mm: add gpu active/reclaim per-node stat counters
Subsystems (10)
Dave Airlie
maintainerPATCH 1/2 mm: add gpu active/reclaim per-node stat counters
From: Dave Airlie <[email protected]>
While discussing memcg intergration with gpu memory allocations, it was pointed out that there was no numa/system counters for GPU memory allocations.
With more integrated memory GPU server systems turning up, and more requirements for memory tracking it seems we should start closing the gap.
Add two counters to track GPU per-node system memory allocations.
The first is currently allocated to GPU objects, and the second is for memory that is stored in GPU page pools that can be reclaimed, by the shrinker.
---
I'd like to get acks to merge this via the drm tree, if possible,
Dave.
--- Documentation/filesystems/proc.rst | 6 ++++++ drivers/base/node.c | 5 +++++ fs/proc/meminfo.c | 6 ++++++ include/linux/mmzone.h | 2 ++ mm/show_mem.c | 9 +++++++-- mm/vmstat.c | 2 ++ 6 files changed, 28 insertions(+), 2 deletions(-)
Documentation/filesystems/proc.rst +6 -0
@@ -1095,6 +1095,8 @@ Example output. You may not have all of these fields. CmaFree: 0 kB Unaccepted: 0 kB Balloon: 0 kB + GPUActive: 0 kB + GPUReclaim: 0 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0
@@ -1273,6 +1275,10 @@ Unaccepted Memory that has not been accepted by the guest Balloon Memory returned to Host by VM Balloon Drivers +GPUActive + Memory allocated to GPU objects +GPUReclaim + Memory in GPU allocator pools that is reclaimable HugePages_Total, HugePages_Free, HugePages_Rsvd, HugePages_Surp, Hugepagesize, Hugetlb See Documentation/admin-guide/mm/hugetlbpage.rst. DirectMap4k, DirectMap2M, DirectMap1G
drivers/base/node.c +5 -0
@@ -463,6 +463,8 @@ static ssize_t node_read_meminfo(struct device *dev,#ifdef CONFIG_UNACCEPTED_MEMORY "Node %d Unaccepted: %8lu kB\n" #endif +"Node %d GPUActive: %8lu kB\n" +"Node %d GPUReclaim: %8lu kB\n" , nid,K (node_page_state (pgdat,NR_FILE_DIRTY )), nid, K(node_page_state(pgdat, NR_WRITEBACK)),
@@ -496,6 +498,9 @@ static ssize_t node_read_meminfo(struct device *dev, , nid,K (sum_zone_node_page_state (nid,NR_UNACCEPTED )) #endif + , + nid,K (node_page_state (pgdat,NR_GPU_ACTIVE )), + nid,K (node_page_state (pgdat,NR_GPU_RECLAIM )) ); len +=hugetlb_report_node_meminfo (buf, len, nid);return len;
fs/proc/meminfo.c +6 -0
@@ -164,6 +164,12 @@ static int meminfo_proc_show(struct seq_file *m, void *v)show_val_kb (m,"Balloon: " ,global_node_page_state (NR_BALLOON_PAGES )); +show_val_kb (m,"GPUActive: " , +global_node_page_state (NR_GPU_ACTIVE )); + +show_val_kb (m,"GPUReclaim: " , +global_node_page_state (NR_GPU_RECLAIM )); +hugetlb_report_meminfo (m);arch_report_meminfo (m);
include/linux/mmzone.h +2 -0
@@ -241,6 +241,8 @@ enum node_stat_item {NR_HUGETLB , #endifNR_BALLOON_PAGES , +NR_GPU_ACTIVE ,/* GPU pages assigned to an object */ +NR_GPU_RECLAIM ,/* GPU pages in shrinkable pool */ NR_VM_NODE_STAT_ITEMS };
mm/show_mem.c +7 -2
@@ -255,7 +255,9 @@ static void show_free_areas(unsigned int filter, nodemask_t *nodemask, int max_z" sec_pagetables:%lukB" " all_unreclaimable? %s" " Balloon:%lukB" -"\n" , +" gpu_active:%lukB" +" gpu_reclaim:%lukB" +"\n" , pgdat->node_id,K (node_page_state (pgdat,NR_ACTIVE_ANON )), K(node_page_state(pgdat, NR_INACTIVE_ANON)),
@@ -281,7 +283,10 @@ static void show_free_areas(unsigned int filter, nodemask_t *nodemask, int max_zK (node_page_state (pgdat,NR_PAGETABLE )),K (node_page_state (pgdat,NR_SECONDARY_PAGETABLE )),str_yes_no (pgdat->kswapd_failures >=MAX_RECLAIM_RETRIES ), - K(node_page_state(pgdat, NR_BALLOON_PAGES))); +K (node_page_state (pgdat,NR_BALLOON_PAGES )), +K (node_page_state (pgdat,NR_GPU_ACTIVE )), + K(node_page_state(pgdat, NR_GPU_RECLAIM))); + }for_each_populated_zone (zone) {
mm/vmstat.c +2 -0
@@ -1281,6 +1281,8 @@ const char * const vmstat_text[] = {"nr_hugetlb" , #endif"nr_balloon_pages" , +"nr_gpu_active" , +"nr_gpu_reclaim" ,/* system-wide enum vm_stat_item counters */ "nr_dirty_threshold" , "nr_dirty_background_threshold",
Signature · 1 line
2.49.0
Dave Airlie
maintainerPATCH 2/2 drm/ttm: use gpu mm stats to track gpu memory allocations.
From: Dave Airlie <[email protected]>
This uses the newly introduced per-node gpu tracking stats, to track GPU memory allocated via TTM and reclaimable memory in the TTM page pools.
These stats will be useful later for system information and later when mem cgroups are integrated.
--- drivers/gpu/drm/ttm/ttm_pool.c | 22 +++++++++++++++++++++- 1 file changed, 21 insertions(+), 1 deletion(-)
drivers/gpu/drm/ttm/ttm_pool.c +21 -1
@@ -130,6 +130,16 @@ static struct list_head shrinker_list;static struct shrinker *mm_shrinker;static DECLARE_RWSEM (pool_shrink_rwsem); +/* helper to get a current valid node id from a pool */ +static int ttm_pool_nid (struct ttm_pool *pool) { +int nid =NUMA_NO_NODE ; +if (pool) + nid = pool->nid; +if (nid ==NUMA_NO_NODE ) + nid =numa_node_id (); +return nid; +} +/* Allocate pages of size 1 << order with the given gfp_flags */ static struct page *ttm_pool_alloc_page (struct ttm_pool *pool,gfp_t gfp_flags,unsigned int order)
@@ -149,8 +159,10 @@ static struct page *ttm_pool_alloc_page(struct ttm_pool *pool, gfp_t gfp_flags,if (!pool->use_dma_alloc) { p =alloc_pages_node (pool->nid, gfp_flags, order); -if (p) +if (p) { p->private = order; +mod_node_page_state (NODE_DATA (ttm_pool_nid (pool)),NR_GPU_ACTIVE , (1 << order)); + }return p; }
@@ -201,6 +213,7 @@ static void ttm_pool_free_page(struct ttm_pool *pool, enum ttm_caching caching,if (!pool || !pool->use_dma_alloc) {__free_pages (p, order); +mod_node_page_state (NODE_DATA (ttm_pool_nid (pool)),NR_GPU_ACTIVE , -(1 << order));return ; }
@@ -275,6 +288,7 @@ static void ttm_pool_unmap(struct ttm_pool *pool, dma_addr_t dma_addr,static void ttm_pool_type_give (struct ttm_pool_type *pt,struct page *p) {unsigned int i, num_pages =1 << pt->order; +int nid =ttm_pool_nid (pt->pool);for (i =0 ; i < num_pages; ++i) {if (PageHighMem (p))
@@ -287,17 +301,23 @@ static void ttm_pool_type_give(struct ttm_pool_type *pt, struct page *p)list_add (&p->lru, &pt->pages);spin_unlock (&pt->lock);atomic_long_add (1 << pt->order, &allocated_pages); + +mod_node_page_state (NODE_DATA (nid),NR_GPU_ACTIVE , -(1 << pt->order)); +mod_node_page_state (NODE_DATA (nid),NR_GPU_RECLAIM , (1 << pt->order)); }/* Take pages from a specific pool_type, return NULL when nothing available */ static struct page *ttm_pool_type_take (struct ttm_pool_type *pt) {struct page *p; +int nid =ttm_pool_nid (pt->pool);spin_lock (&pt->lock); p =list_first_entry_or_null (&pt->pages,typeof (*p), lru);if (p) {atomic_long_sub (1 << pt->order, &allocated_pages); +mod_node_page_state (NODE_DATA (nid),NR_GPU_ACTIVE , (1 << pt->order)); +mod_node_page_state (NODE_DATA (nid),NR_GPU_RECLAIM , -(1 << pt->order));list_del (&p->lru); }spin_unlock (&pt->lock);
Signature · 1 line
2.49.0
Quoted 8 lines · On Wed, 18 Jun 2025 14:06:17 +1000 Dave Airlie <[email protected]> wrote:
While discussing memcg intergration with gpu memory allocations, it was pointed out that there was no numa/system counters for GPU memory allocations.
With more integrated memory GPU server systems turning up, and more requirements for memory tracking it seems we should start closing the gap.
Quoted 6 lines
Add two counters to track GPU per-node system memory allocations.
The first is currently allocated to GPU objects, and the second is for memory that is stored in GPU page pools that can be reclaimed, by the shrinker.
Shakeel Butt
reviewerQuoted 51 lines · Dave Airlie
From: Dave Airlie <[email protected]>
While discussing memcg intergration with gpu memory allocations, it was pointed out that there was no numa/system counters for GPU memory allocations.
With more integrated memory GPU server systems turning up, and more requirements for memory tracking it seems we should start closing the gap.
Add two counters to track GPU per-node system memory allocations.
The first is currently allocated to GPU objects, and the second is for memory that is stored in GPU page pools that can be reclaimed, by the shrinker.
Cc Christian Koenig <[email protected]>Cc Matthew Brost <[email protected]>Cc Johannes Weiner <[email protected]>Cc Andrew Morton <[email protected]>Signed-off-by Dave Airlie <[email protected]>---
I'd like to get acks to merge this via the drm tree, if possible,
Dave.
--- Documentation/filesystems/proc.rst | 6 ++++++ drivers/base/node.c | 5 +++++ fs/proc/meminfo.c | 6 ++++++ include/linux/mmzone.h | 2 ++ mm/show_mem.c | 9 +++++++-- mm/vmstat.c | 2 ++ 6 files changed, 28 insertions(+), 2 deletions(-)Documentation/filesystems/proc.rst +2
-0@@ -1095,6 +1095,8 @@ Example output. You may not have all of these fields. CmaFree: 0 kB Unaccepted: 0 kB Balloon: 0 kB + GPUActive: 0 kB + GPUReclaim: 0 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0
Documentation/filesystems/proc.rst Unaccepted · line 1281
Quoted 7 lines
Documentation/filesystems/proc.rst +4
-0@@ -1273,6 +1275,10 @@ Unaccepted Memory that has not been accepted by the guest Balloon Memory returned to Host by VM Balloon Drivers +GPUActive + Memory allocated to GPU objects +GPUReclaim + Memory in GPU allocator pools that is reclaimable
Can you please explain a bit more about these GPUActive & GPUReclaim? Please correct me if I am wrong, GPUActive is the total memory used by GPU objects and GPUReclaim is the subset of GPUActive which is reclaimable (possibly through shrinkers).
Quoted 57 lines · Shakeel Butt
Quoted 55 lines · Dave Airlie
From: Dave Airlie <[email protected]>
While discussing memcg intergration with gpu memory allocations, it was pointed out that there was no numa/system counters for GPU memory allocations.
With more integrated memory GPU server systems turning up, and more requirements for memory tracking it seems we should start closing the gap.
Add two counters to track GPU per-node system memory allocations.
The first is currently allocated to GPU objects, and the second is for memory that is stored in GPU page pools that can be reclaimed, by the shrinker.
Cc Christian Koenig <[email protected]>Cc Matthew Brost <[email protected]>Cc Johannes Weiner <[email protected]>Cc Andrew Morton <[email protected]>Signed-off-by Dave Airlie <[email protected]>---
I'd like to get acks to merge this via the drm tree, if possible,
Dave.
--- Documentation/filesystems/proc.rst | 6 ++++++ drivers/base/node.c | 5 +++++ fs/proc/meminfo.c | 6 ++++++ include/linux/mmzone.h | 2 ++ mm/show_mem.c | 9 +++++++-- mm/vmstat.c | 2 ++ 6 files changed, 28 insertions(+), 2 deletions(-)Documentation/filesystems/proc.rst +3
-0@@ -1095,6 +1095,8 @@ Example output. You may not have all of these fields. CmaFree: 0 kB Unaccepted: 0 kB Balloon: 0 kB + GPUActive: 0 kB + GPUReclaim: 0 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0@@ -1273,6 +1275,10 @@ Unaccepted Memory that has not been accepted by the guest Balloon Memory returned to Host by VM Balloon Drivers +GPUActive
Quoted 8 lines
Documentation/filesystems/proc.rst Unaccepted · line 1281
Quoted 3 lines
Documentation/filesystems/proc.rst +3
-0+ Memory allocated to GPU objects +GPUReclaim + Memory in GPU allocator pools that is reclaimableCan you please explain a bit more about these GPUActive & GPUReclaim? Please correct me if I am wrong, GPUActive is the total memory used by GPU objects and GPUReclaim is the subset of GPUActive which is reclaimable (possibly through shrinkers).
Currently,
GPUActive is total memory used by active GPU objects. GPUReclaim is the amount of memory (not a subset of Active) that is being stored in GPU reusable pools, that can be retrieved via a simple shrinker. (this memory usually has different page table attributes, uncached or writecombined).
Example workflow:
User allocates cached system RAM for GPU object:
Active increases,
Free cached system RAM,
Active decreases.
User allocates write combined system RAM for GPU object:
Active increases
Free write combined system RAM
Active decreases,
Reclaim increases
User allocates another WC system RAM object:
Reclaim decreases
Active increases
Shrinker shrinks the pool:
Reclaim decreases.
There could be in the future a 3rd type of memory which I'm not sure it if necessary to account at this level, but it would be Active memory that the driver considers discardable, and could be shrunk easily, but I'm not seeing much consistency on usage in drivers for this, or even what use case it is needed for, so I'm not going to address it yet. This could end up in Reclaim, but I'd need to see the use cases for it.
Dave.