Re: [PATCH] clk: rockchip: rk3288: Limit use of USB PHY clock to USB

From: "Heiko Stübner" <heiko@sntech.de>
To: Matthias Kaehlcke <mka@chromium.org>
Cc: Michael Turquette <mturquette@baylibre.com>,
	Stephen Boyd <sboyd@kernel.org>,
	linux-clk@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	linux-rockchip@lists.infradead.org, linux-kernel@vger.kernel.org,
	Douglas Anderson <dianders@chromium.org>
Subject: Re: [PATCH] clk: rockchip: rk3288: Limit use of USB PHY clock to USB
Date: Fri, 12 Apr 2019 11:30:37 +0200	[thread overview]
Message-ID: <2029012.W0xZ3ly3fY@diego> (raw)
In-Reply-To: <20190412001657.GV112750@google.com>

Am Freitag, 12. April 2019, 02:16:57 CEST schrieb Matthias Kaehlcke:
> Hi Heiko,
> 
> On Thu, Apr 11, 2019 at 09:03:07PM +0200, Heiko Stübner wrote:
> > Hi Matthias,
> > 
> > Am Donnerstag, 11. April 2019, 19:59:17 CEST schrieb Matthias Kaehlcke:
> > > The USB PHY clock can be configured as (grand) parent of uart0_sclk and
> > > sclk_gpu. It has been observed that UART0 doesn't work reliably in high
> > > speed mode with the PHY clock as input when certain USB devices are
> > > plugged to the USB HOST1 port (see https://crrev.com/c/320543).
> > > 
> > > Prefix the name of the PHY clock with a '.' in the non-USB muxes to
> > > effectively remove the clock as input from these muxes.
> > > 
> > > Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
> > > ---
> > >  drivers/clk/rockchip/clk-rk3288.c | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/clk/rockchip/clk-rk3288.c b/drivers/clk/rockchip/clk-rk3288.c
> > > index 5a67b7869960..677bc5485201 100644
> > > --- a/drivers/clk/rockchip/clk-rk3288.c
> > > +++ b/drivers/clk/rockchip/clk-rk3288.c
> > > @@ -200,8 +200,8 @@ PNAME(mux_aclk_cpu_src_p)	= { "cpll_aclk_cpu", "gpll_aclk_cpu" };
> > >  PNAME(mux_pll_src_cpll_gpll_p)		= { "cpll", "gpll" };
> > >  PNAME(mux_pll_src_npll_cpll_gpll_p)	= { "npll", "cpll", "gpll" };
> > >  PNAME(mux_pll_src_cpll_gpll_npll_p)	= { "cpll", "gpll", "npll" };
> > > -PNAME(mux_pll_src_cpll_gpll_usb480m_p)	= { "cpll", "gpll", "usbphy480m_src" };
> > > -PNAME(mux_pll_src_cpll_gll_usb_npll_p)	= { "cpll", "gpll", "usbphy480m_src", "npll" };
> > > +PNAME(mux_pll_src_cpll_gpll_usb480m_p)	= { "cpll", "gpll", ".usbphy480m_src" };
> > > +PNAME(mux_pll_src_cpll_gll_usb_npll_p)	= { "cpll", "gpll", ".usbphy480m_src", "npll" };
> > 
> > In general I like to have things like the clock-tree described fully
> > and let the kernel handle correct sourcing ... but:
> > 
> > As you write this seems like a systemic problem when just connecting
> > random peripherals can create unstable clock source frequencies,
> > so I tend to agree here ... but:
> > 
> > Can we please find a more "talking" name for this ... because as with the
> > above someone will find the "." and submit a fix for it ;-) .
> > 
> > So just name it "unstable_dummy" or so?
> 
> I looked for some common pattern, but couldn't find one. I liked the
> '.' since it leaves the name of the clock mostly intact, just hiding
> it (similar to a leading '.' in a Linux file system). But I agree that
> it might not be expressive enough. I still like the idea to keep the
> clock name around for reference, maybe we could name it
> "unstable:usbphy480m_src" or similar. If you don't object I'll send a
> patch with this some time tomorrow.

I've just adapted the patch to use the new parent-name you suggested
and applied it for 5.2 So no need to resend :-) .

Thanks
Heiko