Query: should 'Marvell SATA support' depend on PCI?

Query: should 'Marvell SATA support' depend on PCI?

[Grant Coady]

Hi Alan,

I was reworking the ATA driver menu patch to shrink it (merge 'depends on' 
lines) and noticed that the:

config SATA_MV
        tristate "Marvell SATA support (HIGHLY EXPERIMENTAL)"
        depends on EXPERIMENTAL
        help
          This option enables support for the Marvell Serial ATA family.
          Currently supports 88SX[56]0[48][01] chips.

doesn't have 'depends on PCI && EXPERIMENTAL' like other drivers around 
it, this is okay or should it also depend on PCI?

Grant.

Re: Query: should 'Marvell SATA support' depend on PCI?

[Alan Cox]

> > doesn't have 'depends on PCI && EXPERIMENTAL' like other drivers around 
> > it, this is okay or should it also depend on PCI?
> 
> Well, I think so.  It obviously depends on PCI.

Obviously ? Try looking a bit harder

/**
 *      mv_platform_probe - handle a positive probe of an soc Marvell
 *      host
 *      @pdev: platform device found
 *
 *      LOCKING:
 *      Inherited from caller.
 */
static int mv_platform_probe(struct platform_device *pdev)


Alan

Re: Query: should 'Marvell SATA support' depend on PCI?

[Rafael J. Wysocki]

On Sunday, 10 of August 2008, Grant Coady wrote:
> Hi Alan,
> 
> I was reworking the ATA driver menu patch to shrink it (merge 'depends on' 
> lines) and noticed that the:
> 
> config SATA_MV
>         tristate "Marvell SATA support (HIGHLY EXPERIMENTAL)"
>         depends on EXPERIMENTAL
>         help
>           This option enables support for the Marvell Serial ATA family.
>           Currently supports 88SX[56]0[48][01] chips.
> 
> doesn't have 'depends on PCI && EXPERIMENTAL' like other drivers around 
> it, this is okay or should it also depend on PCI?

Well, I think so.  It obviously depends on PCI.

Thanks,
Rafael

Re: Query: should 'Marvell SATA support' depend on PCI?

[Frans Pop]

Grant Coady wrote:
> I was reworking the ATA driver menu patch to shrink it (merge 'depends
> on' lines) and noticed that the:
> 
> config SATA_MV
>         tristate "Marvell SATA support (HIGHLY EXPERIMENTAL)"
>         depends on EXPERIMENTAL
>         help
>           This option enables support for the Marvell Serial ATA family.
>           Currently supports 88SX[56]0[48][01] chips.
> 
> doesn't have 'depends on PCI && EXPERIMENTAL' like other drivers around
> it, this is okay or should it also depend on PCI?

AFAICT it should not as it can also be implemented as a platform device.

I have SATA_MV on a QNAP TS-109 NAS dervice (arm) and lspci on that gives me:
$ lspci
00:00.0 Memory controller: Marvell Technology Group Ltd. Device 5182 (rev 02)
01:00.0 Memory controller: Marvell Technology Group Ltd. Device 5182 (rev 02)

So the disk controller is not on the PCI bus.
And in /sys/ I see sata_mv listed under devices/platform/.


Cheers,
FJP

Re: Query: should 'Marvell SATA support' depend on PCI?

[Rafael J. Wysocki]

On Sunday, 10 of August 2008, Alan Cox wrote:
> > > doesn't have 'depends on PCI && EXPERIMENTAL' like other drivers around 
> > > it, this is okay or should it also depend on PCI?
> > 
> > Well, I think so.  It obviously depends on PCI.
> 
> Obviously ? Try looking a bit harder
> 
> /**
>  *      mv_platform_probe - handle a positive probe of an soc Marvell
>  *      host
>  *      @pdev: platform device found
>  *
>  *      LOCKING:
>  *      Inherited from caller.
>  */
> static int mv_platform_probe(struct platform_device *pdev)

Sorry, I didn't notice the '#ifdef CONFIG_PCI' in there.

Rafael