[PATCH] mm/mm_init.c: only align start of ZONE_MOVALBE on nodes with memory

[PATCH] mm/mm_init.c: only align start of ZONE_MOVALBE on nodes with memory

[Wei Yang]

At the beginning of find_zone_movable_pfns_for_nodes(), it has properly
set node_states[N_MEMORY] in early_calculate_totalpages().

Instead of iterate on all possible nodes, we can just do the alignment
on nodes with memory.

Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
---
 mm/mm_init.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mm/mm_init.c b/mm/mm_init.c
index 35754a8ae6c0..3ec50b5fda12 100644
--- a/mm/mm_init.c
+++ b/mm/mm_init.c
@@ -547,7 +547,7 @@ static void __init find_zone_movable_pfns_for_nodes(void)
 
 out2:
 	/* Align start of ZONE_MOVABLE on all nids to MAX_ORDER_NR_PAGES */
-	for (nid = 0; nid < MAX_NUMNODES; nid++) {
+	for_each_node_state(nid, N_MEMORY) {
 		unsigned long start_pfn, end_pfn;
 
 		zone_movable_pfn[nid] =
-- 
2.34.1

Re: [PATCH] mm/mm_init.c: only align start of ZONE_MOVALBE on nodes with memory

[David Hildenbrand]

On 11.02.25 09:29, Wei Yang wrote:
> At the beginning of find_zone_movable_pfns_for_nodes(), it has properly
> set node_states[N_MEMORY] in early_calculate_totalpages().
> 
> Instead of iterate on all possible nodes, we can just do the alignment
> on nodes with memory.
> 
> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
> ---
>   mm/mm_init.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index 35754a8ae6c0..3ec50b5fda12 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -547,7 +547,7 @@ static void __init find_zone_movable_pfns_for_nodes(void)
>   
>   out2:
>   	/* Align start of ZONE_MOVABLE on all nids to MAX_ORDER_NR_PAGES */
> -	for (nid = 0; nid < MAX_NUMNODES; nid++) {
> +	for_each_node_state(nid, N_MEMORY) {
>   		unsigned long start_pfn, end_pfn;
>   
>   		zone_movable_pfn[nid] =

I think this should work

Acked-by: David Hildenbrand <david@redhat.com>

-- 
Cheers,

David / dhildenb

Re: [PATCH] mm/mm_init.c: only align start of ZONE_MOVALBE on nodes with memory

[Wei Yang]

On Tue, Feb 11, 2025 at 04:26:26PM +0100, David Hildenbrand wrote:
>On 11.02.25 09:29, Wei Yang wrote:
>> At the beginning of find_zone_movable_pfns_for_nodes(), it has properly
>> set node_states[N_MEMORY] in early_calculate_totalpages().
>> 
>> Instead of iterate on all possible nodes, we can just do the alignment
>> on nodes with memory.
>> 
>> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
>> ---
>>   mm/mm_init.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>> 
>> diff --git a/mm/mm_init.c b/mm/mm_init.c
>> index 35754a8ae6c0..3ec50b5fda12 100644
>> --- a/mm/mm_init.c
>> +++ b/mm/mm_init.c
>> @@ -547,7 +547,7 @@ static void __init find_zone_movable_pfns_for_nodes(void)
>>   out2:
>>   	/* Align start of ZONE_MOVABLE on all nids to MAX_ORDER_NR_PAGES */
>> -	for (nid = 0; nid < MAX_NUMNODES; nid++) {
>> +	for_each_node_state(nid, N_MEMORY) {
>>   		unsigned long start_pfn, end_pfn;
>>   		zone_movable_pfn[nid] =
>
>I think this should work
>
>Acked-by: David Hildenbrand <david@redhat.com>
>

Thanks :-)

>-- 
>Cheers,
>
>David / dhildenb

-- 
Wei Yang
Help you, Help me

Re: [PATCH] mm/mm_init.c: only align start of ZONE_MOVALBE on nodes with memory

[Dev Jain]

On 11/02/25 1:59 pm, Wei Yang wrote:
> At the beginning of find_zone_movable_pfns_for_nodes(), it has properly
> set node_states[N_MEMORY] in early_calculate_totalpages().
> 
> Instead of iterate on all possible nodes, we can just do the alignment
> on nodes with memory.
> 
> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
> ---
>   mm/mm_init.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index 35754a8ae6c0..3ec50b5fda12 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -547,7 +547,7 @@ static void __init find_zone_movable_pfns_for_nodes(void)
>   
>   out2:
>   	/* Align start of ZONE_MOVABLE on all nids to MAX_ORDER_NR_PAGES */
> -	for (nid = 0; nid < MAX_NUMNODES; nid++) {
> +	for_each_node_state(nid, N_MEMORY) {
>   		unsigned long start_pfn, end_pfn;
>   
>   		zone_movable_pfn[nid] =

I have boot tested this, and also read the code and it makes sense: The 
only flag which gives the possibility of ZONE_MOVABLE is N_MEMORY, and 
it gets set in early_calculate_totalpages(), and before the label out2, 
no function plays with node_states[]. So please consider:

Reviewed-by: Dev Jain <dev.jain@arm.com>

Re: [PATCH] mm/mm_init.c: only align start of ZONE_MOVALBE on nodes with memory

[Wei Yang]

On Wed, Feb 12, 2025 at 09:27:08PM +0530, Dev Jain wrote:
>
>
>On 11/02/25 1:59 pm, Wei Yang wrote:
>> At the beginning of find_zone_movable_pfns_for_nodes(), it has properly
>> set node_states[N_MEMORY] in early_calculate_totalpages().
>> 
>> Instead of iterate on all possible nodes, we can just do the alignment
>> on nodes with memory.
>> 
>> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
>> ---
>>   mm/mm_init.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>> 
>> diff --git a/mm/mm_init.c b/mm/mm_init.c
>> index 35754a8ae6c0..3ec50b5fda12 100644
>> --- a/mm/mm_init.c
>> +++ b/mm/mm_init.c
>> @@ -547,7 +547,7 @@ static void __init find_zone_movable_pfns_for_nodes(void)
>>   out2:
>>   	/* Align start of ZONE_MOVABLE on all nids to MAX_ORDER_NR_PAGES */
>> -	for (nid = 0; nid < MAX_NUMNODES; nid++) {
>> +	for_each_node_state(nid, N_MEMORY) {
>>   		unsigned long start_pfn, end_pfn;
>>   		zone_movable_pfn[nid] =
>
>I have boot tested this, and also read the code and it makes sense: The only
>flag which gives the possibility of ZONE_MOVABLE is N_MEMORY, and it gets set
>in early_calculate_totalpages(), and before the label out2, no function plays
>with node_states[]. So please consider:
>
>Reviewed-by: Dev Jain <dev.jain@arm.com>

Thanks :-)

-- 
Wei Yang
Help you, Help me

Re: [PATCH] mm/mm_init.c: only align start of ZONE_MOVALBE on nodes with memory

[Anshuman Khandual]

On 2/11/25 13:59, Wei Yang wrote:
> At the beginning of find_zone_movable_pfns_for_nodes(), it has properly
> set node_states[N_MEMORY] in early_calculate_totalpages().
> 
> Instead of iterate on all possible nodes, we can just do the alignment
> on nodes with memory.
> 
> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
> ---
>  mm/mm_init.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index 35754a8ae6c0..3ec50b5fda12 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -547,7 +547,7 @@ static void __init find_zone_movable_pfns_for_nodes(void)
>  
>  out2:
>  	/* Align start of ZONE_MOVABLE on all nids to MAX_ORDER_NR_PAGES */
> -	for (nid = 0; nid < MAX_NUMNODES; nid++) {
> +	for_each_node_state(nid, N_MEMORY) {
>  		unsigned long start_pfn, end_pfn;
>  
>  		zone_movable_pfn[nid] =

LGTM

Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>

Re: [PATCH] mm/mm_init.c: only align start of ZONE_MOVALBE on nodes with memory

[Mike Rapoport]

On Tue, Feb 11, 2025 at 08:29:00AM +0000, Wei Yang wrote:
> At the beginning of find_zone_movable_pfns_for_nodes(), it has properly
> set node_states[N_MEMORY] in early_calculate_totalpages().
> 
> Instead of iterate on all possible nodes, we can just do the alignment
> on nodes with memory.
> 
> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>

Reviewed-by: Mike Rapoport (Microsoft) <rppt@kernel.org>

> ---
>  mm/mm_init.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index 35754a8ae6c0..3ec50b5fda12 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -547,7 +547,7 @@ static void __init find_zone_movable_pfns_for_nodes(void)
>  
>  out2:
>  	/* Align start of ZONE_MOVABLE on all nids to MAX_ORDER_NR_PAGES */
> -	for (nid = 0; nid < MAX_NUMNODES; nid++) {
> +	for_each_node_state(nid, N_MEMORY) {
>  		unsigned long start_pfn, end_pfn;
>  
>  		zone_movable_pfn[nid] =
> -- 
> 2.34.1
> 

-- 
Sincerely yours,
Mike.