Re: [GIT PATCH] More PCI patches for 2.6.13

[Jeff Garzik <jgarzik@pobox.com>]

Linus Torvalds wrote:
> 
> On Sat, 10 Sep 2005, Andrew Morton wrote:
> 
>>Who do I bug about these longstanding ppc64 warnings, btw ;)
>>
>>drivers/scsi/sata_svw.c: In function `k2_sata_tf_load':
>>drivers/scsi/sata_svw.c:111: warning: passing arg 2 of `eeh_writeb' makes pointer from integer without a cast
>>drivers/scsi/sata_svw.c:116: warning: passing arg 2 of `eeh_writew' makes pointer from integer without a cast
>>drivers/scsi/sata_svw.c:117: warning: passing arg 2 of `eeh_writew' makes pointer from integer without a cast

> I used to have a patch to fix them, and sent it to Jeff ages ago. At that 
> point, he didn't want to use the iomap() functionality, but maybe that has 
> changed.
> 
> Jeff? It requires making almost all the SATA IO base pointers be iomapped: 
> the current

Glad you two asked!  :)

I -do- want to use iomap.  The problem is that no one has yet come up 
with a few that does all the proper resource reservation.  Everybody 
(including myself) did the ioread/iowrite part, but gave up before 
handling all cases of (a) legacy ISA iomap, (b) native PCI IDE iomap, 
and (c) non-standard MMIO iomap.

This "works, but leaks resources" code exists in the 'iomap' branch of
rsync://rsync.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev.git

It languished in this incomplete stage for a while, but very recently 
got moving again:

> commit 374b1873571bf80dc0c1fcceaaad067980f3b9de
> Author: Jeff Garzik <jgarzik@pobox.com>
> Date:   Tue Aug 30 05:42:52 2005 -0400
> 
>     [libata] update several drivers to use pci_iomap()/pci_iounmap()

I actually made a promise to use iomap, since libata will benefit quite 
a bit from ioread/iowrite usage.  I intend to keep that promise... it's 
just taking me a while to get us there.  Now that I figured out the path 
I want to take, we should get there around 2.6.1[56].

	Jeff