Arnd Bergmann wrote:
Signed-off-by: Mike Rapoport <mike@xxxxxxxxxxxxxx>
CC: Arnd Bergmann <arnd@xxxxxxxx>
CC: Russell King <linux@xxxxxxxxxxxxxxxx>
CC: Colin Cross <ccross@xxxxxxxxxx>
CC: Gary King <GKing@xxxxxxxxxx>
Acked-by: Arnd Bergmann <arnd@xxxxxxxx>
On Tuesday 21 September 2010, Mike Rapoport wrote:
/* On TEGRA, many peripherals are very closely packed in
* two 256MB io windows (that actually only use about 64KB
@@ -67,10 +67,20 @@ void tegra_iounmap(volatile void __iomem *addr);
#define IO_ADDRESS(n) ((void __iomem *) IO_TO_VIRT(n))
+#ifdef CONFIG_TEGRA_PCI
+extern void __iomem *tegra_pcie_io_base;
+
+static inline void __iomem *__io(unsigned long addr)
+{
+ return tegra_pcie_io_base + (addr & IO_SPACE_LIMIT);
+}
+#else
static inline void __iomem *__io(unsigned long addr)
{
return (void __iomem *)addr;
}
+#endif
+
#define __io(a) __io(a)
#define __mem_pci(a) (a)
I wonder what the #else path is for. If you don't have PCI, who would
call __io?
No one should call __io without PCI. However, there are drivers that do not
build without __io, e.g.:
CC drivers/serial/8250.o
drivers/serial/8250.c: In function 'hub6_serial_in':
drivers/serial/8250.c:397: error: implicit declaration of function 'outb'
drivers/serial/8250.c:398: error: implicit declaration of function 'inb'
These references to inb and outb will not be called in runtime anyway, but they
are required for the 8250 driver to build.
I believe there are other drivers except 8250 that require to have __io,
although it will be never actually called on, e.g. Tegra-based platform.
> And wouldn't that result in a NULL pointer dereference?
Arnd
--
Sincerely yours,
Mike.
--
To unsubscribe from this list: send the line "unsubscribe linux-tegra" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html