On 16.06.24 03:35, Wei Yang wrote:
On s390, zero page's size relies on total ram pages.
Since we plan to move the accounting into __free_pages_core(),
totalram_pages may not represent the total usable pages on system
at this point when defer_init is enabled.
We can get the total usable pages from memblock directly. The size maybe
not accurate due to the alignment, but enough for the calculation.
Signed-off-by: Wei Yang <richard.weiyang@xxxxxxxxx>
CC: Mike Rapoport (IBM) <rppt@xxxxxxxxxx>
CC: David Hildenbrand <david@xxxxxxxxxx>
---
Not tested on a machine, hope it is fine.
---
arch/s390/mm/init.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c
index e769d2726f4e..d811ffa5e147 100644
--- a/arch/s390/mm/init.c
+++ b/arch/s390/mm/init.c
@@ -65,12 +65,13 @@ static void __init setup_zero_pages(void)
unsigned int order;
struct page *page;
int i;
+ unsigned long total_pages = PHYS_PFN(memblock_phys_mem_size() - memblock_reserved_size());
/* Latest machines require a mapping granularity of 512KB */
order = 7;
/* Limit number of empty zero pages for small memory sizes */
- while (order > 2 && (totalram_pages() >> 10) < (1UL << order))
+ while (order > 2 && (total_pages >> 10) < (1UL << order))
order--;
empty_zero_page = __get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
I suspect that this is good enough as an approximation for that purpose.
Reviewed-by: David Hildenbrand <david@xxxxxxxxxx>
--
Cheers,
David / dhildenb
