* [PATCH] mm: Remove unused exact_nid parameter from memmap_alloc()
@ 2026-09-14 5:51 Kaitao Cheng
2026-09-14 6:33 ` Muchun Song
` (2 more replies)
0 siblings, 3 replies; 5+ messages in thread
From: Kaitao Cheng @ 2026-09-14 5:51 UTC (permalink / raw)
To: Mike Rapoport, Andrew Morton, David Hildenbrand, Lorenzo Stoakes,
Liam R . Howlett, Vlastimil Babka, Suren Baghdasaryan,
Michal Hocko, Muchun Song
Cc: linux-mm, linux-kernel, Kaitao Cheng
From: Kaitao Cheng <chengkaitao@kylinos.cn>
Commit ffe55393137c ("mm/sparse: remove sparse buffer pre-allocation
mechanism") removed sparse_buffer_init(), which was the only caller of
memmap_alloc() that passed true for exact_nid.
All remaining callers pass false, leaving the exact-node allocation
branch unreachable. Remove the exact_nid parameter and unconditionally
use memblock_alloc_try_nid_raw().
This has no functional impact.
Signed-off-by: Kaitao Cheng <chengkaitao@kylinos.cn>
---
mm/mm_init.c | 14 ++++----------
mm/mm_init.h | 4 ++--
mm/sparse-vmemmap.c | 2 +-
mm/sparse.c | 2 +-
4 files changed, 8 insertions(+), 14 deletions(-)
diff --git a/mm/mm_init.c b/mm/mm_init.c
index beeeed4688c1..6af5cabf3e75 100644
--- a/mm/mm_init.c
+++ b/mm/mm_init.c
@@ -1601,7 +1601,7 @@ static void __init free_area_init_core(struct pglist_data *pgdat)
}
void __init *memmap_alloc(phys_addr_t size, phys_addr_t align,
- phys_addr_t min_addr, int nid, bool exact_nid)
+ phys_addr_t min_addr, int nid)
{
void *ptr;
@@ -1609,14 +1609,8 @@ void __init *memmap_alloc(phys_addr_t size, phys_addr_t align,
* Kmemleak will explicitly scan mem_map by traversing all valid
* `struct *page`,so memblock does not need to be added to the scan list.
*/
- if (exact_nid)
- ptr = memblock_alloc_exact_nid_raw(size, align, min_addr,
- MEMBLOCK_ALLOC_NOLEAKTRACE,
- nid);
- else
- ptr = memblock_alloc_try_nid_raw(size, align, min_addr,
- MEMBLOCK_ALLOC_NOLEAKTRACE,
- nid);
+ ptr = memblock_alloc_try_nid_raw(size, align, min_addr,
+ MEMBLOCK_ALLOC_NOLEAKTRACE, nid);
if (ptr && size > 0)
page_init_poison(ptr, size);
@@ -1644,7 +1638,7 @@ static void __init alloc_node_mem_map(struct pglist_data *pgdat)
end = ALIGN(pgdat_end_pfn(pgdat), MAX_ORDER_NR_PAGES);
size = (end - start) * sizeof(struct page);
map = memmap_alloc(size, SMP_CACHE_BYTES, MEMBLOCK_LOW_LIMIT,
- pgdat->node_id, false);
+ pgdat->node_id);
if (!map)
panic("Failed to allocate %ld bytes for node %d memory map\n",
size, pgdat->node_id);
diff --git a/mm/mm_init.h b/mm/mm_init.h
index c9fc35e7e9f1..a8ed5c20851b 100644
--- a/mm/mm_init.h
+++ b/mm/mm_init.h
@@ -31,8 +31,8 @@ static inline void clear_zone_contiguous(struct zone *zone)
void memblock_free_pages(unsigned long pfn, unsigned int order);
-void *memmap_alloc(phys_addr_t size, phys_addr_t align, phys_addr_t min_addr,
- int nid, bool exact_nid);
+void *memmap_alloc(phys_addr_t size, phys_addr_t align,
+ phys_addr_t min_addr, int nid);
void memmap_init_range(unsigned long size, int nid, unsigned long zone,
unsigned long start_pfn, unsigned long zone_end_pfn,
diff --git a/mm/sparse-vmemmap.c b/mm/sparse-vmemmap.c
index f22d815d7af0..c2fa5fbf0972 100644
--- a/mm/sparse-vmemmap.c
+++ b/mm/sparse-vmemmap.c
@@ -53,7 +53,7 @@ static void * __ref __earlyonly_bootmem_alloc(int node,
unsigned long align,
unsigned long goal)
{
- return memmap_alloc(size, align, goal, node, false);
+ return memmap_alloc(size, align, goal, node);
}
void * __meminit vmemmap_alloc_block(unsigned long size, int node)
diff --git a/mm/sparse.c b/mm/sparse.c
index cc28bb41fdb1..f032c98b92d9 100644
--- a/mm/sparse.c
+++ b/mm/sparse.c
@@ -215,7 +215,7 @@ struct page __init *__populate_section_memmap(unsigned long pfn,
{
const unsigned long size = PAGE_ALIGN(sizeof(struct page) * PAGES_PER_SECTION);
- return memmap_alloc(size, size, __pa(MAX_DMA_ADDRESS), nid, false);
+ return memmap_alloc(size, size, __pa(MAX_DMA_ADDRESS), nid);
}
#endif /* !CONFIG_SPARSEMEM_VMEMMAP */
--
2.54.0 (Apple Git-157)
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] mm: Remove unused exact_nid parameter from memmap_alloc()
2026-09-14 5:51 [PATCH] mm: Remove unused exact_nid parameter from memmap_alloc() Kaitao Cheng
@ 2026-09-14 6:33 ` Muchun Song
2026-09-15 4:22 ` Andrew Morton
2026-09-15 8:46 ` Lorenzo Stoakes (ARM)
2 siblings, 0 replies; 5+ messages in thread
From: Muchun Song @ 2026-09-14 6:33 UTC (permalink / raw)
To: Kaitao Cheng
Cc: Mike Rapoport, Andrew Morton, David Hildenbrand, Lorenzo Stoakes,
Liam R . Howlett, Vlastimil Babka, Suren Baghdasaryan,
Michal Hocko, linux-mm, linux-kernel, Kaitao Cheng
> On Sep 14, 2026, at 13:51, Kaitao Cheng <kaitao.cheng@linux.dev> wrote:
>
> From: Kaitao Cheng <chengkaitao@kylinos.cn>
>
> Commit ffe55393137c ("mm/sparse: remove sparse buffer pre-allocation
> mechanism") removed sparse_buffer_init(), which was the only caller of
> memmap_alloc() that passed true for exact_nid.
>
> All remaining callers pass false, leaving the exact-node allocation
> branch unreachable. Remove the exact_nid parameter and unconditionally
> use memblock_alloc_try_nid_raw().
>
> This has no functional impact.
>
> Signed-off-by: Kaitao Cheng <chengkaitao@kylinos.cn>
Acked-by: Muchun Song <muchun.song@linux.dev>
Thanks.
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] mm: Remove unused exact_nid parameter from memmap_alloc()
2026-09-14 5:51 [PATCH] mm: Remove unused exact_nid parameter from memmap_alloc() Kaitao Cheng
2026-09-14 6:33 ` Muchun Song
@ 2026-09-15 4:22 ` Andrew Morton
2026-09-15 8:09 ` Kaitao Cheng
2026-09-15 8:46 ` Lorenzo Stoakes (ARM)
2 siblings, 1 reply; 5+ messages in thread
From: Andrew Morton @ 2026-09-15 4:22 UTC (permalink / raw)
To: Kaitao Cheng
Cc: Mike Rapoport, David Hildenbrand, Lorenzo Stoakes,
Liam R . Howlett, Vlastimil Babka, Suren Baghdasaryan,
Michal Hocko, Muchun Song, linux-mm, linux-kernel, Kaitao Cheng
On Mon, 14 Sep 2026 13:51:30 +0800 Kaitao Cheng <kaitao.cheng@linux.dev> wrote:
> From: Kaitao Cheng <chengkaitao@kylinos.cn>
>
> Commit ffe55393137c ("mm/sparse: remove sparse buffer pre-allocation
> mechanism") removed sparse_buffer_init(), which was the only caller of
> memmap_alloc() that passed true for exact_nid.
>
> All remaining callers pass false, leaving the exact-node allocation
> branch unreachable. Remove the exact_nid parameter and unconditionally
> use memblock_alloc_try_nid_raw().
>
> This has no functional impact.
>
Latest development code has changed quite lot, but this was an easy reject
to fix:
> --- a/mm/sparse-vmemmap.c
> +++ b/mm/sparse-vmemmap.c
> @@ -53,7 +53,7 @@ static void * __ref __earlyonly_bootmem_alloc(int node,
> unsigned long align,
> unsigned long goal)
> {
> - return memmap_alloc(size, align, goal, node, false);
> + return memmap_alloc(size, align, goal, node);
> }
>
> void * __meminit vmemmap_alloc_block(unsigned long size, int node)
--- a/mm/sparse.c~mm-remove-unused-exact_nid-parameter-from-memmap_alloc
+++ a/mm/sparse.c
@@ -215,7 +215,7 @@ struct page __init *__populate_section_m
{
const unsigned long size = PAGE_ALIGN(sizeof(struct page) * PAGES_PER_SECTION);
- return memmap_alloc(size, size, __pa(MAX_DMA_ADDRESS), nid, false);
+ return memmap_alloc(size, size, __pa(MAX_DMA_ADDRESS), nid);
}
#endif /* !CONFIG_SPARSEMEM_VMEMMAP */
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] mm: Remove unused exact_nid parameter from memmap_alloc()
2026-09-15 4:22 ` Andrew Morton
@ 2026-09-15 8:09 ` Kaitao Cheng
0 siblings, 0 replies; 5+ messages in thread
From: Kaitao Cheng @ 2026-09-15 8:09 UTC (permalink / raw)
To: Andrew Morton
Cc: Mike Rapoport, David Hildenbrand, Lorenzo Stoakes,
Liam R . Howlett, Vlastimil Babka, Suren Baghdasaryan,
Michal Hocko, Muchun Song, linux-mm, linux-kernel, Kaitao Cheng
在 2026/9/15 12:22, Andrew Morton 写道:
> On Mon, 14 Sep 2026 13:51:30 +0800 Kaitao Cheng <kaitao.cheng@linux.dev> wrote:
>
>> From: Kaitao Cheng <chengkaitao@kylinos.cn>
>>
>> Commit ffe55393137c ("mm/sparse: remove sparse buffer pre-allocation
>> mechanism") removed sparse_buffer_init(), which was the only caller of
>> memmap_alloc() that passed true for exact_nid.
>>
>> All remaining callers pass false, leaving the exact-node allocation
>> branch unreachable. Remove the exact_nid parameter and unconditionally
>> use memblock_alloc_try_nid_raw().
>>
>> This has no functional impact.
>>
>
> Latest development code has changed quite lot, but this was an easy reject
> to fix:
Thank you very much for helping me fix this issue!
>
>> --- a/mm/sparse-vmemmap.c
>> +++ b/mm/sparse-vmemmap.c
>> @@ -53,7 +53,7 @@ static void * __ref __earlyonly_bootmem_alloc(int node,
>> unsigned long align,
>> unsigned long goal)
>> {
>> - return memmap_alloc(size, align, goal, node, false);
>> + return memmap_alloc(size, align, goal, node);
>> }
>>
>> void * __meminit vmemmap_alloc_block(unsigned long size, int node)
>
>
> --- a/mm/sparse.c~mm-remove-unused-exact_nid-parameter-from-memmap_alloc
> +++ a/mm/sparse.c
> @@ -215,7 +215,7 @@ struct page __init *__populate_section_m
> {
> const unsigned long size = PAGE_ALIGN(sizeof(struct page) * PAGES_PER_SECTION);
>
> - return memmap_alloc(size, size, __pa(MAX_DMA_ADDRESS), nid, false);
> + return memmap_alloc(size, size, __pa(MAX_DMA_ADDRESS), nid);
> }
> #endif /* !CONFIG_SPARSEMEM_VMEMMAP */
>
>
--
Thanks
Kaitao Cheng
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] mm: Remove unused exact_nid parameter from memmap_alloc()
2026-09-14 5:51 [PATCH] mm: Remove unused exact_nid parameter from memmap_alloc() Kaitao Cheng
2026-09-14 6:33 ` Muchun Song
2026-09-15 4:22 ` Andrew Morton
@ 2026-09-15 8:46 ` Lorenzo Stoakes (ARM)
2 siblings, 0 replies; 5+ messages in thread
From: Lorenzo Stoakes (ARM) @ 2026-09-15 8:46 UTC (permalink / raw)
To: Kaitao Cheng
Cc: Mike Rapoport, Andrew Morton, David Hildenbrand,
Liam R . Howlett, Vlastimil Babka, Suren Baghdasaryan,
Michal Hocko, Muchun Song, linux-mm, linux-kernel, Kaitao Cheng
On Mon, Sep 14, 2026 at 01:51:30PM +0800, Kaitao Cheng wrote:
> From: Kaitao Cheng <chengkaitao@kylinos.cn>
>
> Commit ffe55393137c ("mm/sparse: remove sparse buffer pre-allocation
> mechanism") removed sparse_buffer_init(), which was the only caller of
> memmap_alloc() that passed true for exact_nid.
>
> All remaining callers pass false, leaving the exact-node allocation
> branch unreachable. Remove the exact_nid parameter and unconditionally
> use memblock_alloc_try_nid_raw().
>
> This has no functional impact.
>
> Signed-off-by: Kaitao Cheng <chengkaitao@kylinos.cn>
> ---
> mm/mm_init.c | 14 ++++----------
> mm/mm_init.h | 4 ++--
> mm/sparse-vmemmap.c | 2 +-
> mm/sparse.c | 2 +-
> 4 files changed, 8 insertions(+), 14 deletions(-)
>
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index beeeed4688c1..6af5cabf3e75 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -1601,7 +1601,7 @@ static void __init free_area_init_core(struct pglist_data *pgdat)
> }
>
> void __init *memmap_alloc(phys_addr_t size, phys_addr_t align,
> - phys_addr_t min_addr, int nid, bool exact_nid)
> + phys_addr_t min_addr, int nid)
> {
> void *ptr;
>
> @@ -1609,14 +1609,8 @@ void __init *memmap_alloc(phys_addr_t size, phys_addr_t align,
> * Kmemleak will explicitly scan mem_map by traversing all valid
> * `struct *page`,so memblock does not need to be added to the scan list.
> */
> - if (exact_nid)
> - ptr = memblock_alloc_exact_nid_raw(size, align, min_addr,
> - MEMBLOCK_ALLOC_NOLEAKTRACE,
> - nid);
I don't see anything else _in the kernel_ reference
memblock_alloc_exact_nid_raw(), though I do see some stuff in
tools/testing/memblock/tests/alloc_exact_nid_api.c.
So I suggest removing memblock_alloc_exact_nid_raw() and
tools/testing/memblock/tests/alloc_exact_nid_api.c (and updating the testing to
account for it as necessary) too.
Also I see both memblock_alloc_try_nid_raw(), memblock_alloc_exact_nid_raw() and
memblock_alloc_try_nid() call memblock_alloc_internal() with an exact_nid last
parameter too.
Only memblock_alloc_exact_nid_raw(), which you should delete, passes this
parameter as true, so you should drop this parameter too.
AND THEN :) it passes exact_nid to memblock_alloc_range_nid() and... I leave it
to you to go investigate if that ever has exact_nid=true passed there.
Let's clean this up all the way please!
> - else
> - ptr = memblock_alloc_try_nid_raw(size, align, min_addr,
> - MEMBLOCK_ALLOC_NOLEAKTRACE,
> - nid);
> + ptr = memblock_alloc_try_nid_raw(size, align, min_addr,
> + MEMBLOCK_ALLOC_NOLEAKTRACE, nid);
>
> if (ptr && size > 0)
> page_init_poison(ptr, size);
> @@ -1644,7 +1638,7 @@ static void __init alloc_node_mem_map(struct pglist_data *pgdat)
> end = ALIGN(pgdat_end_pfn(pgdat), MAX_ORDER_NR_PAGES);
> size = (end - start) * sizeof(struct page);
> map = memmap_alloc(size, SMP_CACHE_BYTES, MEMBLOCK_LOW_LIMIT,
> - pgdat->node_id, false);
> + pgdat->node_id);
> if (!map)
> panic("Failed to allocate %ld bytes for node %d memory map\n",
> size, pgdat->node_id);
> diff --git a/mm/mm_init.h b/mm/mm_init.h
> index c9fc35e7e9f1..a8ed5c20851b 100644
> --- a/mm/mm_init.h
> +++ b/mm/mm_init.h
> @@ -31,8 +31,8 @@ static inline void clear_zone_contiguous(struct zone *zone)
>
> void memblock_free_pages(unsigned long pfn, unsigned int order);
>
> -void *memmap_alloc(phys_addr_t size, phys_addr_t align, phys_addr_t min_addr,
> - int nid, bool exact_nid);
> +void *memmap_alloc(phys_addr_t size, phys_addr_t align,
> + phys_addr_t min_addr, int nid);
>
> void memmap_init_range(unsigned long size, int nid, unsigned long zone,
> unsigned long start_pfn, unsigned long zone_end_pfn,
> diff --git a/mm/sparse-vmemmap.c b/mm/sparse-vmemmap.c
> index f22d815d7af0..c2fa5fbf0972 100644
> --- a/mm/sparse-vmemmap.c
> +++ b/mm/sparse-vmemmap.c
> @@ -53,7 +53,7 @@ static void * __ref __earlyonly_bootmem_alloc(int node,
> unsigned long align,
> unsigned long goal)
> {
> - return memmap_alloc(size, align, goal, node, false);
> + return memmap_alloc(size, align, goal, node);
> }
>
> void * __meminit vmemmap_alloc_block(unsigned long size, int node)
> diff --git a/mm/sparse.c b/mm/sparse.c
> index cc28bb41fdb1..f032c98b92d9 100644
> --- a/mm/sparse.c
> +++ b/mm/sparse.c
> @@ -215,7 +215,7 @@ struct page __init *__populate_section_memmap(unsigned long pfn,
> {
> const unsigned long size = PAGE_ALIGN(sizeof(struct page) * PAGES_PER_SECTION);
>
> - return memmap_alloc(size, size, __pa(MAX_DMA_ADDRESS), nid, false);
> + return memmap_alloc(size, size, __pa(MAX_DMA_ADDRESS), nid);
> }
> #endif /* !CONFIG_SPARSEMEM_VMEMMAP */
>
> --
> 2.54.0 (Apple Git-157)
>
--
Cheers, Lorenzo
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2026-09-15 8:47 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-09-14 5:51 [PATCH] mm: Remove unused exact_nid parameter from memmap_alloc() Kaitao Cheng
2026-09-14 6:33 ` Muchun Song
2026-09-15 4:22 ` Andrew Morton
2026-09-15 8:09 ` Kaitao Cheng
2026-09-15 8:46 ` Lorenzo Stoakes (ARM)
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®