Message-ID: <20210215130620.3la4bexamozzcvjx@fsr-ub1664-175>
Date:   Mon, 15 Feb 2021 15:06:20 +0200
From:   Abel Vesa <abel.vesa@....com>
To:     Adam Ford <aford173@...il.com>
Cc:     Sascha Hauer <s.hauer@...gutronix.de>,
        arm-soc <linux-arm-kernel@...ts.infradead.org>,
        Adam Ford-BE <aford@...conembedded.com>,
        Aisheng Dong <aisheng.dong@....com>,
        Michael Turquette <mturquette@...libre.com>,
        Stephen Boyd <sboyd@...nel.org>,
        Shawn Guo <shawnguo@...nel.org>,
        Pengutronix Kernel Team <kernel@...gutronix.de>,
        Fabio Estevam <festevam@...il.com>,
        NXP Linux Team <linux-imx@....com>,
        Linus Walleij <linus.walleij@...aro.org>,
        Jerome Brunet <jbrunet@...libre.com>,
        linux-clk <linux-clk@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH V3] clk: imx: Fix reparenting of UARTs not associated
 with sdout

On 21-02-13 08:44:28, Adam Ford wrote:
> On Wed, Feb 3, 2021 at 3:22 PM Adam Ford <aford173@...il.com> wrote:
> >
> > On Thu, Jan 21, 2021 at 4:24 AM Abel Vesa <abel.vesa@....com> wrote:
> > >
> > > On 21-01-21 10:56:17, Sascha Hauer wrote:
> > > > On Wed, Jan 20, 2021 at 06:14:21PM +0200, Abel Vesa wrote:
> > > > > On 21-01-20 16:50:01, Sascha Hauer wrote:
> > > > > > On Wed, Jan 20, 2021 at 05:28:13PM +0200, Abel Vesa wrote:
> > > > > > > On 21-01-20 16:13:05, Sascha Hauer wrote:
> > > > > > > > Hi Abel,
> > > > > > > >
> > > > > > > > On Wed, Jan 20, 2021 at 04:44:54PM +0200, Abel Vesa wrote:
> > > > > > > > > On 21-01-18 08:00:43, Adam Ford wrote:
> > > > > > > > > > On Mon, Jan 18, 2021 at 6:52 AM Abel Vesa <abel.vesa@....com> wrote:
> > > > > > >
> > > > > > > ...
> > > > > > >
> > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > TBH, I'm against the idea of having to call consumer API from a clock provider driver.
> > > > > > > > > > > I'm still investigating a way of moving the uart clock control calls in drivers/serial/imx,
> > > > > > > > > > > where they belong.
> > > > > > > > > >
> > > > > > > > > > That makes sense.
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > > Just a thought. The uart clock used for console remains on from u-boot,
> > > > > > > > > so maybe it's enough to just add the CLK_IGNORE_UNUSED flag to all the
> > > > > > > > > uart root clocks and remove the prepare/enable calls for uart clocks
> > > > > > > > > for good. I don't really have a way to test it right now, but maybe
> > > > > > > > > you could give it a try.
> > > > > > > >
> > > > > > > > That would mean that UART clocks will never be disabled, regardless of
> > > > > > > > whether they are used for console or not. That doesn't sound very
> > > > > > > > appealing.
> > > > > > >
> > > > > > > AFAIK, the only uart clock that is enabled by u-boot is the one used for
> > > > > > > the console. Later on, when the serial driver probes, it will enable it itself.
> > > > > > >
> > > > > > > Unless I'm missing something, this is exactly what we need.
> > > > > >
> > > > > > It might enable it, but with CLK_IGNORE_UNUSED the clock won't be
> > > > > > disabled again when a port is closed after usage
> > > > >
> > > > > OK, tell me what I'm getting wrong in the following scenario:
> > > > >
> > > > > U-boot leaves the console uart clock enabled. All the other ones are disabled.
> > > > >
> > > > > Kernel i.MX clk driver registers the uart clocks with flag CLK_IGNORE_UNUSED.
> > > >
> > > > I was wrong at that point. I originally thought the kernel will never
> > > > disable these clocks, but in fact it only leaves them enabled during the
> > > > clk_disable_unused call.
> > > >
> > > > However, when CLK_IGNORE_UNUSED becomes relevant it's too late already.
> > > > I just chatted with Lucas and he told me what the original problem was
> > > > that his patch solved.
> > > >
> > > > The problem comes when an unrelated device and the earlycon UART have
> > > > the same parent clocks. The parent clock is enabled, but it's reference
> > > > count is zero. Now when the unrelated device probes and toggles its
> > > > clocks then the shared parent clock will be disabled due to the
> > > > reference count being zero. Next time earlycon prints a character the
> > > > system hangs because the UART gates are still enabled, but their parent
> > > > clocks no longer are.
> > > >
> > >
> > > Hmm, that is indeed a problem. That's why I think there should be some
> > > kind of NOCACHE flag for almost all the types of clocks. For example,
> > > in this case, it makes sense for the core to check the bit in the register
> > > and then disable the parent based on that instead of relying on the refcount.
> > > Anyway, that's something that needs to be added in the CCF.
> > >
> > > > Overall I think Lucas' patches are still valid and relevant and with
> > > > Adams patches we even no longer have to enable all UART clocks, but
> > > > only the ones which are actually needed.
> > >
> > > Yeah, for now, I think we can go with Adam's patches. But longterm, I would
> > > like to remove the special case of the uart clocks we have right now in all
> > > the i.MX clock drivers.
> 
> I looked around at other serial drivers, and I found nothing like this
> function for enabling all UART clocks.  There are generic functions
> for registering consoles, earlycon etc, and the serial driver fetches
> the per and igp clocks from the device tree, so I attempted to simply
> remove imx_register_uart_clocks().  I booted an i.MX8M Nano from a
> fully-powered off state, and my serial console came up just fine.

Just because it works, doesn't mean it is safe. To put it simply, the
risk of some  driver disabling a clock that is parent of the uart clock
would render the earlycon broken.

> 
> I checked the clk_summary, and the clock parents are set correctly and
> the respective clock rates appear to be correct (ie, the console is
> working at the desired baud rate, and Bluetooth is happy)
> 
> Since I don't fully understand the serial driver and the clock
> dependencies, I don't want to just simply remove the function without
> discussing it, because I don't know the ramifications.  However, when
> testing on the i.MX8M Nano, things look OK.
> I also tested suspend-resume and the console UART appears to return
> and the Bluetooth UART set to 4Mbps works just fine too.
> 
> I'd like to post a V4 which just removes imx_register_uart_clocks and
> the corresponding calls to it.  I don't know enough about the older
> 32-bit i.MX SoC's, but I have build-tested it, and I can generate a
> patch. Are there any objections and/or concerns?
> 

Please don't remove the imx_register_uart_clocks for now. As much as I
would like it gone, the way the earlycon could end up broken is
so ugly that it would make it a real pain to debug it later on.

> adam
> > >
> >
> > Is the patch I submitted good enough for someone's acked-by or
> > reviewed-by, or are there changes I need to implement?
> >
> > adam
> >
> > > >
> > > > Sascha
> > > >
> > > >
> > > > --
> > > > Pengutronix e.K.                           |                             |
> > > > Steuerwalder Str. 21                       | https://eur01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.pengutronix.de%2F&amp;data=04%7C01%7Cabel.vesa%40nxp.com%7C12862aeac2934a4a94f108d8d02de573%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C637488242849833051%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&amp;sdata=fKDvWYJIU7hBROH6yDM9FYUCHJDK9wMjTeJDMxRff8o%3D&amp;reserved=0  |
> > > > 31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
> > > > Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Powered by blists - more mailing lists