Re: ohci1394 problem (MMIO broken) (was 2.6.25-rc6-git6: Reported regressions from 2.6.24)

[Stefan Richter <stefanr@s5r6.in-berlin.de>]

Linus Torvalds wrote:
> 
> On Sat, 22 Mar 2008, Linus Torvalds wrote:
>> AND no, I don't think our x86-32 ioremap() actually works for this case, 
>> because while the resource data may have the full 64 bits, when the 
>> ioremap() happens it gets truncated to 32 bits.
> 
> Does this patch make any difference?
> 
> (ENTIRELY untested, I checked that it compiles on x86-64, but didn't even 
> test a 32-bit build, I'm hoping whoever sees this issue can also fix up 
> the inevitable small missed pieces)
> 
> 		Linus
> 
> ---
>  arch/x86/mm/ioremap.c   |    6 +++---
>  include/asm-x86/io_32.h |    6 +++---
>  include/asm-x86/io_64.h |    6 +++---
>  lib/iomap.c             |    2 +-
>  4 files changed, 10 insertions(+), 10 deletions(-)
> 
> diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
> index 8fe576b..4afaba0 100644
> --- a/arch/x86/mm/ioremap.c
> +++ b/arch/x86/mm/ioremap.c
> @@ -106,7 +106,7 @@ static int ioremap_change_attr(unsigned long vaddr, unsigned long size,
>   * have to convert them into an offset in a page-aligned mapping, but the
>   * caller shouldn't need to know that small detail.
>   */
> -static void __iomem *__ioremap(unsigned long phys_addr, unsigned long size,
> +static void __iomem *__ioremap(resource_size_t phys_addr, unsigned long size,
>  			       enum ioremap_mode mode)
[...]

On my x86-32 test system, none of the devices gets an MMIO region above 
the 4G map.  (Reconfigured with CONFIG_RESOURCES_64BIT=y while having 
CONFIG_HIGHMEM4G=y, like in Thomas M's .config.)

So the only thing I can confirm is that the patch builds and runs fine 
on x86-64.
-- 
Stefan Richter
-=====-==--- --== =-===
http://arcgraph.de/sr/