Re: [PATCH v3] i2c: lpi2c: cache peripheral clock rate

[Marc Kleine-Budde]

On 02.05.2023 08:50:06, Alexander Stein wrote:
> Hello Wolfgang,
> 
> Am Sonntag, 30. April 2023, 09:05:55 CEST schrieb Wolfram Sang:
> > * PGP Signed by an unknown key
> > 
> > > > IIRC this is a general problem^w limitation of the clock framework,
> > > > clock providers cannot use clocks themselves in certain callback, e.g.
> > > > set_rate.
> > > 
> > > Well, that's essentially impossible when this clock provider is attached
> > > via i2c. i2c transfers potentially need to change or prepare clocks.
> > 
> > So, as I get it, this is not a specific lpi2c problem but affecting any
> > I2C controller driver which uses get_rate() to setup a transfer to a
> > remote I2C clock provider? And this lockdep warning is a false-positive?
> 
> Yes, IMHO this could potentially occur on every I2C controller driver, if a 
> clock provider while holding the clk_prepare_lock, e.g. during registration, 
> issues an i2c transfer.
> I'm not so sure if this is a false-positive, but more like trying to do a 
> nested lock.

It's a general limitation of the clock framework. The clk_prepare_lock()
is taken by clk_prepare() and by clk_get_rate(). So if you implement a
clock provider driver that uses I2C, which might use clk_get_rate() it
might deadlock. However there's a workaround for nested I2C calls:
533ddeb1e86f "(clk: allow reentrant calls into the clk framework,
2013-03-28)". This doesn't work for SPI, as it's using a worker thread.

| https://elixir.bootlin.com/linux/v6.3/source/drivers/clk/clk.c#L128

regards,
Marc

-- 
Pengutronix e.K.                 | Marc Kleine-Budde          |
Embedded Linux                   | https://www.pengutronix.de |
Vertretung Nürnberg              | Phone: +49-5121-206917-129 |
Amtsgericht Hildesheim, HRA 2686 | Fax:   +49-5121-206917-9   |

Attachment: signature.asc
Description: PGP signature