Command Time-Out values in MMC

Command Time-Out values in MMC

[andrzej zaborowski]

Hi,
while browsing the MMC driver sources I noticed that the CTO (Command
Time-Out) register is written a value 0xff in one place (exactly,
here: http://www.kernel.org/git/?p=linux/kernel/git/tmlind/linux-omap-2.6.git;a=blob;h=35708dbf9d8e0f41687caeb70e7dde3f216a86a9;hb=9ca18d5394c5e6b2c74ddc52abaed81df530ff1d;f=drivers/mmc/omap.c#l847),
while the OMAP5910 documentation (exactly, this document:
http://focus.ti.com/dsp/docs/dspsupporttechdocsc.tsp?sectionId=3&tabId=409&familyId=325&abstractName=spru680)
states something like:
[...]
0x01: One clock cycle
0xFD: 253 clock cycles
The 0xFF and 0xFE cannot be used.
[...]

Is this intentional? It seems to cause no problems, but setting 0xfd
instead also didn't break anything for me. Is this limit also
mentioned in documentation for other OMAP boards?
(If this is intentional, please ignore me.)
Regards,
Andrzej Zaborowski

Re: Command Time-Out values in MMC

[Tony Lindgren]

* andrzej zaborowski <balrogg@gmail.com> [060128 11:55]:
> Hi,
> while browsing the MMC driver sources I noticed that the CTO (Command
> Time-Out) register is written a value 0xff in one place (exactly,
> here: http://www.kernel.org/git/?p=linux/kernel/git/tmlind/linux-omap-2.6.git;a=blob;h=35708dbf9d8e0f41687caeb70e7dde3f216a86a9;hb=9ca18d5394c5e6b2c74ddc52abaed81df530ff1d;f=drivers/mmc/omap.c#l847),
> while the OMAP5910 documentation (exactly, this document:
> http://focus.ti.com/dsp/docs/dspsupporttechdocsc.tsp?sectionId=3&tabId=409&familyId=325&abstractName=spru680)
> states something like:
> [...]
> 0x01: One clock cycle
> 0xFD: 253 clock cycles
> The 0xFF and 0xFE cannot be used.
> [...]
> 
> Is this intentional? It seems to cause no problems, but setting 0xfd
> instead also didn't break anything for me. Is this limit also
> mentioned in documentation for other OMAP boards?
> (If this is intentional, please ignore me.)

As far as I remember we're not using the command timeout currently and it
was expiring with some cards and causing false errors.

Basically some blocks on some cards take a really long time to write to.
And if the command timeout expires before the card is done, we get a
false error.

Tony

Re: Command Time-Out values in MMC

[andrzej zaborowski]

On 28/01/06, Tony Lindgren <tony@atomide.com> wrote:
> As far as I remember we're not using the command timeout currently and it
> was expiring with some cards and causing false errors.
>

The value 0 could be used to disable timing out.
Andrzej Zaborowski

Re: Command Time-Out values in MMC

[Tony Lindgren]

* andrzej zaborowski <balrogg@gmail.com> [060128 13:41]:
> On 28/01/06, Tony Lindgren <tony@atomide.com> wrote:
> > As far as I remember we're not using the command timeout currently and it
> > was expiring with some cards and causing false errors.
> >
> 
> The value 0 could be used to disable timing out.

Using a reasonably long timeout and adding handling for it
would probably be the best option.

Tony