Re: [PATCH v4 10/18] iommu/mediatek: Add mt8183 IOMMU support

[Yong Wu]

On Fri, 2018-12-21 at 19:31 +0100, Matthias Brugger wrote:
> 
> On 08/12/2018 09:39, Yong Wu wrote:
> > The M4U IP blocks in mt8183 is MediaTek's generation2 M4U which use
> > the ARM Short-descriptor like mt8173, and most of the HW registers
> > are the same.
> > 
> > Here list main differences between mt8183 and mt8173/mt2712:
> > 1) mt8183 has only one M4U HW like mt8173 while mt2712 has two.
> > 2) mt8183 don't have the "bclk" clock, it use the EMI clock instead.
> > 3) mt8183 can support the dram over 4GB, but it doesn't call this "4GB
> > mode".
> > 4) mt8183 pgtable base register(0x0) extend bit[1:0] which represent
> > the bit[33:32] in the physical address of the pgtable base, But the
> > standard ttbr0[1] means the S bit which is enabled defaultly, Hence,
> > we add a mask.
> > 5) mt8183 HW has a GALS modules, SMI should enable "has_gals" support.
> > 6) the larb-id in smi-common is remapped. M4U should enable
> > larbid_remapped support.
> > 
> > Signed-off-by: Yong Wu <yong.wu@xxxxxxxxxxxx>
> > ---
[...]
> > +static const struct mtk_iommu_plat_data mt8183_data = {
> > +	.m4u_plat            = M4U_MT8183,
> > +	.larbid_remap_enable = true,
> > +	.larbid_remapped     = {0, 4, 5, 6, 7, 2, 3, 1},
> 
> Aren't we reinventing the wheel here?
> Why can't we use larb-id to get the correct id insteaf of providing another data
> structure for the remapping?

Sorry, The remapping id is arbitrary, there is no rule to get it from
the larb-id.

>From Nicolas's comment, I plan to delete "larbid_remap_enable" and only
use "larbid_remap". The other SoCs use the linear mapping here.

In addition, I have to apologize that here will may be improved for
mt2712. There are 2 smi-common(smi-common0 and smi-common1) in mt2712,
actually the remapping relationship for smi-common1 is also different.
If it is really needed, I plan to change it from "larbid_remap" to
"larbid_remap[2]" which 0 is for smi-common0 and 1 is for smi-common1.
Of course, it doesn't affect the iommu functions and only prints the
error log when IOMMU translation fault.

> 
> Regards,
> Matthias

[...]