| Message ID | 20250210063227.41125-3-shivankg@amd.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | Add NUMA mempolicy support for KVM guest-memfd | expand |
On 10.02.25 07:32, Shivank Garg wrote: > Export memory policy related symbols needed by the KVM guest-memfd to > implement NUMA policy support. > > These symbols are required to implement per-memory region NUMA policies > for guest memory, allowing VMMs to control guest memory placement across > NUMA nodes. Probably worth mentioning something like "guest_memfd wants to implement support for NUMA policies just like shmem already does using the shared policy infrastructure. As guest_memfd currently resides in KVM module code, we have to export the relevant symbols. In the future, guest_memfd might be moved to core-mm, at which point the symbols no longer would have to be exported. When/if that happens is still unclear." Acked-by: David Hildenbrand <david@redhat.com>
On 2/12/2025 3:53 PM, David Hildenbrand wrote: > On 10.02.25 07:32, Shivank Garg wrote: >> Export memory policy related symbols needed by the KVM guest-memfd to >> implement NUMA policy support. >> >> These symbols are required to implement per-memory region NUMA policies >> for guest memory, allowing VMMs to control guest memory placement across >> NUMA nodes. > > Probably worth mentioning something like > > "guest_memfd wants to implement support for NUMA policies just like shmem already does using the shared policy infrastructure. As guest_memfd currently resides in KVM module code, we have to export the relevant symbols. > > In the future, guest_memfd might be moved to core-mm, at which point the symbols no longer would have to be exported. When/if that happens is still unclear." > > Acked-by: David Hildenbrand <david@redhat.com> > Thanks for the suggestion. I'll add it. Best Regards, Shivank
On 2/10/25 07:32, Shivank Garg wrote: > Export memory policy related symbols needed by the KVM guest-memfd to > implement NUMA policy support. > > These symbols are required to implement per-memory region NUMA policies > for guest memory, allowing VMMs to control guest memory placement across > NUMA nodes. > > Signed-off-by: Shivank Garg <shivankg@amd.com> I think we should use EXPORT_SYMBOL_GPL() these days. Wasn't there also some way to limit the exports to KVM? > --- > mm/mempolicy.c | 6 ++++++ > 1 file changed, 6 insertions(+) > > diff --git a/mm/mempolicy.c b/mm/mempolicy.c > index bbaadbeeb291..9c15780cfa63 100644 > --- a/mm/mempolicy.c > +++ b/mm/mempolicy.c > @@ -214,6 +214,7 @@ struct mempolicy *get_task_policy(struct task_struct *p) > > return &default_policy; > } > +EXPORT_SYMBOL(get_task_policy); > > static const struct mempolicy_operations { > int (*create)(struct mempolicy *pol, const nodemask_t *nodes); > @@ -347,6 +348,7 @@ void __mpol_put(struct mempolicy *pol) > return; > kmem_cache_free(policy_cache, pol); > } > +EXPORT_SYMBOL(__mpol_put); > > static void mpol_rebind_default(struct mempolicy *pol, const nodemask_t *nodes) > { > @@ -2736,6 +2738,7 @@ struct mempolicy *mpol_shared_policy_lookup(struct shared_policy *sp, > read_unlock(&sp->lock); > return pol; > } > +EXPORT_SYMBOL(mpol_shared_policy_lookup); > > static void sp_free(struct sp_node *n) > { > @@ -3021,6 +3024,7 @@ void mpol_shared_policy_init(struct shared_policy *sp, struct mempolicy *mpol) > mpol_put(mpol); /* drop our incoming ref on sb mpol */ > } > } > +EXPORT_SYMBOL(mpol_shared_policy_init); > > int mpol_set_shared_policy(struct shared_policy *sp, > struct vm_area_struct *vma, struct mempolicy *pol) > @@ -3039,6 +3043,7 @@ int mpol_set_shared_policy(struct shared_policy *sp, > sp_free(new); > return err; > } > +EXPORT_SYMBOL(mpol_set_shared_policy); > > /* Free a backing policy store on inode delete. */ > void mpol_free_shared_policy(struct shared_policy *sp) > @@ -3057,6 +3062,7 @@ void mpol_free_shared_policy(struct shared_policy *sp) > } > write_unlock(&sp->lock); > } > +EXPORT_SYMBOL(mpol_free_shared_policy); > > #ifdef CONFIG_NUMA_BALANCING > static int __initdata numabalancing_override;
On Mon, Feb 17, 2025, Vlastimil Babka wrote: > On 2/10/25 07:32, Shivank Garg wrote: > > Export memory policy related symbols needed by the KVM guest-memfd to > > implement NUMA policy support. > > > > These symbols are required to implement per-memory region NUMA policies > > for guest memory, allowing VMMs to control guest memory placement across > > NUMA nodes. > > > > Signed-off-by: Shivank Garg <shivankg@amd.com> > > I think we should use EXPORT_SYMBOL_GPL() these days. > > Wasn't there also some way to limit the exports to KVM? The infrastructure is still a WIP[1], though when that lands, I definitely plan on tightening down the KVM-induced exports[2]. [1] https://lore.kernel.org/all/20241202145946.108093528@infradead.org [2] https://lore.kernel.org/all/ZzJOoFFPjrzYzKir@google.com
On 2/18/2025 8:52 PM, Sean Christopherson wrote: > On Mon, Feb 17, 2025, Vlastimil Babka wrote: >> On 2/10/25 07:32, Shivank Garg wrote: >>> Export memory policy related symbols needed by the KVM guest-memfd to >>> implement NUMA policy support. >>> >>> These symbols are required to implement per-memory region NUMA policies >>> for guest memory, allowing VMMs to control guest memory placement across >>> NUMA nodes. >>> >>> Signed-off-by: Shivank Garg <shivankg@amd.com> >> >> I think we should use EXPORT_SYMBOL_GPL() these days. Thanks Vlastimil, will use EXPORT_SYMBOL_GPL. >> >> Wasn't there also some way to limit the exports to KVM? > > The infrastructure is still a WIP[1], though when that lands, I definitely plan > on tightening down the KVM-induced exports[2]. > > [1] https://lore.kernel.org/all/20241202145946.108093528@infradead.org > [2] https://lore.kernel.org/all/ZzJOoFFPjrzYzKir@google.com
diff --git a/mm/mempolicy.c b/mm/mempolicy.c index bbaadbeeb291..9c15780cfa63 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -214,6 +214,7 @@ struct mempolicy *get_task_policy(struct task_struct *p) return &default_policy; } +EXPORT_SYMBOL(get_task_policy); static const struct mempolicy_operations { int (*create)(struct mempolicy *pol, const nodemask_t *nodes); @@ -347,6 +348,7 @@ void __mpol_put(struct mempolicy *pol) return; kmem_cache_free(policy_cache, pol); } +EXPORT_SYMBOL(__mpol_put); static void mpol_rebind_default(struct mempolicy *pol, const nodemask_t *nodes) { @@ -2736,6 +2738,7 @@ struct mempolicy *mpol_shared_policy_lookup(struct shared_policy *sp, read_unlock(&sp->lock); return pol; } +EXPORT_SYMBOL(mpol_shared_policy_lookup); static void sp_free(struct sp_node *n) { @@ -3021,6 +3024,7 @@ void mpol_shared_policy_init(struct shared_policy *sp, struct mempolicy *mpol) mpol_put(mpol); /* drop our incoming ref on sb mpol */ } } +EXPORT_SYMBOL(mpol_shared_policy_init); int mpol_set_shared_policy(struct shared_policy *sp, struct vm_area_struct *vma, struct mempolicy *pol) @@ -3039,6 +3043,7 @@ int mpol_set_shared_policy(struct shared_policy *sp, sp_free(new); return err; } +EXPORT_SYMBOL(mpol_set_shared_policy); /* Free a backing policy store on inode delete. */ void mpol_free_shared_policy(struct shared_policy *sp) @@ -3057,6 +3062,7 @@ void mpol_free_shared_policy(struct shared_policy *sp) } write_unlock(&sp->lock); } +EXPORT_SYMBOL(mpol_free_shared_policy); #ifdef CONFIG_NUMA_BALANCING static int __initdata numabalancing_override;
Export memory policy related symbols needed by the KVM guest-memfd to implement NUMA policy support. These symbols are required to implement per-memory region NUMA policies for guest memory, allowing VMMs to control guest memory placement across NUMA nodes. Signed-off-by: Shivank Garg <shivankg@amd.com> --- mm/mempolicy.c | 6 ++++++ 1 file changed, 6 insertions(+)