| Message ID | 1ed042a4c5b8c13c6692e12b2d630d75effbb90a.1552595146.git-series.maxime.ripard@bootlin.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | sunxi: Add DT representation for the MBUS controller | expand |
On 14/03/2019 20:26, Maxime Ripard wrote: > Some SoCs have devices that are using a separate bus from the main bus to > perform DMA. > > These buses might have some restrictions and/or different mapping than from > the CPU side, so we'd need to express those using the usual dma-ranges, but > using a different DT node than the node's parent. > > Now that the generic interconnect bindings are available, we can model an > interconnect with the reserved name "dma" for those use-cases. Nit: it's "dma-mem" now. > Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com> > --- > drivers/of/address.c | 29 ++++++++++++++++++++++++++--- > 1 file changed, 26 insertions(+), 3 deletions(-) > > diff --git a/drivers/of/address.c b/drivers/of/address.c > index 140bd0718067..95e47f2e9198 100644 > --- a/drivers/of/address.c > +++ b/drivers/of/address.c > @@ -678,14 +678,31 @@ u64 of_translate_address(struct device_node *dev, const __be32 *in_addr) > } > EXPORT_SYMBOL(of_translate_address); > > +static struct device_node *__of_get_dma_parent(const struct device_node *np) > +{ > + struct of_phandle_args args; > + int ret, index; > + > + index = of_property_match_string(np, "interconnect-names", "dma-mem"); > + if (index < 0) > + return of_get_parent(np); > + > + ret = of_parse_phandle_with_args(np, "interconnects", > + "#interconnect-cells", > + index, &args); > + if (ret < 0) > + return of_get_parent(np); > + > + return of_node_get(args.np); > +} > + > u64 of_translate_dma_address(struct device_node *dev, const __be32 *in_addr) > { > struct device_node *host; > u64 ret; > > - ret = __of_translate_address(dev, of_get_parent, > + ret = __of_translate_address(dev, __of_get_dma_parent, > in_addr, "dma-ranges", &host); > - Nit: ideally patch #3 wouldn't introduce that blank line in the first place. Otherwise, Reviewed-by: Robin Murphy <robin.murphy@arm.com> > if (host) { > of_node_put(host); > return OF_BAD_ADDR; > @@ -913,9 +930,15 @@ int of_dma_get_range(struct device_node *np, u64 *dma_addr, u64 *paddr, u64 *siz > return -EINVAL; > > while (1) { > + struct device_node *parent; > + > naddr = of_n_addr_cells(node); > nsize = of_n_size_cells(node); > - node = of_get_next_parent(node); > + > + parent = __of_get_dma_parent(node); > + of_node_put(node); > + > + node = parent; > if (!node) > break; > >
diff --git a/drivers/of/address.c b/drivers/of/address.c index 140bd0718067..95e47f2e9198 100644 --- a/drivers/of/address.c +++ b/drivers/of/address.c @@ -678,14 +678,31 @@ u64 of_translate_address(struct device_node *dev, const __be32 *in_addr) } EXPORT_SYMBOL(of_translate_address); +static struct device_node *__of_get_dma_parent(const struct device_node *np) +{ + struct of_phandle_args args; + int ret, index; + + index = of_property_match_string(np, "interconnect-names", "dma-mem"); + if (index < 0) + return of_get_parent(np); + + ret = of_parse_phandle_with_args(np, "interconnects", + "#interconnect-cells", + index, &args); + if (ret < 0) + return of_get_parent(np); + + return of_node_get(args.np); +} + u64 of_translate_dma_address(struct device_node *dev, const __be32 *in_addr) { struct device_node *host; u64 ret; - ret = __of_translate_address(dev, of_get_parent, + ret = __of_translate_address(dev, __of_get_dma_parent, in_addr, "dma-ranges", &host); - if (host) { of_node_put(host); return OF_BAD_ADDR; @@ -913,9 +930,15 @@ int of_dma_get_range(struct device_node *np, u64 *dma_addr, u64 *paddr, u64 *siz return -EINVAL; while (1) { + struct device_node *parent; + naddr = of_n_addr_cells(node); nsize = of_n_size_cells(node); - node = of_get_next_parent(node); + + parent = __of_get_dma_parent(node); + of_node_put(node); + + node = parent; if (!node) break;
Some SoCs have devices that are using a separate bus from the main bus to perform DMA. These buses might have some restrictions and/or different mapping than from the CPU side, so we'd need to express those using the usual dma-ranges, but using a different DT node than the node's parent. Now that the generic interconnect bindings are available, we can model an interconnect with the reserved name "dma" for those use-cases. Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com> --- drivers/of/address.c | 29 ++++++++++++++++++++++++++--- 1 file changed, 26 insertions(+), 3 deletions(-)