Re: [PATCH 15/21] x86, acpi, numa: Don't reserve memory on nodes the kernel resides in.

2013-07-24 Thread Tang Chen
On 07/24/2013 04:59 AM, Tejun Heo wrote: .. +static bool __init kernel_resides_in_range(phys_addr_t base, u64 length) +{ + int i; + struct memblock_type *reserved =&memblock.reserved; + struct memblock_region *region; + phys_addr_t start, end; + + for (i = 0; i<

Re: [PATCH 15/21] x86, acpi, numa: Don't reserve memory on nodes the kernel resides in.

2013-07-23 Thread Tejun Heo
Hello, On Fri, Jul 19, 2013 at 03:59:28PM +0800, Tang Chen wrote: > /* > + * kernel_resides_in_range - Check if kernel resides in a memory range. > + * @base: The base address of the memory range. > + * @length: The length of the memory range. > + * > + * memblock reserves some memory for the ker

[PATCH 15/21] x86, acpi, numa: Don't reserve memory on nodes the kernel resides in.

2013-07-19 Thread Tang Chen
If all the memory ranges in SRAT are hotpluggable, we should not reserve all of them in memblock. Otherwise the kernel won't have enough memory to boot. And also, memblock will reserve some memory at early time, such initrd file, kernel code and data segments, and so on. We cannot avoid these anyw