Re: [RFC PATCH v2 4/7] soc: qcom: Make the Qualcomm UFS/SDCC ICE a dedicated driver

[Eric Biggers <ebiggers@kernel.org>]

On Wed, Mar 08, 2023 at 11:44:33PM +0200, Abel Vesa wrote:
> On 23-03-08 12:01:41, Eric Biggers wrote:
> > On Wed, Mar 08, 2023 at 05:58:35PM +0200, Abel Vesa wrote:
> > >  * Switched QCOM_INLINE_CRYPTO_ENGINE to tristate and made it built-in
> > >    if any of the UFS or the SDHC drivers are built-in. This is to allow
> > >    the API to be available even if the built-in driver doesn't have
> > >    crypto enabled.
> > [...]
> > > diff --git a/drivers/soc/qcom/Kconfig b/drivers/soc/qcom/Kconfig
> > > index a8f283086a21..c584369e9810 100644
> > > --- a/drivers/soc/qcom/Kconfig
> > > +++ b/drivers/soc/qcom/Kconfig
> > > @@ -275,4 +275,10 @@ config QCOM_ICC_BWMON
> > >  	  the fixed bandwidth votes from cpufreq (CPU nodes) thus achieve high
> > >  	  memory throughput even with lower CPU frequencies.
> > >  
> > > +config QCOM_INLINE_CRYPTO_ENGINE
> > > +	tristate
> > > +	depends on SCSI_UFS_CRYPTO || MMC_CRYPTO
> > > +	default y if SCSI_UFS_QCOM=y || MMC_SDHCI_MSM=y
> > > +	select QCOM_SCM
> > 
> > What are the "depends on" and "default y" lines above for?
> > 
> > You're already selecting this from SCSI_UFS_QCOM and MSM_SDHCI_MSM, as I had
> > suggested.  Isn't that enough?
> 
> We have the following:
> (SCSI_UFS_QCOM && SCSI_UFS_CRYPTO) || (MMC_SDHCI_MSM && MMC_CRYPTO)
> 
> So lets take as example the scenario: (m && y) || (y && n).
> 
> The QCOM_INLINE_CRYPTO_ENGINE will be set to 'm' and the sdhci driver
> will not be able to link properly since the ICE API is part of a module.
> 
> Therefore, if just one of SCSI_UFS_QCOM and MMC_SDHCI_MSM is built-in
> and at least one of the crypto options are enabled, set the
> QCOM_INLINE_CRYPTO_ENGINE to 'y' by default in order to make the
> built-in one linkage will not fail.
> 

That does not make sense.  If MMC_CRYPTO is disabled, then the sdhci-msm driver
should not call any of the ICE APIs.  Likewise for ufs-qcom and SCSI_UFS_CRYPTO.

- Eric