RE: [RFC PATCH] mmc: sdhci-of-arasan: Add auto tuning support for ZynqMP Platform

[Manish Narani]

Hi Adrian,

> -----Original Message-----
> From: Adrian Hunter [mailto:adrian.hunter@xxxxxxxxx]
> Sent: Thursday, February 22, 2018 1:50 PM
> To: Manish Narani <MNARANI@xxxxxxxxxx>; michal.simek@xxxxxxxxxx;
> ulf.hansson@xxxxxxxxxx; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; linux-
> mmc@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> devicetree@xxxxxxxxxxxxxxx; mark.rutland@xxxxxxx; robh+dt@xxxxxxxxxx
> Cc: Anirudha Sarangi <anirudh@xxxxxxxxxx>; Srinivas Goud
> <sgoud@xxxxxxxxxx>
> Subject: Re: [RFC PATCH] mmc: sdhci-of-arasan: Add auto tuning support for
> ZynqMP Platform
> 
> On 21/02/18 17:00, Manish Narani wrote:
> > Hi Adrian,
> >
> >> -----Original Message-----
> >> From: Manish Narani
> >> Sent: Wednesday, February 21, 2018 11:39 AM
> >> To: Adrian Hunter <adrian.hunter@xxxxxxxxx>; michal.simek@xxxxxxxxxx;
> >> ulf.hansson@xxxxxxxxxx; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; linux-
> >> mmc@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> >> devicetree@xxxxxxxxxxxxxxx; mark.rutland@xxxxxxx;
> robh+dt@xxxxxxxxxx
> >> Cc: Anirudha Sarangi <anirudh@xxxxxxxxxx>; Srinivas Goud
> >> <sgoud@xxxxxxxxxx>
> >> Subject: RE: [RFC PATCH] mmc: sdhci-of-arasan: Add auto tuning
> >> support for ZynqMP Platform
> >>
> >> Hi Adrian,
> >>
> >>
> >>> -----Original Message-----
> >>> From: Adrian Hunter [mailto:adrian.hunter@xxxxxxxxx]
> >>> Sent: Friday, February 16, 2018 7:37 PM
> >>> To: Manish Narani <MNARANI@xxxxxxxxxx>; michal.simek@xxxxxxxxxx;
> >>> ulf.hansson@xxxxxxxxxx; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; linux-
> >>> mmc@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> >>> devicetree@xxxxxxxxxxxxxxx; mark.rutland@xxxxxxx;
> robh+dt@xxxxxxxxxx
> >>> Cc: Anirudha Sarangi <anirudh@xxxxxxxxxx>; Srinivas Goud
> >>> <sgoud@xxxxxxxxxx>; Manish Narani <MNARANI@xxxxxxxxxx>
> >>> Subject: Re: [RFC PATCH] mmc: sdhci-of-arasan: Add auto tuning
> >>> support for ZynqMP Platform
> >>>
> >>> On 30/01/18 20:14, Manish Narani wrote:
> >>>> This patch adds support of SD auto tuning for ZynqMP platform. Auto
> >>>> tuning sequence sends tuning block to card when operating in UHS-1
> >>>> modes. This resets the DLL and sends CMD19/CMD21 as a part of the
> >>>> auto tuning process. Once the auto tuning process gets completed,
> >>>> reset the DLL to load the newly obtained SDHC tuned tap value.
> >>>
> >>> How is this different from:
> >>> 	1. reset the dll
> >>> 	2. call sdhci_execute_tuning
> >>> 	3. reset the dll
> >>>
> > Below is my take on your above comments:
> > - 'Reset the dll' is a platform specific call inside
> > 'arasan_zynqmp_execute_tuning' which is implemented in
> > sdhci-of-arasan.c
> > - 'arasan_zynqmp_execute_tuning' is called from 'sdhci_execute_tuning'
> > as a platform_execute_tuning routine
> > - So to keep 'DLL reset' routine called from sdhci-of-arasan.c, I have
> > implemented the execute_tuning in sdhci-of-arasan.c
> 
> I meant something like:
> 
> 	if (of_device_is_compatible(pdev->dev.of_node, "xlnx,zynqmp-
> 8.9a"))
> 		host->mmc_host_ops.execute_tuning =
> arasan_zynqmp_execute_tuning;
> 
This will need the removal of 'const' from 
static const struct mmc_host_ops sdhci_ops = {}
in sdhci.c file. Please confirm.

Thanks,
Manish

> 
> static int arasan_zynqmp_execute_tuning(struct mmc_host *mmc, u32
> opcode) {
> 	struct sdhci_host *host = mmc_priv(mmc);
> 	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
> 	struct sdhci_arasan_data *sdhci_arasan =
> sdhci_pltfm_priv(pltfm_host);
> 	int err;
> 
> 	arasan_zynqmp_dll_reset(host, sdhci_arasan->device_id);
> 
> 	err = sdhci_execute_tuning(mmc, opcode);
> 	if (err)
> 		return err;
> 
> 	arasan_zynqmp_dll_reset(host, sdhci_arasan->device_id);
> 
> 	return 0;
> }
> 
> >
> > Alternative way (Please review):
> > - Define a host->quirk2 bit (SDHCI_QUIRK2_DLL_RESET_NEEDED) in
> > sdhci-of-arasan.c indicating DLL reset needed while tuning operation
> > - Call 'dll reset' routine before and after __sdhci_execute_tuning()
> > in sdhci.c when a host->quirk2 bit (SDHCI_QUIRK2_DLL_RESET_NEEDED) is
> > set
> 
> We should try to avoid quirks.