Hi Angelo
Sorry for late due to miss the mail
On Wed, 2022-10-26 at 10:42 +0200, AngeloGioacchino Del Regno wrote:
> Il 26/10/22 09:39, Eddie Huang ha scritto:
> > Add Mediatek mcq resource and runtime configuration function
> > to support MCQ capability
> >
> > Signed-off-by: Eddie Huang <eddie.huang@xxxxxxxxxxxx>
> > ---
> > drivers/ufs/host/ufs-mediatek.c | 37
> > +++++++++++++++++++++++++++++++++++++
> > drivers/ufs/host/ufs-mediatek.h | 7 +++++++
> > 2 files changed, 44 insertions(+)
> >
> > diff --git a/drivers/ufs/host/ufs-mediatek.c
> > b/drivers/ufs/host/ufs-mediatek.c
> > index c958279..3f5fc05 100644
> > --- a/drivers/ufs/host/ufs-mediatek.c
> > +++ b/drivers/ufs/host/ufs-mediatek.c
>
> ..snip..
>
> > +
> > +static int ufs_mtk_config_mcq_resource(struct ufs_hba *hba)
> > +{
> > + hba->mcq_base = hba->mmio_base +
> > + MCQ_QUEUE_OFFSET(hba-
> > >mcq_capabilities);
>
This define in UFSHCI4 spec
ADDR(SQATTR0) = UFS_HCI_BASE + QCFGPTR*200h
> This seems to either be an additional usecase that should be
> implemented into the
> API and not in MediaTek drivers, (as in that case I believe MediaTek
> won't be the
> only user of such usecase)... or just a way to avoid adding the MCQ
> iospace to the
> UFS devicetree node.
>
I can base on next version of UFS MCQ patch [1], and check whether add
new API
[1]: https://www.spinics.net/lists/linux-scsi/msg178322.html
Thanks,
Eddie
