Re: [RFC] AHCI: Remove conflicting Intel Lewisburg 3rd-party SATA RAID device IDs

Re: [RFC] AHCI: Remove conflicting Intel Lewisburg 3rd-party SATA RAID device IDs

[Ocean HY1 He]

Hi Tejun,

According to Intel Lewisburg PCH EDS, SATA RAID controller has two types
of DIDs: RSTe RAID and 3rd Party RAID. RSTe RAID shoud be drived by AHCI
driver and used to construct MDRAID. These 3rd Party RAID IDs are reserved
for other 3rd Party RAID drivers for variety.

So, I think it's a bug to let AHCI driver supporting 3rd Party RAID IDs.
And it should be ok to remove these existed device IDs.

Thanks,
Ocean.

Re: [RFC] AHCI: Remove conflicting Intel Lewisburg 3rd-party SATA RAID device IDs

[Christoph Hellwig]

Hi Ocean,

no, all AHCI devices should be driven by the Linux kernel driver.
We do not offer support for '3rd party' drivers in Linux.

Re: [RFC] AHCI: Remove conflicting Intel Lewisburg 3rd-party SATA RAID device IDs

[One Thousand Gnomes]

On Mon, 29 Feb 2016 09:28:19 +0000
Ocean HY1 He <hehy1@lenovo.com> wrote:

> Hi Tejun,
> 
> According to Intel Lewisburg PCH EDS, SATA RAID controller has two types
> of DIDs: RSTe RAID and 3rd Party RAID. RSTe RAID shoud be drived by AHCI
> driver and used to construct MDRAID. These 3rd Party RAID IDs are reserved
> for other 3rd Party RAID drivers for variety.
> 
> So, I think it's a bug to let AHCI driver supporting 3rd Party RAID IDs.
> And it should be ok to remove these existed device IDs.

We have already had this discussion before. I'm surprised it has come
back again, this time via Intel. If Lenovo wants their RAID supported in
Linux then contribute it open source to the kernel. You've been told this
clearly before.

Alan

[RFC] AHCI: Remove conflicting Intel Lewisburg 3rd-party SATA RAID device IDs

[scott.lawson]

From: Scott Lawson <scott.lawson@intel.com>

These PCI device IDs in the AHCI driver conflict with proprietary drivers,
preventing them from loading.

Signed-off-by: Scott Lawson <scott.lawson@intel.com>
---
 drivers/ata/ahci.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/ata/ahci.c b/drivers/ata/ahci.c
index 546a369..3394492 100644
--- a/drivers/ata/ahci.c
+++ b/drivers/ata/ahci.c
@@ -370,11 +370,9 @@ static const struct pci_device_id ahci_pci_tbl[] = {
 	{ PCI_VDEVICE(INTEL, 0x2826), board_ahci }, /* Lewisburg RAID*/
 	{ PCI_VDEVICE(INTEL, 0xa182), board_ahci }, /* Lewisburg AHCI*/
 	{ PCI_VDEVICE(INTEL, 0xa184), board_ahci }, /* Lewisburg RAID*/
-	{ PCI_VDEVICE(INTEL, 0xa186), board_ahci }, /* Lewisburg RAID*/
 	{ PCI_VDEVICE(INTEL, 0xa18e), board_ahci }, /* Lewisburg RAID*/
 	{ PCI_VDEVICE(INTEL, 0xa202), board_ahci }, /* Lewisburg AHCI*/
 	{ PCI_VDEVICE(INTEL, 0xa204), board_ahci }, /* Lewisburg RAID*/
-	{ PCI_VDEVICE(INTEL, 0xa206), board_ahci }, /* Lewisburg RAID*/
 	{ PCI_VDEVICE(INTEL, 0xa20e), board_ahci }, /* Lewisburg RAID*/
 
 	/* JMicron 360/1/3/5/6, match class to avoid IDE function */
-- 
1.9.1

Re: [RFC] AHCI: Remove conflicting Intel Lewisburg 3rd-party SATA RAID device IDs

[Christoph Hellwig]

On Fri, Feb 26, 2016 at 10:09:29AM -0800, scott.lawson@intel.com wrote:
> From: Scott Lawson <scott.lawson@intel.com>
> 
> These PCI device IDs in the AHCI driver conflict with proprietary drivers,
> preventing them from loading.

And that's a good thing!

Re: [RFC] AHCI: Remove conflicting Intel Lewisburg 3rd-party SATA RAID device IDs

[Tejun Heo]

On Fri, Feb 26, 2016 at 10:09:29AM -0800, scott.lawson@intel.com wrote:
> From: Scott Lawson <scott.lawson@intel.com>
> 
> These PCI device IDs in the AHCI driver conflict with proprietary drivers,
> preventing them from loading.

So, the default ahci driver can't attach if these PCI IDs are in use?
I don't think breaking the upstream kernel for a proprietary driver is
a good idea.

Thanks.

-- 
tejun