On 11/9/2013 6:30 AM, Scott Wood wrote:
On Sat, 2013-11-09 at 06:22 +0530, Prabhakar Kushwaha wrote:
On 11/9/2013 6:14 AM, Scott Wood wrote:
On Thu, 2013-10-31 at 08:57 +0530, Prabhakar Kushwaha wrote:
On 10/31/2013 7:37 AM, York Sun wrote:
Freescale IFC controller has been used for mpc8xxx. It will be used for ARM-based SoC as well. This patch moves the driver to driver/mtd and fix the header file includes.
Signed-off-by: York Sun yorksun@freescale.com
Thanks York for this patch. This was pending from long time.
This patch moves fsl_ifc.c from powerpc to driver/mtd I am not sure having 2 files for IFC controller in driver/mtd/ is correct. Although fsl_ifc.c is for IFC Controller(NOR, NAND machine) and fsl_ifc_nand.c is only for IFC - NAND machine. I am asking this question as same changes will be applicable in Linux also.
Also, We have to move changes of arch/powerpc/cpu/mpc85xx/speed.c to fsl_ifc.c . This will make a IFC driver generic.
As this patch has impact on all IFC platforms, has this been tested on few platforms?
I'm not sure that drivers/mtd is the right place for it as it doesn't expose an mtd interface.
Yes Scott, I am also agree with you. But we are not sure about the right place for it.
This is similar to the question of where to put CPM/QE code. A place will need to be identified for glue code that doesn't fit cleanly into an existing subsystem. drivers/misc is probably the right place, though IIRC on the Linux side the drivers/misc maintainer was generally opposed to its existence and pushed back on new things being added to it.
looks like putting fsl_ifc.c in driver/mtd is not good way of handle it. I should sent on RFC of putting this in driver/misc for Linux and see the response. May be I get a better solution there.
Regards, Prabhakar