On Wed, 20 Aug 2025 10:34:13 +0300
Mike Rapoport <rppt@xxxxxxxxxx> wrote:
> On Tue, Aug 19, 2025 at 05:54:20PM +0100, Jonathan Cameron wrote:
> > On Tue, 12 Aug 2025 15:26:13 +0100
> > <shiju.jose@xxxxxxxxxx> wrote:
> >
> > > From: Shiju Jose <shiju.jose@xxxxxxxxxx>
> > >
> > > In the numa_memblks, a lookup facility is required to retrieve the
> > > physical address range of memory in a NUMA node. ACPI RAS2 memory
> > > features are among the use cases.
> > >
> > > Suggested-by: Jonathan Cameron <jonathan.cameron@xxxxxxxxxx>
> > > Signed-off-by: Shiju Jose <shiju.jose@xxxxxxxxxx>
> >
> > Looks fine to me. Mike, what do you think?
>
> I still don't see why we can't use existing functions like
> get_pfn_range_for_nid() or memblock_search_pfn_nid().
>
> Or even node_start_pfn() and node_spanned_pages().
Good point. No reason anyone would scrub this on memory that hasn't
been hotplugged yet, so no need to use numa-memblk to get the info.
I guess I was thinking of the wrong hammer :)
I'm not sure node_spanned_pages() works though as we need not to include
ranges that might be on another node as we'd give a wrong impression of
what was being scrubbed.
Should be able to use some combination of node_start_pfn() and maybe
memblock_search_pfn_nid() to get it though (that also gets the nid
we already know but meh, no ral harm in that.)
Jonathan
>
> > One passing comment inline.
> >
> > Reviewed-by: Jonathan Cameron <jonathan.cameron@xxxxxxxxxx>
> >
> > > ---
> > > include/linux/numa.h | 9 +++++++++
> > > include/linux/numa_memblks.h | 2 ++
> > > mm/numa.c | 10 ++++++++++
> > > mm/numa_memblks.c | 37 ++++++++++++++++++++++++++++++++++++
> > > 4 files changed, 58 insertions(+)
> > >
> > > diff --git a/include/linux/numa.h b/include/linux/numa.h
> > > index e6baaf6051bc..1d1aabebd26b 100644
> > > --- a/include/linux/numa.h
> > > +++ b/include/linux/numa.h
> > > @@ -41,6 +41,10 @@ int memory_add_physaddr_to_nid(u64 start);
> > > int phys_to_target_node(u64 start);
> > > #endif
> > >
> > > +#ifndef nid_get_mem_physaddr_range
> > > +int nid_get_mem_physaddr_range(int nid, u64 *start, u64 *end);
> > > +#endif
> > > +
> > > int numa_fill_memblks(u64 start, u64 end);
> > >
> > > #else /* !CONFIG_NUMA */
> > > @@ -63,6 +67,11 @@ static inline int phys_to_target_node(u64 start)
> > > return 0;
> > > }
> > >
> > > +static inline int nid_get_mem_physaddr_range(int nid, u64 *start, u64 *end)
> > > +{
> > > + return 0;
> > > +}
> > > +
> > > static inline void alloc_offline_node_data(int nid) {}
> > > #endif
> > >
> > > diff --git a/include/linux/numa_memblks.h b/include/linux/numa_memblks.h
> > > index 991076cba7c5..7b32d96d0134 100644
> > > --- a/include/linux/numa_memblks.h
> > > +++ b/include/linux/numa_memblks.h
> > > @@ -55,6 +55,8 @@ extern int phys_to_target_node(u64 start);
> > > #define phys_to_target_node phys_to_target_node
> > > extern int memory_add_physaddr_to_nid(u64 start);
> > > #define memory_add_physaddr_to_nid memory_add_physaddr_to_nid
> > > +extern int nid_get_mem_physaddr_range(int nid, u64 *start, u64 *end);
> > > +#define nid_get_mem_physaddr_range nid_get_mem_physaddr_range
> > > #endif /* CONFIG_NUMA_KEEP_MEMINFO */
> > >
> > > #endif /* CONFIG_NUMA_MEMBLKS */
> > > diff --git a/mm/numa.c b/mm/numa.c
> > > index 7d5e06fe5bd4..5335af1fefee 100644
> > > --- a/mm/numa.c
> > > +++ b/mm/numa.c
> > > @@ -59,3 +59,13 @@ int phys_to_target_node(u64 start)
> > > }
> > > EXPORT_SYMBOL_GPL(phys_to_target_node);
> > > #endif
> > > +
> > > +#ifndef nid_get_mem_physaddr_range
> > > +int nid_get_mem_physaddr_range(int nid, u64 *start, u64 *end)
> > > +{
> > > + pr_info_once("Unknown target phys addr range for node=%d\n", nid);
> > > +
> > > + return 0;
> > > +}
> > > +EXPORT_SYMBOL_GPL(nid_get_mem_physaddr_range);
> > > +#endif
> > > diff --git a/mm/numa_memblks.c b/mm/numa_memblks.c
> > > index 541a99c4071a..e1e56b7a3499 100644
> > > --- a/mm/numa_memblks.c
> > > +++ b/mm/numa_memblks.c
> > > @@ -590,4 +590,41 @@ int memory_add_physaddr_to_nid(u64 start)
> > > }
> > > EXPORT_SYMBOL_GPL(memory_add_physaddr_to_nid);
> > >
> > > +/**
> > > + * nid_get_mem_physaddr_range - Get the physical address range
> > > + * of the memblk in the NUMA node.
> > > + * @nid: NUMA node ID of the memblk
> > > + * @start: Start address of the memblk
> > > + * @end: End address of the memblk
> > > + *
> > > + * Find the lowest contiguous physical memory address range of the memblk
> > > + * in the NUMA node with the given nid and return the start and end
> > > + * addresses.
> > > + *
> > > + * RETURNS:
> > > + * 0 on success, -errno on failure.
> > > + */
> > > +int nid_get_mem_physaddr_range(int nid, u64 *start, u64 *end)
> > > +{
> > > + struct numa_meminfo *mi = &numa_meminfo;
> > > + int i;
> > > +
> > > + if (!numa_valid_node(nid))
> > > + return -EINVAL;
> > > +
> > > + for (i = 0; i < mi->nr_blks; i++) {
> > > + if (mi->blk[i].nid == nid) {
> > > + *start = mi->blk[i].start;
> > > + /*
> > > + * Assumption: mi->blk[i].end is the last address
> > > + * in the range + 1.
> >
> > This was my fault for asking on internal review if this was documented
> > anywhere. It's kind of implicitly obvious when reading numa_memblk.c
> > because there are a bunch of end - 1 prints.
> > So can probably drop this comment.
> >
> > > + */
> > > + *end = mi->blk[i].end;
> > > + return 0;
> > > + }
> > > + }
> > > +
> > > + return -ENODEV;
> > > +}
> > > +EXPORT_SYMBOL_GPL(nid_get_mem_physaddr_range);
> > > #endif /* CONFIG_NUMA_KEEP_MEMINFO */
> >
>
