Re: io_speed = 0 ???

David Hinds <[email protected]> Fri, 12 Mar 2004 08:04:16 -0800
Newsgroups gmane.linux.pcmcia.devel
Message-ID <[email protected]>
On Fri, Mar 12, 2004 at 12:34:52PM +0100, Steven Scholz wrote:
> Hi there,
> 
> cs.c contains the lines
> 
> /* Access speed for attribute memory windows */
> INT_MODULE_PARM(cis_speed,	300);		/* ns */
> 
> /* Access speed for IO windows */
> INT_MODULE_PARM(io_speed,	0);		/* ns */
> 
> Does that mean, we're not using the timing values store the CIS but 
> using these hardcoded numbers?

The timing for CIS accesses is determined by the PCMCIA specification
(it can't be in the CIS because that would be a catch 22).

The timing for IO accesses is also determined by the PCMCIA spec;
the io_speed parameter permits the user to lengthen IO cycles beyond
the spec.  There are no fields in the CIS for describing timing of IO
cycles.

> Which value should I choose if io_speed is not specified by the user?

As a socket driver developer, you would set the minimum IO cycle times
in accordance with the PCMCIA specification, and then use io_speed to
slow things down more than that.

> In m8xx_pcmcia.c 255ns is choosen if it is not a memory card. Thus 
> resulting in very slow cycle times for PCMCIA ATA cards...

PCMCIA IO *is* slow.  The value of 255 ns is the minimum IO cycle time
permitted by the spec.

-- Dave


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click