[Fwd: Re: kue firmware revision code]

Scott Michel scottm at aero.org
Tue Feb 15 18:43:31 PST 2005

Well, here's Bill Paul's response to whether adjusting my other if_kue.c 
patch is valid... Basicallly, "if it works for you,..."

--- Begin Message ---


scottm at xxxxxxxx


william paul <wpaul at xxxxxxxxxxxxx>


Tue, 15 Feb 2005 12:36:23 -0800 (PST)


scottm at aero.org

Of all the gin joints in all the towns in all the world, Scott Michel
had to walk into mine and say:

> > No no, you've got it backwards. You try changing the code to test for
> > rev > 0x0202 and then you tell _me_ if it's safe. :)
> I did and it **seems** to be OK, but I was asked to ask you for a final 
> blessing. :-)

Most of the time, the only way I know something works right is when
I try it and it doesn't explode.

> > Actually, I don't know. I assume it should work ok. I don't understand
> > why you're getting a revision code of 0x0208 in the first place, since
> > I thought it was generated by the firmware itself, and since I only have
> > the one firmware image (which comes with the driver), I never expected it
> > to change.
> > 
> > Just try it and tell me what happens.
> I'm attempting to grok the way that driver works, in particular, what 
> the usb descriptor really points to relative to the firmware image and 
> the device. If the firmware rev is inside the image *and* it shows up in 
> the USB descriptor, then !BAM!, it's an easy fix.
> So far, I haven't found a 0x02, 0x02 pattern in the image, but I did 
> find a 0x08, 0x02 pattern. So I'm punting a bit based on what I've observed.
> -scooter

The revision code is a major and minor number. It could be the minor
number comes from the NIC rather than the firmware. I really have no
idea what format the firmware image is in: KLSI just gave it to me
pretty much as-is, with a bit of sample code to show how to download
it to the device.

I don't know why the revision code is different for you. When I load
that firmware image into my sample KLSI device, I get 0x0202. The
firmware image is the same one that's been in the tree since the driver
was written. It hasn't been updated in years. So if the revision
code returned after the firmware is loaded really does come from
inside the firmware, it should be the same.

It's possible the revision code is indeed derived at least partly
from the chip, and they've just bumped the rev a bit recently.


-Bill Paul            (510) 749-2329 | Senior Engineer, Master of Unix-Fu
                 wpaul at xxxxxxxxxxxxx | Wind River Systems
              <adamw> you're just BEGGING to face the moose
--- End Message ---

More information about the Submit mailing list