This patch adds the SMBus controller DeviceID for the Intel Panther Point PCH.
Signed-off-by: Seth Heasley <[email protected]>
---
--- linux-2.6.39-rc3/Documentation/i2c/busses/i2c-i801.orig 2011-04-11 17:21:51.000000000 -0700
+++ linux-2.6.39-rc3/Documentation/i2c/busses/i2c-i801 2011-04-14 11:56:59.000000000 -0700
@@ -19,6 +19,7 @@
* Intel 6 Series (PCH)
* Intel Patsburg (PCH)
* Intel DH89xxCC (PCH)
+ * Intel Panther Point (PCH)
Datasheets: Publicly available at the Intel website
On Intel Patsburg and later chipsets, both the normal host SMBus controller
--- linux-2.6.39-rc3/drivers/i2c/busses/Kconfig.orig 2011-04-11 17:21:51.000000000 -0700
+++ linux-2.6.39-rc3/drivers/i2c/busses/Kconfig 2011-04-14 12:02:27.000000000 -0700
@@ -101,6 +101,7 @@
6 Series (PCH)
Patsburg (PCH)
DH89xxCC (PCH)
+ Panther Point (PCH)
This driver can also be built as a module. If so, the module
will be called i2c-i801.
--- linux-2.6.39-rc3/drivers/i2c/busses/i2c-i801.c.orig 2011-04-11 17:21:51.000000000 -0700
+++ linux-2.6.39-rc3/drivers/i2c/busses/i2c-i801.c 2011-04-14 12:03:57.000000000 -0700
@@ -50,6 +50,7 @@
Patsburg (PCH) IDF 0x1d71 32 hard yes yes yes
Patsburg (PCH) IDF 0x1d72 32 hard yes yes yes
DH89xxCC (PCH) 0x2330 32 hard yes yes yes
+ Panther Point (PCH) 0x1e22 32 hard yes yes yes
Features supported by this driver:
Software PEC no
@@ -623,6 +624,7 @@
{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_PATSBURG_SMBUS_IDF1) },
{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_PATSBURG_SMBUS_IDF2) },
{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_DH89XXCC_SMBUS) },
+ { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_PANTHERPOINT_SMBUS) },
{ 0, }
};
Hi Seth,
On Wed, 20 Apr 2011 11:33:39 -0700, Seth Heasley wrote:
> This patch adds the SMBus controller DeviceID for the Intel Panther Point PCH.
With each new chip, we have to add the SMBus device ID to pci_ids.h,
then wait for Jesse to merge that, and only then I can apply the
changes to i2c-i801.c. This approach slows things down needlessly.
It isn't mandatory to add IDs to pci_ids.h when an ID is only used
locally in a device driver. So what I would like to propose is that we
move all PCI_DEVICE_ID_INTEL_*_SMBUS declarations from pci_ids.h to
i2c-i801.c now. Then you can resubmit your Panther Point patches, and
the pci and i2c parts will be independent, so Jesse and myself don't
depend on each other to apply them.
What do you think? Jesse, any objection?
Seth, if you agree, I can take care of the move, or you can send a
patch doing that, whatever you prefer.
--
Jean Delvare
Hi Jean,
>With each new chip, we have to add the SMBus device ID to pci_ids.h,
>then wait for Jesse to merge that, and only then I can apply the
>changes to i2c-i801.c. This approach slows things down needlessly.
>
>It isn't mandatory to add IDs to pci_ids.h when an ID is only used
>locally in a device driver. So what I would like to propose is that we
>move all PCI_DEVICE_ID_INTEL_*_SMBUS declarations from pci_ids.h to
>i2c-i801.c now. Then you can resubmit your Panther Point patches, and
>the pci and i2c parts will be independent, so Jesse and myself don't
>depend on each other to apply them.
>
>Seth, if you agree, I can take care of the move, or you can send a
>patch doing that, whatever you prefer.
I've been thinking about this one myself, and I tend to agree. It'd probably be simpler if you went ahead and made the changes, and I'll be happy to review them.
-Seth
On Fri, 22 Apr 2011 18:44:49 +0200
Jean Delvare <[email protected]> wrote:
> Hi Seth,
>
> On Wed, 20 Apr 2011 11:33:39 -0700, Seth Heasley wrote:
> > This patch adds the SMBus controller DeviceID for the Intel Panther Point PCH.
>
> With each new chip, we have to add the SMBus device ID to pci_ids.h,
> then wait for Jesse to merge that, and only then I can apply the
> changes to i2c-i801.c. This approach slows things down needlessly.
>
> It isn't mandatory to add IDs to pci_ids.h when an ID is only used
> locally in a device driver. So what I would like to propose is that we
> move all PCI_DEVICE_ID_INTEL_*_SMBUS declarations from pci_ids.h to
> i2c-i801.c now. Then you can resubmit your Panther Point patches, and
> the pci and i2c parts will be independent, so Jesse and myself don't
> depend on each other to apply them.
>
> What do you think? Jesse, any objection?
>
> Seth, if you agree, I can take care of the move, or you can send a
> patch doing that, whatever you prefer.
Yeah, that makes sense. I'd be happy to take a patch to pull the
defines out of pci_ids.h and push them into the x86 irq.c and your i2c
code.
--
Jesse Barnes, Intel Open Source Technology Center
On Fri, 22 Apr 2011 09:52:59 -0700, Jesse Barnes wrote:
> On Fri, 22 Apr 2011 18:44:49 +0200
> Jean Delvare <[email protected]> wrote:
>
> > Hi Seth,
> >
> > On Wed, 20 Apr 2011 11:33:39 -0700, Seth Heasley wrote:
> > > This patch adds the SMBus controller DeviceID for the Intel Panther Point PCH.
> >
> > With each new chip, we have to add the SMBus device ID to pci_ids.h,
> > then wait for Jesse to merge that, and only then I can apply the
> > changes to i2c-i801.c. This approach slows things down needlessly.
> >
> > It isn't mandatory to add IDs to pci_ids.h when an ID is only used
> > locally in a device driver. So what I would like to propose is that we
> > move all PCI_DEVICE_ID_INTEL_*_SMBUS declarations from pci_ids.h to
> > i2c-i801.c now. Then you can resubmit your Panther Point patches, and
> > the pci and i2c parts will be independent, so Jesse and myself don't
> > depend on each other to apply them.
> >
> > What do you think? Jesse, any objection?
> >
> > Seth, if you agree, I can take care of the move, or you can send a
> > patch doing that, whatever you prefer.
>
> Yeah, that makes sense. I'd be happy to take a patch to pull the
> defines out of pci_ids.h and push them into the x86 irq.c and your i2c
> code.
I really only had i2c in mind. For irq it's a little different because
it relates to devices which are kind of generic and the IDs could be
(and, I think, are) used in more than one place (such as quirks.c) for
at least some of these devices.
--
Jean Delvare