I don't think you are making things any better by adding
a new ifdef to all the drivers.
> 2. always cast it to u64, which adds unnecessary overhead to 32-bit
> platforms.
Not to HIGHMEM ones. And frankly, trying to super-optimize a driver
because it has two different descriptor format is a mess you as a driver
author choose to get involved in.
Nearly all cards today are 64-bit DMA address descriptors only.
So if anything, this new ifdef will get less and less used over
time.
> 3. use run-time checks all over the place, of the
> "sizeof(dma_addr_t)==sizeof(u64)" kind, which adds unnecessary overhead to
> all platforms.
The compiler optimizes this completely away, it becomes
a compile time test and the unused code block and the test
never make it into the assembler.
So this argument is bogus, as is the define.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/