The quilt patch titled
Subject: mm/fake-numa: allow later numa node hotplug
has been removed from the -mm tree. Its filename was
mm-fake-numa-allow-later-numa-node-hotplug.patch
This patch was dropped because it was merged into the mm-stable branch
of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
------------------------------------------------------
From: Bruno Faccini <bfaccini@xxxxxxxxxx>
Subject: mm/fake-numa: allow later numa node hotplug
Date: Mon, 6 Jan 2025 04:06:59 -0800
Current fake-numa implementation prevents new Numa nodes to be later
hot-plugged by drivers. A common symptom of this limitation is the "node
<X> was absent from the node_possible_map" message by associated warning
in mm/memory_hotplug.c: add_memory_resource().
This comes from the lack of remapping in both pxm_to_node_map[] and
node_to_pxm_map[] tables to take fake-numa nodes into account and thus
triggers collisions with original and physical nodes only-mapping that had
been determined from BIOS tables.
This patch fixes this by doing the necessary node-ids translation in both
pxm_to_node_map[]/node_to_pxm_map[] tables. node_distance[] table has
also been fixed accordingly.
Details:
When trying to use fake-numa feature on our system where new Numa nodes
are being "hot-plugged" upon driver load, this fails with the following
type of message and warning with stack :
node 8 was absent from the node_possible_map WARNING: CPU: 61 PID: 4259 at
mm/memory_hotplug.c:1506 add_memory_resource+0x3dc/0x418
This issue prevents the use of the fake-NUMA debug feature with the
system's full configuration, when it has proven to be sometimes extremely
useful for performance testing of multi-tasked, memory-bound applications,
as it enables better isolation of processes/ranks compared to fat NUMA
nodes.
Usual numactl output after driver has â??hot-pluggedâ??/unveiled some
new Numa nodes with and without memory :
$ numactl --hardware
available: 9 nodes (0-8)
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
65 66 67 68 69 70 71
node 0 size: 490037 MB
node 0 free: 484432 MB
node 1 cpus:
node 1 size: 97280 MB
node 1 free: 97279 MB
node 2 cpus:
node 2 size: 0 MB
node 2 free: 0 MB
node 3 cpus:
node 3 size: 0 MB
node 3 free: 0 MB
node 4 cpus:
node 4 size: 0 MB
node 4 free: 0 MB
node 5 cpus:
node 5 size: 0 MB
node 5 free: 0 MB
node 6 cpus:
node 6 size: 0 MB
node 6 free: 0 MB
node 7 cpus:
node 7 size: 0 MB
node 7 free: 0 MB
node 8 cpus:
node 8 size: 0 MB
node 8 free: 0 MB
node distances:
node 0 1 2 3 4 5 6 7 8
0: 10 80 80 80 80 80 80 80 80
1: 80 10 255 255 255 255 255 255 255
2: 80 255 10 255 255 255 255 255 255
3: 80 255 255 10 255 255 255 255 255
4: 80 255 255 255 10 255 255 255 255
5: 80 255 255 255 255 10 255 255 255
6: 80 255 255 255 255 255 10 255 255
7: 80 255 255 255 255 255 255 10 255
8: 80 255 255 255 255 255 255 255 10
With recent M.Rapoport set of fake-numa patches in mm-everything
and using numa=fake=4 boot parameter :
$ numactl --hardware
available: 4 nodes (0-3)
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
65 66 67 68 69 70 71
node 0 size: 122518 MB
node 0 free: 117141 MB
node 1 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
65 66 67 68 69 70 71
node 1 size: 219911 MB
node 1 free: 219751 MB
node 2 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
65 66 67 68 69 70 71
node 2 size: 122599 MB
node 2 free: 122541 MB
node 3 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
65 66 67 68 69 70 71
node 3 size: 122479 MB
node 3 free: 122408 MB
node distances:
node 0 1 2 3
0: 10 10 10 10
1: 10 10 10 10
2: 10 10 10 10
3: 10 10 10 10
With recent M.Rapoport set of fake-numa patches in mm-everything,
this patch on top, using numa=fake=4 boot parameter :
# numactl â??hardware
available: 12 nodes (0-11)
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
65 66 67 68 69 70 71
node 0 size: 122518 MB
node 0 free: 116429 MB
node 1 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
65 66 67 68 69 70 71
node 1 size: 122631 MB
node 1 free: 122576 MB
node 2 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
65 66 67 68 69 70 71
node 2 size: 122599 MB
node 2 free: 122544 MB
node 3 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
65 66 67 68 69 70 71
node 3 size: 122479 MB
node 3 free: 122419 MB
node 4 cpus:
node 4 size: 97280 MB
node 4 free: 97279 MB
node 5 cpus:
node 5 size: 0 MB
node 5 free: 0 MB
node 6 cpus:
node 6 size: 0 MB
node 6 free: 0 MB
node 7 cpus:
node 7 size: 0 MB
node 7 free: 0 MB
node 8 cpus:
node 8 size: 0 MB
node 8 free: 0 MB
node 9 cpus:
node 9 size: 0 MB
node 9 free: 0 MB
node 10 cpus:
node 10 size: 0 MB
node 10 free: 0 MB
node 11 cpus:
node 11 size: 0 MB
node 11 free: 0 MB
node distances:
node 0 1 2 3 4 5 6 7 8 9 10 11
0: 10 10 10 10 80 80 80 80 80 80 80 80
1: 10 10 10 10 80 80 80 80 80 80 80 80
2: 10 10 10 10 80 80 80 80 80 80 80 80
3: 10 10 10 10 80 80 80 80 80 80 80 80
4: 80 80 80 80 10 255 255 255 255 255 255 255
5: 80 80 80 80 255 10 255 255 255 255 255 255
6: 80 80 80 80 255 255 10 255 255 255 255 255
7: 80 80 80 80 255 255 255 10 255 255 255 255
8: 80 80 80 80 255 255 255 255 10 255 255 255
9: 80 80 80 80 255 255 255 255 255 10 255 255
10: 80 80 80 80 255 255 255 255 255 255 10 255
11: 80 80 80 80 255 255 255 255 255 255 255 10
Link: https://lkml.kernel.org/r/20250106120659.359610-2-bfaccini@xxxxxxxxxx
Signed-off-by: Bruno Faccini <bfaccini@xxxxxxxxxx>
Cc: David Hildenbrand <david@xxxxxxxxxx>
Cc: John Hubbard <jhubbard@xxxxxxxxxx>
Cc: Mike Rapoport (Microsoft) <rppt@xxxxxxxxxx>
Cc: Zi Yan <ziy@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
drivers/acpi/numa/srat.c | 86 +++++++++++++++++++++++++++++++++
include/acpi/acpi_numa.h | 5 +
include/linux/numa_memblks.h | 3 +
mm/numa_emulation.c | 45 ++++++++++++++---
mm/numa_memblks.c | 2
5 files changed, 133 insertions(+), 8 deletions(-)
--- a/drivers/acpi/numa/srat.c~mm-fake-numa-allow-later-numa-node-hotplug
+++ a/drivers/acpi/numa/srat.c
@@ -81,6 +81,92 @@ int acpi_map_pxm_to_node(int pxm)
}
EXPORT_SYMBOL(acpi_map_pxm_to_node);
+#ifdef CONFIG_NUMA_EMU
+/*
+ * Take max_nid - 1 fake-numa nodes into account in both
+ * pxm_to_node_map()/node_to_pxm_map[] tables.
+ */
+int __init fix_pxm_node_maps(int max_nid)
+{
+ static int pxm_to_node_map_copy[MAX_PXM_DOMAINS] __initdata
+ = { [0 ... MAX_PXM_DOMAINS - 1] = NUMA_NO_NODE };
+ static int node_to_pxm_map_copy[MAX_NUMNODES] __initdata
+ = { [0 ... MAX_NUMNODES - 1] = PXM_INVAL };
+ int i, j, index = -1, count = 0;
+ nodemask_t nodes_to_enable;
+
+ if (numa_off || srat_disabled())
+ return -1;
+
+ /* find fake nodes PXM mapping */
+ for (i = 0; i < MAX_NUMNODES; i++) {
+ if (node_to_pxm_map[i] != PXM_INVAL) {
+ for (j = 0; j <= max_nid; j++) {
+ if ((emu_nid_to_phys[j] == i) &&
+ WARN(node_to_pxm_map_copy[j] != PXM_INVAL,
+ "Node %d is already binded to PXM %d\n",
+ j, node_to_pxm_map_copy[j]))
+ return -1;
+ if (emu_nid_to_phys[j] == i) {
+ node_to_pxm_map_copy[j] =
+ node_to_pxm_map[i];
+ if (j > index)
+ index = j;
+ count++;
+ }
+ }
+ }
+ }
+ if (WARN(index != max_nid, "%d max nid when expected %d\n",
+ index, max_nid))
+ return -1;
+
+ nodes_clear(nodes_to_enable);
+
+ /* map phys nodes not used for fake nodes */
+ for (i = 0; i < MAX_NUMNODES; i++) {
+ if (node_to_pxm_map[i] != PXM_INVAL) {
+ for (j = 0; j <= max_nid; j++)
+ if (emu_nid_to_phys[j] == i)
+ break;
+ /* fake nodes PXM mapping has been done */
+ if (j <= max_nid)
+ continue;
+ /* find first hole */
+ for (j = 0;
+ j < MAX_NUMNODES &&
+ node_to_pxm_map_copy[j] != PXM_INVAL;
+ j++)
+ ;
+ if (WARN(j == MAX_NUMNODES,
+ "Number of nodes exceeds MAX_NUMNODES\n"))
+ return -1;
+ node_to_pxm_map_copy[j] = node_to_pxm_map[i];
+ node_set(j, nodes_to_enable);
+ count++;
+ }
+ }
+
+ /* creating reverse mapping in pxm_to_node_map[] */
+ for (i = 0; i < MAX_NUMNODES; i++)
+ if (node_to_pxm_map_copy[i] != PXM_INVAL &&
+ pxm_to_node_map_copy[node_to_pxm_map_copy[i]] == NUMA_NO_NODE)
+ pxm_to_node_map_copy[node_to_pxm_map_copy[i]] = i;
+
+ /* overwrite with new mapping */
+ for (i = 0; i < MAX_NUMNODES; i++) {
+ node_to_pxm_map[i] = node_to_pxm_map_copy[i];
+ pxm_to_node_map[i] = pxm_to_node_map_copy[i];
+ }
+
+ /* enable other nodes found in PXM for hotplug */
+ nodes_or(numa_nodes_parsed, nodes_to_enable, numa_nodes_parsed);
+
+ pr_debug("found %d total number of nodes\n", count);
+ return 0;
+}
+#endif
+
static void __init
acpi_table_print_srat_entry(struct acpi_subtable_header *header)
{
--- a/include/acpi/acpi_numa.h~mm-fake-numa-allow-later-numa-node-hotplug
+++ a/include/acpi/acpi_numa.h
@@ -17,11 +17,16 @@ extern int node_to_pxm(int);
extern int acpi_map_pxm_to_node(int);
extern unsigned char acpi_srat_revision;
extern void disable_srat(void);
+extern int fix_pxm_node_maps(int max_nid);
extern void bad_srat(void);
extern int srat_disabled(void);
#else /* CONFIG_ACPI_NUMA */
+static inline int fix_pxm_node_maps(int max_nid)
+{
+ return 0;
+}
static inline void disable_srat(void)
{
}
--- a/include/linux/numa_memblks.h~mm-fake-numa-allow-later-numa-node-hotplug
+++ a/include/linux/numa_memblks.h
@@ -29,7 +29,10 @@ int __init numa_cleanup_meminfo(struct n
int __init numa_memblks_init(int (*init_func)(void),
bool memblock_force_top_down);
+extern int numa_distance_cnt;
+
#ifdef CONFIG_NUMA_EMU
+extern int emu_nid_to_phys[MAX_NUMNODES];
int numa_emu_cmdline(char *str);
void __init numa_emu_update_cpu_to_node(int *emu_nid_to_phys,
unsigned int nr_emu_nids);
--- a/mm/numa_emulation.c~mm-fake-numa-allow-later-numa-node-hotplug
+++ a/mm/numa_emulation.c
@@ -8,11 +8,12 @@
#include <linux/memblock.h>
#include <linux/numa_memblks.h>
#include <asm/numa.h>
+#include <acpi/acpi_numa.h>
#define FAKE_NODE_MIN_SIZE ((u64)32 << 20)
#define FAKE_NODE_MIN_HASH_MASK (~(FAKE_NODE_MIN_SIZE - 1UL))
-static int emu_nid_to_phys[MAX_NUMNODES];
+int emu_nid_to_phys[MAX_NUMNODES];
static char *emu_cmdline __initdata;
int __init numa_emu_cmdline(char *str)
@@ -379,6 +380,7 @@ void __init numa_emulation(struct numa_m
size_t phys_size = numa_dist_cnt * numa_dist_cnt * sizeof(phys_dist[0]);
int max_emu_nid, dfl_phys_nid;
int i, j, ret;
+ nodemask_t physnode_mask = numa_nodes_parsed;
if (!emu_cmdline)
goto no_emu;
@@ -395,7 +397,6 @@ void __init numa_emulation(struct numa_m
* split the system RAM into N fake nodes.
*/
if (strchr(emu_cmdline, 'U')) {
- nodemask_t physnode_mask = numa_nodes_parsed;
unsigned long n;
int nid = 0;
@@ -465,9 +466,6 @@ void __init numa_emulation(struct numa_m
*/
max_emu_nid = setup_emu2phys_nid(&dfl_phys_nid);
- /* commit */
- *numa_meminfo = ei;
-
/* Make sure numa_nodes_parsed only contains emulated nodes */
nodes_clear(numa_nodes_parsed);
for (i = 0; i < ARRAY_SIZE(ei.blk); i++)
@@ -475,10 +473,21 @@ void __init numa_emulation(struct numa_m
ei.blk[i].nid != NUMA_NO_NODE)
node_set(ei.blk[i].nid, numa_nodes_parsed);
- numa_emu_update_cpu_to_node(emu_nid_to_phys, ARRAY_SIZE(emu_nid_to_phys));
+ /* fix pxm_to_node_map[] and node_to_pxm_map[] to avoid collision
+ * with faked numa nodes, particularly during later memory hotplug
+ * handling, and also update numa_nodes_parsed accordingly.
+ */
+ ret = fix_pxm_node_maps(max_emu_nid);
+ if (ret < 0)
+ goto no_emu;
+
+ /* commit */
+ *numa_meminfo = ei;
+
+ numa_emu_update_cpu_to_node(emu_nid_to_phys, max_emu_nid + 1);
/* make sure all emulated nodes are mapped to a physical node */
- for (i = 0; i < ARRAY_SIZE(emu_nid_to_phys); i++)
+ for (i = 0; i < max_emu_nid + 1; i++)
if (emu_nid_to_phys[i] == NUMA_NO_NODE)
emu_nid_to_phys[i] = dfl_phys_nid;
@@ -501,12 +510,34 @@ void __init numa_emulation(struct numa_m
numa_set_distance(i, j, dist);
}
}
+ for (i = 0; i < numa_distance_cnt; i++) {
+ for (j = 0; j < numa_distance_cnt; j++) {
+ int physi, physj;
+ u8 dist;
+
+ /* distance between fake nodes is already ok */
+ if (emu_nid_to_phys[i] != NUMA_NO_NODE &&
+ emu_nid_to_phys[j] != NUMA_NO_NODE)
+ continue;
+ if (emu_nid_to_phys[i] != NUMA_NO_NODE)
+ physi = emu_nid_to_phys[i];
+ else
+ physi = i - max_emu_nid;
+ if (emu_nid_to_phys[j] != NUMA_NO_NODE)
+ physj = emu_nid_to_phys[j];
+ else
+ physj = j - max_emu_nid;
+ dist = phys_dist[physi * numa_dist_cnt + physj];
+ numa_set_distance(i, j, dist);
+ }
+ }
/* free the copied physical distance table */
memblock_free(phys_dist, phys_size);
return;
no_emu:
+ numa_nodes_parsed = physnode_mask;
/* No emulation. Build identity emu_nid_to_phys[] for numa_add_cpu() */
for (i = 0; i < ARRAY_SIZE(emu_nid_to_phys); i++)
emu_nid_to_phys[i] = i;
--- a/mm/numa_memblks.c~mm-fake-numa-allow-later-numa-node-hotplug
+++ a/mm/numa_memblks.c
@@ -7,7 +7,7 @@
#include <linux/numa.h>
#include <linux/numa_memblks.h>
-static int numa_distance_cnt;
+int numa_distance_cnt;
static u8 *numa_distance;
nodemask_t numa_nodes_parsed __initdata;
_
Patches currently in -mm which might be from bfaccini@xxxxxxxxxx are
