blob: d7d4384578268208e7f12310919782ced2ede48f (
plain)
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
|
From ab79ec5b6389507b4970d68862abb95d0b2b94c9 Mon Sep 17 00:00:00 2001
From: Christian Ehrhardt <christian.ehrhardt@canonical.com>
Date: Jun 17 2019 10:37:48 +0000
Subject: fix sparse node ids
CPU-ids can be sparse due to disabling a subset of CPUs.
On ppc64le this even will make the node_ids sparse, this is actually pretty
common on ppc64 when SMT is disabled.
Numad has the assumption of cpu/node-ids always being linear and due to that
accesses the 'node' array out of bounds. That triggers crashes like the
following:
Thread 1 "numad" received signal SIGSEGV, Segmentation fault.
#0 0x00000fb6cd2779f4 in bind_process_and_migrate_memory (p=0xfb6fc1e0f70)
at numad.c:998
#1 0x00000fb6cd27d148 in manage_loads () at numad.c:2225
#2 0x00000fb6cd2734dc in main (argc=<optimized out>, argv=<optimized out>)
at numad.c:2654
Instead of directly indexing with node_id we need to detect which array
element has the matching node_id and use that.
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
---
diff --git a/numad.c b/numad.c
index a6a7a5d..524bf61 100644
--- a/numad.c
+++ b/numad.c
@@ -995,7 +995,18 @@ int bind_process_and_migrate_memory(process_data_p p) {
int node_id = 0;
while (nodes) {
if (ID_IS_IN_LIST(node_id, p->node_list_p)) {
- OR_LISTS(cpu_bind_list_p, cpu_bind_list_p, node[node_id].cpu_list_p);
+ int id = -1;
+ for (int node_ix = 0; (node_ix < num_nodes); node_ix++) {
+ if (node[node_ix].node_id == node_id) {
+ id = node_ix;
+ break;
+ }
+ }
+ if (id == -1) {
+ numad_log(LOG_CRIT, "Node %d is requested, but unknown\n", node_id);
+ exit(EXIT_FAILURE);
+ }
+ OR_LISTS(cpu_bind_list_p, cpu_bind_list_p, node[id].cpu_list_p);
nodes -= 1;
}
node_id += 1;
|
