Re: musb is broken in master

[Yegor Yefremov <yegorslists@googlemail.com>]

Hi Ahmad, all,

On Tue, Feb 18, 2020 at 12:48 PM Ahmad Fatoum <a.fatoum@pengutronix.de> wrote:
>
> On 2/18/20 12:19 PM, Yegor Yefremov wrote:
> > Hello Ahmad,
> >
> > On Tue, Feb 18, 2020 at 11:57 AM Ahmad Fatoum <a.fatoum@pengutronix.de> wrote:
> >>
> >> Hello Yegor,
> >>
> >> On 2/18/20 10:58 AM, Yegor Yefremov wrote:
> >>> I have bisected the musb driver on am335x (baltos system
> >>> arch/arm/dts/am335x-baltos-minimal.dts) from v2019.11.0 and this is
> >>> the result:
> >>>
> >>> 574eed3f6fcf056aa4c9e46c4b5224e3f7844d8d is the first bad commit
> >>> commit 574eed3f6fcf056aa4c9e46c4b5224e3f7844d8d
> >>> Author: Sascha Hauer <s.hauer@pengutronix.de>
> >>> Date:   Thu Dec 19 05:46:54 2019 +0100
> >>>
> >>>     dts: update to v5.5-rc1
> >>>
> >>>     Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
> >>>
> >>> :040000 040000 4f6416a5af14d6328ef2c8ac46d9ac4a17a37404
> >>> 525e33dda1e735ef2f633f4d99c501ba938a7263 M      dts
> >>>
> >>> I need only the host functionality to mount a USB drive. My defconfig
> >>> can be found here [1].
> >>>
> >>> Any idea what could have happened?
> >>
> >> git log v5.4..v5.5-rc1 arch/arm/boot/dts/am335x*.dtsi
> >> gives me kernel commit 12afc0cf81 ("ARM: dts: Drop pointless status changing for am3 musb").
> >> Does reverting it fix the issue for you?
> >
> > Yes, it does. Though the revert result is a little bit ugly i.e. at
> > least one conflict.
>
> yes, it's not meant as permanent solution.
> Which status = "okay" is the one that 'fixes' it?
>
> > But when I start the kernel 5.6-rc2 that already incorporates these
> > DTS changes, I have no problems with USB. So far only the barebox is
> > affected by this change.
>
> Possibly a device driver in barebox for the now-disabled peripheral
> turns on a clock, enables a regulator, toggles a reset or initializes
> some register that also affects the USB working at all.
>
> Best course of action would be to check which driver it is, check whether
> Linux uses it, comment out stuff to find what the magic sauce is that
> the driver does and add that to the driver that misses it.
>
> (sometimes it's just that the device tree binding changed. In that case,
> you can either patch in a missing property in arch/arm/dts or patch
> barebox drivers to accept the new binding as well)

Can it be that musb was broken because of this change in the mainline
linux kernel [1]?

The kernel uses ti-sysc mechanism to get/configure the related hw
blocks. If I grep for ti-sysc in kernel, I get a lot of hits. But in
barebox, it is only drivers/bus/ti-sysc.c that I find. Seems like a
rather big update is required.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/arm/boot/dts/am33xx.dtsi?h=v5.6-rc3&id=0782e8572ce43f521ed6ff15e4a7ab9aa5acdc85

Yegor

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox