* [v3 RFC PATCH 5/8] ipv4: Remove flowi from struct rtable.
From: David Miller @ 2011-03-03 1:02 UTC (permalink / raw)
To: netdev
The only necessary parts are the src/dst addresses, the
interface indexes, the TOS, and the mark.
The rest is unnecessary bloat, which amounts to nearly
50 bytes on 64-bit.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/route.h | 24 +++++++-----
net/ipv4/icmp.c | 2 +-
net/ipv4/ipmr.c | 52 ++++++++++++++++++++-----
net/ipv4/route.c | 95 ++++++++++++++++++++++++++--------------------
net/ipv4/xfrm4_policy.c | 7 +++-
5 files changed, 116 insertions(+), 64 deletions(-)
diff --git a/include/net/route.h b/include/net/route.h
index ef85013..ebf50a9 100644
--- a/include/net/route.h
+++ b/include/net/route.h
@@ -53,16 +53,20 @@ struct fib_info;
struct rtable {
struct dst_entry dst;
- /* Cache lookup keys */
- struct flowi fl;
+ /* Lookup key. */
+ __be32 rt_key_dst;
+ __be32 rt_key_src;
int rt_genid;
unsigned rt_flags;
__u16 rt_type;
+ __u8 rt_tos;
__be32 rt_dst; /* Path destination */
__be32 rt_src; /* Path source */
int rt_iif;
+ int rt_oif;
+ __u32 rt_mark;
/* Info on neighbour */
__be32 rt_gateway;
@@ -76,12 +80,12 @@ struct rtable {
static inline bool rt_is_input_route(struct rtable *rt)
{
- return rt->fl.iif != 0;
+ return rt->rt_iif != 0;
}
static inline bool rt_is_output_route(struct rtable *rt)
{
- return rt->fl.iif == 0;
+ return rt->rt_iif == 0;
}
struct ip_rt_acct {
@@ -199,12 +203,12 @@ static inline struct rtable *ip_route_newports(struct rtable *rt,
__be16 dport, struct sock *sk)
{
if (sport != orig_sport || dport != orig_dport) {
- struct flowi fl = { .oif = rt->fl.oif,
- .mark = rt->fl.mark,
- .fl4_dst = rt->fl.fl4_dst,
- .fl4_src = rt->fl.fl4_src,
- .fl4_tos = rt->fl.fl4_tos,
- .proto = rt->fl.proto,
+ struct flowi fl = { .oif = rt->rt_oif,
+ .mark = rt->rt_mark,
+ .fl4_dst = rt->rt_key_dst,
+ .fl4_src = rt->rt_key_src,
+ .fl4_tos = rt->rt_tos,
+ .proto = protocol,
.fl_ip_sport = sport,
.fl_ip_dport = dport };
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index 994a785..1771ce6 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -563,7 +563,7 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
rcu_read_lock();
if (rt_is_input_route(rt) &&
net->ipv4.sysctl_icmp_errors_use_inbound_ifaddr)
- dev = dev_get_by_index_rcu(net, rt->fl.iif);
+ dev = dev_get_by_index_rcu(net, rt->rt_iif);
if (dev)
saddr = inet_select_addr(dev, 0, RT_SCOPE_LINK);
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index 26ca2f2..9d5f634 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -1813,12 +1813,22 @@ int ip_mr_input(struct sk_buff *skb)
if (IPCB(skb)->flags & IPSKB_FORWARDED)
goto dont_forward;
- err = ipmr_fib_lookup(net, &skb_rtable(skb)->fl, &mrt);
- if (err < 0) {
- kfree_skb(skb);
- return err;
+ {
+ struct rtable *rt = skb_rtable(skb);
+ struct flowi fl = {
+ .fl4_dst = rt->rt_key_dst,
+ .fl4_src = rt->rt_key_src,
+ .fl4_tos = rt->rt_tos,
+ .oif = rt->rt_oif,
+ .iif = rt->rt_iif,
+ .mark = rt->rt_mark,
+ };
+ err = ipmr_fib_lookup(net, &fl, &mrt);
+ if (err < 0) {
+ kfree_skb(skb);
+ return err;
+ }
}
-
if (!local) {
if (IPCB(skb)->opt.router_alert) {
if (ip_call_ra_chain(skb))
@@ -1946,9 +1956,19 @@ int pim_rcv_v1(struct sk_buff *skb)
pim = igmp_hdr(skb);
- if (ipmr_fib_lookup(net, &skb_rtable(skb)->fl, &mrt) < 0)
- goto drop;
-
+ {
+ struct rtable *rt = skb_rtable(skb);
+ struct flowi fl = {
+ .fl4_dst = rt->rt_key_dst,
+ .fl4_src = rt->rt_key_src,
+ .fl4_tos = rt->rt_tos,
+ .oif = rt->rt_oif,
+ .iif = rt->rt_iif,
+ .mark = rt->rt_mark,
+ };
+ if (ipmr_fib_lookup(net, &fl, &mrt) < 0)
+ goto drop;
+ }
if (!mrt->mroute_do_pim ||
pim->group != PIM_V1_VERSION || pim->code != PIM_V1_REGISTER)
goto drop;
@@ -1978,9 +1998,19 @@ static int pim_rcv(struct sk_buff *skb)
csum_fold(skb_checksum(skb, 0, skb->len, 0))))
goto drop;
- if (ipmr_fib_lookup(net, &skb_rtable(skb)->fl, &mrt) < 0)
- goto drop;
-
+ {
+ struct rtable *rt = skb_rtable(skb);
+ struct flowi fl = {
+ .fl4_dst = rt->rt_key_dst,
+ .fl4_src = rt->rt_key_src,
+ .fl4_tos = rt->rt_tos,
+ .oif = rt->rt_oif,
+ .iif = rt->rt_iif,
+ .mark = rt->rt_mark,
+ };
+ if (ipmr_fib_lookup(net, &fl, &mrt) < 0)
+ goto drop;
+ }
if (__pim_rcv(mrt, skb, sizeof(*pim))) {
drop:
kfree_skb(skb);
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index e59241f..9e302dc 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -629,7 +629,7 @@ static struct dst_entry *ipv4_negative_advice(struct dst_entry *dst)
} else if (rt->rt_flags & RTCF_REDIRECTED) {
#if RT_CACHE_DEBUG >= 1
printk(KERN_DEBUG "ipv4_negative_advice: redirect to %pI4/%02x dropped\n",
- &rt->rt_dst, rt->fl.fl4_tos);
+ &rt->rt_dst, rt->rt_tos);
#endif
ip_rt_put(rt);
ret = NULL;
@@ -977,8 +977,17 @@ void ip_rt_get_source(u8 *addr, struct rtable *rt)
if (rt_is_output_route(rt))
src = rt->rt_src;
else {
+ struct flowi fl = {
+ .fl4_dst = rt->rt_key_dst,
+ .fl4_src = rt->rt_key_src,
+ .fl4_tos = rt->rt_tos,
+ .oif = rt->rt_oif,
+ .iif = rt->rt_iif,
+ .mark = rt->rt_mark,
+ };
+
rcu_read_lock();
- if (fib_lookup(dev_net(rt->dst.dev), &rt->fl, &res) == 0)
+ if (fib_lookup(dev_net(rt->dst.dev), &fl, &res) == 0)
src = FIB_RES_PREFSRC(res);
else
src = inet_select_addr(rt->dst.dev, rt->rt_gateway,
@@ -1028,7 +1037,8 @@ static unsigned int ipv4_default_mtu(const struct dst_entry *dst)
return mtu;
}
-static void rt_init_metrics(struct rtable *rt, struct fib_info *fi)
+static void rt_init_metrics(struct rtable *rt, const struct flowi *oldflp,
+ struct fib_info *fi)
{
struct inet_peer *peer;
int create = 0;
@@ -1036,7 +1046,7 @@ static void rt_init_metrics(struct rtable *rt, struct fib_info *fi)
/* If a peer entry exists for this destination, we must hook
* it up in order to get at cached metrics.
*/
- if (rt->fl.flags & FLOWI_FLAG_PRECOW_METRICS)
+ if (oldflp && (oldflp->flags & FLOWI_FLAG_PRECOW_METRICS))
create = 1;
rt_bind_peer(rt, create);
@@ -1063,7 +1073,8 @@ static void rt_init_metrics(struct rtable *rt, struct fib_info *fi)
}
}
-static void rt_set_nexthop(struct rtable *rt, const struct fib_result *res,
+static void rt_set_nexthop(struct rtable *rt, const struct flowi *oldflp,
+ const struct fib_result *res,
struct fib_info *fi, u16 type, u32 itag)
{
struct dst_entry *dst = &rt->dst;
@@ -1072,7 +1083,7 @@ static void rt_set_nexthop(struct rtable *rt, const struct fib_result *res,
if (FIB_RES_GW(*res) &&
FIB_RES_NH(*res).nh_scope == RT_SCOPE_LINK)
rt->rt_gateway = FIB_RES_GW(*res);
- rt_init_metrics(rt, fi);
+ rt_init_metrics(rt, oldflp, fi);
#ifdef CONFIG_IP_ROUTE_CLASSID
dst->tclassid = FIB_RES_NH(*res).nh_tclassid;
#endif
@@ -1144,20 +1155,19 @@ static int ip_route_input_mc(struct sk_buff *skb, __be32 daddr, __be32 saddr,
rth->dst.output = ip_rt_bug;
- rth->fl.fl4_dst = daddr;
+ rth->rt_key_dst = daddr;
rth->rt_dst = daddr;
- rth->fl.fl4_tos = tos;
- rth->fl.mark = skb->mark;
- rth->fl.fl4_src = saddr;
+ rth->rt_tos = tos;
+ rth->rt_mark = skb->mark;
+ rth->rt_key_src = saddr;
rth->rt_src = saddr;
#ifdef CONFIG_IP_ROUTE_CLASSID
rth->dst.tclassid = itag;
#endif
- rth->rt_iif =
- rth->fl.iif = dev->ifindex;
+ rth->rt_iif = dev->ifindex;
rth->dst.dev = init_net.loopback_dev;
dev_hold(rth->dst.dev);
- rth->fl.oif = 0;
+ rth->rt_oif = 0;
rth->rt_gateway = daddr;
rth->rt_spec_dst= spec_dst;
rth->rt_genid = rt_genid(dev_net(dev));
@@ -1281,25 +1291,24 @@ static int __mkroute_input(struct sk_buff *skb,
goto cleanup;
}
- rth->fl.fl4_dst = daddr;
+ rth->rt_key_dst = daddr;
rth->rt_dst = daddr;
- rth->fl.fl4_tos = tos;
- rth->fl.mark = skb->mark;
- rth->fl.fl4_src = saddr;
+ rth->rt_tos = tos;
+ rth->rt_mark = skb->mark;
+ rth->rt_key_src = saddr;
rth->rt_src = saddr;
rth->rt_gateway = daddr;
- rth->rt_iif =
- rth->fl.iif = in_dev->dev->ifindex;
+ rth->rt_iif = in_dev->dev->ifindex;
rth->dst.dev = (out_dev)->dev;
dev_hold(rth->dst.dev);
- rth->fl.oif = 0;
+ rth->rt_oif = 0;
rth->rt_spec_dst= spec_dst;
rth->dst.input = ip_forward;
rth->dst.output = ip_output;
rth->rt_genid = rt_genid(dev_net(rth->dst.dev));
- rt_set_nexthop(rth, res, res->fi, res->type, itag);
+ rt_set_nexthop(rth, NULL, res, res->fi, res->type, itag);
rth->rt_flags = flags;
@@ -1449,17 +1458,16 @@ local_input:
rth->dst.output= ip_rt_bug;
rth->rt_genid = rt_genid(net);
- rth->fl.fl4_dst = daddr;
+ rth->rt_key_dst = daddr;
rth->rt_dst = daddr;
- rth->fl.fl4_tos = tos;
- rth->fl.mark = skb->mark;
- rth->fl.fl4_src = saddr;
+ rth->rt_tos = tos;
+ rth->rt_mark = skb->mark;
+ rth->rt_key_src = saddr;
rth->rt_src = saddr;
#ifdef CONFIG_IP_ROUTE_CLASSID
rth->dst.tclassid = itag;
#endif
- rth->rt_iif =
- rth->fl.iif = dev->ifindex;
+ rth->rt_iif = dev->ifindex;
rth->dst.dev = net->loopback_dev;
dev_hold(rth->dst.dev);
rth->rt_gateway = daddr;
@@ -1613,11 +1621,11 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
if (!rth)
return ERR_PTR(-ENOBUFS);
- rth->fl.fl4_dst = oldflp->fl4_dst;
- rth->fl.fl4_tos = tos;
- rth->fl.fl4_src = oldflp->fl4_src;
- rth->fl.oif = oldflp->oif;
- rth->fl.mark = oldflp->mark;
+ rth->rt_key_dst = oldflp->fl4_dst;
+ rth->rt_tos = tos;
+ rth->rt_key_src = oldflp->fl4_src;
+ rth->rt_oif = oldflp->oif;
+ rth->rt_mark = oldflp->mark;
rth->rt_dst = fl->fl4_dst;
rth->rt_src = fl->fl4_src;
rth->rt_iif = oldflp->oif ? : dev_out->ifindex;
@@ -1655,7 +1663,7 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
#endif
}
- rt_set_nexthop(rth, res, fi, type, 0);
+ rt_set_nexthop(rth, oldflp, res, fi, type, 0);
rth->rt_flags = flags;
return rth;
@@ -1887,7 +1895,12 @@ struct dst_entry *ipv4_blackhole_route(struct net *net, struct dst_entry *dst_or
if (new->dev)
dev_hold(new->dev);
- rt->fl = ort->fl;
+ rt->rt_key_dst = ort->rt_key_dst;
+ rt->rt_key_src = ort->rt_key_src;
+ rt->rt_tos = ort->rt_tos;
+ rt->rt_iif = ort->rt_iif;
+ rt->rt_oif = ort->rt_oif;
+ rt->rt_mark = ort->rt_mark;
rt->rt_genid = rt_genid(net);
rt->rt_flags = ort->rt_flags;
@@ -1950,7 +1963,7 @@ static int rt_fill_info(struct net *net,
r->rtm_family = AF_INET;
r->rtm_dst_len = 32;
r->rtm_src_len = 0;
- r->rtm_tos = rt->fl.fl4_tos;
+ r->rtm_tos = rt->rt_tos;
r->rtm_table = RT_TABLE_MAIN;
NLA_PUT_U32(skb, RTA_TABLE, RT_TABLE_MAIN);
r->rtm_type = rt->rt_type;
@@ -1962,9 +1975,9 @@ static int rt_fill_info(struct net *net,
NLA_PUT_BE32(skb, RTA_DST, rt->rt_dst);
- if (rt->fl.fl4_src) {
+ if (rt->rt_key_src) {
r->rtm_src_len = 32;
- NLA_PUT_BE32(skb, RTA_SRC, rt->fl.fl4_src);
+ NLA_PUT_BE32(skb, RTA_SRC, rt->rt_key_src);
}
if (rt->dst.dev)
NLA_PUT_U32(skb, RTA_OIF, rt->dst.dev->ifindex);
@@ -1974,7 +1987,7 @@ static int rt_fill_info(struct net *net,
#endif
if (rt_is_input_route(rt))
NLA_PUT_BE32(skb, RTA_PREFSRC, rt->rt_spec_dst);
- else if (rt->rt_src != rt->fl.fl4_src)
+ else if (rt->rt_src != rt->rt_key_src)
NLA_PUT_BE32(skb, RTA_PREFSRC, rt->rt_src);
if (rt->rt_dst != rt->rt_gateway)
@@ -1983,8 +1996,8 @@ static int rt_fill_info(struct net *net,
if (rtnetlink_put_metrics(skb, dst_metrics_ptr(&rt->dst)) < 0)
goto nla_put_failure;
- if (rt->fl.mark)
- NLA_PUT_BE32(skb, RTA_MARK, rt->fl.mark);
+ if (rt->rt_mark)
+ NLA_PUT_BE32(skb, RTA_MARK, rt->rt_mark);
error = rt->dst.error;
expires = (rt->peer && rt->peer->pmtu_expires) ?
@@ -2018,7 +2031,7 @@ static int rt_fill_info(struct net *net,
}
} else
#endif
- NLA_PUT_U32(skb, RTA_IIF, rt->fl.iif);
+ NLA_PUT_U32(skb, RTA_IIF, rt->rt_iif);
}
if (rtnl_put_cacheinfo(skb, &rt->dst, id, ts, tsage,
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index 45b8214..c70c42e 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -70,7 +70,12 @@ static int xfrm4_fill_dst(struct xfrm_dst *xdst, struct net_device *dev,
{
struct rtable *rt = (struct rtable *)xdst->route;
- xdst->u.rt.fl = *fl;
+ rt->rt_key_dst = fl->fl4_dst;
+ rt->rt_key_src = fl->fl4_src;
+ rt->rt_tos = fl->fl4_tos;
+ rt->rt_iif = fl->iif;
+ rt->rt_oif = fl->oif;
+ rt->rt_mark = fl->mark;
xdst->u.dst.dev = dev;
dev_hold(dev);
--
1.7.4.1
^ permalink raw reply related
* [v3 RFC PATCH 6/8] net: Make dst_alloc() take more explicit initializations.
From: David Miller @ 2011-03-03 1:02 UTC (permalink / raw)
To: netdev
Now the dst->dev, dev->obsolete, and dst->flags values can
be specified as well.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/dst.h | 3 ++-
net/core/dst.c | 18 +++++++++++++-----
net/decnet/dn_route.c | 13 ++-----------
net/ipv4/route.c | 48 +++++++++++++++++++-----------------------------
net/ipv6/route.c | 29 +++++++++++------------------
net/xfrm/xfrm_policy.c | 2 +-
6 files changed, 48 insertions(+), 65 deletions(-)
diff --git a/include/net/dst.h b/include/net/dst.h
index 2a46cba..d9b4e78 100644
--- a/include/net/dst.h
+++ b/include/net/dst.h
@@ -352,7 +352,8 @@ static inline struct dst_entry *skb_dst_pop(struct sk_buff *skb)
}
extern int dst_discard(struct sk_buff *skb);
-extern void *dst_alloc(struct dst_ops * ops, int initial_ref);
+extern void *dst_alloc(struct dst_ops * ops, struct net_device *dev,
+ int initial_ref, int initial_obsolete, int flags);
extern void __dst_free(struct dst_entry * dst);
extern struct dst_entry *dst_destroy(struct dst_entry * dst);
diff --git a/net/core/dst.c b/net/core/dst.c
index 91104d3..9505778 100644
--- a/net/core/dst.c
+++ b/net/core/dst.c
@@ -166,7 +166,8 @@ EXPORT_SYMBOL(dst_discard);
const u32 dst_default_metrics[RTAX_MAX];
-void *dst_alloc(struct dst_ops *ops, int initial_ref)
+void *dst_alloc(struct dst_ops *ops, struct net_device *dev,
+ int initial_ref, int initial_obsolete, int flags)
{
struct dst_entry *dst;
@@ -177,12 +178,19 @@ void *dst_alloc(struct dst_ops *ops, int initial_ref)
dst = kmem_cache_zalloc(ops->kmem_cachep, GFP_ATOMIC);
if (!dst)
return NULL;
- atomic_set(&dst->__refcnt, initial_ref);
dst->ops = ops;
- dst->lastuse = jiffies;
- dst->path = dst;
- dst->input = dst->output = dst_discard;
+ dst->dev = dev;
+ if (dev)
+ dev_hold(dev);
dst_init_metrics(dst, dst_default_metrics, true);
+ dst->path = dst;
+ dst->input = dst_discard;
+ dst->output = dst_discard;
+
+ dst->obsolete = initial_obsolete;
+ atomic_set(&dst->__refcnt, initial_ref);
+ dst->lastuse = jiffies;
+ dst->flags = flags;
#if RT_CACHE_DEBUG >= 2
atomic_inc(&dst_total);
#endif
diff --git a/net/decnet/dn_route.c b/net/decnet/dn_route.c
index 484fdbf..a3d2c3f 100644
--- a/net/decnet/dn_route.c
+++ b/net/decnet/dn_route.c
@@ -1122,13 +1122,10 @@ make_route:
if (dev_out->flags & IFF_LOOPBACK)
flags |= RTCF_LOCAL;
- rt = dst_alloc(&dn_dst_ops, 0);
+ rt = dst_alloc(&dn_dst_ops, dev_out, 1, 0, DST_HOST);
if (rt == NULL)
goto e_nobufs;
- atomic_set(&rt->dst.__refcnt, 1);
- rt->dst.flags = DST_HOST;
-
rt->fl.fld_src = oldflp->fld_src;
rt->fl.fld_dst = oldflp->fld_dst;
rt->fl.oif = oldflp->oif;
@@ -1143,8 +1140,6 @@ make_route:
rt->rt_dst_map = fl.fld_dst;
rt->rt_src_map = fl.fld_src;
- rt->dst.dev = dev_out;
- dev_hold(dev_out);
rt->dst.neighbour = neigh;
neigh = NULL;
@@ -1392,7 +1387,7 @@ static int dn_route_input_slow(struct sk_buff *skb)
}
make_route:
- rt = dst_alloc(&dn_dst_ops, 0);
+ rt = dst_alloc(&dn_dst_ops, out_dev, 0, 0, DST_HOST);
if (rt == NULL)
goto e_nobufs;
@@ -1412,9 +1407,7 @@ make_route:
rt->fl.iif = in_dev->ifindex;
rt->fl.mark = fl.mark;
- rt->dst.flags = DST_HOST;
rt->dst.neighbour = neigh;
- rt->dst.dev = out_dev;
rt->dst.lastuse = jiffies;
rt->dst.output = dn_rt_bug;
switch(res.type) {
@@ -1433,8 +1426,6 @@ make_route:
rt->dst.input = dst_discard;
}
rt->rt_flags = flags;
- if (rt->dst.dev)
- dev_hold(rt->dst.dev);
err = dn_rt_set_next_hop(rt, &res);
if (err)
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 9e302dc..9fcf8ce 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -1103,21 +1103,17 @@ static void rt_set_nexthop(struct rtable *rt, const struct flowi *oldflp,
rt->rt_type = type;
}
-static struct rtable *rt_dst_alloc(bool nopolicy, bool noxfrm)
+static struct rtable *rt_dst_alloc(struct net_device *dev,
+ bool nopolicy, bool noxfrm)
{
- struct rtable *rt = dst_alloc(&ipv4_dst_ops, 1);
- if (rt) {
- rt->dst.obsolete = -1;
-
- /* To avoid expensive rcu stuff for this uncached dst, we set
- * DST_NOCACHE so that dst_release() can free dst without
- * waiting a grace period.
- */
- rt->dst.flags = DST_NOCACHE | DST_HOST |
- (nopolicy ? DST_NOPOLICY : 0) |
- (noxfrm ? DST_NOXFRM : 0);
- }
- return rt;
+ /* To avoid expensive rcu stuff for this uncached dst, we set
+ * DST_NOCACHE so that dst_release() can free dst without
+ * waiting a grace period.
+ */
+ return dst_alloc(&ipv4_dst_ops, dev, 1, -1,
+ DST_NOCACHE | DST_HOST |
+ (nopolicy ? DST_NOPOLICY : 0) |
+ (noxfrm ? DST_NOXFRM : 0));
}
/* called in rcu_read_lock() section */
@@ -1149,7 +1145,8 @@ static int ip_route_input_mc(struct sk_buff *skb, __be32 daddr, __be32 saddr,
if (err < 0)
goto e_err;
}
- rth = rt_dst_alloc(IN_DEV_CONF_GET(in_dev, NOPOLICY), false);
+ rth = rt_dst_alloc(init_net.loopback_dev,
+ IN_DEV_CONF_GET(in_dev, NOPOLICY), false);
if (!rth)
goto e_nobufs;
@@ -1165,8 +1162,6 @@ static int ip_route_input_mc(struct sk_buff *skb, __be32 daddr, __be32 saddr,
rth->dst.tclassid = itag;
#endif
rth->rt_iif = dev->ifindex;
- rth->dst.dev = init_net.loopback_dev;
- dev_hold(rth->dst.dev);
rth->rt_oif = 0;
rth->rt_gateway = daddr;
rth->rt_spec_dst= spec_dst;
@@ -1284,7 +1279,8 @@ static int __mkroute_input(struct sk_buff *skb,
}
}
- rth = rt_dst_alloc(IN_DEV_CONF_GET(in_dev, NOPOLICY),
+ rth = rt_dst_alloc(out_dev->dev,
+ IN_DEV_CONF_GET(in_dev, NOPOLICY),
IN_DEV_CONF_GET(out_dev, NOXFRM));
if (!rth) {
err = -ENOBUFS;
@@ -1299,8 +1295,6 @@ static int __mkroute_input(struct sk_buff *skb,
rth->rt_src = saddr;
rth->rt_gateway = daddr;
rth->rt_iif = in_dev->dev->ifindex;
- rth->dst.dev = (out_dev)->dev;
- dev_hold(rth->dst.dev);
rth->rt_oif = 0;
rth->rt_spec_dst= spec_dst;
@@ -1451,7 +1445,8 @@ brd_input:
RT_CACHE_STAT_INC(in_brd);
local_input:
- rth = rt_dst_alloc(IN_DEV_CONF_GET(in_dev, NOPOLICY), false);
+ rth = rt_dst_alloc(net->loopback_dev,
+ IN_DEV_CONF_GET(in_dev, NOPOLICY), false);
if (!rth)
goto e_nobufs;
@@ -1468,8 +1463,6 @@ local_input:
rth->dst.tclassid = itag;
#endif
rth->rt_iif = dev->ifindex;
- rth->dst.dev = net->loopback_dev;
- dev_hold(rth->dst.dev);
rth->rt_gateway = daddr;
rth->rt_spec_dst= spec_dst;
rth->dst.input= ip_local_deliver;
@@ -1616,7 +1609,8 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
fi = NULL;
}
- rth = rt_dst_alloc(IN_DEV_CONF_GET(in_dev, NOPOLICY),
+ rth = rt_dst_alloc(dev_out,
+ IN_DEV_CONF_GET(in_dev, NOPOLICY),
IN_DEV_CONF_GET(in_dev, NOXFRM));
if (!rth)
return ERR_PTR(-ENOBUFS);
@@ -1629,10 +1623,6 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
rth->rt_dst = fl->fl4_dst;
rth->rt_src = fl->fl4_src;
rth->rt_iif = oldflp->oif ? : dev_out->ifindex;
- /* get references to the devices that are to be hold by the routing
- cache entry */
- rth->dst.dev = dev_out;
- dev_hold(dev_out);
rth->rt_gateway = fl->fl4_dst;
rth->rt_spec_dst= fl->fl4_src;
@@ -1880,7 +1870,7 @@ static struct dst_ops ipv4_dst_blackhole_ops = {
struct dst_entry *ipv4_blackhole_route(struct net *net, struct dst_entry *dst_orig)
{
- struct rtable *rt = dst_alloc(&ipv4_dst_blackhole_ops, 1);
+ struct rtable *rt = dst_alloc(&ipv4_dst_blackhole_ops, NULL, 1, 0, 0);
struct rtable *ort = (struct rtable *) dst_orig;
if (rt) {
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 053a92e..9a685f5 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -220,9 +220,10 @@ static struct rt6_info ip6_blk_hole_entry_template = {
#endif
/* allocate dst with ip6_dst_ops */
-static inline struct rt6_info *ip6_dst_alloc(struct dst_ops *ops)
+static inline struct rt6_info *ip6_dst_alloc(struct dst_ops *ops,
+ struct net_device *dev)
{
- return (struct rt6_info *)dst_alloc(ops, 0);
+ return (struct rt6_info *)dst_alloc(ops, dev, 0, 0, 0);
}
static void ip6_dst_destroy(struct dst_entry *dst)
@@ -872,10 +873,10 @@ EXPORT_SYMBOL(ip6_route_output);
struct dst_entry *ip6_blackhole_route(struct net *net, struct dst_entry *dst_orig)
{
- struct rt6_info *rt = dst_alloc(&ip6_dst_blackhole_ops, 1);
- struct rt6_info *ort = (struct rt6_info *) dst_orig;
+ struct rt6_info *rt, *ort = (struct rt6_info *) dst_orig;
struct dst_entry *new = NULL;
+ rt = dst_alloc(&ip6_dst_blackhole_ops, ort->dst.dev, 1, 0, 0);
if (rt) {
new = &rt->dst;
@@ -884,9 +885,6 @@ struct dst_entry *ip6_blackhole_route(struct net *net, struct dst_entry *dst_ori
new->output = dst_discard;
dst_copy_metrics(new, &ort->dst);
- new->dev = ort->dst.dev;
- if (new->dev)
- dev_hold(new->dev);
rt->rt6i_idev = ort->rt6i_idev;
if (rt->rt6i_idev)
in6_dev_hold(rt->rt6i_idev);
@@ -1029,13 +1027,12 @@ struct dst_entry *icmp6_dst_alloc(struct net_device *dev,
if (unlikely(idev == NULL))
return NULL;
- rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops);
+ rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops, dev);
if (unlikely(rt == NULL)) {
in6_dev_put(idev);
goto out;
}
- dev_hold(dev);
if (neigh)
neigh_hold(neigh);
else {
@@ -1044,7 +1041,6 @@ struct dst_entry *icmp6_dst_alloc(struct net_device *dev,
neigh = NULL;
}
- rt->rt6i_dev = dev;
rt->rt6i_idev = idev;
rt->rt6i_nexthop = neigh;
atomic_set(&rt->dst.__refcnt, 1);
@@ -1203,7 +1199,7 @@ int ip6_route_add(struct fib6_config *cfg)
goto out;
}
- rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops);
+ rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops, NULL);
if (rt == NULL) {
err = -ENOMEM;
@@ -1712,7 +1708,8 @@ void rt6_pmtu_discovery(struct in6_addr *daddr, struct in6_addr *saddr,
static struct rt6_info * ip6_rt_copy(struct rt6_info *ort)
{
struct net *net = dev_net(ort->rt6i_dev);
- struct rt6_info *rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops);
+ struct rt6_info *rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops,
+ ort->dst.dev);
if (rt) {
rt->dst.input = ort->dst.input;
@@ -1720,9 +1717,6 @@ static struct rt6_info * ip6_rt_copy(struct rt6_info *ort)
dst_copy_metrics(&rt->dst, &ort->dst);
rt->dst.error = ort->dst.error;
- rt->dst.dev = ort->dst.dev;
- if (rt->dst.dev)
- dev_hold(rt->dst.dev);
rt->rt6i_idev = ort->rt6i_idev;
if (rt->rt6i_idev)
in6_dev_hold(rt->rt6i_idev);
@@ -1992,7 +1986,8 @@ struct rt6_info *addrconf_dst_alloc(struct inet6_dev *idev,
int anycast)
{
struct net *net = dev_net(idev->dev);
- struct rt6_info *rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops);
+ struct rt6_info *rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops,
+ net->loopback_dev);
struct neighbour *neigh;
if (rt == NULL) {
@@ -2002,13 +1997,11 @@ struct rt6_info *addrconf_dst_alloc(struct inet6_dev *idev,
return ERR_PTR(-ENOMEM);
}
- dev_hold(net->loopback_dev);
in6_dev_hold(idev);
rt->dst.flags = DST_HOST;
rt->dst.input = ip6_input;
rt->dst.output = ip6_output;
- rt->rt6i_dev = net->loopback_dev;
rt->rt6i_idev = idev;
dst_metric_set(&rt->dst, RTAX_HOPLIMIT, -1);
rt->dst.obsolete = -1;
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index b1932a6..f9bb423 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -1345,7 +1345,7 @@ static inline struct xfrm_dst *xfrm_alloc_dst(struct net *net, int family)
default:
BUG();
}
- xdst = dst_alloc(dst_ops, 0);
+ xdst = dst_alloc(dst_ops, NULL, 0, 0, 0);
xfrm_policy_put_afinfo(afinfo);
if (likely(xdst))
--
1.7.4.1
^ permalink raw reply related
* [v3 RFC PATCH 7/8] net: Use non-zero allocations in dst_alloc().
From: David Miller @ 2011-03-03 1:02 UTC (permalink / raw)
To: netdev
Make dst_alloc() and it's users explicitly initialize the entire
entry.
The zero'ing done by kmem_cache_zalloc() was almost entirely
redundant.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/core/dst.c | 20 ++++++++++--
net/decnet/dn_route.c | 2 +
net/ipv4/route.c | 82 +++++++++++++++++++++++++++--------------------
net/ipv6/route.c | 8 ++++-
net/xfrm/xfrm_policy.c | 1 +
5 files changed, 74 insertions(+), 39 deletions(-)
diff --git a/net/core/dst.c b/net/core/dst.c
index 9505778..30f0093 100644
--- a/net/core/dst.c
+++ b/net/core/dst.c
@@ -175,22 +175,36 @@ void *dst_alloc(struct dst_ops *ops, struct net_device *dev,
if (ops->gc(ops))
return NULL;
}
- dst = kmem_cache_zalloc(ops->kmem_cachep, GFP_ATOMIC);
+ dst = kmem_cache_alloc(ops->kmem_cachep, GFP_ATOMIC);
if (!dst)
return NULL;
- dst->ops = ops;
+ dst->child = NULL;
dst->dev = dev;
if (dev)
dev_hold(dev);
+ dst->ops = ops;
dst_init_metrics(dst, dst_default_metrics, true);
+ dst->expires = 0UL;
dst->path = dst;
+ dst->neighbour = NULL;
+ dst->hh = NULL;
+#ifdef CONFIG_XFRM
+ dst->xfrm = NULL;
+#endif
dst->input = dst_discard;
dst->output = dst_discard;
-
+ dst->error = 0;
dst->obsolete = initial_obsolete;
+ dst->header_len = 0;
+ dst->trailer_len = 0;
+#ifdef CONFIG_IP_ROUTE_CLASSID
+ dst->tclassid = 0;
+#endif
atomic_set(&dst->__refcnt, initial_ref);
+ dst->__use = 0;
dst->lastuse = jiffies;
dst->flags = flags;
+ dst->next = NULL;
#if RT_CACHE_DEBUG >= 2
atomic_inc(&dst_total);
#endif
diff --git a/net/decnet/dn_route.c b/net/decnet/dn_route.c
index a3d2c3f..12dc7e4 100644
--- a/net/decnet/dn_route.c
+++ b/net/decnet/dn_route.c
@@ -1126,6 +1126,7 @@ make_route:
if (rt == NULL)
goto e_nobufs;
+ memset(&rt->fl, 0, sizeof(rt->fl));
rt->fl.fld_src = oldflp->fld_src;
rt->fl.fld_dst = oldflp->fld_dst;
rt->fl.oif = oldflp->oif;
@@ -1391,6 +1392,7 @@ make_route:
if (rt == NULL)
goto e_nobufs;
+ memset(&rt->fl, 0, sizeof(rt->fl));
rt->rt_saddr = fl.fld_src;
rt->rt_daddr = fl.fld_dst;
rt->rt_gateway = fl.fld_dst;
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 9fcf8ce..4f3ac67 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -1100,7 +1100,6 @@ static void rt_set_nexthop(struct rtable *rt, const struct flowi *oldflp,
#endif
set_class_tag(rt, itag);
#endif
- rt->rt_type = type;
}
static struct rtable *rt_dst_alloc(struct net_device *dev,
@@ -1150,24 +1149,27 @@ static int ip_route_input_mc(struct sk_buff *skb, __be32 daddr, __be32 saddr,
if (!rth)
goto e_nobufs;
+#ifdef CONFIG_IP_ROUTE_CLASSID
+ rth->dst.tclassid = itag;
+#endif
rth->dst.output = ip_rt_bug;
rth->rt_key_dst = daddr;
- rth->rt_dst = daddr;
- rth->rt_tos = tos;
- rth->rt_mark = skb->mark;
rth->rt_key_src = saddr;
+ rth->rt_genid = rt_genid(dev_net(dev));
+ rth->rt_flags = RTCF_MULTICAST;
+ rth->rt_type = RTN_MULTICAST;
+ rth->rt_tos = tos;
+ rth->rt_dst = daddr;
rth->rt_src = saddr;
-#ifdef CONFIG_IP_ROUTE_CLASSID
- rth->dst.tclassid = itag;
-#endif
rth->rt_iif = dev->ifindex;
rth->rt_oif = 0;
+ rth->rt_mark = skb->mark;
rth->rt_gateway = daddr;
rth->rt_spec_dst= spec_dst;
- rth->rt_genid = rt_genid(dev_net(dev));
- rth->rt_flags = RTCF_MULTICAST;
- rth->rt_type = RTN_MULTICAST;
+ rth->rt_peer_genid = 0;
+ rth->peer = NULL;
+ rth->fi = NULL;
if (our) {
rth->dst.input= ip_local_deliver;
rth->rt_flags |= RTCF_LOCAL;
@@ -1288,24 +1290,27 @@ static int __mkroute_input(struct sk_buff *skb,
}
rth->rt_key_dst = daddr;
- rth->rt_dst = daddr;
- rth->rt_tos = tos;
- rth->rt_mark = skb->mark;
rth->rt_key_src = saddr;
+ rth->rt_genid = rt_genid(dev_net(rth->dst.dev));
+ rth->rt_flags = flags;
+ rth->rt_type = res->type;
+ rth->rt_tos = tos;
+ rth->rt_dst = daddr;
rth->rt_src = saddr;
- rth->rt_gateway = daddr;
rth->rt_iif = in_dev->dev->ifindex;
rth->rt_oif = 0;
+ rth->rt_mark = skb->mark;
+ rth->rt_gateway = daddr;
rth->rt_spec_dst= spec_dst;
+ rth->rt_peer_genid = 0;
+ rth->peer = NULL;
+ rth->fi = NULL;
rth->dst.input = ip_forward;
rth->dst.output = ip_output;
- rth->rt_genid = rt_genid(dev_net(rth->dst.dev));
rt_set_nexthop(rth, NULL, res, res->fi, res->type, itag);
- rth->rt_flags = flags;
-
*result = rth;
err = 0;
cleanup:
@@ -1450,29 +1455,33 @@ local_input:
if (!rth)
goto e_nobufs;
+ rth->dst.input= ip_local_deliver;
rth->dst.output= ip_rt_bug;
- rth->rt_genid = rt_genid(net);
+#ifdef CONFIG_IP_ROUTE_CLASSID
+ rth->dst.tclassid = itag;
+#endif
rth->rt_key_dst = daddr;
- rth->rt_dst = daddr;
- rth->rt_tos = tos;
- rth->rt_mark = skb->mark;
rth->rt_key_src = saddr;
+ rth->rt_genid = rt_genid(net);
+ rth->rt_flags = flags|RTCF_LOCAL;
+ rth->rt_type = res.type;
+ rth->rt_tos = tos;
+ rth->rt_dst = daddr;
rth->rt_src = saddr;
-#ifdef CONFIG_IP_ROUTE_CLASSID
- rth->dst.tclassid = itag;
-#endif
rth->rt_iif = dev->ifindex;
+ rth->rt_oif = 0;
+ rth->rt_mark = skb->mark;
rth->rt_gateway = daddr;
rth->rt_spec_dst= spec_dst;
- rth->dst.input= ip_local_deliver;
- rth->rt_flags = flags|RTCF_LOCAL;
+ rth->rt_peer_genid = 0;
+ rth->peer = NULL;
+ rth->fi = NULL;
if (res.type == RTN_UNREACHABLE) {
rth->dst.input= ip_error;
rth->dst.error= -err;
rth->rt_flags &= ~RTCF_LOCAL;
}
- rth->rt_type = res.type;
rth = rt_finalize(rth, skb);
err = 0;
if (IS_ERR(rth))
@@ -1615,19 +1624,24 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
if (!rth)
return ERR_PTR(-ENOBUFS);
+ rth->dst.output = ip_output;
+
rth->rt_key_dst = oldflp->fl4_dst;
- rth->rt_tos = tos;
rth->rt_key_src = oldflp->fl4_src;
- rth->rt_oif = oldflp->oif;
- rth->rt_mark = oldflp->mark;
+ rth->rt_genid = rt_genid(dev_net(dev_out));
+ rth->rt_flags = flags;
+ rth->rt_type = type;
+ rth->rt_tos = tos;
rth->rt_dst = fl->fl4_dst;
rth->rt_src = fl->fl4_src;
rth->rt_iif = oldflp->oif ? : dev_out->ifindex;
+ rth->rt_oif = oldflp->oif;
+ rth->rt_mark = oldflp->mark;
rth->rt_gateway = fl->fl4_dst;
rth->rt_spec_dst= fl->fl4_src;
-
- rth->dst.output=ip_output;
- rth->rt_genid = rt_genid(dev_net(dev_out));
+ rth->rt_peer_genid = 0;
+ rth->peer = NULL;
+ rth->fi = NULL;
RT_CACHE_STAT_INC(out_slow_tot);
@@ -1654,8 +1668,6 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
}
rt_set_nexthop(rth, oldflp, res, fi, type, 0);
-
- rth->rt_flags = flags;
return rth;
}
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 9a685f5..13f717d 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -223,7 +223,11 @@ static struct rt6_info ip6_blk_hole_entry_template = {
static inline struct rt6_info *ip6_dst_alloc(struct dst_ops *ops,
struct net_device *dev)
{
- return (struct rt6_info *)dst_alloc(ops, dev, 0, 0, 0);
+ struct rt6_info *rt = dst_alloc(ops, dev, 0, 0, 0);
+
+ memset(&rt->rt6i_table, 0, sizeof(*rt) - sizeof(struct dst_entry));
+
+ return rt;
}
static void ip6_dst_destroy(struct dst_entry *dst)
@@ -878,6 +882,8 @@ struct dst_entry *ip6_blackhole_route(struct net *net, struct dst_entry *dst_ori
rt = dst_alloc(&ip6_dst_blackhole_ops, ort->dst.dev, 1, 0, 0);
if (rt) {
+ memset(&rt->rt6i_table, 0, sizeof(*rt) - sizeof(struct dst_entry));
+
new = &rt->dst;
new->__use = 1;
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index f9bb423..86dcc94 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -1346,6 +1346,7 @@ static inline struct xfrm_dst *xfrm_alloc_dst(struct net *net, int family)
BUG();
}
xdst = dst_alloc(dst_ops, NULL, 0, 0, 0);
+ memset(&xdst->u.rt6.rt6i_table, 0, sizeof(*xdst) - sizeof(struct dst_entry));
xfrm_policy_put_afinfo(afinfo);
if (likely(xdst))
--
1.7.4.1
^ permalink raw reply related
* [v3 RFC PATCH 8/8] ipv4: Kill rt_key_{src,dst} from struct rtable.
From: David Miller @ 2011-03-03 1:02 UTC (permalink / raw)
To: netdev
They are always used in contexts where they can be reconstituted,
or where the finally resolved rt->rt_{src,dst} is semantically
equivalent.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/route.h | 8 ++------
net/ipv4/ipmr.c | 33 ++++++++++++++++++---------------
net/ipv4/route.c | 24 +++++++-----------------
net/ipv4/xfrm4_policy.c | 2 --
4 files changed, 27 insertions(+), 40 deletions(-)
diff --git a/include/net/route.h b/include/net/route.h
index ebf50a9..7e9d428 100644
--- a/include/net/route.h
+++ b/include/net/route.h
@@ -53,10 +53,6 @@ struct fib_info;
struct rtable {
struct dst_entry dst;
- /* Lookup key. */
- __be32 rt_key_dst;
- __be32 rt_key_src;
-
int rt_genid;
unsigned rt_flags;
__u16 rt_type;
@@ -205,8 +201,8 @@ static inline struct rtable *ip_route_newports(struct rtable *rt,
if (sport != orig_sport || dport != orig_dport) {
struct flowi fl = { .oif = rt->rt_oif,
.mark = rt->rt_mark,
- .fl4_dst = rt->rt_key_dst,
- .fl4_src = rt->rt_key_src,
+ .fl4_dst = rt->rt_dst,
+ .fl4_src = rt->rt_src,
.fl4_tos = rt->rt_tos,
.proto = protocol,
.fl_ip_sport = sport,
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index 9d5f634..c8dd0b6 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -1815,13 +1815,14 @@ int ip_mr_input(struct sk_buff *skb)
{
struct rtable *rt = skb_rtable(skb);
+ struct iphdr *iph = ip_hdr(skb);
struct flowi fl = {
- .fl4_dst = rt->rt_key_dst,
- .fl4_src = rt->rt_key_src,
- .fl4_tos = rt->rt_tos,
+ .fl4_dst = iph->daddr,
+ .fl4_src = iph->saddr,
+ .fl4_tos = iph->tos,
.oif = rt->rt_oif,
- .iif = rt->rt_iif,
- .mark = rt->rt_mark,
+ .iif = skb->dev->ifindex,
+ .mark = skb->mark,
};
err = ipmr_fib_lookup(net, &fl, &mrt);
if (err < 0) {
@@ -1958,13 +1959,14 @@ int pim_rcv_v1(struct sk_buff *skb)
{
struct rtable *rt = skb_rtable(skb);
+ struct iphdr *iph = ip_hdr(skb);
struct flowi fl = {
- .fl4_dst = rt->rt_key_dst,
- .fl4_src = rt->rt_key_src,
- .fl4_tos = rt->rt_tos,
+ .fl4_dst = iph->daddr,
+ .fl4_src = iph->saddr,
+ .fl4_tos = iph->tos,
.oif = rt->rt_oif,
- .iif = rt->rt_iif,
- .mark = rt->rt_mark,
+ .iif = skb->dev->ifindex,
+ .mark = skb->mark,
};
if (ipmr_fib_lookup(net, &fl, &mrt) < 0)
goto drop;
@@ -2000,13 +2002,14 @@ static int pim_rcv(struct sk_buff *skb)
{
struct rtable *rt = skb_rtable(skb);
+ struct iphdr *iph = ip_hdr(skb);
struct flowi fl = {
- .fl4_dst = rt->rt_key_dst,
- .fl4_src = rt->rt_key_src,
- .fl4_tos = rt->rt_tos,
+ .fl4_dst = iph->daddr,
+ .fl4_src = iph->saddr,
+ .fl4_tos = iph->tos,
.oif = rt->rt_oif,
- .iif = rt->rt_iif,
- .mark = rt->rt_mark,
+ .iif = skb->dev->ifindex,
+ .mark = skb->mark,
};
if (ipmr_fib_lookup(net, &fl, &mrt) < 0)
goto drop;
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 4f3ac67..04b8954 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -978,8 +978,8 @@ void ip_rt_get_source(u8 *addr, struct rtable *rt)
src = rt->rt_src;
else {
struct flowi fl = {
- .fl4_dst = rt->rt_key_dst,
- .fl4_src = rt->rt_key_src,
+ .fl4_dst = rt->rt_dst,
+ .fl4_src = rt->rt_src,
.fl4_tos = rt->rt_tos,
.oif = rt->rt_oif,
.iif = rt->rt_iif,
@@ -1154,8 +1154,6 @@ static int ip_route_input_mc(struct sk_buff *skb, __be32 daddr, __be32 saddr,
#endif
rth->dst.output = ip_rt_bug;
- rth->rt_key_dst = daddr;
- rth->rt_key_src = saddr;
rth->rt_genid = rt_genid(dev_net(dev));
rth->rt_flags = RTCF_MULTICAST;
rth->rt_type = RTN_MULTICAST;
@@ -1289,8 +1287,6 @@ static int __mkroute_input(struct sk_buff *skb,
goto cleanup;
}
- rth->rt_key_dst = daddr;
- rth->rt_key_src = saddr;
rth->rt_genid = rt_genid(dev_net(rth->dst.dev));
rth->rt_flags = flags;
rth->rt_type = res->type;
@@ -1461,8 +1457,6 @@ local_input:
rth->dst.tclassid = itag;
#endif
- rth->rt_key_dst = daddr;
- rth->rt_key_src = saddr;
rth->rt_genid = rt_genid(net);
rth->rt_flags = flags|RTCF_LOCAL;
rth->rt_type = res.type;
@@ -1626,8 +1620,6 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
rth->dst.output = ip_output;
- rth->rt_key_dst = oldflp->fl4_dst;
- rth->rt_key_src = oldflp->fl4_src;
rth->rt_genid = rt_genid(dev_net(dev_out));
rth->rt_flags = flags;
rth->rt_type = type;
@@ -1897,8 +1889,6 @@ struct dst_entry *ipv4_blackhole_route(struct net *net, struct dst_entry *dst_or
if (new->dev)
dev_hold(new->dev);
- rt->rt_key_dst = ort->rt_key_dst;
- rt->rt_key_src = ort->rt_key_src;
rt->rt_tos = ort->rt_tos;
rt->rt_iif = ort->rt_iif;
rt->rt_oif = ort->rt_oif;
@@ -1947,7 +1937,7 @@ struct rtable *ip_route_output_flow(struct net *net, struct flowi *flp,
}
EXPORT_SYMBOL_GPL(ip_route_output_flow);
-static int rt_fill_info(struct net *net,
+static int rt_fill_info(struct net *net, __be32 src,
struct sk_buff *skb, u32 pid, u32 seq, int event,
int nowait, unsigned int flags)
{
@@ -1977,9 +1967,9 @@ static int rt_fill_info(struct net *net,
NLA_PUT_BE32(skb, RTA_DST, rt->rt_dst);
- if (rt->rt_key_src) {
+ if (src) {
r->rtm_src_len = 32;
- NLA_PUT_BE32(skb, RTA_SRC, rt->rt_key_src);
+ NLA_PUT_BE32(skb, RTA_SRC, src);
}
if (rt->dst.dev)
NLA_PUT_U32(skb, RTA_OIF, rt->dst.dev->ifindex);
@@ -1989,7 +1979,7 @@ static int rt_fill_info(struct net *net,
#endif
if (rt_is_input_route(rt))
NLA_PUT_BE32(skb, RTA_PREFSRC, rt->rt_spec_dst);
- else if (rt->rt_src != rt->rt_key_src)
+ else if (rt->rt_src != src)
NLA_PUT_BE32(skb, RTA_PREFSRC, rt->rt_src);
if (rt->rt_dst != rt->rt_gateway)
@@ -2131,7 +2121,7 @@ static int inet_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr* nlh, void
if (rtm->rtm_flags & RTM_F_NOTIFY)
rt->rt_flags |= RTCF_NOTIFY;
- err = rt_fill_info(net, skb, NETLINK_CB(in_skb).pid, nlh->nlmsg_seq,
+ err = rt_fill_info(net, src, skb, NETLINK_CB(in_skb).pid, nlh->nlmsg_seq,
RTM_NEWROUTE, 0, 0);
if (err <= 0)
goto errout_free;
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index c70c42e..f0bc2c8 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -70,8 +70,6 @@ static int xfrm4_fill_dst(struct xfrm_dst *xdst, struct net_device *dev,
{
struct rtable *rt = (struct rtable *)xdst->route;
- rt->rt_key_dst = fl->fl4_dst;
- rt->rt_key_src = fl->fl4_src;
rt->rt_tos = fl->fl4_tos;
rt->rt_iif = fl->iif;
rt->rt_oif = fl->oif;
--
1.7.4.1
^ permalink raw reply related
* Re: [BUG] VPN broken in net-next
From: David Miller @ 2011-03-03 1:03 UTC (permalink / raw)
Cc: netdev
In-Reply-To: <20110302165653.4b854489@nehalam>
From: Stephen Hemminger <shemminger@vyatta.com>
Date: Wed, 2 Mar 2011 16:56:53 -0800
> On Wed, 02 Mar 2011 16:50:09 -0800 (PST)
> David Miller <davem@davemloft.net> wrote:
>
>> From: Stephen Hemminger <shemminger@vyatta.com>
>> Date: Wed, 2 Mar 2011 16:46:37 -0800
>>
>> > The addresses (that matter) when VPN is up are:
>>
>> I really need to know what addresses interfaces have the time of the
>> __ip_dev_find() call which, if I'm not mistaken, is before the VPN is
>> up.
>
>
> inet 127.0.0.1/8 scope host lo
> inet 192.168.1.11/24 brd 192.168.1.255 scope global eth0
> inet 192.168.100.1/24 brd 192.168.100.255 scope global virbr0
> inet 192.168.99.1/24 brd 192.168.99.255 scope global virbr1
I see nothing providing 10.0.whatever that __ip_dev_find() is being
asked to resolve.
I think we were allowing the route lookup pptp is trying to do at
connect time erroneously, and it should elide the explicit source
address specification in the flow.
See my other email.
^ permalink raw reply
* Re: RFC v1: sysctl: add sysctl header cookie, share tables between nets
From: David Miller @ 2011-03-03 1:06 UTC (permalink / raw)
To: lucian.grijincu; +Cc: adobriyan, ebiederm, tavi, netdev, linux-kernel
In-Reply-To: <1298659961-23863-1-git-send-email-lucian.grijincu@gmail.com>
From: Lucian Adrian Grijincu <lucian.grijincu@gmail.com>
Date: Fri, 25 Feb 2011 20:52:32 +0200
> This is a new approach to the "share sysctl tables" RFC series I
> posted earlier this month.
I do not disagree conceptually with these changes from a networking
perspective, but I am not a sysctl layer expert so I don't know if the
generic sysctl bits are a good idea or not.
^ permalink raw reply
* Re: fcoe: correct checking for bonding
From: Robert Love @ 2011-03-03 1:09 UTC (permalink / raw)
To: James.Bottomley, Jiri Pirko
Cc: linux-scsi@vger.kernel.org, davem@davemloft.net,
devel@open-fcoe.org, netdev@vger.kernel.org, fubar@us.ibm.com,
joe.eykholt@gmail.com
In-Reply-To: <20110302095550.GA2858@psychotron.brq.redhat.com>
On Wed, 2011-03-02 at 01:55 -0800, Jiri Pirko wrote:
> Or perhaps this should be applied to net-next?
>
I think this should go through scsi-misc as all the other
libfc/libfcoe/fcoe patches do.
> Wed, Mar 02, 2011 at 07:05:35AM CET, jpirko@redhat.com wrote:
> >Check for bonding master and refuse to use that.
> >
> >Signed-off-by: Jiri Pirko <jpirko@redhat.com>
> >---
> > drivers/scsi/fcoe/fcoe.c | 4 +---
> > 1 files changed, 1 insertions(+), 3 deletions(-)
> >
> >diff --git a/drivers/scsi/fcoe/fcoe.c b/drivers/scsi/fcoe/fcoe.c
> >index 9f9600b..3becc6a 100644
> >--- a/drivers/scsi/fcoe/fcoe.c
> >+++ b/drivers/scsi/fcoe/fcoe.c
> >@@ -285,9 +285,7 @@ static int fcoe_interface_setup(struct fcoe_interface *fcoe,
> > }
> >
> > /* Do not support for bonding device */
> >- if ((netdev->priv_flags & IFF_MASTER_ALB) ||
> >- (netdev->priv_flags & IFF_SLAVE_INACTIVE) ||
> >- (netdev->priv_flags & IFF_MASTER_8023AD)) {
> >+ if (netdev->priv_flags & IFF_BONDING && netdev->flags & IFF_MASTER) {
> > FCOE_NETDEV_DBG(netdev, "Bonded interfaces not supported\n");
> > return -EOPNOTSUPP;
> > }
> >--
> >1.7.3.4
> >
James, feel free to pick up this patch. I don't have anything in my fcoe
tree right now that it would conflict with. I'll also put it in my tree
and resend if you don't put it into scsi-misc directly.
Acked-by: Robert Love <robert.w.love@intel.com>
^ permalink raw reply
* Re: [BUG] VPN broken in net-next
From: Stephen Hemminger @ 2011-03-03 1:16 UTC (permalink / raw)
To: David Miller; +Cc: netdev
In-Reply-To: <20110302.170346.13723622.davem@davemloft.net>
On Wed, 02 Mar 2011 17:03:46 -0800 (PST)
David Miller <davem@davemloft.net> wrote:
> From: Stephen Hemminger <shemminger@vyatta.com>
> Date: Wed, 2 Mar 2011 16:56:53 -0800
>
> > On Wed, 02 Mar 2011 16:50:09 -0800 (PST)
> > David Miller <davem@davemloft.net> wrote:
> >
> >> From: Stephen Hemminger <shemminger@vyatta.com>
> >> Date: Wed, 2 Mar 2011 16:46:37 -0800
> >>
> >> > The addresses (that matter) when VPN is up are:
> >>
> >> I really need to know what addresses interfaces have the time of the
> >> __ip_dev_find() call which, if I'm not mistaken, is before the VPN is
> >> up.
> >
> >
> > inet 127.0.0.1/8 scope host lo
> > inet 192.168.1.11/24 brd 192.168.1.255 scope global eth0
> > inet 192.168.100.1/24 brd 192.168.100.255 scope global virbr0
> > inet 192.168.99.1/24 brd 192.168.99.255 scope global virbr1
>
> I see nothing providing 10.0.whatever that __ip_dev_find() is being
> asked to resolve.
>
> I think we were allowing the route lookup pptp is trying to do at
> connect time erroneously, and it should elide the explicit source
> address specification in the flow.
>
The VPN connection comes up the problem is that no packets pass over
it successfully. The 10.X address is the other side of the VPN.
I tried this, but it didn't work.
--- a/drivers/net/pptp.c 2011-03-02 17:01:55.353313682 -0800
+++ b/drivers/net/pptp.c 2011-03-02 17:02:05.381146980 -0800
@@ -473,7 +473,6 @@ static int pptp_connect(struct socket *s
.nl_u = {
.ip4_u = {
.daddr = opt->dst_addr.sin_addr.s_addr,
- .saddr = opt->src_addr.sin_addr.s_addr,
.tos = RT_CONN_FLAGS(sk) } },
.proto = IPPROTO_GRE };
security_sk_classify_flow(sk, &fl);
^ permalink raw reply
* Re: [RFC][PATCH] Export DNSSL RA option to userspace
From: Pierre Ynard @ 2011-03-03 1:27 UTC (permalink / raw)
To: netdev; +Cc: pierre-list
In-Reply-To: <20101212144300.68e0fb16@mjolnir.ossman.eu>
Hello,
Thanks for the patch!
> From a5b60e80eeed87763c811ecfb0d7aa5695d0a2bf Mon Sep 17 00:00:00 2001
> From: Pierre Ossman <pierre@xxxxxxxxx>
> Date: Sun, 12 Dec 2010 00:06:48 +0100
> Subject: [PATCH 1/2] ipv6: export DNS search list option to userspace
>
> Like DNS resolver addresses, the suffixes to be used in DNS lookups need
> to be configured by userspace. Make sure userspace has access to the
> option containing that information.
>
> Signed-off-by: Pierre Ossman <pierre@xxxxxxxxx>
> ---
> include/net/ndisc.h | 3 ++-
> net/ipv6/ndisc.c | 8 +++++++-
> 2 files changed, 9 insertions(+), 2 deletions(-)
Acked-by: Pierre Ynard <linkfanel@yahoo.fr>
--
Pierre Ynard
"Une âme dans un corps, c'est comme un dessin sur une feuille de papier."
^ permalink raw reply
* Re: [RFC][PATCH] Export all RA options that we don't handle to userspace
From: Pierre Ynard @ 2011-03-03 1:30 UTC (permalink / raw)
To: netdev; +Cc: pierre-list
In-Reply-To: <20101212144706.1d2be015@mjolnir.ossman.eu>
> Second patch that exports everything. If noone objects to this model,
> then merge the two patches and just use the commit message from the
> second one.
I'm a bit wary about that, we don't know what new option might come out
in the future, that we really don't want to forward to userspace. I
suggest we remain conservative here.
Regards,
--
Pierre Ynard
"Une âme dans un corps, c'est comme un dessin sur une feuille de papier."
^ permalink raw reply
* Re: [RFC][PATCH] Export DNSSL RA option to userspace
From: Pierre Ynard @ 2011-03-03 1:36 UTC (permalink / raw)
To: netdev; +Cc: pierre-list
In-Reply-To: <20101212150701.20d4fd6d@mjolnir.ossman.eu>
> I've also noticed a problem in the nduseropt code that I'm not sure how
> to solve (given that this is now a stable userspace interface). Both
> RFC5006 and RFC6106 state the following:
>
> Note: An RDNSS address or a DNSSL domain name MUST be used only as
> long as both the RA router Lifetime (advertised by a Router
> Advertisement message [RFC4861]) and the corresponding option
> Lifetime have not expired.
>
> But the RA router lifetime is not included in the information sent.
> Normally this is probably not an issue as the RDNSS and DNSSL lifetime
> will be shorter than the router lifetime. One exception is when the
> router is disabled at which point it will send a RA with router
> lifetime to 0 (RFC4861 section 6.2.5). That means userspace will not be
> informed that the DNS information should be removed immediately*.
>
> Is there any way we can safely extend the interface with this
> information? I'm not familiar enough with it myself yet to determine if
> it's possible...
You could define a new netlink attribute to carry the lifetime of
the router in that RA. But that seems pointless to me, since in my
understanding, you have to track the lifetime of that router in future
RAs too, independently of whether or not they contain an RDNSS/DNSSL
option or not. So the nduseropt code path is not the right place to
track this.
I suppose it's possible to use a netlink socket to track events related
to the corresponding router and/or directly query its lifetime, in
the same way as any networking tool does, and remove the option when
it expires. That's what I planned to do in rdnssd, although I haven't
looked into details.
Regards,
--
Pierre Ynard
"Une âme dans un corps, c'est comme un dessin sur une feuille de papier."
^ permalink raw reply
* Re: [PATCH] e1000: fix race condition while driver unload/reset using kref count
From: Prasanna Panchamukhi @ 2011-03-03 1:47 UTC (permalink / raw)
To: Brandeburg, Jesse
Cc: Allan, Bruce W, Kirsher, Jeffrey T, Pieper, Jeffrey E,
e1000-devel@lists.sourceforge.net, netdev@vger.kernel.org
In-Reply-To: <alpine.WNT.2.00.1103021337560.8204@JBRANDEB-DESK2.amr.corp.intel.com>
[Sorry, resending it, as it did not make it to netdev]
On 03/02/2011 02:50 PM, Brandeburg, Jesse wrote:
>
> On Wed, 2 Mar 2011, prasanna.panchamukhi@riverbed.com wrote:
>
>> This race conditions occurs with reentrant e1000_down(), which gets
>> called by multiple threads while driver unload or reset.
>> This patch fixes the race condition when one thread tries to destroy
>> the memory allocated for tx buffer_info, while another thread still
>> happen to access tx buffer_info.
>> This patch fixes the above race condition using kref count.
> I'm very interested in any test cases that you might have come up with to
> reproduce this issue.
This patch is an alternative approach to the below commit:
commit 338c15e470d818f215d651505dc169d4e92f36a4
Author: Jesse Brandeburg <jesse.brandeburg@intel.com>
Date: Wed Sep 22 18:22:42 2010 +0000
e1000: fix occasional panic on unload
Its a very rare event.
One of the ways is to tweak the driver to reset more frequently(5 times
per second) & do ifconfig up/down in tight loop.
Another way could be to do a module load/unload or keep rebooting the
system in tight loop.
> The patch itself looks interesting, and probably okay, but we really need
> a reproduction case.
>
Yeah, I understand your point.
> Also, do we need to adjust the rtnl_lock stuff or do this for rx
> buffer_info structs?
kref counting buffer_info struct is good enough.
> I'm concerned that maybe we don't understand the full flow of events
> leading up to this failure.
possible race condition because of e1000_down() being re-entrant. Multiple threads can end up calling
e1000_down() either during driver unload or reboot/reset.
below is the kernel traces
Oct 10 11:59:48 localhost kernel: EIP: 0060:[put_page+2/117] Tainted: PF
B VLI
Oct 10 11:59:48 localhost kernel: EIP: 0060:[<e0148432>] Tainted: PF B
VLI
Oct 10 11:59:48 localhost kernel: EFLAGS: 00010202 (2.6.9-34.EL-rbt-1784SMP)
Oct 10 11:59:48 localhost kernel: EIP is at put_page+0x2/0x75
Oct 10 11:59:48 localhost kernel: eax: 12a051a5 ebx: 00000002 ecx: e04b3300
edx: 12a051a5
Oct 10 11:59:48 localhost kernel: esi: f72e8b80 edi: 00000cb7 ebp: f7d09000
esp: e2158ebc
Oct 10 11:59:48 localhost kernel: ds: 007b es: 007b ss: 0068
Oct 10 11:59:48 localhost kernel: Process events/1 (pid: 7, threadinfo=e2158000
task=e2182b70)
Oct 10 11:59:48 localhost kernel: Stack: e03cea9c f72e8b80 f8b6db70 e03ceada
00000000 e03ceb98 e02d8e80 e2158ee0
Oct 10 11:59:48 localhost kernel: f88c0000 00000000 f72e8b80 f7ce2180
e02c4a88 f7ce2180 f7d092a0 e02c4ab4
Oct 10 11:59:48 localhost kernel: f7d092a0 00000000 0103f0f8 f7d09000
e02c4b5d f88c0000 f7d092a0 e02c2247
Oct 10 11:59:48 localhost kernel: Call Trace:
Oct 10 11:59:48 localhost kernel: [skb_release_data+57/111]
skb_release_data+0x39/0x6f
Oct 10 11:59:48 localhost kernel: [<e03cea9c>] skb_release_data+0x39/0x6f
Oct 10 11:59:48 localhost kernel: [kfree_skbmem+8/21] kfree_skbmem+0x8/0x15
Oct 10 11:59:48 localhost kernel: [<e03ceada>] kfree_skbmem+0x8/0x15
Oct 10 11:59:48 localhost kernel: [__kfree_skb+177/343] __kfree_skb+0xb1/0x157
Oct 10 11:59:48 localhost kernel: [<e03ceb98>] __kfree_skb+0xb1/0x157
Oct 10 11:59:48 localhost kernel: [e1000_get_phy_info_m88+56/281]
e1000_get_phy_info_m88+0x38/0x119
Oct 10 11:59:48 localhost kernel: [<e02d8e80>]
e1000_get_phy_info_m88+0x38/0x119
Oct 10 11:59:48 localhost kernel: [e1000_unmap_and_free_tx_resource+156/165]
e1000_unmap_and_free_tx_resource+0x9c/0xa5
Oct 10 11:59:48 localhost kernel: [<e02c4a88>]
e1000_unmap_and_free_tx_resource+0x9c/0xa5
Oct 10 11:59:48 localhost kernel: [e1000_clean_tx_ring+35/162]
e1000_clean_tx_ring+0x23/0xa2
Oct 10 11:59:48 localhost kernel: [<e02c4ab4>] e1000_clean_tx_ring+0x23/0xa2
Oct 10 11:59:48 localhost kernel: [e1000_clean_all_tx_rings+42/56]
e1000_clean_all_tx_rings+0x2a/0x38
Oct 10 11:59:48 localhost kernel: [<e02c4b5d>]
e1000_clean_all_tx_rings+0x2a/0x38
Oct 10 11:59:48 localhost kernel: [e1000_down+356/380] e1000_down+0x164/0x17c
Oct 10 11:59:48 localhost kernel: [<e02c2247>] e1000_down+0x164/0x17c
Oct 10 11:59:48 localhost kernel: [e1000_reinit_locked+120/146]
e1000_reinit_locked+0x78/0x92
Oct 10 11:59:48 localhost kernel: [<e02c22d7>] e1000_reinit_locked+0x78/0x92
Oct 10 11:59:48 localhost kernel: [e1000_reset_task+21/33]
e1000_reset_task+0x15/0x21
Oct 10 11:59:48 localhost kernel: [<e02c74a3>] e1000_reset_task+0x15/0x21
Oct 10 11:59:48 localhost kernel: [worker_thread+432/570]
worker_thread+0x1b0/0x23a
Oct 10 11:59:48 localhost kernel: [<e0132f8d>] worker_thread+0x1b0/0x23a
Oct 10 11:59:48 localhost kernel: [schedule+808/2803] schedule+0x328/0xaf3
Oct 10 11:59:48 localhost kernel: [<e0431e18>] schedule+0x328/0xaf3
Oct 10 11:59:48 localhost kernel: [e1000_reset_task+0/33]
Thanks
Prasanna
> Jesse
>
>> Signed-off-by: Prasanna S. Panchamukhi<prasanna.panchamukhi@riverbed.com>
>> ---
>> drivers/net/e1000/e1000.h | 2 +
>> drivers/net/e1000/e1000_main.c | 41 +++++++++++++++++++++++++++++++++++++--
>> 2 files changed, 40 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/net/e1000/e1000.h b/drivers/net/e1000/e1000.h
>> index a881dd0..36f55b3 100644
>> --- a/drivers/net/e1000/e1000.h
>> +++ b/drivers/net/e1000/e1000.h
>> @@ -168,6 +168,8 @@ struct e1000_tx_ring {
>> unsigned int next_to_clean;
>> /* array of buffer information structs */
>> struct e1000_buffer *buffer_info;
>> + spinlock_t bufinfo_lock; /* protect access to buffer_info */
>> + struct kref bufinfo_refcount; /* refcount access to buffer info */
>>
>> u16 tdh;
>> u16 tdt;
>> diff --git a/drivers/net/e1000/e1000_main.c b/drivers/net/e1000/e1000_main.c
>> index beec573..336d3e1 100644
>> --- a/drivers/net/e1000/e1000_main.c
>> +++ b/drivers/net/e1000/e1000_main.c
>> @@ -1531,6 +1531,8 @@ setup_tx_desc_die:
>>
>> txdr->next_to_use = 0;
>> txdr->next_to_clean = 0;
>> + spin_lock_init(&txdr->bufinfo_lock);
>> + kref_init(&txdr->bufinfo_refcount);
>>
>> return 0;
>> }
>> @@ -1880,6 +1882,22 @@ static void e1000_configure_rx(struct e1000_adapter *adapter)
>> ew32(RCTL, rctl);
>> }
>>
>> +/*
>> + * Free tx buffer info resources, only when no other thread is
>> + * accessing it. Access to buffer_info is refcounted by bufinfo_refcount,
>> + * hence memory allocated must be destroyed when bufinfo_refcount
>> + * becomes zero. This routine gets executed when bufinfo_refcount
>> + * becomes zero.
>> + */
>> +static void e1000_free_tx_buffer_info(struct kref *ref)
>> +{
>> + struct e1000_tx_ring *tx_ring =
>> + container_of(ref, struct e1000_tx_ring, bufinfo_refcount);
>> +
>> + vfree(tx_ring->buffer_info);
>> + tx_ring->buffer_info = NULL;
>> +}
>> +
>> /**
>> * e1000_free_tx_resources - Free Tx Resources per Queue
>> * @adapter: board private structure
>> @@ -1895,8 +1913,9 @@ static void e1000_free_tx_resources(struct e1000_adapter *adapter,
>>
>> e1000_clean_tx_ring(adapter, tx_ring);
>>
>> - vfree(tx_ring->buffer_info);
>> - tx_ring->buffer_info = NULL;
>> + spin_lock(&tx_ring->bufinfo_lock);
>> + kref_put(&tx_ring->bufinfo_refcount, e1000_free_tx_buffer_info);
>> + spin_unlock(&tx_ring->bufinfo_lock);
>>
>> dma_free_coherent(&pdev->dev, tx_ring->size, tx_ring->desc,
>> tx_ring->dma);
>> @@ -1954,8 +1973,20 @@ static void e1000_clean_tx_ring(struct e1000_adapter *adapter,
>> unsigned long size;
>> unsigned int i;
>>
>> - /* Free all the Tx ring sk_buffs */
>> + spin_lock(&tx_ring->bufinfo_lock);
>> + /*
>> + * Check buffer_info is not NULL, and
>> + * increment the refcount to prevent
>> + * the buffer getting freed underneath.
>> + */
>> + if (tx_ring->buffer_info == NULL) {
>> + spin_unlock(&tx_ring->bufinfo_lock);
>> + return;
>> + }
>> + kref_get(&tx_ring->bufinfo_refcount);
>> + spin_unlock(&tx_ring->bufinfo_lock);
>>
>> + /* Free all the Tx ring sk_buffs */
>> for (i = 0; i< tx_ring->count; i++) {
>> buffer_info =&tx_ring->buffer_info[i];
>> e1000_unmap_and_free_tx_resource(adapter, buffer_info);
>> @@ -1968,6 +1999,10 @@ static void e1000_clean_tx_ring(struct e1000_adapter *adapter,
>>
>> memset(tx_ring->desc, 0, tx_ring->size);
>>
>> + spin_lock(&tx_ring->bufinfo_lock);
>> + kref_put(&tx_ring->bufinfo_refcount, e1000_free_tx_buffer_info);
>> + spin_unlock(&tx_ring->bufinfo_lock);
>> +
>> tx_ring->next_to_use = 0;
>> tx_ring->next_to_clean = 0;
>> tx_ring->last_tx_tso = 0;
>>
^ permalink raw reply
* Re: [PATCH] e1000: fix race condition while driver unload/reset using kref count
From: Jeff Kirsher @ 2011-03-03 1:56 UTC (permalink / raw)
To: prasanna.panchamukhi@riverbed.com
Cc: Allan, Bruce W, Brandeburg, Jesse, Pieper, Jeffrey E,
e1000-devel@lists.sourceforge.net, netdev@vger.kernel.org
In-Reply-To: <1299091597-28409-1-git-send-email-prasanna.panchamukhi@riverbed.com>
[-- Attachment #1: Type: text/plain, Size: 801 bytes --]
On Wed, 2011-03-02 at 10:46 -0800, prasanna.panchamukhi@riverbed.com
wrote:
> This race conditions occurs with reentrant e1000_down(), which gets
> called by multiple threads while driver unload or reset.
> This patch fixes the race condition when one thread tries to destroy
> the memory allocated for tx buffer_info, while another thread still
> happen to access tx buffer_info.
> This patch fixes the above race condition using kref count.
>
> Signed-off-by: Prasanna S. Panchamukhi
> <prasanna.panchamukhi@riverbed.com>
> ---
> drivers/net/e1000/e1000.h | 2 +
> drivers/net/e1000/e1000_main.c | 41
> +++++++++++++++++++++++++++++++++++++--
> 2 files changed, 40 insertions(+), 3 deletions(-)
Thanks Prasanna! I have added the patch to my e1000 queue of patches.
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 490 bytes --]
^ permalink raw reply
* linux-next: manual merge of the net tree with the net-current tree
From: Stephen Rothwell @ 2011-03-03 2:40 UTC (permalink / raw)
To: David Miller, netdev
Cc: linux-next, linux-kernel, Dmitry Kravkov, Yaniv Rosner,
Eilon Greenstein
[-- Attachment #1: Type: text/plain, Size: 517 bytes --]
Hi all,
Today's linux-next merge of the net tree got a conflict in
drivers/net/bnx2x/bnx2x.h between commit
b746f7e52fe33ce66ea0cf6127838eff507839ff ("bnx2x: update driver version
to 1.62.00-6") from the net-current tree and commit
6b28ff3be829a851378551245fd6b3f9bf93b0ad ("bnx2x: Update bnx2x version to
1.62.11-0") from the net tree.
Obvious, really :-) I used the version from the net tree.
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/
[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]
^ permalink raw reply
* linux-next: manual merge of the net tree with the acpi tree
From: Stephen Rothwell @ 2011-03-03 2:44 UTC (permalink / raw)
To: David Miller, netdev
Cc: linux-next, linux-kernel, Huang Ying, Len Brown, Ben Hutchings
Hi all,
Today's linux-next merge of the net tree got a conflict in lib/Makefile
between commit 248852b614cb7713552188a9c2552ba774945733 ("lib, Add
lock-less NULL terminated single list") from the acpi tree and commitc
39649c331c70952700f99832b03f87e9d7f5b4b ("lib: cpu_rmap: CPU affinity
reverse-mapping") from the net tree.
Just overlapping additions. I fixed it up (see below) and can carry the
fix as necessary.
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
diff --cc lib/Makefile
index 7cb0608,b73ba01..0000000
--- a/lib/Makefile
+++ b/lib/Makefile
@@@ -110,8 -110,8 +110,10 @@@ obj-$(CONFIG_ATOMIC64_SELFTEST) += atom
obj-$(CONFIG_AVERAGE) += average.o
+obj-$(CONFIG_LLIST) += llist.o
+
+ obj-$(CONFIG_CPU_RMAP) += cpu_rmap.o
+
hostprogs-y := gen_crc32table
clean-files := crc32table.h
^ permalink raw reply
* [PATCH] mv643xx_eth: Use netdev_<level> and pr_<level>
From: Joe Perches @ 2011-03-03 3:50 UTC (permalink / raw)
To: Lennert Buytenhek; +Cc: netdev
Use the current logging styles.
Signed-off-by: Joe Perches <joe@perches.com>
---
Hey Lennart.
I don't know if this module is still being maintained.
If so, here's a conversion to the current logging styles.
drivers/net/mv643xx_eth.c | 74 ++++++++++++++++++++-------------------------
1 files changed, 33 insertions(+), 41 deletions(-)
diff --git a/drivers/net/mv643xx_eth.c b/drivers/net/mv643xx_eth.c
index 02076e1..34425b9 100644
--- a/drivers/net/mv643xx_eth.c
+++ b/drivers/net/mv643xx_eth.c
@@ -35,6 +35,8 @@
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
#include <linux/init.h>
#include <linux/dma-mapping.h>
#include <linux/in.h>
@@ -627,9 +629,8 @@ err:
if ((cmd_sts & (RX_FIRST_DESC | RX_LAST_DESC)) !=
(RX_FIRST_DESC | RX_LAST_DESC)) {
if (net_ratelimit())
- dev_printk(KERN_ERR, &mp->dev->dev,
- "received packet spanning "
- "multiple descriptors\n");
+ netdev_err(mp->dev,
+ "received packet spanning multiple descriptors\n");
}
if (cmd_sts & ERROR_SUMMARY)
@@ -868,15 +869,14 @@ static netdev_tx_t mv643xx_eth_xmit(struct sk_buff *skb, struct net_device *dev)
if (has_tiny_unaligned_frags(skb) && __skb_linearize(skb)) {
txq->tx_dropped++;
- dev_printk(KERN_DEBUG, &dev->dev,
- "failed to linearize skb with tiny "
- "unaligned fragment\n");
+ netdev_printk(KERN_DEBUG, dev,
+ "failed to linearize skb with tiny unaligned fragment\n");
return NETDEV_TX_BUSY;
}
if (txq->tx_ring_size - txq->tx_desc_count < MAX_SKB_FRAGS + 1) {
if (net_ratelimit())
- dev_printk(KERN_ERR, &dev->dev, "tx queue full?!\n");
+ netdev_err(dev, "tx queue full?!\n");
kfree_skb(skb);
return NETDEV_TX_OK;
}
@@ -959,7 +959,7 @@ static int txq_reclaim(struct tx_queue *txq, int budget, int force)
skb = __skb_dequeue(&txq->tx_skb);
if (cmd_sts & ERROR_SUMMARY) {
- dev_printk(KERN_INFO, &mp->dev->dev, "tx error\n");
+ netdev_info(mp->dev, "tx error\n");
mp->dev->stats.tx_errors++;
}
@@ -1122,20 +1122,20 @@ static int smi_bus_read(struct mii_bus *bus, int addr, int reg)
int ret;
if (smi_wait_ready(msp)) {
- printk(KERN_WARNING "mv643xx_eth: SMI bus busy timeout\n");
+ pr_warn("SMI bus busy timeout\n");
return -ETIMEDOUT;
}
writel(SMI_OPCODE_READ | (reg << 21) | (addr << 16), smi_reg);
if (smi_wait_ready(msp)) {
- printk(KERN_WARNING "mv643xx_eth: SMI bus busy timeout\n");
+ pr_warn("SMI bus busy timeout\n");
return -ETIMEDOUT;
}
ret = readl(smi_reg);
if (!(ret & SMI_READ_VALID)) {
- printk(KERN_WARNING "mv643xx_eth: SMI bus read not valid\n");
+ pr_warn("SMI bus read not valid\n");
return -ENODEV;
}
@@ -1148,7 +1148,7 @@ static int smi_bus_write(struct mii_bus *bus, int addr, int reg, u16 val)
void __iomem *smi_reg = msp->base + SMI_REG;
if (smi_wait_ready(msp)) {
- printk(KERN_WARNING "mv643xx_eth: SMI bus busy timeout\n");
+ pr_warn("SMI bus busy timeout\n");
return -ETIMEDOUT;
}
@@ -1156,7 +1156,7 @@ static int smi_bus_write(struct mii_bus *bus, int addr, int reg, u16 val)
(addr << 16) | (val & 0xffff), smi_reg);
if (smi_wait_ready(msp)) {
- printk(KERN_WARNING "mv643xx_eth: SMI bus busy timeout\n");
+ pr_warn("SMI bus busy timeout\n");
return -ETIMEDOUT;
}
@@ -1566,9 +1566,8 @@ mv643xx_eth_set_ringparam(struct net_device *dev, struct ethtool_ringparam *er)
if (netif_running(dev)) {
mv643xx_eth_stop(dev);
if (mv643xx_eth_open(dev)) {
- dev_printk(KERN_ERR, &dev->dev,
- "fatal error on re-opening device after "
- "ring param change\n");
+ netdev_err(dev,
+ "fatal error on re-opening device after ring param change\n");
return -ENOMEM;
}
}
@@ -1874,7 +1873,7 @@ static int rxq_init(struct mv643xx_eth_private *mp, int index)
}
if (rxq->rx_desc_area == NULL) {
- dev_printk(KERN_ERR, &mp->dev->dev,
+ netdev_err(mp->dev,
"can't allocate rx ring (%d bytes)\n", size);
goto out;
}
@@ -1884,8 +1883,7 @@ static int rxq_init(struct mv643xx_eth_private *mp, int index)
rxq->rx_skb = kmalloc(rxq->rx_ring_size * sizeof(*rxq->rx_skb),
GFP_KERNEL);
if (rxq->rx_skb == NULL) {
- dev_printk(KERN_ERR, &mp->dev->dev,
- "can't allocate rx skb ring\n");
+ netdev_err(mp->dev, "can't allocate rx skb ring\n");
goto out_free;
}
@@ -1944,8 +1942,7 @@ static void rxq_deinit(struct rx_queue *rxq)
}
if (rxq->rx_desc_count) {
- dev_printk(KERN_ERR, &mp->dev->dev,
- "error freeing rx ring -- %d skbs stuck\n",
+ netdev_err(mp->dev, "error freeing rx ring -- %d skbs stuck\n",
rxq->rx_desc_count);
}
@@ -1987,7 +1984,7 @@ static int txq_init(struct mv643xx_eth_private *mp, int index)
}
if (txq->tx_desc_area == NULL) {
- dev_printk(KERN_ERR, &mp->dev->dev,
+ netdev_err(mp->dev,
"can't allocate tx ring (%d bytes)\n", size);
return -ENOMEM;
}
@@ -2093,7 +2090,7 @@ static void handle_link_event(struct mv643xx_eth_private *mp)
if (netif_carrier_ok(dev)) {
int i;
- printk(KERN_INFO "%s: link down\n", dev->name);
+ netdev_info(dev, "link down\n");
netif_carrier_off(dev);
@@ -2124,10 +2121,8 @@ static void handle_link_event(struct mv643xx_eth_private *mp)
duplex = (port_status & FULL_DUPLEX) ? 1 : 0;
fc = (port_status & FLOW_CONTROL_ENABLED) ? 1 : 0;
- printk(KERN_INFO "%s: link up, %d Mb/s, %s duplex, "
- "flow control %sabled\n", dev->name,
- speed, duplex ? "full" : "half",
- fc ? "en" : "dis");
+ netdev_info(dev, "link up, %d Mb/s, %s duplex, flow control %sabled\n",
+ speed, duplex ? "full" : "half", fc ? "en" : "dis");
if (!netif_carrier_ok(dev))
netif_carrier_on(dev);
@@ -2337,7 +2332,7 @@ static int mv643xx_eth_open(struct net_device *dev)
err = request_irq(dev->irq, mv643xx_eth_irq,
IRQF_SHARED, dev->name, dev);
if (err) {
- dev_printk(KERN_ERR, &dev->dev, "can't assign irq\n");
+ netdev_err(dev, "can't assign irq\n");
return -EAGAIN;
}
@@ -2483,9 +2478,8 @@ static int mv643xx_eth_change_mtu(struct net_device *dev, int new_mtu)
*/
mv643xx_eth_stop(dev);
if (mv643xx_eth_open(dev)) {
- dev_printk(KERN_ERR, &dev->dev,
- "fatal error on re-opening device after "
- "MTU change\n");
+ netdev_err(dev,
+ "fatal error on re-opening device after MTU change\n");
}
return 0;
@@ -2508,7 +2502,7 @@ static void mv643xx_eth_tx_timeout(struct net_device *dev)
{
struct mv643xx_eth_private *mp = netdev_priv(dev);
- dev_printk(KERN_INFO, &dev->dev, "tx timeout\n");
+ netdev_info(dev, "tx timeout\n");
schedule_work(&mp->tx_timeout_task);
}
@@ -2603,8 +2597,8 @@ static int mv643xx_eth_shared_probe(struct platform_device *pdev)
int ret;
if (!mv643xx_eth_version_printed++)
- printk(KERN_NOTICE "MV-643xx 10/100/1000 ethernet "
- "driver version %s\n", mv643xx_eth_driver_version);
+ pr_notice("MV-643xx 10/100/1000 ethernet driver version %s\n",
+ mv643xx_eth_driver_version);
ret = -EINVAL;
res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
@@ -2871,14 +2865,12 @@ static int mv643xx_eth_probe(struct platform_device *pdev)
pd = pdev->dev.platform_data;
if (pd == NULL) {
- dev_printk(KERN_ERR, &pdev->dev,
- "no mv643xx_eth_platform_data\n");
+ dev_err(&pdev->dev, "no mv643xx_eth_platform_data\n");
return -ENODEV;
}
if (pd->shared == NULL) {
- dev_printk(KERN_ERR, &pdev->dev,
- "no mv643xx_eth_platform_data->shared\n");
+ dev_err(&pdev->dev, "no mv643xx_eth_platform_data->shared\n");
return -ENODEV;
}
@@ -2957,11 +2949,11 @@ static int mv643xx_eth_probe(struct platform_device *pdev)
if (err)
goto out;
- dev_printk(KERN_NOTICE, &dev->dev, "port %d with MAC address %pM\n",
- mp->port_num, dev->dev_addr);
+ netdev_notice(dev, "port %d with MAC address %pM\n",
+ mp->port_num, dev->dev_addr);
if (mp->tx_desc_sram_size > 0)
- dev_printk(KERN_NOTICE, &dev->dev, "configured with sram\n");
+ netdev_notice(dev, "configured with sram\n");
return 0;
^ permalink raw reply related
* [PATCH] sctp: do not mark chunk abandoned if peer has no PRSCTP capable
From: Wei Yongjun @ 2011-03-03 4:20 UTC (permalink / raw)
To: netdev@vger.kernel.org, lksctp, David Miller, Vlad Yasevich
Chunk is marked abandoned if the chunk is expires, and it not be
retransmited even if the peer has no PRSCTP capable, but the peer
will still wait for retransmit it to update CTSN.
This patch disable mark chunk abandoned if peer has no PRSCTP
capable.
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
---
net/sctp/chunk.c | 3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/net/sctp/chunk.c b/net/sctp/chunk.c
index 6c85564..0d4832d 100644
--- a/net/sctp/chunk.c
+++ b/net/sctp/chunk.c
@@ -347,6 +347,9 @@ int sctp_chunk_abandoned(struct sctp_chunk *chunk)
{
struct sctp_datamsg *msg = chunk->msg;
+ if (!chunk->asoc->peer.prsctp_capable)
+ return 0;
+
if (!msg->can_abandon)
return 0;
--
1.6.5.2
^ permalink raw reply related
* [PATCH 2/2] bonding: COW before overwriting the destination MAC address
From: Changli Gao @ 2011-03-03 4:25 UTC (permalink / raw)
To: Jay Vosburgh; +Cc: David S. Miller, netdev, Changli Gao
When there is a ptype handler holding a clone of this skb, whose
destination MAC addresse is overwritten, the owner of this handler may
get a corrupted packet.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
---
drivers/net/bonding/bond_main.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 912b416..211be52 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -1513,6 +1513,11 @@ static struct sk_buff *bond_handle_frame(struct sk_buff *skb)
skb->pkt_type == PACKET_HOST) {
u16 *dest = (u16 *) eth_hdr(skb)->h_dest;
+ if (unlikely(skb_cow_head(skb,
+ skb->data - skb_mac_header(skb)))) {
+ kfree_skb(skb);
+ return NULL;
+ }
memcpy(dest, bond_dev->dev_addr, ETH_ALEN);
}
^ permalink raw reply related
* [PATCH 1/2] bonding: remove the unused dummy functions when net poll controller isn't enabled
From: Changli Gao @ 2011-03-03 4:26 UTC (permalink / raw)
To: Jay Vosburgh; +Cc: David S. Miller, netdev, Changli Gao
These two functions are only used when net poll controller is enabled.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
---
drivers/net/bonding/bond_main.c | 8 --------
1 file changed, 8 deletions(-)
diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 0592e6d..912b416 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -1380,14 +1380,6 @@ static inline void slave_disable_netpoll(struct slave *slave)
static void bond_netpoll_cleanup(struct net_device *bond_dev)
{
}
-static int bond_netpoll_setup(struct net_device *dev, struct netpoll_info *ni)
-{
- return 0;
-}
-static struct netpoll_info *bond_netpoll_info(struct bonding *bond)
-{
- return NULL;
-}
#endif
/*---------------------------------- IOCTL ----------------------------------*/
^ permalink raw reply related
* inetpeer with create==0
From: David Miller @ 2011-03-03 4:45 UTC (permalink / raw)
To: eric.dumazet; +Cc: netdev
Eric, I was profiling the non-routing-cache case and something that stuck
out is the case of calling inet_getpeer() with create==0.
If an entry is not found, we have to redo the lookup under a spinlock
to make certain that a concurrent writer rebalancing the tree does
not "hide" an existing entry from us.
This makes the case of a create==0 lookup for a not-present entry
really expensive. It is on the order of 600 cpu cycles on my
Niagara2.
I added a hack to not do the relookup under the lock when create==0
and it now costs less than 300 cycles.
This is now a pretty common operation with the way we handle COW'd
metrics, so I think it's definitely worth optimizing.
I looked at the generic radix tree implementation, and it supports
full RCU lookups in parallel with insert/delete. It handles the race
case without the relookup under lock because it creates fixed paths
to "slots" where nodes live using shifts and masks. So if a path
to a slot ever existed, it will always exist.
Take a look at lib/radix-tree.c and include/linux/radix-tree.h if
you are curious.
I think we should do something similar for inetpeer. Currently we
cannot just use the existing generic radix-tree code because it only
supports indexes as large as "unsigned long" and we need to handle
128-bit ipv6 addresses.
^ permalink raw reply
* Re: [PATCH 1/5] tg3: use usleep_range not msleep for small sleeps
From: David Miller @ 2011-03-03 4:55 UTC (permalink / raw)
To: martinez.javier; +Cc: mcarlson, mchan, grant.likely, netdev, kernelnewbies
In-Reply-To: <1298999069-12740-2-git-send-email-martinez.javier@gmail.com>
From: Javier Martinez Canillas <martinez.javier@gmail.com>
Date: Tue, 1 Mar 2011 18:04:26 +0100
>
> Signed-off-by: Javier Martinez Canillas <martinez.javier@gmail.com>
Why?
^ permalink raw reply
* Re: [PATCH 3/5] tg3: Enclose macro with complex values in parenthesis
From: David Miller @ 2011-03-03 4:56 UTC (permalink / raw)
To: martinez.javier; +Cc: mcarlson, mchan, grant.likely, netdev, kernelnewbies
In-Reply-To: <1298999069-12740-3-git-send-email-martinez.javier@gmail.com>
From: Javier Martinez Canillas <martinez.javier@gmail.com>
Date: Tue, 1 Mar 2011 18:04:27 +0100
> Macros with complex values should be enclosed in parenthesis.
> Change accordingly.
>
> Signed-off-by: Javier Martinez Canillas <martinez.javier@gmail.com>
This seems unnecessary, what problem do the current macro definitions
create?
^ permalink raw reply
* Re: inetpeer with create==0
From: Changli Gao @ 2011-03-03 5:30 UTC (permalink / raw)
To: David Miller; +Cc: eric.dumazet, netdev
In-Reply-To: <20110302.204545.193730647.davem@davemloft.net>
On Thu, Mar 3, 2011 at 12:45 PM, David Miller <davem@davemloft.net> wrote:
>
>
> I looked at the generic radix tree implementation, and it supports
> full RCU lookups in parallel with insert/delete. It handles the race
> case without the relookup under lock because it creates fixed paths
> to "slots" where nodes live using shifts and masks. So if a path
> to a slot ever existed, it will always exist.
>
> Take a look at lib/radix-tree.c and include/linux/radix-tree.h if
> you are curious.
>
> I think we should do something similar for inetpeer. Currently we
> cannot just use the existing generic radix-tree code because it only
> supports indexes as large as "unsigned long" and we need to handle
> 128-bit ipv6 addresses.
I am just wondering why we need a trie(radix tree) here, we don't have
to do LPM since the lengths of the keys are always fixed(ether 4 or
16). Maybe a hash table is enough and simpler, and it is RCU
compatible.
--
Regards,
Changli Gao(xiaosuo@gmail.com)
^ permalink raw reply
* Re: inetpeer with create==0
From: David Miller @ 2011-03-03 5:36 UTC (permalink / raw)
To: xiaosuo; +Cc: eric.dumazet, netdev
In-Reply-To: <AANLkTimwaUaC7Xu2djSNR9k98sYbSM0V88sCAc34R6Ac@mail.gmail.com>
From: Changli Gao <xiaosuo@gmail.com>
Date: Thu, 3 Mar 2011 13:30:03 +0800
> I am just wondering why we need a trie(radix tree) here, we don't have
> to do LPM since the lengths of the keys are always fixed(ether 4 or
> 16). Maybe a hash table is enough and simpler, and it is RCU
> compatible.
Because trie eliminates all of the issues of having to size a hash
table, dynamically resize it, etc.
Trie gives well bounded performance dependent solely upon size of
the table, rather than access patterns, distribution of keys, and
how perfect hash function is.
We used to use a hash table for the page cache too, but now we use
a trie structure there as well, and that uses full long word sized
keys and the generic raidx-tree code.
Using hash tables is really foolish for potentially large data sets.
^ permalink raw reply
* bonding...
From: David Miller @ 2011-03-03 5:49 UTC (permalink / raw)
To: fubar; +Cc: netdev
Hey, if someone could step up and help with bonding maintainence
in some tangible way, I'd really appreciate it.
Currently the situation is that many people work on bonding patches,
and whilst I do try and wait for some ACKs to arrive, I am the person
who has to sort out when changes are ready, decide to apply them, and
poke for review when things fall through the cracks.
Sometimes patches go for weeks without ACKs, and in that situation
I have to either try to understand the changes myself, or wait
potentially forever for someone with bonding knowledge to take a
good look at the patch and properly review it.
It was nearly 2 weeks before Oleg V. Ukhno's 802.3ad round-robin patch
got looked at by anyone with bonding knowledge. And it only happened
because I got tired of seeing his poor patch rot in patchwork
and had to explicitly asked for review the other day.
This is unacceptable, people are submitting multiple bonding patches
every single day now. It needs a clueful bonding person looking at
these submissions on a constant basis.
This is a serious problem and is backlogging the netdev patch queue.
So if someone would become an active bonding patch-accumulator, and
send me sets of patches that are ready to apply, I would really
appreciate it.
Thanks.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox