The pci-rcar driver is enabled for compile tests, and this has
shown that the driver cannot build without CONFIG_OF,
following the inclusion of f8f2fe7355fb "PCI: rcar: Use new OF
interrupt mapping when possible":
drivers/pci/host/pcie-rcar.c: In function 'pci_dma_range_parser_init':
drivers/pci/host/pcie-rcar.c:1039:2: error: implicit declaration of function 'of_n_addr_cells' [-Werror=implicit-function-declaration]
parser->pna = of_n_addr_cells(node);
^
As pointed out by Ben Dooks and Geert Uytterhoeven, this is actually
supposed to build fine, which we can achieve if we make the
declaration of of_irq_parse_and_map_pci conditional on CONFIG_OF
and provide an empty inline function otherwise, as we do for
a lot of other of interfaces.
This lets us build the rcar_pci driver again without CONFIG_OF
for build testing. All platforms using this driver select OF,
so this doesn't change anything for the users.
Cc: Bjorn Helgaas <[email protected]>
Cc: Magnus Damm <[email protected]>
Cc: Geert Uytterhoeven <[email protected]>
Cc: Ben Dooks <[email protected]>
Cc: [email protected]
Cc: [email protected]
Acked-by: Geert Uytterhoeven <[email protected]>
Reviewed-by: Frank Rowand <[email protected]>
Fixes: c25da4778803 ("PCI: rcar: Add Renesas R-Car PCIe driver")
Signed-off-by: Arnd Bergmann <[email protected]>
---
I originally sent the patch in 2014 when things initially broke
and then forgot about while it was applied in my randconfig patch
stack. I sent it again in 2016 and then in July 2017 without reply.
It's still broken and I think this is the correct fix.
---
include/linux/of.h | 3 +++
1 file changed, 3 insertions(+)
diff --git a/include/linux/of.h b/include/linux/of.h
index cfc34117fc92..fb700d1e1fab 100644
--- a/include/linux/of.h
+++ b/include/linux/of.h
@@ -734,6 +734,9 @@ static inline struct device_node *of_get_cpu_node(int cpu,
return NULL;
}
+static inline int of_n_addr_cells(struct device_node *np) { return 0; }
+static inline int of_n_size_cells(struct device_node *np) { return 0; }
+
static inline int of_property_read_u64(const struct device_node *np,
const char *propname, u64 *out_value)
{
--
2.9.0
The ipmmu-vmsa driver fails in compile-testing on non-OF platforms:
drivers/iommu/ipmmu-vmsa.o: In function `ipmmu_of_xlate':
ipmmu-vmsa.c:(.text+0x740): undefined reference to `of_find_device_by_node'
It would be reasonable to assume that this interface works but
returns failure on non-OF builds, like it does on machines that
have been booted in another way, so this adds another inline
function helper.
Fixes: 7b2d59611fef ("iommu/ipmmu-vmsa: Replace local utlb code with fwspec ids")
Signed-off-by: Arnd Bergmann <[email protected]>
---
include/linux/of_platform.h | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/include/linux/of_platform.h b/include/linux/of_platform.h
index e0d1946270f3..fb908e598348 100644
--- a/include/linux/of_platform.h
+++ b/include/linux/of_platform.h
@@ -57,7 +57,14 @@ extern const struct of_device_id of_default_bus_match_table[];
extern struct platform_device *of_device_alloc(struct device_node *np,
const char *bus_id,
struct device *parent);
+#ifdef CONFIG_OF
extern struct platform_device *of_find_device_by_node(struct device_node *np);
+#else
+static inline struct platform_device *of_find_device_by_node(struct device_node *np)
+{
+ return NULL;
+}
+#endif
/* Platform devices and busses creation */
extern struct platform_device *of_platform_device_create(struct device_node *np,
--
2.9.0
On Mon, Sep 11, 2017 at 3:07 PM, Arnd Bergmann <[email protected]> wrote:
> The pci-rcar driver is enabled for compile tests, and this has
> shown that the driver cannot build without CONFIG_OF,
> following the inclusion of f8f2fe7355fb "PCI: rcar: Use new OF
> interrupt mapping when possible":
>
> drivers/pci/host/pcie-rcar.c: In function 'pci_dma_range_parser_init':
> drivers/pci/host/pcie-rcar.c:1039:2: error: implicit declaration of function 'of_n_addr_cells' [-Werror=implicit-function-declaration]
> parser->pna = of_n_addr_cells(node);
> ^
> As pointed out by Ben Dooks and Geert Uytterhoeven, this is actually
> supposed to build fine, which we can achieve if we make the
> declaration of of_irq_parse_and_map_pci conditional on CONFIG_OF
> and provide an empty inline function otherwise, as we do for
> a lot of other of interfaces.
>
> This lets us build the rcar_pci driver again without CONFIG_OF
> for build testing. All platforms using this driver select OF,
> so this doesn't change anything for the users.
Sorry, I keep missing this. It's a flaw in my mail filtering...
There's a patch series[1] to factor out pci_dma_range_parser_init of
the driver which should also fix this. I'd prefer not to take this
because ideally all address parsing should be within the DT core code.
Rob
[1] https://patchwork.kernel.org/patch/9927541/
Hi Rob,
On Mon, Sep 18, 2017 at 11:28 PM, Rob Herring <[email protected]> wrote:
> On Mon, Sep 11, 2017 at 3:07 PM, Arnd Bergmann <[email protected]> wrote:
>> The pci-rcar driver is enabled for compile tests, and this has
>> shown that the driver cannot build without CONFIG_OF,
>> following the inclusion of f8f2fe7355fb "PCI: rcar: Use new OF
>> interrupt mapping when possible":
>>
>> drivers/pci/host/pcie-rcar.c: In function 'pci_dma_range_parser_init':
>> drivers/pci/host/pcie-rcar.c:1039:2: error: implicit declaration of function 'of_n_addr_cells' [-Werror=implicit-function-declaration]
>> parser->pna = of_n_addr_cells(node);
>> ^
>> As pointed out by Ben Dooks and Geert Uytterhoeven, this is actually
>> supposed to build fine, which we can achieve if we make the
>> declaration of of_irq_parse_and_map_pci conditional on CONFIG_OF
>> and provide an empty inline function otherwise, as we do for
>> a lot of other of interfaces.
>>
>> This lets us build the rcar_pci driver again without CONFIG_OF
>> for build testing. All platforms using this driver select OF,
>> so this doesn't change anything for the users.
>
> Sorry, I keep missing this. It's a flaw in my mail filtering...
>
> There's a patch series[1] to factor out pci_dma_range_parser_init of
> the driver which should also fix this. I'd prefer not to take this
> because ideally all address parsing should be within the DT core code.
>
> Rob
>
> [1] https://patchwork.kernel.org/patch/9927541/
Yes, that looks like a better solution.
(For those who didn't see the rest of the series: of_pci_dma_range_parser_init()
in provided in of/address.c, and does have a dummy)
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds