Re: PCI: Remove users of pci_enable_device_bars()

Re: PCI: Remove users of pci_enable_device_bars()

[Jeff Garzik]

Linux Kernel Mailing List wrote:
> Commit:     0948391641918b95d8d96c15089eb5ac156850b3

>     PCI: Remove users of pci_enable_device_bars()
>     
>     This patch converts users of pci_enable_device_bars() to the new
>     pci_enable_device_{io,mem} interface.
>     
>     The new API fits nicely, except maybe for the QLA case where a bit of
>     code re-organization might be a good idea but I prefer sticking to the
>     simple patch as I don't have hardware to test on.
>     
>     I'll also need some feedback on the cs5520 change.
>     
>     Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
>     Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

> --- a/drivers/ata/pata_cs5520.c
> +++ b/drivers/ata/pata_cs5520.c
> @@ -229,7 +229,7 @@ static int __devinit cs5520_init_one(struct pci_dev *pdev, const struct pci_devi
>  		return -ENOMEM;
>  
>  	/* Perform set up for DMA */
> -	if (pci_enable_device_bars(pdev, 1<<2)) {
> +	if (pci_enable_device_io(pdev)) {
>  		printk(KERN_ERR DRV_NAME ": unable to configure BAR2.\n");
>  		return -ENODEV;
>  	}
> diff --git a/drivers/ide/pci/cs5520.c b/drivers/ide/pci/cs5520.c
> index 6ec00b8..10adc49 100644
> --- a/drivers/ide/pci/cs5520.c
> +++ b/drivers/ide/pci/cs5520.c
> @@ -156,8 +156,14 @@ static int __devinit cs5520_init_one(struct pci_dev *dev, const struct pci_devic
>  	ide_setup_pci_noise(dev, d);
>  
>  	/* We must not grab the entire device, it has 'ISA' space in its
> -	   BARS too and we will freak out other bits of the kernel */
> -	if (pci_enable_device_bars(dev, 1<<2)) {
> +	 * BARS too and we will freak out other bits of the kernel
> +	 *
> +	 * pci_enable_device_bars() is going away. I replaced it with
> +	 * IO only enable for now but I'll need confirmation this is
> +	 * allright for that device. If not, it will need some kind of
> +	 * quirk. --BenH.
> +	 */
> +	if (pci_enable_device_io(dev)) {
>  		printk(KERN_WARNING "%s: Unable to enable 55x0.\n", d->name);
>  		return -ENODEV;
>  	}

So...  did the feedback mentioned in the commit log ever appear?  :)

	Jeff

Re: PCI: Remove users of pci_enable_device_bars()

[Bartlomiej Zolnierkiewicz]

On Saturday 02 February 2008, Jeff Garzik wrote:
> Linux Kernel Mailing List wrote:
> > Commit:     0948391641918b95d8d96c15089eb5ac156850b3
> 
> >     PCI: Remove users of pci_enable_device_bars()
> >     
> >     This patch converts users of pci_enable_device_bars() to the new
> >     pci_enable_device_{io,mem} interface.
> >     
> >     The new API fits nicely, except maybe for the QLA case where a bit of
> >     code re-organization might be a good idea but I prefer sticking to the
> >     simple patch as I don't have hardware to test on.
> >     
> >     I'll also need some feedback on the cs5520 change.
> >     
> >     Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> >     Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
> 
> > --- a/drivers/ata/pata_cs5520.c
> > +++ b/drivers/ata/pata_cs5520.c
> > @@ -229,7 +229,7 @@ static int __devinit cs5520_init_one(struct pci_dev *pdev, const struct pci_devi
> >  		return -ENOMEM;
> >  
> >  	/* Perform set up for DMA */
> > -	if (pci_enable_device_bars(pdev, 1<<2)) {
> > +	if (pci_enable_device_io(pdev)) {
> >  		printk(KERN_ERR DRV_NAME ": unable to configure BAR2.\n");
> >  		return -ENODEV;
> >  	}
> > diff --git a/drivers/ide/pci/cs5520.c b/drivers/ide/pci/cs5520.c
> > index 6ec00b8..10adc49 100644
> > --- a/drivers/ide/pci/cs5520.c
> > +++ b/drivers/ide/pci/cs5520.c
> > @@ -156,8 +156,14 @@ static int __devinit cs5520_init_one(struct pci_dev *dev, const struct pci_devic
> >  	ide_setup_pci_noise(dev, d);
> >  
> >  	/* We must not grab the entire device, it has 'ISA' space in its
> > -	   BARS too and we will freak out other bits of the kernel */
> > -	if (pci_enable_device_bars(dev, 1<<2)) {
> > +	 * BARS too and we will freak out other bits of the kernel
> > +	 *
> > +	 * pci_enable_device_bars() is going away. I replaced it with
> > +	 * IO only enable for now but I'll need confirmation this is
> > +	 * allright for that device. If not, it will need some kind of
> > +	 * quirk. --BenH.
> > +	 */
> > +	if (pci_enable_device_io(dev)) {
> >  		printk(KERN_WARNING "%s: Unable to enable 55x0.\n", d->name);
> >  		return -ENODEV;
> >  	}
> 
> So...  did the feedback mentioned in the commit log ever appear?  :)

Since Alan has commented on it:

http://lkml.org/lkml/2007/12/17/422

"5520 in fact is always enabled as it is the host bridge.
pci_enable_device_io will do just fine. The 5520 fun is if you disable it
the system hangs."

I moved on assuming that either submitter or integrator would follow up...
but it seems that it didn't happen. :(

PS Could somebody remove the stale comment from cs5520.c?

Thanks,
Bart

Re: PCI: Remove users of pci_enable_device_bars()

[Greg KH]

On Sat, Feb 02, 2008 at 09:01:10PM +0100, Bartlomiej Zolnierkiewicz wrote:
> On Saturday 02 February 2008, Jeff Garzik wrote:
> > Linux Kernel Mailing List wrote:
> > > Commit:     0948391641918b95d8d96c15089eb5ac156850b3
> > 
> > >     PCI: Remove users of pci_enable_device_bars()
> > >     
> > >     This patch converts users of pci_enable_device_bars() to the new
> > >     pci_enable_device_{io,mem} interface.
> > >     
> > >     The new API fits nicely, except maybe for the QLA case where a bit of
> > >     code re-organization might be a good idea but I prefer sticking to the
> > >     simple patch as I don't have hardware to test on.
> > >     
> > >     I'll also need some feedback on the cs5520 change.
> > >     
> > >     Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> > >     Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
> > 
> > > --- a/drivers/ata/pata_cs5520.c
> > > +++ b/drivers/ata/pata_cs5520.c
> > > @@ -229,7 +229,7 @@ static int __devinit cs5520_init_one(struct pci_dev *pdev, const struct pci_devi
> > >  		return -ENOMEM;
> > >  
> > >  	/* Perform set up for DMA */
> > > -	if (pci_enable_device_bars(pdev, 1<<2)) {
> > > +	if (pci_enable_device_io(pdev)) {
> > >  		printk(KERN_ERR DRV_NAME ": unable to configure BAR2.\n");
> > >  		return -ENODEV;
> > >  	}
> > > diff --git a/drivers/ide/pci/cs5520.c b/drivers/ide/pci/cs5520.c
> > > index 6ec00b8..10adc49 100644
> > > --- a/drivers/ide/pci/cs5520.c
> > > +++ b/drivers/ide/pci/cs5520.c
> > > @@ -156,8 +156,14 @@ static int __devinit cs5520_init_one(struct pci_dev *dev, const struct pci_devic
> > >  	ide_setup_pci_noise(dev, d);
> > >  
> > >  	/* We must not grab the entire device, it has 'ISA' space in its
> > > -	   BARS too and we will freak out other bits of the kernel */
> > > -	if (pci_enable_device_bars(dev, 1<<2)) {
> > > +	 * BARS too and we will freak out other bits of the kernel
> > > +	 *
> > > +	 * pci_enable_device_bars() is going away. I replaced it with
> > > +	 * IO only enable for now but I'll need confirmation this is
> > > +	 * allright for that device. If not, it will need some kind of
> > > +	 * quirk. --BenH.
> > > +	 */
> > > +	if (pci_enable_device_io(dev)) {
> > >  		printk(KERN_WARNING "%s: Unable to enable 55x0.\n", d->name);
> > >  		return -ENODEV;
> > >  	}
> > 
> > So...  did the feedback mentioned in the commit log ever appear?  :)
> 
> Since Alan has commented on it:
> 
> http://lkml.org/lkml/2007/12/17/422
> 
> "5520 in fact is always enabled as it is the host bridge.
> pci_enable_device_io will do just fine. The 5520 fun is if you disable it
> the system hangs."
> 
> I moved on assuming that either submitter or integrator would follow up...
> but it seems that it didn't happen. :(
> 
> PS Could somebody remove the stale comment from cs5520.c?

Ben, care to send me a patch for this, as it's your change?

thanks,

greg k-h

Re: PCI: Remove users of pci_enable_device_bars()

[Benjamin Herrenschmidt]

> > Since Alan has commented on it:
> > 
> > http://lkml.org/lkml/2007/12/17/422
> > 
> > "5520 in fact is always enabled as it is the host bridge.
> > pci_enable_device_io will do just fine. The 5520 fun is if you disable it
> > the system hangs."
> > 
> > I moved on assuming that either submitter or integrator would follow up...
> > but it seems that it didn't happen. :(
> > 
> > PS Could somebody remove the stale comment from cs5520.c?
> 
> Ben, care to send me a patch for this, as it's your change?

Ah sure, will do tomorrow.

Cheers,
Ben.