > > correctly
> > > >    positioned and wi-fi works fine after modprobing rsi_sdio.
> > 
> > Did you remove it form the blacklist:
> > 
> >    
> > https://source.puri.sm/Librem5/librem5-devkit-tools/blob/master/etc/modprobe.d/librem5-devkit.conf#L4
> > 
> > Cheers,
> >  -- Guido
>