Setting 'limit' variable to 0 might seem like it means "no limit". But
in the memblock API, 0 actually means the 'MEMBLOCK_ALLOC_ACCESSIBLE'
enum, which limits the physical address range based on
'memblock.current_limit'. This can be confusing.
To make things clearer, I suggest renaming the variable to
'limit_or_flag'. This name shows that the variable can either be a
number for limits or an enum for a flag. This way, readers will easily
understand what kind of value is being passed to the memblock API and
how it works without needing to look into the API details.
Signed-off-by: Leesoo Ahn <lsahn@xxxxxxxxxx>
---
mm/sparse.c | 11 ++++++-----
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/mm/sparse.c b/mm/sparse.c
index de40b2c73406..80e50ba26f24 100644
--- a/mm/sparse.c
+++ b/mm/sparse.c
@@ -333,7 +333,7 @@ sparse_early_usemaps_alloc_pgdat_section(struct pglist_data *pgdat,
unsigned long size)
{
struct mem_section_usage *usage;
- unsigned long goal, limit;
+ unsigned long goal, limit_or_flag;
int nid;
/*
* A page may contain usemaps for other sections preventing the
@@ -346,12 +346,13 @@ sparse_early_usemaps_alloc_pgdat_section(struct pglist_data *pgdat,
* this problem.
*/
goal = pgdat_to_phys(pgdat) & (PAGE_SECTION_MASK << PAGE_SHIFT);
- limit = goal + (1UL << PA_SECTION_SHIFT);
+ limit_or_flag = goal + (1UL << PA_SECTION_SHIFT);
nid = early_pfn_to_nid(goal >> PAGE_SHIFT);
again:
- usage = memblock_alloc_try_nid(size, SMP_CACHE_BYTES, goal, limit, nid);
- if (!usage && limit) {
- limit = 0;
+ usage = memblock_alloc_try_nid(size, SMP_CACHE_BYTES, goal,
+ limit_or_flag, nid);
+ if (!usage && (limit_or_flag != MEMBLOCK_ALLOC_ACCESSIBLE)) {
+ limit_or_flag = MEMBLOCK_ALLOC_ACCESSIBLE;
goto again;
}
return usage;
--
2.34.1
