* [PATCH 2/6] forcedeth: new ethtool stat "tx_timeout" to account for tx_timeouts
From: David Decotigny @ 2011-05-19 0:14 UTC (permalink / raw)
To: David S. Miller, Joe Perches, Szymon Janc, netdev, linux-kernel
Cc: kernel-net-upstream, Sameer Nanda, David Decotigny
In-Reply-To: <1305764080-24853-1-git-send-email-decot@google.com>
From: Sameer Nanda <snanda@google.com>
This change publishes a new ethtool stats: tx_timeout that counts the
number of times the tx_timeout callback was triggered.
Signed-off-by: David Decotigny <decot@google.com>
---
drivers/net/forcedeth.c | 4 ++++
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/drivers/net/forcedeth.c b/drivers/net/forcedeth.c
index 895471d..112dc0b 100644
--- a/drivers/net/forcedeth.c
+++ b/drivers/net/forcedeth.c
@@ -632,6 +632,7 @@ static const struct nv_ethtool_str nv_estats_str[] = {
{ "rx_packets" },
{ "rx_errors_total" },
{ "tx_errors_total" },
+ { "tx_timeout" },
/* version 2 stats */
{ "tx_deferral" },
@@ -672,6 +673,7 @@ struct nv_ethtool_stats {
u64 rx_packets;
u64 rx_errors_total;
u64 tx_errors_total;
+ u64 tx_timeout;
/* version 2 stats */
u64 tx_deferral;
@@ -2526,6 +2528,8 @@ static void nv_tx_timeout(struct net_device *dev)
spin_lock_irq(&np->lock);
+ np->estats.tx_timeout++;
+
/* 1) stop tx engine */
nv_stop_tx(dev);
--
1.7.3.1
^ permalink raw reply related
* [PATCH 1/6] forcedeth: Improve stats counters
From: David Decotigny @ 2011-05-19 0:14 UTC (permalink / raw)
To: David S. Miller, Joe Perches, Szymon Janc, netdev, linux-kernel
Cc: kernel-net-upstream, Mandeep Baines, David Decotigny
From: Mandeep Baines <msb@google.com>
Rx byte count was off; instead use the hardware's count. Tx packet
count was counting pre-TSO packets; instead count on-the-wire packets.
Report hardware dropped frame count as rx_fifo_errors.
- The count of transmitted packets reported by the forcedeth driver
reports pre-TSO (TCP Segmentation Offload) packet counts and not the
count of the number of packets sent on the wire. This change fixes
the forcedeth driver to report the correct count. Fixed the code by
copying the count stored in the NIC H/W to the value reported by the
driver.
- Count rx_drop_frame errors as rx_fifo_errors:
We see a lot of rx_drop_frame errors if we disable the rx bottom-halves
for too long. Normally, rx_fifo_errors would be counted in this case.
The rx_drop_frame error count is private to forcedeth and is not
reported by ifconfig or sysfs. The rx_fifo_errors count is currently
unused in the forcedeth driver. It is reported by ifconfig as overruns.
This change reports rx_drop_frame errors as rx_fifo_errors.
Signed-off-by: David Decotigny <decot@google.com>
---
drivers/net/forcedeth.c | 4 ++++
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/drivers/net/forcedeth.c b/drivers/net/forcedeth.c
index d09e8b0..895471d 100644
--- a/drivers/net/forcedeth.c
+++ b/drivers/net/forcedeth.c
@@ -1684,6 +1684,7 @@ static void nv_get_hw_stats(struct net_device *dev)
np->estats.tx_pause += readl(base + NvRegTxPause);
np->estats.rx_pause += readl(base + NvRegRxPause);
np->estats.rx_drop_frame += readl(base + NvRegRxDropFrame);
+ np->estats.rx_errors_total += np->estats.rx_drop_frame;
}
if (np->driver_data & DEV_HAS_STATISTICS_V3) {
@@ -1708,11 +1709,14 @@ static struct net_device_stats *nv_get_stats(struct net_device *dev)
nv_get_hw_stats(dev);
/* copy to net_device stats */
+ dev->stats.tx_packets = np->estats.tx_packets;
+ dev->stats.rx_bytes = np->estats.rx_bytes;
dev->stats.tx_bytes = np->estats.tx_bytes;
dev->stats.tx_fifo_errors = np->estats.tx_fifo_errors;
dev->stats.tx_carrier_errors = np->estats.tx_carrier_errors;
dev->stats.rx_crc_errors = np->estats.rx_crc_errors;
dev->stats.rx_over_errors = np->estats.rx_over_errors;
+ dev->stats.rx_fifo_errors = np->estats.rx_drop_frame;
dev->stats.rx_errors = np->estats.rx_errors_total;
dev->stats.tx_errors = np->estats.tx_errors_total;
}
--
1.7.3.1
^ permalink raw reply related
* Re: [PATCH net-next] bridge: add notification over netlink when STP changes state
From: Stephen Hemminger @ 2011-05-18 23:48 UTC (permalink / raw)
To: Ben Greear; +Cc: David Miller, bridge, netdev
In-Reply-To: <4DD3F168.1060308@candelatech.com>
I updated rstp daemon to handle hearing itself (and some other issues).
Not sure how many users there are of that code.
See:
git://git.kernel.org/pub/scm/linux/kernel/git/shemminger/rstp.git
^ permalink raw reply
* Re: [PATCH 1/2] forcedeth: make module parameters readable in /sys/module
From: David Decotigny @ 2011-05-18 23:21 UTC (permalink / raw)
To: Stephen Hemminger
Cc: David S. Miller, Joe Perches, Szymon Janc, netdev, linux-kernel,
kernel-net-upstream
In-Reply-To: <20110518150346.508d6406@nehalam>
Hi Stephen,
On Wed, May 18, 2011 at 3:03 PM, Stephen Hemminger
<shemminger@vyatta.com> wrote:
> Although this makes more info for developer, it also means more
> stuff in sysfs taking more memory and not providing any real value
Right. I'll drop this patch for now.
I do agree that this creates unnecessary pressure on some systems. But
on other systems, it is useful to know how all the loaded modules were
actually configured to track down something weird.
Regarding /sys/modules/X/parameters, the only knobs I know of are
CONFIG_SYSFS and the last argument to module_param(). If this is
correct, then it seems tricky to accommodate for both use cases above
without being crude (eg. disable sysfs, edit/sed the sources to change
last argument of module_param() macro, ...). Please correct me if I am
wrong. But if I am not, how about having a new knob allowing to
precisely control the presence/absence of /sys/module (another
CONFIG_*)? That way, the last argument to module_param() can be
interpreted as the permission when /sys/module is supported, and would
be ignored otherwise: module authors could gradually change their
perm=S_IRUGO and family without caring much about memory footprint.
And both use cases above could be supported without having to disable
sysfs altogether or change the sources.
I would be happy to work on this if it makes any sense.
Regards,
^ permalink raw reply
* [PATCH] drivers/net: ks8842 Fix crash on received packet when in PIO mode.
From: Dennis Aberilla @ 2011-05-18 22:59 UTC (permalink / raw)
To: info; +Cc: davem, netdev
This patch fixes a kernel crash during packet reception due to not enough allocated bytes for the skb. This applies to the driver when running in PIO mode in an ISA bus setup.
Signed-off-by: Dennis Aberilla <denzzzhome@yahoo.com>
---
diff --git a/drivers/net/ks8842.c b/drivers/net/ks8842.c
index f0d8346..9069aee 100644
--- a/drivers/net/ks8842.c
+++ b/drivers/net/ks8842.c
@@ -662,7 +662,7 @@ static void ks8842_rx_frame(struct net_device *netdev,
/* check the status */
if ((status & RXSR_VALID) && !(status & RXSR_ERROR)) {
- struct sk_buff *skb = netdev_alloc_skb_ip_align(netdev, len);
+ struct sk_buff *skb = netdev_alloc_skb_ip_align(netdev, len + 2);
if (skb) {
--
Using mutt this time.
|Dennis
=======================================================================
This email, including any attachments, is only for the intended
addressee. It is subject to copyright, is confidential and may be
the subject of legal or other privilege, none of which is waived or
lost by reason of this transmission.
If the receiver is not the intended addressee, please accept our
apologies, notify us by return, delete all copies and perform no
other act on the email.
Unfortunately, we cannot warrant that the email has not been
altered or corrupted during transmission.
=======================================================================
^ permalink raw reply related
* [PATCH 4/4] ipv4: Pass explicit destination address to rt_bind_peer().
From: David Miller @ 2011-05-18 22:56 UTC (permalink / raw)
To: netdev
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/route.h | 4 ++--
net/ipv4/icmp.c | 8 ++++----
net/ipv4/route.c | 16 ++++++++--------
net/ipv4/tcp_ipv4.c | 2 +-
4 files changed, 15 insertions(+), 15 deletions(-)
diff --git a/include/net/route.h b/include/net/route.h
index 600d0f2..db7b343 100644
--- a/include/net/route.h
+++ b/include/net/route.h
@@ -290,14 +290,14 @@ static inline struct rtable *ip_route_newports(struct flowi4 *fl4, struct rtable
return rt;
}
-extern void rt_bind_peer(struct rtable *rt, int create);
+extern void rt_bind_peer(struct rtable *rt, __be32 daddr, int create);
static inline struct inet_peer *rt_get_peer(struct rtable *rt, __be32 daddr)
{
if (rt->peer)
return rt->peer;
- rt_bind_peer(rt, 0);
+ rt_bind_peer(rt, daddr, 0);
return rt->peer;
}
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index 3f47585..5395e45 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -234,7 +234,7 @@ static inline void icmp_xmit_unlock(struct sock *sk)
*/
static inline bool icmpv4_xrlim_allow(struct net *net, struct rtable *rt,
- int type, int code)
+ struct flowi4 *fl4, int type, int code)
{
struct dst_entry *dst = &rt->dst;
bool rc = true;
@@ -253,7 +253,7 @@ static inline bool icmpv4_xrlim_allow(struct net *net, struct rtable *rt,
/* Limit if icmp type is enabled in ratemask. */
if ((1 << type) & net->ipv4.sysctl_icmp_ratemask) {
if (!rt->peer)
- rt_bind_peer(rt, 1);
+ rt_bind_peer(rt, fl4->daddr, 1);
rc = inet_peer_xrlim_allow(rt->peer,
net->ipv4.sysctl_icmp_ratelimit);
}
@@ -363,7 +363,7 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
rt = ip_route_output_key(net, &fl4);
if (IS_ERR(rt))
goto out_unlock;
- if (icmpv4_xrlim_allow(net, rt, icmp_param->data.icmph.type,
+ if (icmpv4_xrlim_allow(net, rt, &fl4, icmp_param->data.icmph.type,
icmp_param->data.icmph.code))
icmp_push_reply(icmp_param, &fl4, &ipc, &rt);
ip_rt_put(rt);
@@ -603,7 +603,7 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
if (IS_ERR(rt))
goto out_unlock;
- if (!icmpv4_xrlim_allow(net, rt, type, code))
+ if (!icmpv4_xrlim_allow(net, rt, &fl4, type, code))
goto ende;
/* RFC says return as much as we can without exceeding 576 bytes. */
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 9c5ad86..b24d58e 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -156,7 +156,7 @@ static u32 *ipv4_cow_metrics(struct dst_entry *dst, unsigned long old)
u32 *p = NULL;
if (!rt->peer)
- rt_bind_peer(rt, 1);
+ rt_bind_peer(rt, rt->rt_dst, 1);
peer = rt->peer;
if (peer) {
@@ -1193,11 +1193,11 @@ static u32 rt_peer_genid(void)
return atomic_read(&__rt_peer_genid);
}
-void rt_bind_peer(struct rtable *rt, int create)
+void rt_bind_peer(struct rtable *rt, __be32 daddr, int create)
{
struct inet_peer *peer;
- peer = inet_getpeer_v4(rt->rt_dst, create);
+ peer = inet_getpeer_v4(daddr, create);
if (peer && cmpxchg(&rt->peer, NULL, peer) != NULL)
inet_putpeer(peer);
@@ -1231,7 +1231,7 @@ void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more)
if (rt) {
if (rt->peer == NULL)
- rt_bind_peer(rt, 1);
+ rt_bind_peer(rt, rt->rt_dst, 1);
/* If peer is attached to destination, it is never detached,
so that we need not to grab a lock to dereference it.
@@ -1377,7 +1377,7 @@ void ip_rt_send_redirect(struct sk_buff *skb)
rcu_read_unlock();
if (!rt->peer)
- rt_bind_peer(rt, 1);
+ rt_bind_peer(rt, rt->rt_dst, 1);
peer = rt->peer;
if (!peer) {
icmp_send(skb, ICMP_REDIRECT, ICMP_REDIR_HOST, rt->rt_gateway);
@@ -1445,7 +1445,7 @@ static int ip_error(struct sk_buff *skb)
}
if (!rt->peer)
- rt_bind_peer(rt, 1);
+ rt_bind_peer(rt, rt->rt_dst, 1);
peer = rt->peer;
send = true;
@@ -1552,7 +1552,7 @@ static void ip_rt_update_pmtu(struct dst_entry *dst, u32 mtu)
dst_confirm(dst);
if (!rt->peer)
- rt_bind_peer(rt, 1);
+ rt_bind_peer(rt, rt->rt_dst, 1);
peer = rt->peer;
if (peer) {
if (mtu < ip_rt_min_pmtu)
@@ -1609,7 +1609,7 @@ static struct dst_entry *ipv4_dst_check(struct dst_entry *dst, u32 cookie)
struct inet_peer *peer;
if (!rt->peer)
- rt_bind_peer(rt, 0);
+ rt_bind_peer(rt, rt->rt_dst, 0);
peer = rt->peer;
if (peer && peer->pmtu_expires)
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 1d290ac..3c8d9b6 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -1781,7 +1781,7 @@ struct inet_peer *tcp_v4_get_peer(struct sock *sk, bool *release_it)
*release_it = true;
} else {
if (!rt->peer)
- rt_bind_peer(rt, 1);
+ rt_bind_peer(rt, inet->inet_daddr, 1);
peer = rt->peer;
*release_it = false;
}
--
1.7.4.4
^ permalink raw reply related
* [PATCH 3/4] ipv4: Pass explicit destination address to rt_get_peer().
From: David Miller @ 2011-05-18 22:56 UTC (permalink / raw)
To: netdev
This will next trickle down to rt_bind_peer().
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/route.h | 2 +-
net/ipv4/tcp_ipv4.c | 6 +++---
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/include/net/route.h b/include/net/route.h
index ba0e084..600d0f2 100644
--- a/include/net/route.h
+++ b/include/net/route.h
@@ -292,7 +292,7 @@ static inline struct rtable *ip_route_newports(struct flowi4 *fl4, struct rtable
extern void rt_bind_peer(struct rtable *rt, int create);
-static inline struct inet_peer *rt_get_peer(struct rtable *rt)
+static inline struct inet_peer *rt_get_peer(struct rtable *rt, __be32 daddr)
{
if (rt->peer)
return rt->peer;
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 7e0bc60..1d290ac 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -206,7 +206,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
if (tcp_death_row.sysctl_tw_recycle &&
!tp->rx_opt.ts_recent_stamp && fl4->daddr == daddr) {
- struct inet_peer *peer = rt_get_peer(rt);
+ struct inet_peer *peer = rt_get_peer(rt, fl4->daddr);
/*
* VJ's idea. We save last timestamp seen from
* the destination in peer table, when entering state
@@ -1353,8 +1353,8 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
if (tmp_opt.saw_tstamp &&
tcp_death_row.sysctl_tw_recycle &&
(dst = inet_csk_route_req(sk, &fl4, req)) != NULL &&
- (peer = rt_get_peer((struct rtable *)dst)) != NULL &&
- peer->daddr.addr.a4 == saddr) {
+ fl4.daddr == saddr &&
+ (peer = rt_get_peer((struct rtable *)dst, fl4.daddr)) != NULL) {
inet_peer_refcheck(peer);
if ((u32)get_seconds() - peer->tcp_ts_stamp < TCP_PAWS_MSL &&
(s32)(peer->tcp_ts - req->ts_recent) >
--
1.7.4.4
^ permalink raw reply related
* [PATCH 2/4] ipv4: Make caller provide flowi4 key to inet_csk_route_req().
From: David Miller @ 2011-05-18 22:56 UTC (permalink / raw)
To: netdev
This way the caller can get at the fully resolved fl4->{daddr,saddr}
etc.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/inet_connection_sock.h | 1 +
net/dccp/ipv4.c | 3 ++-
net/ipv4/inet_connection_sock.c | 10 +++++-----
net/ipv4/tcp_ipv4.c | 6 ++++--
4 files changed, 12 insertions(+), 8 deletions(-)
diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h
index 96546ca..e6db62e 100644
--- a/include/net/inet_connection_sock.h
+++ b/include/net/inet_connection_sock.h
@@ -249,6 +249,7 @@ extern int inet_csk_bind_conflict(const struct sock *sk,
extern int inet_csk_get_port(struct sock *sk, unsigned short snum);
extern struct dst_entry* inet_csk_route_req(struct sock *sk,
+ struct flowi4 *fl4,
const struct request_sock *req);
extern struct dst_entry* inet_csk_route_child_sock(struct sock *sk,
struct sock *newsk,
diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c
index 46b15e9..8c36adf 100644
--- a/net/dccp/ipv4.c
+++ b/net/dccp/ipv4.c
@@ -497,8 +497,9 @@ static int dccp_v4_send_response(struct sock *sk, struct request_sock *req,
int err = -1;
struct sk_buff *skb;
struct dst_entry *dst;
+ struct flowi4 fl4;
- dst = inet_csk_route_req(sk, req);
+ dst = inet_csk_route_req(sk, &fl4, req);
if (dst == NULL)
goto out;
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 3a2ba56..61fac4c 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -350,24 +350,24 @@ void inet_csk_reset_keepalive_timer(struct sock *sk, unsigned long len)
EXPORT_SYMBOL(inet_csk_reset_keepalive_timer);
struct dst_entry *inet_csk_route_req(struct sock *sk,
+ struct flowi4 *fl4,
const struct request_sock *req)
{
struct rtable *rt;
const struct inet_request_sock *ireq = inet_rsk(req);
struct ip_options_rcu *opt = inet_rsk(req)->opt;
struct net *net = sock_net(sk);
- struct flowi4 fl4;
- flowi4_init_output(&fl4, sk->sk_bound_dev_if, sk->sk_mark,
+ flowi4_init_output(fl4, sk->sk_bound_dev_if, sk->sk_mark,
RT_CONN_FLAGS(sk), RT_SCOPE_UNIVERSE,
sk->sk_protocol, inet_sk_flowi_flags(sk),
(opt && opt->opt.srr) ? opt->opt.faddr : ireq->rmt_addr,
ireq->loc_addr, ireq->rmt_port, inet_sk(sk)->inet_sport);
- security_req_classify_flow(req, flowi4_to_flowi(&fl4));
- rt = ip_route_output_flow(net, &fl4, sk);
+ security_req_classify_flow(req, flowi4_to_flowi(fl4));
+ rt = ip_route_output_flow(net, fl4, sk);
if (IS_ERR(rt))
goto no_route;
- if (opt && opt->opt.is_strictroute && fl4.daddr != rt->rt_gateway)
+ if (opt && opt->opt.is_strictroute && fl4->daddr != rt->rt_gateway)
goto route_err;
return &rt->dst;
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index f67fb34..7e0bc60 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -769,11 +769,12 @@ static int tcp_v4_send_synack(struct sock *sk, struct dst_entry *dst,
struct request_values *rvp)
{
const struct inet_request_sock *ireq = inet_rsk(req);
+ struct flowi4 fl4;
int err = -1;
struct sk_buff * skb;
/* First, grab a route. */
- if (!dst && (dst = inet_csk_route_req(sk, req)) == NULL)
+ if (!dst && (dst = inet_csk_route_req(sk, &fl4, req)) == NULL)
return -1;
skb = tcp_make_synack(sk, dst, req, rvp);
@@ -1338,6 +1339,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
req->cookie_ts = tmp_opt.tstamp_ok;
} else if (!isn) {
struct inet_peer *peer = NULL;
+ struct flowi4 fl4;
/* VJ's idea. We save last timestamp seen
* from the destination in peer table, when entering
@@ -1350,7 +1352,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
*/
if (tmp_opt.saw_tstamp &&
tcp_death_row.sysctl_tw_recycle &&
- (dst = inet_csk_route_req(sk, req)) != NULL &&
+ (dst = inet_csk_route_req(sk, &fl4, req)) != NULL &&
(peer = rt_get_peer((struct rtable *)dst)) != NULL &&
peer->daddr.addr.a4 == saddr) {
inet_peer_refcheck(peer);
--
1.7.4.4
^ permalink raw reply related
* [PATCH 1/4] ipv4: Kill RT_CACHE_DEBUG
From: David Miller @ 2011-05-18 22:56 UTC (permalink / raw)
To: netdev
It's way past it's usefulness. And this gets rid of a bunch
of stray ->rt_{dst,src} references.
Even the comment documenting the macro was inaccurate (stated
default was 1 when it's 0).
If reintroduced, it should be done properly, with dynamic debug
facilities.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/dst.h | 7 -------
net/core/dst.c | 22 ----------------------
net/ipv4/route.c | 22 ----------------------
3 files changed, 0 insertions(+), 51 deletions(-)
diff --git a/include/net/dst.h b/include/net/dst.h
index 2588a9a..07a0402 100644
--- a/include/net/dst.h
+++ b/include/net/dst.h
@@ -16,13 +16,6 @@
#include <net/neighbour.h>
#include <asm/processor.h>
-/*
- * 0 - no debugging messages
- * 1 - rare events and bugs (default)
- * 2 - trace mode.
- */
-#define RT_CACHE_DEBUG 0
-
#define DST_GC_MIN (HZ/10)
#define DST_GC_INC (HZ/2)
#define DST_GC_MAX (120*HZ)
diff --git a/net/core/dst.c b/net/core/dst.c
index 30f0093..da47a29 100644
--- a/net/core/dst.c
+++ b/net/core/dst.c
@@ -33,9 +33,6 @@
* 3) This list is guarded by a mutex,
* so that the gc_task and dst_dev_event() can be synchronized.
*/
-#if RT_CACHE_DEBUG >= 2
-static atomic_t dst_total = ATOMIC_INIT(0);
-#endif
/*
* We want to keep lock & list close together
@@ -69,10 +66,6 @@ static void dst_gc_task(struct work_struct *work)
unsigned long expires = ~0L;
struct dst_entry *dst, *next, head;
struct dst_entry *last = &head;
-#if RT_CACHE_DEBUG >= 2
- ktime_t time_start = ktime_get();
- struct timespec elapsed;
-#endif
mutex_lock(&dst_gc_mutex);
next = dst_busy_list;
@@ -146,15 +139,6 @@ loop:
spin_unlock_bh(&dst_garbage.lock);
mutex_unlock(&dst_gc_mutex);
-#if RT_CACHE_DEBUG >= 2
- elapsed = ktime_to_timespec(ktime_sub(ktime_get(), time_start));
- printk(KERN_DEBUG "dst_total: %d delayed: %d work_perf: %d"
- " expires: %lu elapsed: %lu us\n",
- atomic_read(&dst_total), delayed, work_performed,
- expires,
- elapsed.tv_sec * USEC_PER_SEC +
- elapsed.tv_nsec / NSEC_PER_USEC);
-#endif
}
int dst_discard(struct sk_buff *skb)
@@ -205,9 +189,6 @@ void *dst_alloc(struct dst_ops *ops, struct net_device *dev,
dst->lastuse = jiffies;
dst->flags = flags;
dst->next = NULL;
-#if RT_CACHE_DEBUG >= 2
- atomic_inc(&dst_total);
-#endif
dst_entries_add(ops, 1);
return dst;
}
@@ -267,9 +248,6 @@ again:
dst->ops->destroy(dst);
if (dst->dev)
dev_put(dst->dev);
-#if RT_CACHE_DEBUG >= 2
- atomic_dec(&dst_total);
-#endif
kmem_cache_free(dst->ops->kmem_cachep, dst);
dst = child;
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index cb93c32..9c5ad86 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -968,10 +968,6 @@ static int rt_garbage_collect(struct dst_ops *ops)
break;
expire >>= 1;
-#if RT_CACHE_DEBUG >= 2
- printk(KERN_DEBUG "expire>> %u %d %d %d\n", expire,
- dst_entries_get_fast(&ipv4_dst_ops), goal, i);
-#endif
if (dst_entries_get_fast(&ipv4_dst_ops) < ip_rt_max_size)
goto out;
@@ -992,10 +988,6 @@ work_done:
dst_entries_get_fast(&ipv4_dst_ops) < ipv4_dst_ops.gc_thresh ||
dst_entries_get_slow(&ipv4_dst_ops) < ipv4_dst_ops.gc_thresh)
expire = ip_rt_gc_timeout;
-#if RT_CACHE_DEBUG >= 2
- printk(KERN_DEBUG "expire++ %u %d %d %d\n", expire,
- dst_entries_get_fast(&ipv4_dst_ops), goal, rover);
-#endif
out: return 0;
}
@@ -1179,16 +1171,6 @@ restart:
rt->dst.rt_next = rt_hash_table[hash].chain;
-#if RT_CACHE_DEBUG >= 2
- if (rt->dst.rt_next) {
- struct rtable *trt;
- printk(KERN_DEBUG "rt_cache @%02x: %pI4",
- hash, &rt->rt_dst);
- for (trt = rt->dst.rt_next; trt; trt = trt->dst.rt_next)
- printk(" . %pI4", &trt->rt_dst);
- printk("\n");
- }
-#endif
/*
* Since lookup is lockfree, we must make sure
* previous writes to rt are committed to memory
@@ -1347,10 +1329,6 @@ static struct dst_entry *ipv4_negative_advice(struct dst_entry *dst)
unsigned hash = rt_hash(rt->rt_key_dst, rt->rt_key_src,
rt->rt_oif,
rt_genid(dev_net(dst->dev)));
-#if RT_CACHE_DEBUG >= 1
- printk(KERN_DEBUG "ipv4_negative_advice: redirect to %pI4/%02x dropped\n",
- &rt->rt_dst, rt->rt_key_tos);
-#endif
rt_del(hash, rt);
ret = NULL;
} else if (rt->peer &&
--
1.7.4.4
^ permalink raw reply related
* [PATCH 0/4] More rt->rt_dst whittling...
From: David Miller @ 2011-05-18 22:56 UTC (permalink / raw)
To: netdev
After this patch series all of the real references to rt->rt_dst that
remain are having to do with inetpeer lookups.
Already, rt->rt_src has no fundamental references any longer, and with
the routing cache it can be safely removed.
The last step of detaching the final rt->rt_dst uses will need a little
bit of work, because we currently don't propagate the originating
lookup flow key into the various dst_ops such as ->cow_metrics,
->default_mtu, etc. and those do PEER lookups which need the destination
address in use.
^ permalink raw reply
* RE: [RFC V4 PATCH] rtnetlink: Compute and store minimum ifinfo dump size
From: Rose, Gregory V @ 2011-05-18 22:49 UTC (permalink / raw)
To: David Miller
Cc: netdev@vger.kernel.org, bhutchings@solarflare.com,
eric.dumazet@gmail.com
In-Reply-To: <20110518.184741.476096648010572360.davem@davemloft.net>
> -----Original Message-----
> From: David Miller [mailto:davem@davemloft.net]
> Sent: Wednesday, May 18, 2011 3:48 PM
> To: Rose, Gregory V
> Cc: netdev@vger.kernel.org; bhutchings@solarflare.com;
> eric.dumazet@gmail.com
> Subject: Re: [RFC V4 PATCH] rtnetlink: Compute and store minimum ifinfo
> dump size
>
> From: "Rose, Gregory V" <gregory.v.rose@intel.com>
> Date: Wed, 18 May 2011 15:35:58 -0700
>
> > It looked to me like rtmsg_ifinfo is called when each device is
> > registered through register_netdevice() and rtmsg_ifinfo in turn
> > calls the if_nlmsg_size function returning the ifinfo dump size for
> > each device. In my testing it looked like the proper maximum size
> > was being set after I loaded the drivers for the SR-IOV capable
> > devices.
>
> Aha, indeed you're right. I missed this part. Thanks for explaining.
>
> So as far as I can tell it should work, I'll let others review it to
> see if they are ok with this approach.
Ok, sure. We'll see if Eric catches anything, he's got an eagle eye.
;^)
- Greg
^ permalink raw reply
* Re: [RFC V4 PATCH] rtnetlink: Compute and store minimum ifinfo dump size
From: David Miller @ 2011-05-18 22:47 UTC (permalink / raw)
To: gregory.v.rose; +Cc: netdev, bhutchings, eric.dumazet
In-Reply-To: <43F901BD926A4E43B106BF17856F0755018E321C21@orsmsx508.amr.corp.intel.com>
From: "Rose, Gregory V" <gregory.v.rose@intel.com>
Date: Wed, 18 May 2011 15:35:58 -0700
> It looked to me like rtmsg_ifinfo is called when each device is
> registered through register_netdevice() and rtmsg_ifinfo in turn
> calls the if_nlmsg_size function returning the ifinfo dump size for
> each device. In my testing it looked like the proper maximum size
> was being set after I loaded the drivers for the SR-IOV capable
> devices.
Aha, indeed you're right. I missed this part. Thanks for explaining.
So as far as I can tell it should work, I'll let others review it to
see if they are ok with this approach.
Thanks.
^ permalink raw reply
* Re: [Patch net-next-2.6] rfkill: remove CONFIG_RFKILL_INPUT
From: Johannes Berg @ 2011-05-18 22:47 UTC (permalink / raw)
To: Américo Wang
Cc: Linux Kernel Network Developers,
linux-wireless-u79uwXL29TY76Z2rM5mHXA, David S. Miller,
Andrew Morton
In-Reply-To: <BANLkTin9EbYhOHiU-EwGDTsXegks_hAT_A-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
On Mon, 2011-05-16 at 18:44 +0800, Américo Wang wrote:
> This is scheduled to be removed in 2.6.33, now we are at 2.6.39-rc2.
> So I think it is safe to remove it completely.
We still don't have urfkill widely deployed, so NAK. 2.6.33 was wishful
thinking.
johannes
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* RE: [RFC V4 PATCH] rtnetlink: Compute and store minimum ifinfo dump size
From: Rose, Gregory V @ 2011-05-18 22:35 UTC (permalink / raw)
To: David Miller
Cc: netdev@vger.kernel.org, bhutchings@solarflare.com,
eric.dumazet@gmail.com
In-Reply-To: <20110518.180913.273729651812642114.davem@davemloft.net>
> -----Original Message-----
> From: David Miller [mailto:davem@davemloft.net]
> Sent: Wednesday, May 18, 2011 3:09 PM
> To: Rose, Gregory V
> Cc: netdev@vger.kernel.org; bhutchings@solarflare.com;
> eric.dumazet@gmail.com
> Subject: Re: [RFC V4 PATCH] rtnetlink: Compute and store minimum ifinfo
> dump size
>
> From: Greg Rose <gregory.v.rose@intel.com>
> Date: Wed, 18 May 2011 08:54:11 -0700
>
> > The message size allocated for rtnl info dumps was limited to a single
> > page. This is not enough for additional interface info available with
> > devices that support SR-IOV. Calculate the amount of data required so
> > the dump can allocate enough data to satisfy the request.
> >
> > V2 added the calcit function to the rtnl_register calls so that
> > dump functions could get the minimum dump allocation size if they
> > needed to.
> >
> > V3 leverages the fact that the netdev register function ends up
> > calling if_nlmsg_size. We collect the minimum dump allocation size
> > there and keep it in a module static variable so that the calcit
> > function doesn't have to search for the device on every info dump.
> >
> > V4 fixes the title.
> >
> > Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
>
> This only works if someone does a plain GETLINK request on the largest
> space requiring device before the dump request happens.
It looked to me like rtmsg_ifinfo is called when each device is registered through register_netdevice() and rtmsg_ifinfo in turn calls the if_nlmsg_size function returning the ifinfo dump size for each device. In my testing it looked like the proper maximum size was being set after I loaded the drivers for the SR-IOV capable devices.
Perhaps I missed something? Is register_netdevice not always called before a device is used?
- Greg
^ permalink raw reply
* Re: [PATCH] Check the value of doi before referencing it
From: Paul Moore @ 2011-05-18 22:32 UTC (permalink / raw)
To: huzaifas; +Cc: David Miller, netdev, kaber, yoshfuji, jmorris, pekkas, kuznet
In-Reply-To: <20110518.010453.1295824792800350412.davem@davemloft.net>
On Wednesday, May 18, 2011 1:04:53 AM David Miller wrote:
> From: Huzaifa Sidhpurwala <huzaifas@redhat.com>
> Date: Wed, 18 May 2011 09:59:40 +0530
>
> > Value of doi is not checked before referencing it.
> > Though this does not cause any null pointer dereference since
> > all the callers of cipso_v4_doi_add check the value of doi
> > before calling the function, but it would be a good programming
> > practice to do so anyways :)
> >
> > Signed-off-by: Huzaifa Sidhpurwala <huzaifas@redhat.com>
>
> I don't think we should fix bugs that do not exist.
I agree with David.
If there were a large number of callers or cipso_v4_doi_add() was a more
general function there might be some merit in performing more sanity checks on
the values passed to the function. However, as it stands, cipso_v4_doi_add()
is a fairly specialized function which is called by a small number of
functions all of which are internal to NetLabel.
--
paul moore
linux @ hp
^ permalink raw reply
* Re: [PATCH 2/2] bna: Add Generic Netlink Interface
From: Debashis Dutt @ 2011-05-18 22:28 UTC (permalink / raw)
To: David Lamparter
Cc: Rasesh Mody, davem, netdev, adapter_linux_open_src_team,
Debashis Dutt
In-Reply-To: <20110518120005.GC3762520@jupiter.n2.diac24.net>
On Wed, May 18, 2011 at 5:00 AM, David Lamparter <equinox@diac24.net> wrote:
> On Tue, May 17, 2011 at 09:57:01PM -0700, Rasesh Mody wrote:
>> This patch adds the generic netlink communication interface to BNA driver. The
>> in-kernel generic netlink infrastructure can be used to collect hardware
>> specific control information and control attributes. The driver makes use of
>> the "doit" handler provided by the generic netlink layer to accomplish this.
> [...]
>> +struct bnad_genl_ioc_info {
>> + int status;
>> + u16 bnad_num;
>> + u16 rsvd;
>> + char serialnum[64];
>> + char hwpath[BFA_STRING_32];
>> + char adapter_hwpath[BFA_STRING_32];
>> + char guid[BFA_ADAPTER_SYM_NAME_LEN*2];
>> + char name[BFA_ADAPTER_SYM_NAME_LEN];
>> + char port_name[BFA_ADAPTER_SYM_NAME_LEN];
>> + char eth_name[BFA_ADAPTER_SYM_NAME_LEN];
>> + u64 rsvd1[4];
>> + mac_t mac;
>> + mac_t factory_mac; /* Factory mac address */
>> + mac_t current_mac; /* Currently assigned mac address */
>> + enum bfa_ioc_type ioc_type;
>> + u16 pvid; /* Port vlan id */
>> + u16 rsvd2;
>> + u32 host;
>> + u32 bandwidth; /* For PF support */
>> + u32 rsvd3;
>> +};
>> +
>> +struct bnad_genl_ioc_attr {
>> + int status;
>> + u16 bnad_num;
>> + u16 rsvd;
>> + struct bfa_ioc_attr ioc_attr;
>> +};
>
> These things all look like they're better put into sysfs. Why would you
> create a genl protocol just to query some presumably static attributes?
>
>
> -David
>
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
Hi David,
There could be different ways of doing it, but the reason we chose to go
the generic netlink route are as follows:
1) The recommended format of sysfs (as per sysfs.txt in kernel documentation)
is one single line of ASCII text per file.
As a result this:
a) if there are a lot of attributes to be collected,
the driver will end up in creating that many sysfs files.
b) Reading / writing each attribute will result in a set of
[open()/read()/write()/close()] calls.
This is not very scalable, particularly if we want to expand on this
interface for better management functionality in the future.
Instead, generic netlink provides a much better way of multiplexing data
over a single socket.
2) Asynchronous event notifications can be much easily handled using the generic
netlink interface.
Thanks
--Debashis
^ permalink raw reply
* Re: [PATCH v3] ipconfig wait for carrier
From: David Miller @ 2011-05-18 22:14 UTC (permalink / raw)
To: micha; +Cc: netdev
In-Reply-To: <4DD36E54.9040706@neli.hopto.org>
From: Micha Nelissen <micha@neli.hopto.org>
Date: Wed, 18 May 2011 08:59:32 +0200
> Op 2011-05-18 8:37, David Miller schreef:
>> From: Micha Nelissen<micha@neli.hopto.org>
>> Date: Wed, 18 May 2011 08:32:35 +0200
>>
>>> I'm confused. Against which tree/commit do you want it then?
>>
>> Linus's current tree would be fine as would:
>>
>> git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6.git
>
> Ok I see, thanks. The patch should apply just fine to your tree, there
> is only a spelling change since 2.6.38 which does not conflict.
Please fix ic_is_init_dev() to return a proper boolean "false" instead
of "0" when IFF_LOOPBACK is set.
^ permalink raw reply
* Re: [PATCH net-next] bridge: add notification over netlink when STP changes state
From: David Miller @ 2011-05-18 22:10 UTC (permalink / raw)
To: shemminger; +Cc: bridge, netdev
In-Reply-To: <20110518081718.0b1fc390@nehalam>
From: Stephen Hemminger <shemminger@vyatta.com>
Date: Wed, 18 May 2011 08:17:18 -0700
> The first netlink code in the bridge module was to notify
> user space implementations of Spanning Tree Protocol about
> new ports. It did not handle the case of kernel mode STP
> changing states which could be useful for monitoring.
>
> This patch causes RTM_NEWLINK message to occur on kernel
> transitions. It does not send message if request was from user
> space STP, since that would cause reflection and break existing
> API.
>
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
I agree with Ben that hiding events in any way is going to
be a severe limitation, long term.
I do not want to apply this and keep going down this road.
Sorry.
^ permalink raw reply
* Re: [RFC V4 PATCH] rtnetlink: Compute and store minimum ifinfo dump size
From: David Miller @ 2011-05-18 22:09 UTC (permalink / raw)
To: gregory.v.rose; +Cc: netdev, bhutchings, eric.dumazet
In-Reply-To: <20110518155410.28491.62723.stgit@gitlad.jf.intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
Date: Wed, 18 May 2011 08:54:11 -0700
> The message size allocated for rtnl info dumps was limited to a single
> page. This is not enough for additional interface info available with
> devices that support SR-IOV. Calculate the amount of data required so
> the dump can allocate enough data to satisfy the request.
>
> V2 added the calcit function to the rtnl_register calls so that
> dump functions could get the minimum dump allocation size if they
> needed to.
>
> V3 leverages the fact that the netdev register function ends up
> calling if_nlmsg_size. We collect the minimum dump allocation size
> there and keep it in a module static variable so that the calcit
> function doesn't have to search for the device on every info dump.
>
> V4 fixes the title.
>
> Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
This only works if someone does a plain GETLINK request on the largest
space requiring device before the dump request happens.
Otherwise the value used by the calcit function won't be properly
prepared yet.
^ permalink raw reply
* Re: [PATCH 1/2] forcedeth: make module parameters readable in /sys/module
From: Stephen Hemminger @ 2011-05-18 22:03 UTC (permalink / raw)
To: David Decotigny
Cc: David S. Miller, Joe Perches, Szymon Janc, netdev, linux-kernel,
kernel-net-upstream
In-Reply-To: <1305753000-14933-1-git-send-email-decot@google.com>
On Wed, 18 May 2011 14:09:59 -0700
David Decotigny <decot@google.com> wrote:
> This change allows to publish the values of the module parameters in
> /sys/module.
>
>
Although this makes more info for developer, it also means more
stuff in sysfs taking more memory and not providing any real value
that can't be found by looking at the /etc/modprobe.d for any parameters
user might have set.
^ permalink raw reply
* Re: [PATCH 2/2] forcedeth: allow to silence tx_timeout debug messages
From: David Miller @ 2011-05-18 22:01 UTC (permalink / raw)
To: decot; +Cc: joe, szymon, netdev, linux-kernel, kernel-net-upstream, snanda
In-Reply-To: <BANLkTinG4K-bWS1qqBr_G9mK5OHDYknszg@mail.gmail.com>
From: David Decotigny <decot@google.com>
Date: Wed, 18 May 2011 14:43:58 -0700
> On Wed, May 18, 2011 at 2:16 PM, David Miller <davem@davemloft.net> wrote:
>> You can add a knob which when enabled gives the old verbosity
>> back for diagnostic purposes.
>
> That was the intent of this patch: it adds a debug_tx_timeout module
> parameter to act as the knob. I can rephrase the description of the
> patch, it didn't make this so clear.
> Or are you suggesting I should implement this with another kind of knob?
I'm saying implement things differently.
Reduce the verbosity of the TX timeout message but still print some
amount of information, but provide a new knob that turns on the
existing verbose messages.
^ permalink raw reply
* [PATCH RFC] virtio_net: fix patch: virtio_net: limit xmit polling
From: Michael S. Tsirkin @ 2011-05-18 22:01 UTC (permalink / raw)
To: rusty, habanero, Shirley Ma, Krishna Kumar2, kvm, steved,
Tom Lendacky <tahm@
Cc: Rusty Russell, Michael S. Tsirkin, virtualization, netdev,
linux-kernel
The patch virtio_net: limit xmit polling
got the logic reversed: it polled while we had
capacity not while ring was empty.
Fix it up and clean up a bit by using a for loop.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
OK, turns out that patch was borken. Here's
a fix that survived stress test on my box.
Pushed on my branch, I'll send a rebased series
with Rusty's comments addressed ASAP.
drivers/net/virtio_net.c | 10 ++++++----
1 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 9982bd7..c8cd22d 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -514,12 +514,14 @@ static bool free_old_xmit_skbs(struct virtnet_info *vi, int capacity)
struct sk_buff *skb;
unsigned int len;
bool c;
+ int n;
+
/* We try to free up at least 2 skbs per one sent, so that we'll get
* all of the memory back if they are used fast enough. */
- int n = 2;
-
- while ((c = virtqueue_get_capacity(vi->svq) >= capacity) && --n > 0 &&
- (skb = virtqueue_get_buf(vi->svq, &len)) != NULL) {
+ for (n = 0;
+ ((c = virtqueue_get_capacity(vi->svq)) < capacity || n < 2) &&
+ ((skb = virtqueue_get_buf(vi->svq, &len)));
+ ++n) {
pr_debug("Sent skb %p\n", skb);
vi->dev->stats.tx_bytes += skb->len;
vi->dev->stats.tx_packets++;
--
1.7.5.53.gc233e
^ permalink raw reply related
* Re: [PATCH 2/2] forcedeth: allow to silence tx_timeout debug messages
From: David Decotigny @ 2011-05-18 21:43 UTC (permalink / raw)
To: David Miller
Cc: Joe Perches, szymon, netdev, linux-kernel, kernel-net-upstream,
Sameer Nanda
In-Reply-To: <20110518.171603.2048423328863525352.davem@davemloft.net>
David,
No problem, I will send the patch series correctly numbered. Sorry for that.
Regarding your comment about this debug info change:
On Wed, May 18, 2011 at 2:16 PM, David Miller <davem@davemloft.net> wrote:
> You can add a knob which when enabled gives the old verbosity
> back for diagnostic purposes.
That was the intent of this patch: it adds a debug_tx_timeout module
parameter to act as the knob. I can rephrase the description of the
patch, it didn't make this so clear.
Or are you suggesting I should implement this with another kind of knob?
Thanks!
^ permalink raw reply
* Re: [PATCH net-next-2.6] be2net: Enable SR-IOV for Lancer
From: David Miller @ 2011-05-18 21:31 UTC (permalink / raw)
To: padmanabh.ratnakar; +Cc: netdev, mammatha.edhala
In-Reply-To: <55d7e67f-84d8-4b3a-9d39-a68fab9a2a54@exht1.ad.emulex.com>
From: Padmanabh Ratnakar <padmanabh.ratnakar@Emulex.Com>
Date: Wed, 18 May 2011 18:56:22 +0530
> From: Mammatha Edhala <mammatha.edhala@emulex.com>
>
> Enable SR-IOV for Lancer
>
> Signed-off-by: Mammatha Edhala <mammatha.edhala@emulex.com>
> Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Applied.
^ permalink raw reply
* Re: [PATCH net-next] bnx2x: add support for retrieving dcb peer configuration
From: David Miller @ 2011-05-18 21:29 UTC (permalink / raw)
To: shmulikr; +Cc: eilong, netdev
In-Reply-To: <1305723331.14458.13.camel@lb-tlvb-shmulik.il.broadcom.com>
From: "Shmulik Ravid" <shmulikr@broadcom.com>
Date: Wed, 18 May 2011 15:55:31 +0300
> This patch adds support to the bnx2x for retrieving dcb peer (remote)
> configuration from the embedded DCBX stack.
>
> Signed-off-by: Shmulik Ravid <shmulikr@broadcom.com>
> Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Applied, 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