| Message ID | 156140037770.2951909.3387200938880485927.stgit@dwillia2-desk3.amr.corp.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | EFI Specific Purpose Memory Support | expand |
On Mon, Jun 24, 2019 at 8:33 PM Dan Williams <dan.j.williams@intel.com> wrote: > > There are multiple scenarios where the HMAT may contain information > about proximity domains that are not currently online. Rather than fail > to report any HMAT data just elide those offline domains. > > If and when those domains are later onlined they can be added to the > HMEM reporting at that point. > > This was found while testing EFI_MEMORY_SP support which reserves > "specific purpose" memory from the general allocation pool. If that > reservation results in an empty numa-node then the node is not marked > online leading a spurious: > > "acpi/hmat: Ignoring HMAT: Invalid table" > > ...result for HMAT parsing. > > Reviewed-by: Dave Hansen <dave.hansen@linux.intel.com> > Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > --- > drivers/acpi/numa/hmat.c | 14 +++++++++++--- > 1 file changed, 11 insertions(+), 3 deletions(-) > > diff --git a/drivers/acpi/numa/hmat.c b/drivers/acpi/numa/hmat.c > index 96b7d39a97c6..2c220cb7b620 100644 > --- a/drivers/acpi/numa/hmat.c > +++ b/drivers/acpi/numa/hmat.c > @@ -96,9 +96,6 @@ static __init void alloc_memory_target(unsigned int mem_pxm) > { > struct memory_target *target; > > - if (pxm_to_node(mem_pxm) == NUMA_NO_NODE) > - return; > - > target = find_mem_target(mem_pxm); > if (target) > return; > @@ -588,6 +585,17 @@ static __init void hmat_register_targets(void) > struct memory_target *target; > > list_for_each_entry(target, &targets, node) { > + int nid = pxm_to_node(target->memory_pxm); > + > + /* > + * Skip offline nodes. This can happen when memory > + * marked EFI_MEMORY_SP, "specific purpose", is applied > + * to all the memory in a promixity domain leading to > + * the node being marked offline / unplugged, or if > + * memory-only "hotplug" node is offline. > + */ > + if (nid == NUMA_NO_NODE || !node_online(nid)) > + continue; > hmat_register_target_initiators(target); > hmat_register_target_perf(target); > } >
diff --git a/drivers/acpi/numa/hmat.c b/drivers/acpi/numa/hmat.c index 96b7d39a97c6..2c220cb7b620 100644 --- a/drivers/acpi/numa/hmat.c +++ b/drivers/acpi/numa/hmat.c @@ -96,9 +96,6 @@ static __init void alloc_memory_target(unsigned int mem_pxm) { struct memory_target *target; - if (pxm_to_node(mem_pxm) == NUMA_NO_NODE) - return; - target = find_mem_target(mem_pxm); if (target) return; @@ -588,6 +585,17 @@ static __init void hmat_register_targets(void) struct memory_target *target; list_for_each_entry(target, &targets, node) { + int nid = pxm_to_node(target->memory_pxm); + + /* + * Skip offline nodes. This can happen when memory + * marked EFI_MEMORY_SP, "specific purpose", is applied + * to all the memory in a promixity domain leading to + * the node being marked offline / unplugged, or if + * memory-only "hotplug" node is offline. + */ + if (nid == NUMA_NO_NODE || !node_online(nid)) + continue; hmat_register_target_initiators(target); hmat_register_target_perf(target); }
There are multiple scenarios where the HMAT may contain information about proximity domains that are not currently online. Rather than fail to report any HMAT data just elide those offline domains. If and when those domains are later onlined they can be added to the HMEM reporting at that point. This was found while testing EFI_MEMORY_SP support which reserves "specific purpose" memory from the general allocation pool. If that reservation results in an empty numa-node then the node is not marked online leading a spurious: "acpi/hmat: Ignoring HMAT: Invalid table" ...result for HMAT parsing. Reviewed-by: Dave Hansen <dave.hansen@linux.intel.com> Signed-off-by: Dan Williams <dan.j.williams@intel.com> --- drivers/acpi/numa/hmat.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-)