OpenE2K
/
linux
6
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

microblaze: Use klimit instead of _end for memory init 

For noMMU system when you use larger rootfs image
there is problem with using _end label because
we increase klimit but in memory initialization
we use still _end which is wrong. Larger mtd rootfs
was rewritten by init_bootmem_node.

MMU kernel use static initialization where klimit
is setup to _end. There is no any other hanling
with klimit.

Signed-off-by: Michal Simek <monstr@monstr.eu>
This commit is contained in:
Michal Simek 2009-08-11 12:36:12 +02:00
parent 2856ed35ea
commit 8f37b6c985
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
arch/microblaze/mm/init.c
View File

@ -130,13 +130,13 @@ void __init setup_memory(void)
* (in case the address isn't page-aligned).
*/
#ifndef CONFIG_MMU
map_size = init_bootmem_node(NODE_DATA(0), PFN_UP(TOPHYS((u32)_end)),
map_size = init_bootmem_node(NODE_DATA(0), PFN_UP(TOPHYS((u32)klimit)),
min_low_pfn, max_low_pfn);
#else
map_size = init_bootmem_node(&contig_page_data,
PFN_UP(TOPHYS((u32)_end)), min_low_pfn, max_low_pfn);
PFN_UP(TOPHYS((u32)klimit)), min_low_pfn, max_low_pfn);
#endif
lmb_reserve(PFN_UP(TOPHYS((u32)_end)) << PAGE_SHIFT, map_size);
lmb_reserve(PFN_UP(TOPHYS((u32)klimit)) << PAGE_SHIFT, map_size);
/* free bootmem is whole main memory */
free_bootmem(memory_start, memory_size);