Re: [PATCH 2/4] dmaengine: sh: rcar-dmac: SMMU doesn't need IPMMU workaround

[Wolfram Sang <wsa+renesas@sang-engineering.com>]

[-- Attachment #1: Type: text/plain, Size: 1633 bytes --]


> >  struct rcar_dmac_of_data {
> >         u32 chan_offset_base;
> >         u32 chan_offset_stride;
> > +       unsigned int has_ipmmu:1;
> 
> broken_ipmmu_microTLB_0?

Yes, that's a better name.

> > @@ -2000,11 +2003,13 @@ static void rcar_dmac_shutdown(struct platform_device *pdev)
> >  static const struct rcar_dmac_of_data rcar_dmac_data = {
> >         .chan_offset_base = 0x8000,
> >         .chan_offset_stride = 0x80,
> > +       .has_ipmmu = 1,
> 
> The issue this fixes was seen on R-Car H2 and M2-W.
> R-Car H3 ES1.0 already existed at that time, so R-Car Gen3 might be
> affected or not.

Yes...

> 
> >  };
> >
> >  static const struct rcar_dmac_of_data rcar_gen4_dmac_data = {
> >         .chan_offset_base = 0x0,
> >         .chan_offset_stride = 0x1000,
> > +       .has_ipmmu = 1,
> 
> Let's hope this issue was fixed in R-Car Gen4.  Perhaps someone can
> verify this using the test Laurent shared[1] a long time ago?

... I also envisioned future updates once we find out if Gen3/4 are
affected. But for now, enabling Gen5 has priority and this is what this
series does. Rest can be done incrementally.

> >  static const struct of_device_id rcar_dmac_of_ids[] = {
> 
> [1] https://lore.kernel.org/1422137953-30156-1-git-send-email-laurent.pinchart+renesas@ideasonboard.com

Ah, I didn't find this one. The commit/branch still exists today :)

https://git.linuxtv.org/pinchartl/media.git/commit/?h=dma/iommu/debug&id=86db1ee12532f8de2bca02c0bab77ad7619ef7f2

Unsurprisingly, cherry-picking it (10 years old) to v7.3-rc1 is a mess.
I suggest to postpone this...


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]