Re: [PATCH V3 2/2] mmc: sdhci-msm: Re-initialize DLL if MCLK is gated dynamically

[Veerabhadrarao Badiganti <vbadigan@codeaurora.org>]

On 11/2/2018 2:03 AM, Doug Anderson wrote:
> Hi,
>
> On Thu, Nov 1, 2018 at 5:08 AM Veerabhadrarao Badiganti
> <vbadigan@codeaurora.org> wrote:
>>   static const struct sdhci_msm_variant_info sdhci_msm_v5_var = {
>>          .mci_removed = true,
>> +       .restore_dll_config = false,
>> +       .var_ops = &v5_var_ops,
>> +       .offset = &sdhci_msm_v5_offset,
>> +};
>> +
>> +static const struct sdhci_msm_variant_info sdm845_sdhci_var = {
>> +       .mci_removed = true,
>> +       .restore_dll_config = true,
>>          .var_ops = &v5_var_ops,
>>          .offset = &sdhci_msm_v5_offset,
>>   };
> One last thing: are there actually any "v5" controllers that _don't_
> require restoring the DLL?  Since "sdm845" is currently the only v5
> controller maybe just set "restore_dll_config = true" for all v5
> controllers and when there's a new v5 controller that doesn't need it
> then match off the SoC-specific compatible string.  As per my review
> of the bindings patch IMO you should include both the "v5" and the
> SoC-specific string for SDM845 (and all future SoCs) so you could make
> the generic v5 case do this...
>

Yes. QCS404 is one of the target which uses "V5" controller and it 
doesn't need restoring of DLL.
I checked your comments on bindings patch. Will update it in the next 
patchset.

> -Doug

Thanks,
Veera