Re: [PATCH 1/6] mmc: tmio: split core functionality, DMA and MFD glue

[Magnus Damm <magnus.damm@gmail.com>]

On Sun, Mar 13, 2011 at 7:48 AM, Guennadi Liakhovetski
<g.liakhovetski@gmx.de> wrote:
> On Sun, 13 Mar 2011, Magnus Damm wrote:
>
>> On Fri, Mar 11, 2011 at 4:51 PM, Guennadi Liakhovetski
>> <g.liakhovetski@gmx.de> wrote:
>> > TMIO MMC chips contain an SD / SDIO IP core from Panasonic, similar to
>> > the one, used in MN5774 and other MN57xx controllers. These IP cores are
>> > included in many multifunction devices, in sh-mobile chips from Renesas,
>> > in the latter case they can also use DMA. Some sh-mobile implementations
>> > also have some other specialities, that MFD-based solutions don't have.
>> > This makes supporting all these features in a monolithic driver inconveniet
>> > and error-prone. This patch splits the driver into 3 parts: the core,
>> > the MFD glue and the DMA support. In case of a modular build, two modules
>> > will be built: mmc_tmio_core and mmc_tmio.
>> >
>> > Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
>> > ---
>>
>> Thanks for your work on this, I'm happy to see these patches because
>> they make it easier for me to add multi-interrupt source to the SDHI
>> portion - something that is needed for sh73a0.
>>
>> >  drivers/mmc/host/Kconfig        |    4 +
>> >  drivers/mmc/host/Makefile       |    3 +
>> >  drivers/mmc/host/tmio_mmc.c     | 1303 +--------------------------------------
>> >  drivers/mmc/host/tmio_mmc.h     |  152 +++++
>> >  drivers/mmc/host/tmio_mmc_dma.c |  302 +++++++++
>> >  drivers/mmc/host/tmio_mmc_pio.c |  923 +++++++++++++++++++++++++++
>> >  6 files changed, 1402 insertions(+), 1285 deletions(-)
>> >  create mode 100644 drivers/mmc/host/tmio_mmc.h
>> >  create mode 100644 drivers/mmc/host/tmio_mmc_dma.c
>> >  create mode 100644 drivers/mmc/host/tmio_mmc_pio.c
>>
>> Something is wrong with this patch. The Makefile adds tmio_mmc_core.c
>> but it's not part of the diffstat above. Perhaps you're posting an
>> older version of the patch?
>
> No, the Makefile is adding tmio_mmc_core.o:
>
> diff --git a/drivers/mmc/host/Makefile b/drivers/mmc/host/Makefile
> index 30aa686..79c42dd 100644
> --- a/drivers/mmc/host/Makefile
> +++ b/drivers/mmc/host/Makefile
> @@ -29,6 +29,9 @@ endif
>  obj-$(CONFIG_MMC_S3C)          += s3cmci.o
>  obj-$(CONFIG_MMC_SDRICOH_CS)   += sdricoh_cs.o
>  obj-$(CONFIG_MMC_TMIO)         += tmio_mmc.o
> +obj-$(CONFIG_MMC_TMIO_CORE)    += tmio_mmc_core.o
> +tmio_mmc_core-y                        := tmio_mmc_pio.o
> +tmio_mmc_core-$(CONFIG_TMIO_MMC_DMA)   += tmio_mmc_dma.o
>  obj-$(CONFIG_MMC_CB710)        += cb710-mmc.o
>  obj-$(CONFIG_MMC_VIA_SDMMC)    += via-sdmmc.o
>  obj-$(CONFIG_SDH_BFIN)         += bfin_sdh.o
>
> which is linked from two source files. Looks correct to me.

Ok, sorry, my mistake.

What's the reason behind creating tmio_mmc_core.o instead of simply
having tmio_mmc_pio.o + tmio_mmc_dma.o? That must be the most
straightforward solution, no?

Thanks,

/ magnus