Re: [PATCH 0/4] R-Car SATA driver and platform data

[Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>]

Hello.

On 18-02-2013 5:23, Simon Horman wrote:

>>     Here's the set of 4 patches against the Simon Horman's renesas.git repo, 'next' branch:

>> [1/4] ARM: mach-shmobile: r8a7779: SATA DT configuration
>> [2/4] ARM: mach-shmobile: r8a7779: add SATA support
>> [3/4] libata: add R-Car SATA driver
>> [4/4] ARM: mach-shmobile: marzen: add SATA support

>> Patch 3 also applies to the 'upstream' branch of Jeff Garzik's libata-dev.git repo.

> Hi Sergei, Hi All,

> I do not expect patch 3 of this series to got through
> the renesas tree, however, I have included it and the other three patches
> in this series to a topic branch in the renesas tree, topic/r-car-stat.

    I probably make the patch 3 patch 1 in the next posting.

> The base of the topic branch is:
> * a merge of renesas.git next and libata-dev.git upstream, and;
> * a boot fix for Marzen which I plan to add to next once it is reviewed:
>    "ARM: shmobile: r8a7779: Correct TMU clock support again" (v3).

> With this in place I built a kernel using the marzen_defconfig and booted it.
> However, I wonder if I am missing a kernel option, or perhaps something
> else of a minor nature as I see the following:

> # dmesg | grep sata
> sata_rcar sata_rcar: coherent DMA mask is unset
> sata_rcar sata_rcar: failed to start port 0 (errno=-12)
> sata_rcar: probe of sata_rcar failed with error -12

    No, it's entirely our fault -- Vladimir didn't test the driver in non-DT 
mode with DMA support added (due to the above mentioned boot fix missing), and 
I didn't notice the missing intializers for 
'sata_device.dev.[coherent_]dma_mask' when reviewing his patches. V2 will be 
posted soon.

WBR, Sergei