Skip to contentSkip to conversation

prevent the read ahead of /proc/slabinfo in ss

PATCH In review Bryton Lee 3 participants 4 messages Feb 9, 2015 – Feb 10, 2015

Subsystems (2)
lore
Share & tools
lore

Bryton Lee

author
#0 lore raw

PATCH 1/1 prevent the read ahead of /proc/slabinfo in ss

ss reads ahead of /proc/slabinfo whatever slabstat will be used or not in future.
this will cause huge system delay when the kernel hires SLAB allocator(SLUB allocator is ok). when program reads /proc/slabinfo, it will call s_show() in SLAB allocator level, and s_show() calls spin_lock_irq(&l3->list_lock) then iterate on whole three lists. if one slab has about 900 million objects (for example dentry), it will cause more than 1000ms delay.

so this patch prevents the read ahead of /proc/slabinfo, ss runs with most parameters not using slabstat at all.

---
 misc/ss.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
misc/ss.c +7 -2
index f434f57..987dd70 100644
@@ -617,6 +617,7 @@ struct slabstat
 };
 
 struct slabstat slabstat;
+int slabstat_valid = 0;
 
 static const char *slabstat_ids[] =
 {
@@ -2128,6 +2129,9 @@ static int tcp_show(struct filter *f, int socktype)
 	 * it is able to give us some memory for snapshot.
 	 */
 	if (1) {
+		if (!slabstat_valid)
+			get_slabstat(&slabstat);
+
 		int guess = slabstat.socks+slabstat.tcp_syns;
 		if (f->states&(1<<SS_TIME_WAIT))
 			guess += slabstat.tcp_tws;
@@ -3157,6 +3161,9 @@ static int print_summary(void)
 	if (get_snmp_int("Tcp:", "CurrEstab", &sn.tcp_estab) < 0)
 		perror("ss: get_snmpstat");
 
+	if (!slabstat_valid)
+		get_slabstat(&slabstat);
+
 	printf("Total: %d (kernel %d)\n", s.socks, slabstat.socks);
 
 	printf("TCP:   %d (estab %d, closed %d, orphaned %d, synrecv %d, timewait %d/%d), ports %d\n",
@@ -3504,8 +3511,6 @@ int main(int argc, char *argv[])
 	argc -= optind;
 	argv += optind;
 
-	get_slabstat(&slabstat);
-
 	if (do_summary) {
 		print_summary();
 		if (do_default && argc == 0)
Signature · 1 line
2.0.5
Acked-by on v1 1/1
#1 lore raw
Quoted 7 lines · Bryton Lee

ss reads ahead of /proc/slabinfo whatever slabstat will be used or not in future.
this will cause huge system delay when the kernel hires SLAB allocator(SLUB allocator is ok). when program reads /proc/slabinfo, it will call s_show() in SLAB allocator level, and s_show() calls spin_lock_irq(&l3->list_lock) then iterate on whole three lists. if one slab has about 900 million objects (for example dentry), it will cause more than 1000ms delay.

so this patch prevents the read ahead of /proc/slabinfo, ss runs with most parameters not using slabstat at all.

Indeed, thanks !

maintainer
#2 lore raw
In reply to Bryton Lee

From: Bryton Lee <[email protected]>
Date: Mon, 9 Feb 2015 19:37:10 +0800

misc/ss.c struct slabstat · line 623

Quoted 8 lines
misc/ss.c +1 -0
@@ -617,6 +617,7 @@ struct slabstat
 };
 
 struct slabstat slabstat;
+int slabstat_valid = 0;
 
 static const char *slabstat_ids[] =
 {

Nothing sets this to a non-zero value. If nothing is going to set it to a non-zero value, the code it guards should simple be removed instead.

Otherwise, nothing uses this variable outside of this file, and if that's intentional it should be static.

author
#3 lore raw
Quoted 6 lines · David Miller

From: Bryton Lee <[email protected]>
Date: Mon, 9 Feb 2015 19:37:10 +0800

Quoted 3 lines
(quoted diff) +0 -0
@@ -617,6 +617,7 @@ struct slabstat
 };
 
Quoted 9 lines

misc/ss.c struct slabstat · line 623

Quoted 5 lines
misc/ss.c +1 -0
 struct slabstat slabstat;
+int slabstat_valid = 0;
 
 static const char *slabstat_ids[] =
 {

Nothing sets this to a non-zero value. If nothing is going to set it to a non-zero value, the code it guards should simple be removed instead.

when ss runs witch -s paramater slabstat will be used, and if kernel not support NETLINK ss -t will use slabstat too. so it's not nothing sets this to a non-zero value.

Quoted 2 lines

Otherwise, nothing uses this variable outside of this file, and if that's intentional it should be static.

Yes, these should be changed to static, I will submit these change soon. thanks!

Signature · 3 lines
Best Regards

Bryton.Lee