netdev - [PATCH net-next] pkt_sched: fq: more robust memory allocation

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <1386894912.19078.130.camel@edumazet-glaptop2.roam.corp.google.com>
Date:	Thu, 12 Dec 2013 16:35:12 -0800
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	David Miller <davem@...emloft.net>
Cc:	netdev <netdev@...r.kernel.org>
Subject: [PATCH net-next] pkt_sched: fq: more robust memory allocation

From: Eric Dumazet <edumazet@...gle.com>

This patch brings NUMA support and automatic fallback to vmalloc()
in case kmalloc() failed to allocate FQ hash table.

NUMA support depends on XPS being setup for the device before
qdisc allocation. After a XPS change, it might be worth creating
qdisc hierarchy again.

Signed-off-by: Eric Dumazet <edumazet@...gle.com>
---
 net/sched/sch_fq.c |   30 ++++++++++++++++++++++++------
 1 file changed, 24 insertions(+), 6 deletions(-)

diff --git a/net/sched/sch_fq.c b/net/sched/sch_fq.c
index 95d843961907..c44ff5016b17 100644
--- a/net/sched/sch_fq.c
+++ b/net/sched/sch_fq.c
@@ -47,6 +47,7 @@
 #include <linux/rbtree.h>
 #include <linux/hash.h>
 #include <linux/prefetch.h>
+#include <linux/vmalloc.h>
 #include <net/netlink.h>
 #include <net/pkt_sched.h>
 #include <net/sock.h>
@@ -578,15 +579,32 @@ static void fq_rehash(struct fq_sched_data *q,
 	q->stat_gc_flows += fcnt;
 }
 
-static int fq_resize(struct fq_sched_data *q, u32 log)
+static void fq_free(void *addr)
 {
+	if (addr && is_vmalloc_addr(addr))
+		vfree(addr);
+	else
+		kfree(addr);
+}
+
+static int fq_resize(struct Qdisc *sch, u32 log)
+{
+	struct fq_sched_data *q = qdisc_priv(sch);
+	size_t sz = sizeof(struct rb_root) << log;
 	struct rb_root *array;
+	int node;
 	u32 idx;
 
 	if (q->fq_root && log == q->fq_trees_log)
 		return 0;
 
-	array = kmalloc(sizeof(struct rb_root) << log, GFP_KERNEL);
+	/* If XPS was setup, we can allocate memory on right NUMA node */
+	node = netdev_queue_numa_node_read(sch->dev_queue);
+
+	array = kmalloc_node(sz, GFP_KERNEL | __GFP_REPEAT | __GFP_NOWARN,
+			     node);
+	if (!array)
+		array = vmalloc_node(sz, node);
 	if (!array)
 		return -ENOMEM;
 
@@ -595,7 +613,7 @@ static int fq_resize(struct fq_sched_data *q, u32 log)
 
 	if (q->fq_root) {
 		fq_rehash(q, q->fq_root, q->fq_trees_log, array, log);
-		kfree(q->fq_root);
+		fq_free(q->fq_root);
 	}
 	q->fq_root = array;
 	q->fq_trees_log = log;
@@ -676,7 +694,7 @@ static int fq_change(struct Qdisc *sch, struct nlattr *opt)
 	}
 
 	if (!err)
-		err = fq_resize(q, fq_log);
+		err = fq_resize(sch, fq_log);
 
 	while (sch->q.qlen > sch->limit) {
 		struct sk_buff *skb = fq_dequeue(sch);
@@ -697,7 +715,7 @@ static void fq_destroy(struct Qdisc *sch)
 	struct fq_sched_data *q = qdisc_priv(sch);
 
 	fq_reset(sch);
-	kfree(q->fq_root);
+	fq_free(q->fq_root);
 	qdisc_watchdog_cancel(&q->watchdog);
 }
 
@@ -723,7 +741,7 @@ static int fq_init(struct Qdisc *sch, struct nlattr *opt)
 	if (opt)
 		err = fq_change(sch, opt);
 	else
-		err = fq_resize(q, q->fq_trees_log);
+		err = fq_resize(sch, q->fq_trees_log);
 
 	return err;
 }


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html