2009-03-11 03:41:28

by David Brownell

[permalink] [raw]
Subject: Re: [PATCH] OMAP: HSMMC: Initialize hsmmc controller registers when resuming

On Monday 23 February 2009, Adrian Hunter wrote:
> Although I have not tested it, I very much doubt
> dual-voltage cards work. ?That is because VMMC1_185V
> is zero, which has the side-effect of turning the
> regulator off (see arch/arm/mach-omap2/mmc-twl4030.c)

And a second reason to know they don't quite work ... in
the file drivers/mmc/host/omap_hsmmc.c, omap_mmc_set_ios()
sets the voltage for MMC_POWER_OFF (0) or MMC_POWER_UP (1_,
which gives the initial setting -- e.g. 3.15 Volts, so it
can enumerate at the high range.

But after enumerating the card at that voltage, checking
the OCR values, and concluding that the slot and card can
both run at 1.85V ... the MMC_POWER_ON (2) code is used.
But the driver completely ignores it ... the low voltage
(more power-efficient!) voltage range never kicks in.

It'd be nice to have a nice unambiguous set_voltage()
request from the MMC core. The set_ios() thing has
always been confusing.

- Dave


2009-03-11 06:50:40

by Pierre Ossman

[permalink] [raw]
Subject: Re: [PATCH] OMAP: HSMMC: Initialize hsmmc controller registers when resuming

On Tue, 10 Mar 2009 19:33:50 -0800
David Brownell <[email protected]> wrote:

>
> It'd be nice to have a nice unambiguous set_voltage()
> request from the MMC core. The set_ios() thing has
> always been confusing.
>

set_ios() should be taken out back. But someone has to have the time to
reimplement things. :/

--
-- Pierre Ossman

Linux kernel, MMC maintainer http://www.kernel.org
rdesktop, core developer http://www.rdesktop.org

WARNING: This correspondence is being monitored by the
Swedish government. Make sure your server uses encryption
for SMTP traffic and consider using PGP for end-to-end
encryption.