Script 'mail_helper' called by obssrc
Hello community,
here is the log from the commit of package mcelog for openSUSE:Factory checked
in at 2025-04-15 16:45:07
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Comparing /work/SRC/openSUSE:Factory/mcelog (Old)
and /work/SRC/openSUSE:Factory/.mcelog.new.1907 (New)
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Package is "mcelog"
Tue Apr 15 16:45:07 2025 rev:71 rq:1269206 version:204
Changes:
--------
--- /work/SRC/openSUSE:Factory/mcelog/mcelog.changes 2025-01-29
16:09:34.725475361 +0100
+++ /work/SRC/openSUSE:Factory/.mcelog.new.1907/mcelog.changes 2025-04-15
16:48:10.480803822 +0200
@@ -1,0 +2,8 @@
+Fri Apr 11 12:34:24 UTC 2025 - tr...@suse.de
+
+- Update to version 204:
+ * Enable offline retries by default
+ * Add ability to retry failed page offlines with an exponential backoff
+ * Fix misspelling in variable name
+
+-------------------------------------------------------------------
Old:
----
mcelog-202.obscpio
New:
----
mcelog-204.obscpio
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Other differences:
------------------
++++++ mcelog.spec ++++++
--- /var/tmp/diff_new_pack.1EDHAv/_old 2025-04-15 16:48:11.476845529 +0200
+++ /var/tmp/diff_new_pack.1EDHAv/_new 2025-04-15 16:48:11.476845529 +0200
@@ -21,7 +21,7 @@
%define _fillupdir %{_localstatedir}/adm/fillup-templates
%endif
Name: mcelog
-Version: 202
+Version: 204
Release: 0
Summary: Log Machine Check Events
License: GPL-2.0-only
++++++ _servicedata ++++++
--- /var/tmp/diff_new_pack.1EDHAv/_old 2025-04-15 16:48:11.532847874 +0200
+++ /var/tmp/diff_new_pack.1EDHAv/_new 2025-04-15 16:48:11.536848042 +0200
@@ -7,6 +7,6 @@
<param
name="url">https://github.com/andikleen/mcelog.git</param>
<param
name="changesrevision">1f3a769c8fb736815a56ea104b7b751c5565cb88</param></service><service
name="tar_scm">
<param
name="url">https://git.kernel.org/pub/scm/utils/cpu/mce/mcelog.git</param>
- <param
name="changesrevision">6b3fe165f35ef9897a6573042851922b3a3e4e4d</param></service></servicedata>
+ <param
name="changesrevision">ad244c6b60fcbd08a5d73497bfb3487983801598</param></service></servicedata>
(No newline at EOF)
++++++ mcelog-202.obscpio -> mcelog-204.obscpio ++++++
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/mcelog-202/leaky-bucket.c
new/mcelog-204/leaky-bucket.c
--- old/mcelog-202/leaky-bucket.c 2024-12-02 19:00:44.000000000 +0100
+++ new/mcelog-204/leaky-bucket.c 2025-02-15 20:29:39.000000000 +0100
@@ -27,12 +27,12 @@
}
void bucket_age(const struct bucket_conf *c, struct leaky_bucket *b,
- time_t now)
+ time_t now, unsigned char capacity_multiplier)
{
long diff;
diff = now - b->tstamp;
if (diff >= c->agetime) {
- unsigned age = (diff / (double)c->agetime) * c->capacity;
+ unsigned age = (diff / (double)c->agetime) * c->capacity *
capacity_multiplier;
b->tstamp = now;
if (age > b->count)
b->count = 0;
@@ -44,13 +44,13 @@
/* Account increase in leaky bucket. Return 1 if bucket overflowed. */
int __bucket_account(const struct bucket_conf *c, struct leaky_bucket *b,
- unsigned inc, time_t t)
+ unsigned inc, time_t t, unsigned char capacity_multiplier)
{
if (c->capacity == 0)
return 0;
- bucket_age(c, b, t);
+ bucket_age(c, b, t, capacity_multiplier);
b->count += inc;
- if (b->count >= c->capacity) {
+ if (b->count >= c->capacity * capacity_multiplier) {
b->excess += b->count;
/* should disable overflow completely in the same time unit */
b->count = 0;
@@ -62,7 +62,7 @@
int bucket_account(const struct bucket_conf *c, struct leaky_bucket *b,
unsigned inc)
{
- return __bucket_account(c, b, inc, bucket_time());
+ return __bucket_account(c, b, inc, bucket_time(), 1);
}
static int timeconv(char unit, int *out)
@@ -89,7 +89,6 @@
xasprintf(&buf, "not enabled");
} else {
int unit = 0;
- //bucket_age(c, b, bucket_time());
timeconv(c->tunit, &unit);
xasprintf(&buf, "%u in %u%c", b->count + b->excess,
c->agetime/unit, c->tunit);
@@ -198,7 +197,7 @@
for (i = 1; i <= TOTAL_EVENTS; i++) {
event_time = start_time + i * SECONDS_PER_EVENT;
- ret = __bucket_account(&c, &b, 1, event_time);
+ ret = __bucket_account(&c, &b, 1, event_time, 1);
#ifdef TEST_LEAKY_BUCKET_DEBUG
if (ret)
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/mcelog-202/leaky-bucket.h
new/mcelog-204/leaky-bucket.h
--- old/mcelog-202/leaky-bucket.h 2024-12-02 19:00:44.000000000 +0100
+++ new/mcelog-204/leaky-bucket.h 2025-02-15 20:29:39.000000000 +0100
@@ -23,12 +23,12 @@
int bucket_account(const struct bucket_conf *c, struct leaky_bucket *b,
unsigned inc);
int __bucket_account(const struct bucket_conf *c, struct leaky_bucket *b,
- unsigned inc, time_t time);
+ unsigned inc, time_t time, unsigned char
capacity_multiplier);
char *bucket_output(const struct bucket_conf *c, struct leaky_bucket *b);
int bucket_conf_init(struct bucket_conf *c, const char *rate);
void bucket_init(struct leaky_bucket *b);
time_t bucket_time(void);
void bucket_age(const struct bucket_conf *c, struct leaky_bucket *b,
- time_t now);
+ time_t now, unsigned char capacity_multiplier);
#endif
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/mcelog-202/mcelog.conf new/mcelog-204/mcelog.conf
--- old/mcelog-202/mcelog.conf 2024-12-02 19:00:44.000000000 +0100
+++ new/mcelog-204/mcelog.conf 2025-02-15 20:29:39.000000000 +0100
@@ -159,6 +159,9 @@
# Threshold for the correct memory errors trigger script.
memory-ce-threshold = 2 / 24h
+# Retry failed offlines with exponential backoff
+memory-ce-offline-retry = yes
+
# Trigger script for corrected errors.
# memory-ce-trigger = page-error-trigger
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/mcelog-202/memdb.c new/mcelog-204/memdb.c
--- old/mcelog-202/memdb.c 2024-12-02 19:00:44.000000000 +0100
+++ new/mcelog-204/memdb.c 2025-02-15 20:29:39.000000000 +0100
@@ -196,7 +196,7 @@
{
if (corr_err_cnt && --corr_err_cnt > 0) {
md->ce.count += corr_err_cnt;
- if (__bucket_account(&t->ce_bucket_conf, &md->ce.bucket,
corr_err_cnt, m->time)) {
+ if (__bucket_account(&t->ce_bucket_conf, &md->ce.bucket,
corr_err_cnt, m->time, 1)) {
char *msg;
xasprintf(&msg, "Fallback %s memory error count %d
exceeded threshold",
t->type, corr_err_cnt);
@@ -217,11 +217,11 @@
if (m->status & MCI_STATUS_UC) {
md->uc.count++;
- if (__bucket_account(&t->uc_bucket_conf, &md->uc.bucket, 1,
m->time))
+ if (__bucket_account(&t->uc_bucket_conf, &md->uc.bucket, 1,
m->time, 1))
memdb_trigger(msg, md, m->time, &md->uc,
&t->uc_bucket_conf, NULL, false, reporter);
} else {
md->ce.count++;
- if (__bucket_account(&t->ce_bucket_conf, &md->ce.bucket, 1,
m->time))
+ if (__bucket_account(&t->ce_bucket_conf, &md->ce.bucket, 1,
m->time, 1))
memdb_trigger(msg, md, m->time, &md->ce,
&t->ce_bucket_conf, NULL, false, reporter);
}
free(msg);
@@ -278,7 +278,7 @@
int all = (flags & DUMP_ALL);
char *s;
- bucket_age(bc, &e->bucket, bucket_time());
+ bucket_age(bc, &e->bucket, bucket_time(), 1);
if (e->count || e->bucket.count || all)
fprintf(f, "%s:\n", name);
if (e->count || all) {
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/mcelog-202/page.c new/mcelog-204/page.c
--- old/mcelog-202/page.c 2024-12-02 19:00:44.000000000 +0100
+++ new/mcelog-204/page.c 2025-02-15 20:29:39.000000000 +0100
@@ -41,14 +41,16 @@
#define PAGE_SHIFT 12
#define PAGE_SIZE (1UL << PAGE_SHIFT)
+#define OFFLINE_RETRY_EXP_BACKOFF 2
+#define NO_OFFLINE_RETRY 1
enum { PAGE_ONLINE = 0, PAGE_OFFLINE = 1, PAGE_OFFLINE_FAILED = 2 };
struct mempage {
struct rb_node nd;
- /* one char used by rb_node */
char offlined;
char triggered;
+ unsigned char offline_threshold_multiplier;
// 1(32bit)-5(64bit) bytes of padding to play with here
u64 addr;
struct err_type ce;
@@ -74,12 +76,13 @@
static int corr_err_counters;
static struct mempage_cluster *mp_cluster;
-static struct mempage_replacement mp_repalcement;
+static struct mempage_replacement mp_replacement;
static struct rb_root mempage_root;
static LIST_HEAD(mempage_cluster_lru_list);
static struct bucket_conf page_trigger_conf;
static struct bucket_conf mp_replacement_trigger_conf;
static char *page_error_pre_soft_trigger, *page_error_post_soft_trigger;
+static unsigned offline_retry_backoff_base = NO_OFFLINE_RETRY;
static const char *page_state[] = {
[PAGE_ONLINE] = "online",
@@ -111,6 +114,7 @@
mp = &mp_cluster->mp[mp_cluster->mp_used++];
mp->offlined = PAGE_ONLINE;
mp->triggered = 0;
+ mp->offline_threshold_multiplier = NO_OFFLINE_RETRY;
mp->ce.count = 0;
return mp;
@@ -241,6 +245,7 @@
if (memory_offline(addr) < 0) {
Lprintf("Offlining page %llx failed: %s\n", addr,
strerror(errno));
mp->offlined = PAGE_OFFLINE_FAILED;
+ mp->offline_threshold_multiplier *= offline_retry_backoff_base;
} else
mp->offlined = PAGE_OFFLINE;
}
@@ -323,6 +328,7 @@
bucket_init(&mp->ce.bucket);
mempage_insert(addr, mp);
mempage_cluster_lru_list_insert(to_cluster(mp));
+ mp->offline_threshold_multiplier = NO_OFFLINE_RETRY;
corr_err_counters++;
} else if (!mp) {
mp = mempage_replace();
@@ -331,14 +337,14 @@
mempage_cluster_lru_list_update(to_cluster(mp));
/* Report how often the replacement of counter 'mp' happened */
- ++mp_repalcement.count;
- if (__bucket_account(&mp_replacement_trigger_conf,
&mp_repalcement.bucket, 1, t)) {
- thresh = bucket_output(&mp_replacement_trigger_conf,
&mp_repalcement.bucket);
+ ++mp_replacement.count;
+ if (__bucket_account(&mp_replacement_trigger_conf,
&mp_replacement.bucket, 1, t, 1)) {
+ thresh = bucket_output(&mp_replacement_trigger_conf,
&mp_replacement.bucket);
xasprintf(&msg, "Replacements of page correctable error
counter exceed threshold %s", thresh);
free(thresh);
thresh = NULL;
- counter_trigger(msg, t, &mp_repalcement,
&mp_replacement_trigger_conf, false);
+ counter_trigger(msg, t, &mp_replacement,
&mp_replacement_trigger_conf, false);
free(msg);
msg = NULL;
}
@@ -346,10 +352,11 @@
mempage_cluster_lru_list_update(to_cluster(mp));
}
++mp->ce.count;
- if (__bucket_account(&page_trigger_conf, &mp->ce.bucket, 1, t)) {
+ if (__bucket_account(&page_trigger_conf, &mp->ce.bucket, 1, t,
mp->offline_threshold_multiplier)) {
struct memdimm *md;
- if (mp->offlined != PAGE_ONLINE)
+ if ((offline_retry_backoff_base == OFFLINE_RETRY_EXP_BACKOFF &&
mp->offlined == PAGE_OFFLINE) ||
+ (offline_retry_backoff_base == NO_OFFLINE_RETRY &&
mp->offlined != PAGE_ONLINE))
return;
/* Only do triggers and messages for online pages */
thresh = bucket_output(&page_trigger_conf, &mp->ce.bucket);
@@ -432,6 +439,8 @@
config_trigger("page", "memory-ce", &page_trigger_conf);
config_trigger("page", "memory-ce-counter-replacement",
&mp_replacement_trigger_conf);
+ if (config_bool("page", "memory-ce-offline-retry") == 1)
+ offline_retry_backoff_base = OFFLINE_RETRY_EXP_BACKOFF;
n = config_choice("page", "memory-ce-action", offline_choice);
if (n >= 0)
offline = n;
@@ -461,5 +470,5 @@
if (n != max_corr_err_counters)
Lprintf("Round up max-corr-err-counters from %d to %d\n", n,
max_corr_err_counters);
- bucket_init(&mp_repalcement.bucket);
+ bucket_init(&mp_replacement.bucket);
}
++++++ mcelog.obsinfo ++++++
--- /var/tmp/diff_new_pack.1EDHAv/_old 2025-04-15 16:48:11.708855244 +0200
+++ /var/tmp/diff_new_pack.1EDHAv/_new 2025-04-15 16:48:11.712855412 +0200
@@ -1,5 +1,5 @@
name: mcelog
-version: 202
-mtime: 1733162444
-commit: 6b3fe165f35ef9897a6573042851922b3a3e4e4d
+version: 204
+mtime: 1739647779
+commit: ad244c6b60fcbd08a5d73497bfb3487983801598
