Re: [Acx100-devel] TNETW1450: Good News

[Andreas M. <an...@us...>]

Hi,

On Wed, Jan 25, 2006 at 08:17:14AM +0200, Denis Vlasenko wrote:
> > Actually this is a bad idea. Keeping PCI and USB drivers separate is
> > what has been done inside the kernel for a long time now and I actually
> > don't know of any driver that combined USB and PCI parts.
> 
> This does not prove that combined driver is a bad thing to do.

Indeed.

> > The ifdefs are 
> > still bad and should be avoided where ever possible.
> 
> What ifdefs? These?
> 
> # grep -r '^#if.*PCI' .
> ./pci.c:#ifndef PCI_D0
> ./common.c:#if defined(CONFIG_ACX_PCI)
> ./common.c:#if defined(CONFIG_ACX_PCI)
> ./acx_struct.h:#if defined(CONFIG_ACX_PCI)
> ./acx_struct.h:#ifdef ACX_PCI
> ./acx_struct.h:#ifdef ACX_PCI
> 
> # grep -r '^#if.*USB' .
> ./common.c:#if defined(CONFIG_ACX_USB)
> ./common.c:#if defined(CONFIG_ACX_USB)
> ./acx_struct.h:#if defined(CONFIG_ACX_USB)
> ./acx_struct.h:#ifdef ACX_USB
> ./acx_struct.h:#ifdef ACX_USB
> 
> Are you saying that we have too much ifdefs about PCI/USB?

Heh, no need for bragging about your wonderful driver cleanup, we know that
you're the Best Coder Ever :)

BTW, I did some more work yesterday, and it turned out that firmware upload
will take one or two more days since it's a bit more complicated.
However even firmware upload has obvious similarities to current code.

BTW, I now have two different firmware images: Ver 2.4.0.70 (used in all
Windows drivers that I know of) and Ver 2.4.0.68.

Oh, and in that respect:
I think that we might want to do one final driver renaming :-\

Suggestions would be:
module acx -> wlan_tnetw1xxx or wlan_tnetw or maybe just wlan_ti (but that one
might conflict with a driver for some next generation of TI cards)

Firmware images:
tiacx1CCcRR.bin -> tnetw1100c11.bin, tnetw1100usbc0D.bin, tnetw1130c16.bin,
                   tnetw1450c1b.bin

The reason being the new TNETW1450 (and I was a bit unhappy about the acx
part anyway which isn't used anywhere any more and is rather non-recognizable
in the module name case).

This additional change would be a bit of a pain, of course, so I'm not sure
whether we should do it. And of course we should do it after careful
consideration of exact names only.

Comments?

Andreas