Re: [BUG FIX] [PATCH] fork_init() max_low_pfn fixes potential OOM bug on big highmem machine

[Coywolf Qi Hunt <coywolf@greatcn.org>]

Russell King wrote:

>On Mon, Jun 28, 2004 at 11:55:29PM +0800, Coywolf Qi Hunt wrote:
>  
>
>>Hello all,
>>
>>On machine with 16G(or 8G if 4k stacks) or more memory, high max_threads 
>>could let system run out of low memory.
>>This patch decides max_threads by the amount of low memory instead of 
>>the total physical memory.
>>Systems without high memory would not be affected.
>>    
>>
>
>This is wrong - max_low_pfn can be high on systems where physical RAM
>doesn't start at address 0.  Such is very common on ARM platforms,
>where RAM is located at 0xa0000000 or 0xc0000000 physical, which
>leads to any calculation based upon max_low_pfn to believe we have
>more than 3GB of RAM when we may only have 64MB or so.
>
>I think we may need a num_lowpages for this...
>
Actually there's physical DRAM offset: PHY_OFFSET, defined on ARM only. 
max_low_pfn happens to be the same as `num_lowpages'.
These assignments seems illogical in naming. But just happen to let this 
patch work.  Other platforms may still break.

[coywolf@everest ~/linux-2.6.7/arch]$ grep max_low_pfn arm* -rn
arm/mm/init.c:235:      max_low_pfn = memend_pfn - O_PFN_DOWN(PHYS_OFFSET);
arm26/mm/init.c:187:    max_low_pfn = memend_pfn - PFN_DOWN(PHYS_OFFSET);
arm26/mm/mm-memc.c:157: page_nr = max_low_pfn;

-- 
Coywolf Qi Hunt
Admin of http://GreatCN.org and http://LoveCN.org