Re: SDHCI: mdelay() in hot path in esdhc_pltfm_set_clock looses CAN (!) frames

[Sascha Hauer <s.hauer@pengutronix.de>]

On Wed, Jul 08, 2015 at 08:22:57AM +0200, Holger Schurig wrote:
> Wow, at last some reaction. And I thought nobody cares ...
> 
> BTW, removing CONFIG_MMC_CLKGATE helped a bit, because the pointless
> clock-off-clock-on while the device is booting (or accessing multiple
> sectors within a short time) isn't going to happen anymore.
> 
> But really, a mdelay(1) in a driver ...

Look for the other mdelays in sdhci.c. There are up to a hundred of
them in sdhci_reset() which is called under a spin_lock_irqsave. I
remember that with the wrong type of card detection the MMC core
frequently polls for a card and if none is found the sdhci core was
resetted which falls into this mdelay polling loop in sdhci_reset. So if
you care about realtime you better have a card inserted...

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |