Re: wrong usage of MAX_DMA_ADDRESS in bootmem.h

On Tue, Sep 30, 2008 at 02:56:26PM -0500, Christoph Lameter wrote:
Nicolas Pitre wrote:
I have implemented highmem for ARM. To catch wrong usage of __pa() and

One issue is that bootmem.h uses __pa(MAX_DMA_ADDRESS). However
MAX_DMA_ADDRESS on ARM is defined as 0xffffffff because there is usually
no restriction on the maximum DMA-able address.

RMK suggested that those places should be using ISA_DMA_THRESHOLD
instead -- here's an excert of our conversation on this topic:


ok so do

#define MAX_DMA_ADDRESS ISA_DMA_THRESHOLD

Not correct. MAX_DMA_ADDRESS is a virtual address. ISA_DMA_THRESHOLD
is the last byte of _physical_ memory which ISA DMA can transfer:

include/asm-x86/scatterlist.h:#define ISA_DMA_THRESHOLD (0x00ffffff)

So what you've just suggested is completely insane.

>> I suspect all those places which are doing __pa(MAX_DMA_ADDRESS) really
want to be using ISA_DMA_THRESHOLD - that's something to raise on LKML
if it's causing problems.

MAX_DMA_ADDRESS is the highest address used for ZONE_DMA / GFP_DMA

Incorrect. MAX_DMA_ADDRESS is the highest possible virtual DMA address:

include/asm-x86/dma.h:#define MAX_DMA_ADDRESS (PAGE_OFFSET + 0x1000000)

Think about what that means on an x86 machine with less than 16MB of RAM.

Does ISA_DMA_THRESHOLD have any meaning on ARM? If you use old ISA stuff then
you need CONFIG_ZONE_DMA and therefore also MAX_DMA_ADDRESS.

As we have already covered in the past, CONFIG_ZONE_DMA has to always
be enabled on ARM because ARM always puts all memory in the first zone.
To do otherwise introduces lots of special cases, and I steadfastly
refuse to make the memory initialisation any more complicated than it
already is.

And besides, this has nothing to do with that issue.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Relevant Pages