The following patch eliminates pcidev.driver_data, in favor of
pcidev.dev.driver_data, thereby shrinking pcidev by four bytes.
In the future, I hope this patch will make it simpler to have
a facility where drivers can ask the generic device layer to do the
kmalloc of their private memory area, but that's another patch.
Applying this patch now will help discourage anyone from
building drivers that rely on having two different private pointers
in struct pci_dev and struct device.
I oringally posted this patch against 2.5.44 along with a
separate patch that changed the few device drivers that directly
referenced pcidev.dev to use pci_{get,set}_drvdata(). The latter
patches got into 2.5.45 via Jeff Garzik. At 2.5.45, I reposted this
patch and Greg Kroah-Hartman said that he would submit it to you in
"the next round of pci cleanups I'm going to be sending to Linus", but
it seems to have fallen through the cracks since then. This patch has
been posted to lkml twice before and nobody has stated any objections
to this patch.
Although I do not make much use of 2.5.48 due to the modules
problems, I have been running with this patch since 2.5.44 with no
problems.
Please apply. Thanks in advance.
--
Adam J. Richter __ ______________ 575 Oroville Road
[email protected] \ / Milpitas, California 95035
+1 408 309-6081 | g g d r a s i l United States of America
"Free Software For The Rest Of Us."
On Tue, Nov 19, 2002 at 09:16:26PM -0800, Adam J. Richter wrote:
> At 2.5.45, I reposted this
> patch and Greg Kroah-Hartman said that he would submit it to you in
> "the next round of pci cleanups I'm going to be sending to Linus", but
> it seems to have fallen through the cracks since then.
I've been on "vacation" this week, and wanted to get my pcibios cleanup
changes done first (which I just sent off.) Sorry for the delay.
This patch looks good, and if Linus doesn't take it directly, I'll put
it up in a tree to send to him later :)
Sorry again for the delay,
greg k-h