Re: [PATCH] clks: imx7: fix initial clock setup with deep probe enabled

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Sep 02, 2022 at 04:02:32PM +0200, Ahmad Fatoum wrote:
> Hello,
> 
> On 31.08.22 10:21, Johannes Zink wrote:
> > Hi Sascha,
> > 
> > On Tue, 2022-08-30 at 17:05 +0200, Sascha Hauer wrote:
> >> On Tue, Aug 30, 2022 at 10:39:37AM +0200, Johannes Zink wrote:
> >>> We register the i.MX7 clock controller driver at core_initcall
> >>> level and
> >>> then do some initial clock setup/reparenting at postcore_initcall
> >>> level.
> >>> This doesn't work as expected when deep probe is enabled, because
> >>> while
> >>> the driver is registered at core_initcall level, it's only probed
> >>> later on, currently at postcore_initcall level because it's a
> >>> dependency
> >>> of the timer for which of_ensure_device_probed is called.
> >>>
> >>> As the initial clock setup is also at postcore_initcall level, it's
> >>> no
> >>> longer guaranteed that the code executes in the same order. Fix
> >>> this by
> >>> directly doing the setup at the end of the probe function.
> >>
> >> Does this still work with deep probe disabled?
> >>
> >> I am asking because this effectively reverts bce79428773 ("clk:
> >> i.MX7: do
> >> clock reparenting when all clocks are initialized"). Switching all
> >> i.MX7 boards to deep probe might be a solution as well.
> 
> fixed clocks are registered via CLK_OF_DECLARE, so by the time the
> imx7 clock probe runs, they will always be available, independent of
> whether deep probe is enabled.

So this problem is already fixed, great.

> 
> > Though I would not _expect_ any problems here, I would like to check
> > that more thoroughly. This will probably take a bit, please stand by.
> 
> I looked at this with Johannes and I am puzzled how the problem
> that bce79428773 aimed to fix occurred. Could you elaborate?

Have a look at the tree at bce79428773. Back then the i.MX7 clock driver
was registered at core_initcall level, the same level that of_probe()
was running which first called of_platform_populate() and then
of_clk_init().

Sascha

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |




[Index of Archives]     [Linux Embedded]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux