Re: [PATCH V2 14/14] ARM: OMAP4: hwmod data: Clean up the data file

[Sricharan R <r.sricharan@ti.com>]

Hi,
On Saturday 08 June 2013 12:14 AM, Santosh Shilimkar wrote:
> On Friday 07 June 2013 02:37 PM, Tony Lindgren wrote:
>> * Santosh Shilimkar <santosh.shilimkar@ti.com> [130607 11:20]:
>>> On Friday 07 June 2013 01:50 PM, Tony Lindgren wrote:
>>>> * Tony Lindgren <tony@atomide.com> [130607 09:35]:
>>>>> * Paul Walmsley <paul@pwsan.com> [130607 05:38]:
>>>>>> On Fri, 7 Jun 2013, Sricharan R wrote:
>>>>>>
>>>>>>> - The IO resource information like dma request lines, irq number and
>>>>>>>   ocp address space can be populated via dt blob. So such data is stripped
>>>>>>>   from OMAP4 SOC hwmod data file.
>>>>>>>
>>>>>>> - The devices which are still missing the device tree bindings,
>>>>>>>   address space entries are not removed yet. When such devices add
>>>>>>>   the dt bindings, respective address space data can be deleted.
>>>>>>>
>>>>>>> - Also other unnecessary hwmods like firewalls are removed as a part of this.
>>>>>>>   Since emif was getting registered only because of this firewalls links,
>>>>>>>   the mpu->emif direct link is added now.
>>>>>>>
>>>>>>> The above update, results in reduction of about ~1650 lines of code.
>>>>>>>
>>>>>>> Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
>>>>>>> Signed-off-by: Sricharan R <r.sricharan@ti.com>
>>>>>> Acked-by: Paul Walmsley <paul@pwsan.com>
>>>>>>
>>>>>> Can't test this one since I don't have an OMAP4 DT config set up in the testbed
>>>>>> yet.  Maybe will add that to the testbed after the v3.10 release.
>>>>> OK thanks, applying into omap-for-v3.11/cleanup.
>>>> I had to undo the following parts to avoid regressions on omap4sdp.
>>>> Can you please follow up on fixing the related issues so the fixup
>>>> won't be needed?
>>>>
>>>> Seems to work now the same way as earlier for both omap4sdp and blaze
>>>> es, except for DSS, which seems to be a separate issue as posted by
>>>> Tomi. Pushed out now to omap-for-v3.11/cleanup.
>>>>
>>> Thats strange. You shouldn't need those fixes since that data is
>>> expected to be coming from DT.
>>>
>>> Sricharan, Can you please try out the patch against Tony's
>>> V3.11/clean-up branch and see whats missing there.
>> Looking at the diff output with and without, it seems to be that the
>> DMA channels won't get configured.
>>
> That should work since DMA engine changes are in mainline and the
> DMA channel information is already in DT.
>
> regards,
> Santosh
>
 The issue is that mmc, spi drivers try to get the 'dma' request number from
 get_platform_resource, which fails if hwmod has not populated the data.

 There are patches  already from Balaji [1] for hsmmc, Matt Porter [2] mcspi
 for adapting the drivers for Dma engine. Dma engine gets the data from DT.

      [1] http://comments.gmane.org/gmane.linux.kernel.mmc/20378
      [2] http://www.spinics.net/lists/linux-omap/msg87634.html

       [1] is already in  mmc-next for 3.10 branch.

  Then one more patch for spi [3] is required for completing this. This is similar
   to the patch done for mmc to skip the get_platform_resource call in case of DT.

  [3] http://www.mail-archive.com/linux-omap@vger.kernel.org/msg90112.html

Regards,
 Sricharan