Re: ARM: SoC fixes for 4.15

From: gregkh
Date: Sat Jan 20 2018 - 03:35:49 EST


On Fri, Jan 19, 2018 at 05:12:15PM +0100, Arnd Bergmann wrote:
> The following changes since commit b2cd1df66037e7c4697c7e40496bf7e4a5e16a2d:
>
> Linux 4.15-rc7 (2018-01-07 14:22:41 -0800)
>
> are available in the git repository at:
>
> ssh://git@xxxxxxxxxxxxxxxxxxx/pub/scm/linux/kernel/git/arm/arm-soc.git
> tags/armsoc-fixes
>
> for you to fetch changes up to b7563e2796f8b23c98afcfea7363194227fa089d:
>
> phy: work around 'phys' references to usb-nop-xceiv devices
> (2018-01-19 16:12:37 +0100)
>
> ----------------------------------------------------------------
> ARM: SoC fixes for 4.15
>
> We have various small DT fixes, and one important regression fix:
>
> The recent device tree bugfixes that were intended to address issues that
> 'dtc' started warning about in 4.15 fixed various USB PHY device nodes,
> but it turns out that we had code that depended on those nodes being
> incorrect and the probe failing with a particular error code. With the
> workaround we can also deal with correct device nodes.
>
> The DT fixes include:
> - Allwinner A10 and A20 had the display pipeline set up incorrectly
> (introduced in v4.15)
> - The Altera PMU lacked an interrupt-parent (never worked)
> - Pin muxing on the Openblocks A7 (never worked)
> - Clocks might get set up wrong on Armada 7K/8K (4.15 regression)
>
> We now have additional device tree patches to address all the remaining
> warnings introduced in 4.15, but decided to queue them for 4.16 instead,
> to avoid risking another regression like the USB PHY thing mentioned
> above.
>
> ----------------------------------------------------------------
>
> [note: I originally sent the phy-core patch to Greg with Kishon's
> ack, as drivers/phy stuff tends to go that route but got no reply
> over the last week, so I'm including it here instead. As the bug
> was introduced through arm-soc, this seems like the best way
> to now make sure we fix it before the release.]

No objection from me for this.

thanks,

greg k-h