Re: [git patches] net driver fixes

On Tue, 3 Jul 2007 00:13:29 +0200
Adrian Bunk [off-list ref] wrote:

On Mon, Jul 02, 2007 at 10:54:01AM -0400, Jeff Garzik wrote:

quoted

...
maximilian attems (1):
      starfire list alpha as 64 bit arch
...

--- a/drivers/net/starfire.c
+++ b/drivers/net/starfire.c

@@ -152,7 +152,7 @@ static int full_duplex[MAX_UNITS] = {0, };
  * This SUCKS.
  * We need a much better method to determine if dma_addr_t is 64-bit.
  */
-#if (defined(__i386__) && defined(CONFIG_HIGHMEM64G)) || defined(__x86_64__) || defined (__ia64__) || defined(__mips64__) || (defined(__mips__) && defined(CONFIG_HIGHMEM) && defined(CONFIG_64BIT_PHYS_ADDR))
+#if (defined(__i386__) && defined(CONFIG_HIGHMEM64G)) || defined(__x86_64__) || defined (__ia64__) || defined(__alpha__) || defined(__mips64__) || (defined(__mips__) && defined(CONFIG_HIGHMEM) && defined(CONFIG_64BIT_PHYS_ADDR))
 /* 64-bit dma_addr_t */
 #define ADDR_64BITS	/* This chip uses 64 bit addresses. */
 #define netdrv_addr_t u64

...

The patch is correct and definitely the best solution at this time of 
the 2.6.22 development cycle.

But the comment in the context exactly matches what I thought when
I saw this code...

Does anyone have a suggestion how to do this better?

cu
Adrian

Yeah, write code that handles it in compiler. Like:
	if (sizeof(dma_addr_t) > sizeof(u32)) {
		do something...
	}
That means that it gets checked even if optimizer removes it as
unneeded.

-- 
Stephen Hemminger [off-list ref]