Re: [PATCH v7 5/6] cache: Add L2 cache management for Andes AX45MP RISC-V core

[Lad, Prabhakar]

Hi Conor,

Thank you for the review.

On Fri, Mar 31, 2023 at 1:45 PM Conor Dooley <conor.dooley@xxxxxxxxxxxxx> wrote:
>
> On Thu, Mar 30, 2023 at 09:42:16PM +0100, Prabhakar wrote:
>
> > +STANDALONE CACHE CONTROLLER DRIVERS
>
> > +F:   include/cache
>
> This can go since the file no longer exists.
>
Agreed I will drop this.

> > +config AX45MP_L2_CACHE
> > +     bool "Andes Technology AX45MP L2 Cache controller"
> > +     depends on RISCV && RISCV_DMA_NONCOHERENT
>
> This can just be depends on RISCV_DMA_NONCOHERENT, since that's only
> defined on RISC-V.
>
Agreed.

> > +static void ax45mp_get_l2_line_size(struct platform_device *pdev)
> > +{
> > +     struct device_node *np = pdev->dev.of_node;
> > +     struct device *dev = &pdev->dev;
> > +     int ret;
> > +
> > +     ret = of_property_read_u32(np, "cache-line-size", &ax45mp_priv->ax45mp_cache_line_size);
> > +     if (ret) {
> > +             dev_err(dev, "Failed to get cache-line-size, defaulting to 64 bytes\n");
> > +             ax45mp_priv->ax45mp_cache_line_size = AX45MP_CACHE_LINE_SIZE;
> > +     }
> > +
> > +     if (ax45mp_priv->ax45mp_cache_line_size != AX45MP_CACHE_LINE_SIZE) {
> > +             dev_err(dev, "Expected cache-line-size to be 64 bytes (found:%u). Defaulting to 64 bytes\n",
> > +                     ax45mp_priv->ax45mp_cache_line_size);
> > +             ax45mp_priv->ax45mp_cache_line_size = AX45MP_CACHE_LINE_SIZE;
> > +     }
>
> I forget, why are we doing this defaulting rather than falling over
> immediately if we detect the property is missing or wrong?
>
No reason as such on not failing on property not existing/Invalid. I
will bail out in an error case now.

> > +}
>
> > +static const struct riscv_cache_ops ax45mp_cmo_ops = {
> > +     .clean_range = &ax45mp_cpu_dma_wb_range,
> > +     .inv_range = &ax45mp_cpu_dma_inval_range,
> > +     .flush_range = &ax45mp_cpu_dma_flush_range,
> > +};
>
> I think it would be nice if your driver functions matched the names used
> by the ops. (and as I said on the other patch, I think the ops should
> match the cross-arch naming.
>
Agreed, will do.

> Otherwise, looks grand - although I think I was mostly happy with the
> last revision too.a
>
I know you had provided the RB for the last version ;)

Cheers,
Prabhakar