* Re: [PATCH RFC 5/5] qedf: Add FIP request handling
From: Hannes Reinecke @ 2016-12-28 9:11 UTC (permalink / raw)
To: Dupuis, Chad, martin.petersen-QHcLZuEGTsvQT0dZR+AlfA
Cc: fcoe-devel-s9riP+hp16TNLxjTenLetw, netdev-u79uwXL29TY76Z2rM5mHXA,
yuval.mintz-YGCgFSpz5w/QT0dZR+AlfA,
linux-scsi-u79uwXL29TY76Z2rM5mHXA,
QLogic-Storage-Upstream-YGCgFSpz5w/QT0dZR+AlfA
In-Reply-To: <1482520628-24207-6-git-send-email-chad.dupuis-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
On 12/23/2016 08:17 PM, Dupuis, Chad wrote:
> From: "Dupuis, Chad" <chad.dupuis-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
>
> This patch adds handling for FIP requests and responses that are handled by
> the driver itself and not by libfcoe.
>
> Signed-off-by: Nilesh Javali <nilesh.javali-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
> Signed-off-by: Manish Rangankar <manish.rangankar-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
> Signed-off-by: Saurav Kashyap <saurav.kashyap-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
> Signed-off-by: Chad Dupuis <chad.dupuis-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
> ---
> drivers/scsi/qedf/qedf_fip.c | 267 +++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 267 insertions(+)
> create mode 100644 drivers/scsi/qedf/qedf_fip.c
>
> diff --git a/drivers/scsi/qedf/qedf_fip.c b/drivers/scsi/qedf/qedf_fip.c
> new file mode 100644
> index 0000000..4f185c6
> --- /dev/null
> +++ b/drivers/scsi/qedf/qedf_fip.c
> @@ -0,0 +1,267 @@
> +/*
> + * QLogic FCoE Offload Driver
> + * Copyright (c) 2016 Cavium Inc.
> + *
> + * This software is available under the terms of the GNU General Public License
> + * (GPL) Version 2, available from the file COPYING in the main directory of
> + * this source tree.
> + */
> +#include <linux/if_ether.h>
> +#include <linux/if_vlan.h>
> +#include "qedf.h"
> +
> +extern const struct qed_fcoe_ops *qed_ops;
> +/*
> + * FIP VLAN functions that will eventually move to libfcoe.
> + */
> +
> +void qedf_fcoe_send_vlan_req(struct qedf_ctx *qedf)
> +{
> + struct sk_buff *skb;
> + char *eth_fr;
> + int fr_len;
> + struct fip_vlan *vlan;
> +#define MY_FIP_ALL_FCF_MACS ((__u8[6]) { 1, 0x10, 0x18, 1, 0, 2 })
> + static u8 my_fcoe_all_fcfs[ETH_ALEN] = MY_FIP_ALL_FCF_MACS;
Do you support VN2VN, too?
Cheers,
Hannes
--
Dr. Hannes Reinecke zSeries & Storage
hare-l3A5Bk7waGM@public.gmane.org +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
^ permalink raw reply
* [PATCH v3] net: dev_weight: TX/RX orthogonality
From: Matthias Tafelmeier @ 2016-12-28 9:42 UTC (permalink / raw)
To: netdev; +Cc: hagen, fw, edumazet, daniel
In-Reply-To: <20161227164758.GA10870@localhost.localdomain>
Oftenly, introducing side effects on packet processing on the other half
of the stack by adjusting one of TX/RX via sysctl is not desirable.
There are cases of demand for asymmetric, orthogonal configurability.
This holds true especially for nodes where RPS for RFS usage on top is
configured and therefore use the 'old dev_weight'. This is quite a
common base configuration setup nowadays, even with NICs of superior processing
support (e.g. aRFS).
A good example use case are nodes acting as noSQL data bases with a
large number of tiny requests and rather fewer but large packets as responses.
It's affordable to have large budget and rx dev_weights for the
requests. But as a side effect having this large a number on TX
processed in one run can overwhelm drivers.
This patch therefore introduces an independent configurability via sysctl to
userland.
---
Documentation/sysctl/net.txt | 21 +++++++++++++++++++++
include/linux/netdevice.h | 2 ++
net/core/dev.c | 4 +++-
net/core/sysctl_net_core.c | 14 ++++++++++++++
net/sched/sch_generic.c | 2 +-
5 files changed, 41 insertions(+), 2 deletions(-)
diff --git a/Documentation/sysctl/net.txt b/Documentation/sysctl/net.txt
index f0480f7..53cef32 100644
--- a/Documentation/sysctl/net.txt
+++ b/Documentation/sysctl/net.txt
@@ -61,6 +61,27 @@ The maximum number of packets that kernel can handle on a NAPI interrupt,
it's a Per-CPU variable.
Default: 64
+dev_weight_rx_bias
+--------------
+
+RPS (e.g. RFS, aRFS) processing is competing with the registered NAPI poll function
+of the driver for the per softirq cycle netdev_budget. This parameter influences
+the proportion of the configured netdev_budget that is spent on RPS based packet
+processing during RX softirq cycles. It is further meant for making current
+dev_weight adaptable for asymmetric CPU needs on RX/TX side of the network stack.
+(see dev_weight_tx_bias) It is effective on a per CPU basis. Determination is based
+on dev_weight and is calculated multiplicative (dev_weight * dev_weight_rx_bias).
+Default: 1
+
+dev_weight_tx_bias
+--------------
+
+Scales the maximum number of packets that can be processed during a TX softirq cycle.
+Effective on a per CPU basis. Allows scaling of current dev_weight for asymmetric
+net stack processing needs. Be careful to avoid making TX softirq processing a CPU hog.
+Calculation is based on dev_weight (dev_weight * dev_weight_tx_bias).
+Default: 1
+
default_qdisc
--------------
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 994f742..46b4b66 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -3795,6 +3795,8 @@ void netdev_stats_to_stats64(struct rtnl_link_stats64 *stats64,
extern int netdev_max_backlog;
extern int netdev_tstamp_prequeue;
extern int weight_p;
+extern int dev_weight_rx_bias;
+extern int dev_weight_tx_bias;
bool netdev_has_upper_dev(struct net_device *dev, struct net_device *upper_dev);
struct net_device *netdev_upper_get_next_dev_rcu(struct net_device *dev,
diff --git a/net/core/dev.c b/net/core/dev.c
index 8db5a0b..7ce1736 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -3428,6 +3428,8 @@ EXPORT_SYMBOL(netdev_max_backlog);
int netdev_tstamp_prequeue __read_mostly = 1;
int netdev_budget __read_mostly = 300;
int weight_p __read_mostly = 64; /* old backlog weight */
+int dev_weight_rx_bias __read_mostly = 1; /* bias for backlog weight */
+int dev_weight_tx_bias __read_mostly = 1; /* bias for output_queue quota */
/* Called with irq disabled */
static inline void ____napi_schedule(struct softnet_data *sd,
@@ -4833,7 +4835,7 @@ static int process_backlog(struct napi_struct *napi, int quota)
net_rps_action_and_irq_enable(sd);
}
- napi->weight = weight_p;
+ napi->weight = weight_p * dev_weight_rx_bias;
while (again) {
struct sk_buff *skb;
diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c
index 2a46e40..2197388 100644
--- a/net/core/sysctl_net_core.c
+++ b/net/core/sysctl_net_core.c
@@ -276,6 +276,20 @@ static struct ctl_table net_core_table[] = {
.proc_handler = proc_dointvec
},
{
+ .procname = "dev_weight_rx_bias",
+ .data = &dev_weight_rx_bias,
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = proc_dointvec
+ },
+ {
+ .procname = "dev_weight_tx_bias",
+ .data = &dev_weight_tx_bias,
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = proc_dointvec
+ },
+ {
.procname = "netdev_max_backlog",
.data = &netdev_max_backlog,
.maxlen = sizeof(int),
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index 6eb9c8e..19374ef 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -247,7 +247,7 @@ static inline int qdisc_restart(struct Qdisc *q, int *packets)
void __qdisc_run(struct Qdisc *q)
{
- int quota = weight_p;
+ int quota = weight_p * dev_weight_tx_bias;
int packets;
while (qdisc_restart(q, &packets)) {
--
2.7.4
^ permalink raw reply related
* [PATCH 2/2] ipv4: Namespaceify tcp_max_syn_backlog knob
From: Haishuang Yan @ 2016-12-28 9:52 UTC (permalink / raw)
To: David S. Miller, Alexey Kuznetsov, James Morris, Nikolay Borisov
Cc: netdev, linux-kernel, Haishuang Yan
In-Reply-To: <1482918753-1235-1-git-send-email-yanhaishuang@cmss.chinamobile.com>
Different namespace application might require different maximal
number of remembered connection requests.
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
---
include/net/netns/ipv4.h | 1 +
include/net/request_sock.h | 4 +---
net/core/request_sock.c | 2 --
net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
net/ipv4/tcp.c | 2 --
net/ipv4/tcp_input.c | 4 ++--
net/ipv4/tcp_ipv4.c | 7 +++++--
7 files changed, 16 insertions(+), 18 deletions(-)
diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 99becaf..96b15a2 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -122,6 +122,7 @@ struct netns_ipv4 {
unsigned int sysctl_tcp_notsent_lowat;
int sysctl_tcp_tw_reuse;
struct inet_timewait_death_row tcp_death_row;
+ int sysctl_max_syn_backlog;
int sysctl_igmp_max_memberships;
int sysctl_igmp_max_msf;
diff --git a/include/net/request_sock.h b/include/net/request_sock.h
index 6ebe13e..a12a5d2 100644
--- a/include/net/request_sock.h
+++ b/include/net/request_sock.h
@@ -1,7 +1,7 @@
/*
* NET Generic infrastructure for Network protocols.
*
- * Definitions for request_sock
+ * Definitions for request_sock
*
* Authors: Arnaldo Carvalho de Melo <acme@conectiva.com.br>
*
@@ -123,8 +123,6 @@ static inline void reqsk_put(struct request_sock *req)
reqsk_free(req);
}
-extern int sysctl_max_syn_backlog;
-
/*
* For a TCP Fast Open listener -
* lock - protects the access to all the reqsk, which is co-owned by
diff --git a/net/core/request_sock.c b/net/core/request_sock.c
index 5d26056..9b8727c 100644
--- a/net/core/request_sock.c
+++ b/net/core/request_sock.c
@@ -34,8 +34,6 @@
* and it will increase in proportion to the memory of machine.
* Note : Dont forget somaxconn that may limit backlog too.
*/
-int sysctl_max_syn_backlog = 256;
-EXPORT_SYMBOL(sysctl_max_syn_backlog);
void reqsk_queue_alloc(struct request_sock_queue *queue)
{
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 66f8f1b..134d8e1 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -324,13 +324,6 @@ static struct ctl_table ipv4_table[] = {
.proc_handler = proc_dointvec
},
{
- .procname = "tcp_max_syn_backlog",
- .data = &sysctl_max_syn_backlog,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = proc_dointvec
- },
- {
.procname = "inet_peer_threshold",
.data = &inet_peer_threshold,
.maxlen = sizeof(int),
@@ -960,6 +953,13 @@ static struct ctl_table ipv4_net_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec
},
+ {
+ .procname = "tcp_max_syn_backlog",
+ .data = &init_net.ipv4.sysctl_max_syn_backlog,
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = proc_dointvec
+ },
#ifdef CONFIG_IP_ROUTE_MULTIPATH
{
.procname = "fib_multipath_use_neigh",
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 91938c9..f0637a9 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -3378,9 +3378,7 @@ void __init tcp_init(void)
cnt = tcp_hashinfo.ehash_mask + 1;
-
sysctl_tcp_max_orphans = cnt / 2;
- sysctl_max_syn_backlog = max(128, cnt / 256);
tcp_init_mem();
/* Set per-socket limits to no more than 1/128 the pressure threshold */
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index c614802..ec6d843 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -6377,8 +6377,8 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops,
}
/* Kill the following clause, if you dislike this way. */
else if (!net->ipv4.sysctl_tcp_syncookies &&
- (sysctl_max_syn_backlog - inet_csk_reqsk_queue_len(sk) <
- (sysctl_max_syn_backlog >> 2)) &&
+ (net->ipv4.sysctl_max_syn_backlog - inet_csk_reqsk_queue_len(sk) <
+ (net->ipv4.sysctl_max_syn_backlog >> 2)) &&
!tcp_peer_is_proven(req, dst, false,
tmp_opt.saw_tstamp)) {
/* Without syncookies last quarter of
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 56b5f49..7e4be4f 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2419,7 +2419,7 @@ static void __net_exit tcp_sk_exit(struct net *net)
static int __net_init tcp_sk_init(struct net *net)
{
- int res, cpu;
+ int res, cpu, cnt;
net->ipv4.tcp_sk = alloc_percpu(struct sock *);
if (!net->ipv4.tcp_sk)
@@ -2458,10 +2458,13 @@ static int __net_init tcp_sk_init(struct net *net)
net->ipv4.sysctl_tcp_notsent_lowat = UINT_MAX;
net->ipv4.sysctl_tcp_tw_reuse = 0;
+ cnt = tcp_hashinfo.ehash_mask + 1;
net->ipv4.tcp_death_row.sysctl_tw_recycle = 0;
- net->ipv4.tcp_death_row.sysctl_max_tw_buckets = (tcp_hashinfo.ehash_mask + 1) / 2;
+ net->ipv4.tcp_death_row.sysctl_max_tw_buckets = (cnt + 1) / 2;
net->ipv4.tcp_death_row.hashinfo = &tcp_hashinfo;
+ net->ipv4.sysctl_max_syn_backlog = max(128, cnt / 256);
+
return 0;
fail:
tcp_sk_exit(net);
--
2.5.0
^ permalink raw reply related
* [PATCH 1/2] ipv4: Namespaceify tcp_tw_recycle and tcp_max_tw_buckets knob
From: Haishuang Yan @ 2016-12-28 9:52 UTC (permalink / raw)
To: David S. Miller, Alexey Kuznetsov, James Morris, Nikolay Borisov
Cc: netdev, linux-kernel, Haishuang Yan
Different namespace application might require fast recycling
TIME-WAIT sockets independently of the host.
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
---
include/net/inet_timewait_sock.h | 13 +------------
include/net/netns/ipv4.h | 11 +++++++++++
include/net/tcp.h | 1 -
net/ipv4/af_inet.c | 2 --
net/ipv4/inet_timewait_sock.c | 3 +--
net/ipv4/proc.c | 2 +-
net/ipv4/sysctl_net_ipv4.c | 28 ++++++++++++++--------------
net/ipv4/tcp.c | 3 ++-
net/ipv4/tcp_input.c | 2 +-
net/ipv4/tcp_ipv4.c | 12 ++++++++----
net/ipv4/tcp_minisocks.c | 14 +++++---------
net/ipv6/tcp_ipv6.c | 7 ++++---
12 files changed, 48 insertions(+), 50 deletions(-)
diff --git a/include/net/inet_timewait_sock.h b/include/net/inet_timewait_sock.h
index c9b3eb7..6a75d67 100644
--- a/include/net/inet_timewait_sock.h
+++ b/include/net/inet_timewait_sock.h
@@ -29,16 +29,6 @@
#include <linux/atomic.h>
-struct inet_hashinfo;
-
-struct inet_timewait_death_row {
- atomic_t tw_count;
-
- struct inet_hashinfo *hashinfo ____cacheline_aligned_in_smp;
- int sysctl_tw_recycle;
- int sysctl_max_tw_buckets;
-};
-
struct inet_bind_bucket;
/*
@@ -125,8 +115,7 @@ static inline void inet_twsk_reschedule(struct inet_timewait_sock *tw, int timeo
void inet_twsk_deschedule_put(struct inet_timewait_sock *tw);
-void inet_twsk_purge(struct inet_hashinfo *hashinfo,
- struct inet_timewait_death_row *twdr, int family);
+void inet_twsk_purge(struct inet_hashinfo *hashinfo, int family);
static inline
struct net *twsk_net(const struct inet_timewait_sock *twsk)
diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 0378e88..99becaf 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -27,6 +27,16 @@ struct ping_group_range {
kgid_t range[2];
};
+struct inet_hashinfo;
+
+struct inet_timewait_death_row {
+ atomic_t tw_count;
+
+ struct inet_hashinfo *hashinfo ____cacheline_aligned_in_smp;
+ int sysctl_tw_recycle;
+ int sysctl_max_tw_buckets;
+};
+
struct netns_ipv4 {
#ifdef CONFIG_SYSCTL
struct ctl_table_header *forw_hdr;
@@ -111,6 +121,7 @@ struct netns_ipv4 {
int sysctl_tcp_fin_timeout;
unsigned int sysctl_tcp_notsent_lowat;
int sysctl_tcp_tw_reuse;
+ struct inet_timewait_death_row tcp_death_row;
int sysctl_igmp_max_memberships;
int sysctl_igmp_max_msf;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 6061963..1da0aa7 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -231,7 +231,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
*/
#define TFO_SERVER_WO_SOCKOPT1 0x400
-extern struct inet_timewait_death_row tcp_death_row;
/* sysctl variables for tcp */
extern int sysctl_tcp_timestamps;
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 1830e6f..29b1dd9 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -1831,8 +1831,6 @@ static int __init inet_init(void)
ip_init();
- tcp_v4_init();
-
/* Setup TCP slab cache for open requests. */
tcp_init();
diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c
index ddcd56c..f8aff2c 100644
--- a/net/ipv4/inet_timewait_sock.c
+++ b/net/ipv4/inet_timewait_sock.c
@@ -257,8 +257,7 @@ void __inet_twsk_schedule(struct inet_timewait_sock *tw, int timeo, bool rearm)
}
EXPORT_SYMBOL_GPL(__inet_twsk_schedule);
-void inet_twsk_purge(struct inet_hashinfo *hashinfo,
- struct inet_timewait_death_row *twdr, int family)
+void inet_twsk_purge(struct inet_hashinfo *hashinfo, int family)
{
struct inet_timewait_sock *tw;
struct sock *sk;
diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c
index 7143ca1..0247ca0 100644
--- a/net/ipv4/proc.c
+++ b/net/ipv4/proc.c
@@ -65,7 +65,7 @@ static int sockstat_seq_show(struct seq_file *seq, void *v)
socket_seq_show(seq);
seq_printf(seq, "TCP: inuse %d orphan %d tw %d alloc %d mem %ld\n",
sock_prot_inuse_get(net, &tcp_prot), orphans,
- atomic_read(&tcp_death_row.tw_count), sockets,
+ atomic_read(&net->ipv4.tcp_death_row.tw_count), sockets,
proto_memory_allocated(&tcp_prot));
seq_printf(seq, "UDP: inuse %d mem %ld\n",
sock_prot_inuse_get(net, &udp_prot),
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 22cbd61..66f8f1b 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -290,13 +290,6 @@ static struct ctl_table ipv4_table[] = {
.proc_handler = proc_dointvec
},
{
- .procname = "tcp_max_tw_buckets",
- .data = &tcp_death_row.sysctl_max_tw_buckets,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = proc_dointvec
- },
- {
.procname = "tcp_fastopen",
.data = &sysctl_tcp_fastopen,
.maxlen = sizeof(int),
@@ -310,13 +303,6 @@ static struct ctl_table ipv4_table[] = {
.proc_handler = proc_tcp_fastopen_key,
},
{
- .procname = "tcp_tw_recycle",
- .data = &tcp_death_row.sysctl_tw_recycle,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = proc_dointvec
- },
- {
.procname = "tcp_abort_on_overflow",
.data = &sysctl_tcp_abort_on_overflow,
.maxlen = sizeof(int),
@@ -960,6 +946,20 @@ static struct ctl_table ipv4_net_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec
},
+ {
+ .procname = "tcp_max_tw_buckets",
+ .data = &init_net.ipv4.tcp_death_row.sysctl_max_tw_buckets,
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = proc_dointvec
+ },
+ {
+ .procname = "tcp_tw_recycle",
+ .data = &init_net.ipv4.tcp_death_row.sysctl_tw_recycle,
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = proc_dointvec
+ },
#ifdef CONFIG_IP_ROUTE_MULTIPATH
{
.procname = "fib_multipath_use_neigh",
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 1ef3165..91938c9 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -3334,6 +3334,7 @@ void __init tcp_init(void)
percpu_counter_init(&tcp_sockets_allocated, 0, GFP_KERNEL);
percpu_counter_init(&tcp_orphan_count, 0, GFP_KERNEL);
+ inet_hashinfo_init(&tcp_hashinfo);
tcp_hashinfo.bind_bucket_cachep =
kmem_cache_create("tcp_bind_bucket",
sizeof(struct inet_bind_bucket), 0,
@@ -3378,7 +3379,6 @@ void __init tcp_init(void)
cnt = tcp_hashinfo.ehash_mask + 1;
- tcp_death_row.sysctl_max_tw_buckets = cnt / 2;
sysctl_tcp_max_orphans = cnt / 2;
sysctl_max_syn_backlog = max(128, cnt / 256);
@@ -3399,6 +3399,7 @@ void __init tcp_init(void)
pr_info("Hash tables configured (established %u bind %u)\n",
tcp_hashinfo.ehash_mask + 1, tcp_hashinfo.bhash_size);
+ tcp_v4_init();
tcp_metrics_init();
BUG_ON(tcp_register_congestion_control(&tcp_reno) != 0);
tcp_tasklet_init();
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 6c79075..c614802 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -6363,7 +6363,7 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops,
* timewait bucket, so that all the necessary checks
* are made in the function processing timewait state.
*/
- if (tcp_death_row.sysctl_tw_recycle) {
+ if (net->ipv4.tcp_death_row.sysctl_tw_recycle) {
bool strict;
dst = af_ops->route_req(sk, &fl, req, &strict);
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index fe9da4f..56b5f49 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -146,6 +146,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
struct rtable *rt;
int err;
struct ip_options_rcu *inet_opt;
+ struct inet_timewait_death_row *tcp_death_row = &sock_net(sk)->ipv4.tcp_death_row;
if (addr_len < sizeof(struct sockaddr_in))
return -EINVAL;
@@ -196,7 +197,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
tp->write_seq = 0;
}
- if (tcp_death_row.sysctl_tw_recycle &&
+ if (tcp_death_row->sysctl_tw_recycle &&
!tp->rx_opt.ts_recent_stamp && fl4->daddr == daddr)
tcp_fetch_timewait_stamp(sk, &rt->dst);
@@ -215,7 +216,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
* complete initialization after this.
*/
tcp_set_state(sk, TCP_SYN_SENT);
- err = inet_hash_connect(&tcp_death_row, sk);
+ err = inet_hash_connect(tcp_death_row, sk);
if (err)
goto failure;
@@ -2457,6 +2458,10 @@ static int __net_init tcp_sk_init(struct net *net)
net->ipv4.sysctl_tcp_notsent_lowat = UINT_MAX;
net->ipv4.sysctl_tcp_tw_reuse = 0;
+ net->ipv4.tcp_death_row.sysctl_tw_recycle = 0;
+ net->ipv4.tcp_death_row.sysctl_max_tw_buckets = (tcp_hashinfo.ehash_mask + 1) / 2;
+ net->ipv4.tcp_death_row.hashinfo = &tcp_hashinfo;
+
return 0;
fail:
tcp_sk_exit(net);
@@ -2466,7 +2471,7 @@ static int __net_init tcp_sk_init(struct net *net)
static void __net_exit tcp_sk_exit_batch(struct list_head *net_exit_list)
{
- inet_twsk_purge(&tcp_hashinfo, &tcp_death_row, AF_INET);
+ inet_twsk_purge(&tcp_hashinfo, AF_INET);
}
static struct pernet_operations __net_initdata tcp_sk_ops = {
@@ -2477,7 +2482,6 @@ static struct pernet_operations __net_initdata tcp_sk_ops = {
void __init tcp_v4_init(void)
{
- inet_hashinfo_init(&tcp_hashinfo);
if (register_pernet_subsys(&tcp_sk_ops))
panic("Failed to create the TCP control socket.\n");
}
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 28ce5ee..06fde26 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -29,12 +29,6 @@
int sysctl_tcp_abort_on_overflow __read_mostly;
-struct inet_timewait_death_row tcp_death_row = {
- .sysctl_max_tw_buckets = NR_FILE * 2,
- .hashinfo = &tcp_hashinfo,
-};
-EXPORT_SYMBOL_GPL(tcp_death_row);
-
static bool tcp_in_window(u32 seq, u32 end_seq, u32 s_win, u32 e_win)
{
if (seq == s_win)
@@ -100,6 +94,7 @@ tcp_timewait_state_process(struct inet_timewait_sock *tw, struct sk_buff *skb,
struct tcp_options_received tmp_opt;
struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw);
bool paws_reject = false;
+ struct inet_timewait_death_row *tcp_death_row = &sock_net((struct sock*)tw)->ipv4.tcp_death_row;
tmp_opt.saw_tstamp = 0;
if (th->doff > (sizeof(*th) >> 2) && tcptw->tw_ts_recent_stamp) {
@@ -153,7 +148,7 @@ tcp_timewait_state_process(struct inet_timewait_sock *tw, struct sk_buff *skb,
tcptw->tw_ts_recent = tmp_opt.rcv_tsval;
}
- if (tcp_death_row.sysctl_tw_recycle &&
+ if (tcp_death_row->sysctl_tw_recycle &&
tcptw->tw_ts_recent_stamp &&
tcp_tw_remember_stamp(tw))
inet_twsk_reschedule(tw, tw->tw_timeout);
@@ -264,11 +259,12 @@ void tcp_time_wait(struct sock *sk, int state, int timeo)
const struct tcp_sock *tp = tcp_sk(sk);
struct inet_timewait_sock *tw;
bool recycle_ok = false;
+ struct inet_timewait_death_row *tcp_death_row = &sock_net(sk)->ipv4.tcp_death_row;
- if (tcp_death_row.sysctl_tw_recycle && tp->rx_opt.ts_recent_stamp)
+ if (tcp_death_row->sysctl_tw_recycle && tp->rx_opt.ts_recent_stamp)
recycle_ok = tcp_remember_stamp(sk);
- tw = inet_twsk_alloc(sk, &tcp_death_row, state);
+ tw = inet_twsk_alloc(sk, tcp_death_row, state);
if (tw) {
struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 73bc8fc6..a4cdf6a 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -123,6 +123,7 @@ static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
struct dst_entry *dst;
int addr_type;
int err;
+ struct inet_timewait_death_row *tcp_death_row = &sock_net(sk)->ipv4.tcp_death_row;
if (addr_len < SIN6_LEN_RFC2133)
return -EINVAL;
@@ -258,7 +259,7 @@ static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
sk->sk_gso_type = SKB_GSO_TCPV6;
ip6_dst_store(sk, dst, NULL, NULL);
- if (tcp_death_row.sysctl_tw_recycle &&
+ if (tcp_death_row->sysctl_tw_recycle &&
!tp->rx_opt.ts_recent_stamp &&
ipv6_addr_equal(&fl6.daddr, &sk->sk_v6_daddr))
tcp_fetch_timewait_stamp(sk, dst);
@@ -273,7 +274,7 @@ static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
inet->inet_dport = usin->sin6_port;
tcp_set_state(sk, TCP_SYN_SENT);
- err = inet6_hash_connect(&tcp_death_row, sk);
+ err = inet6_hash_connect(tcp_death_row, sk);
if (err)
goto late_failure;
@@ -1948,7 +1949,7 @@ static void __net_exit tcpv6_net_exit(struct net *net)
static void __net_exit tcpv6_net_exit_batch(struct list_head *net_exit_list)
{
- inet_twsk_purge(&tcp_hashinfo, &tcp_death_row, AF_INET6);
+ inet_twsk_purge(&tcp_hashinfo, AF_INET6);
}
static struct pernet_operations tcpv6_net_ops = {
--
2.5.0
^ permalink raw reply related
* Re: George's crazy full state idea (Re: HalfSipHash Acceptable Usage)
From: George Spelvin @ 2016-12-28 10:04 UTC (permalink / raw)
To: daniel, hannes, linux
Cc: ak, davem, David.Laight, ebiggers3, eric.dumazet, Jason,
kernel-hardening, linux-crypto, linux-kernel, luto, netdev, tom,
tytso, vegard.nossum
In-Reply-To: <1482902602.2234.1.camel@stressinduktion.org>
Hannes Frederic Sowa wrote:
> We call extract_crng when we run out of batched entropy and reseed. How
> often we call down to extract_crng depends on how much entropy we
> extracted by calls to get_random_int/long, so the number of calls into
> those functions matter.
>
> In extract_crng we have a timer which reseeds every 300s the CPRNG and
> either uses completely new entropy from the CRNG or calls down into the
> CPRNG while also doing backtracing protection (which feeds chacha's
> block size / 2 back into chacha, if I read the code correctly, thus
> 1024 bits, which should be enough).
In the current code, _extract_crng checks to see if more than 300 s
have elapsed since last time it was reseeded, and if so, reseeds with
fresh entropy.
In addition, on every read (or get_random_bytes), if the request leaves
enough ranfom bits in the last ChaCha block, it feeds back 256 bits
(the ChaCha block size is 16*32 = 512 bits) for anti-backtracking.
If the last read happened to not fit under that limit (size % 512 >
256), *and* there are no calls for RNG output for a long time, there is
no upper limit to how long the old ChaCha key can hang around.
> On Fri, 2016-12-23 at 20:17 -0500, George Spelvin wrote:
>> For example, two mix-backs of 64 bits gives you 65 bit security, not 128.
>> (Because each mixback can be guessed and verified separately.)
> Exactly, but the full reseed after running out of entropy is strong
> enough to not be defeated by your argumentation. Neither the reseed
> from the CRNG.
Yes, I was just reacting to your original statement:
>>>>> couldn't we simply use 8 bytes of the 64 byte
>>>>> return block to feed it directly back into the state chacha?
It's not the idea that's bad, just the proposed quantity.
>> If you want that, I have a pile of patches to prandom I really
>> should push upstream. Shall I refresh them and send them to you?
> I would like to have a look at them in the new year, certainly! I can
> also take care about the core prandom patches, but don't know if I have
> time to submit the others to the different subsystems.
>
> Maybe, if David would be okay with that, we can submit all patches
> through his tree, as he is also the dedicated maintainer for prandom.
Amazing, thank you very much! They're just minor cleanups, nothing
too exciting. I'll put it in the queue to make sure they're up to
date.
^ permalink raw reply
* [PATCH v3] stmmac: enable rx queues
From: Joao Pinto @ 2016-12-28 11:15 UTC (permalink / raw)
To: davem; +Cc: netdev, Joao Pinto
When the hardware is synthesized with multiple queues, all queues are
disabled for default. This patch adds the rx queues configuration.
This patch was successfully tested in a Synopsys QoS Reference design.
Signed-off-by: Joao Pinto <jpinto@synopsys.com>
---
changes v2 -> v3 (Seraphin Bonnaffe):
- GMAC_RX_QUEUE_CLEAR macro simplified
changes v1 -> v2 (Niklas Cassel and Seraphin Bonnaffe):
- Instead of using number of DMA channels, lets use number of queues
- Create 2 flavors of RX queue enable Macros: AV and DCB (AV by default)
- Make sure that the RX queue related bits are cleared before setting
- Check if rx_queue_enable is available before executing
drivers/net/ethernet/stmicro/stmmac/common.h | 5 +++++
drivers/net/ethernet/stmicro/stmmac/dwmac4.h | 8 ++++++++
drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c | 12 ++++++++++++
drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c | 5 +++++
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 22 ++++++++++++++++++++++
5 files changed, 52 insertions(+)
diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h b/drivers/net/ethernet/stmicro/stmmac/common.h
index b13a144..6c96291 100644
--- a/drivers/net/ethernet/stmicro/stmmac/common.h
+++ b/drivers/net/ethernet/stmicro/stmmac/common.h
@@ -323,6 +323,9 @@ struct dma_features {
/* TX and RX number of channels */
unsigned int number_rx_channel;
unsigned int number_tx_channel;
+ /* TX and RX number of queues */
+ unsigned int number_rx_queues;
+ unsigned int number_tx_queues;
/* Alternate (enhanced) DESC mode */
unsigned int enh_desc;
};
@@ -454,6 +457,8 @@ struct stmmac_ops {
void (*core_init)(struct mac_device_info *hw, int mtu);
/* Enable and verify that the IPC module is supported */
int (*rx_ipc)(struct mac_device_info *hw);
+ /* Enable RX Queues */
+ void (*rx_queue_enable)(struct mac_device_info *hw, u32 queue);
/* Dump MAC registers */
void (*dump_regs)(struct mac_device_info *hw);
/* Handle extra events on specific interrupts hw dependent */
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4.h b/drivers/net/ethernet/stmicro/stmmac/dwmac4.h
index 3e8d4fe..b524598 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac4.h
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4.h
@@ -22,6 +22,7 @@
#define GMAC_HASH_TAB_32_63 0x00000014
#define GMAC_RX_FLOW_CTRL 0x00000090
#define GMAC_QX_TX_FLOW_CTRL(x) (0x70 + x * 4)
+#define GMAC_RXQ_CTRL0 0x000000a0
#define GMAC_INT_STATUS 0x000000b0
#define GMAC_INT_EN 0x000000b4
#define GMAC_PCS_BASE 0x000000e0
@@ -44,6 +45,11 @@
#define GMAC_MAX_PERFECT_ADDRESSES 128
+/* MAC RX Queue Enable */
+#define GMAC_RX_QUEUE_CLEAR(queue) ~(GENMASK(1, 0) << ((queue) * 2))
+#define GMAC_RX_AV_QUEUE_ENABLE(queue) BIT((queue) * 2)
+#define GMAC_RX_DCB_QUEUE_ENABLE(queue) BIT(((queue) * 2) + 1)
+
/* MAC Flow Control RX */
#define GMAC_RX_FLOW_CTRL_RFE BIT(0)
@@ -133,6 +139,8 @@ enum power_event {
/* MAC HW features2 bitmap */
#define GMAC_HW_FEAT_TXCHCNT GENMASK(21, 18)
#define GMAC_HW_FEAT_RXCHCNT GENMASK(15, 12)
+#define GMAC_HW_FEAT_TXQCNT GENMASK(9, 6)
+#define GMAC_HW_FEAT_RXQCNT GENMASK(3, 0)
/* MAC HW ADDR regs */
#define GMAC_HI_DCS GENMASK(18, 16)
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c b/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c
index eaed7cb..ecfbf57 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c
@@ -59,6 +59,17 @@ static void dwmac4_core_init(struct mac_device_info *hw, int mtu)
writel(value, ioaddr + GMAC_INT_EN);
}
+static void dwmac4_rx_queue_enable(struct mac_device_info *hw, u32 queue)
+{
+ void __iomem *ioaddr = hw->pcsr;
+ u32 value = readl(ioaddr + GMAC_RXQ_CTRL0);
+
+ value &= GMAC_RX_QUEUE_CLEAR(queue);
+ value |= GMAC_RX_AV_QUEUE_ENABLE(queue);
+
+ writel(value, ioaddr + GMAC_RXQ_CTRL0);
+}
+
static void dwmac4_dump_regs(struct mac_device_info *hw)
{
void __iomem *ioaddr = hw->pcsr;
@@ -392,6 +403,7 @@ static void dwmac4_debug(void __iomem *ioaddr, struct stmmac_extra_stats *x)
static const struct stmmac_ops dwmac4_ops = {
.core_init = dwmac4_core_init,
.rx_ipc = dwmac4_rx_ipc_enable,
+ .rx_queue_enable = dwmac4_rx_queue_enable,
.dump_regs = dwmac4_dump_regs,
.host_irq_status = dwmac4_irq_status,
.flow_ctrl = dwmac4_flow_ctrl,
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c b/drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c
index 8196ab5..377d1b4 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c
@@ -303,6 +303,11 @@ static void dwmac4_get_hw_feature(void __iomem *ioaddr,
((hw_cap & GMAC_HW_FEAT_RXCHCNT) >> 12) + 1;
dma_cap->number_tx_channel =
((hw_cap & GMAC_HW_FEAT_TXCHCNT) >> 18) + 1;
+ /* TX and RX number of queues */
+ dma_cap->number_rx_queues =
+ ((hw_cap & GMAC_HW_FEAT_RXQCNT) >> 0) + 1;
+ dma_cap->number_tx_queues =
+ ((hw_cap & GMAC_HW_FEAT_TXQCNT) >> 6) + 1;
/* IEEE 1588-2002 */
dma_cap->time_stamp = 0;
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index 3e40578..bc9cff9 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -1271,6 +1271,24 @@ static void free_dma_desc_resources(struct stmmac_priv *priv)
}
/**
+ * stmmac_mac_enable_rx_queues - Enable MAC rx queues
+ * @priv: driver private structure
+ * Description: It is used for enabling the rx queues in the MAC
+ */
+static void stmmac_mac_enable_rx_queues(struct stmmac_priv *priv)
+{
+ int rx_count = priv->dma_cap.number_rx_queues;
+ int queue = 0;
+
+ /* If GMAC does not have multiqueues, then this is not necessary*/
+ if (rx_count == 1)
+ return;
+
+ for (queue = 0; queue < rx_count; queue++)
+ priv->hw->mac->rx_queue_enable(priv->hw, queue);
+}
+
+/**
* stmmac_dma_operation_mode - HW DMA operation mode
* @priv: driver private structure
* Description: it is used for configuring the DMA operation mode register in
@@ -1691,6 +1709,10 @@ static int stmmac_hw_setup(struct net_device *dev, bool init_ptp)
/* Initialize the MAC Core */
priv->hw->mac->core_init(priv->hw, dev->mtu);
+ /* Initialize MAC RX Queues */
+ if (priv->hw->mac->rx_queue_enable)
+ stmmac_mac_enable_rx_queues(priv);
+
ret = priv->hw->mac->rx_ipc(priv->hw);
if (!ret) {
netdev_warn(priv->dev, "RX IPC Checksum Offload disabled\n");
--
2.9.3
^ permalink raw reply related
* [PATCH RESEND net-next 0/5] SCTP cleanups
From: Marcelo Ricardo Leitner @ 2016-12-28 11:26 UTC (permalink / raw)
To: netdev; +Cc: linux-sctp, Vlad Yasevich, Neil Horman
Some cleanups/simplifications I've been collecting.
Resending now with net-next open.
Marcelo Ricardo Leitner (5):
sctp: reduce indent level at sctp_sf_tabort_8_4_8
sctp: reduce indent level in sctp_sf_shut_8_4_5
sctp: simplify addr copy
sctp: remove return value from sctp_packet_init/config
sctp: sctp_chunk_length_valid should return bool
include/net/sctp/structs.h | 7 ++-
net/sctp/ipv6.c | 16 +++---
net/sctp/output.c | 14 ++----
net/sctp/protocol.c | 18 +++----
net/sctp/sm_statefuns.c | 122 ++++++++++++++++++++++-----------------------
5 files changed, 81 insertions(+), 96 deletions(-)
--
2.9.3
^ permalink raw reply
* [PATCH RESEND net-next 1/5] sctp: reduce indent level at sctp_sf_tabort_8_4_8
From: Marcelo Ricardo Leitner @ 2016-12-28 11:26 UTC (permalink / raw)
To: netdev; +Cc: linux-sctp, Vlad Yasevich, Neil Horman
In-Reply-To: <cover.1482924092.git.marcelo.leitner@gmail.com>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
---
net/sctp/sm_statefuns.c | 44 +++++++++++++++++++++-----------------------
1 file changed, 21 insertions(+), 23 deletions(-)
diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
index 8ec20a64a3f8055a0c3576627c5ec5dad7e99ca8..32587b1f84e729221965e270607fea7ef93a7430 100644
--- a/net/sctp/sm_statefuns.c
+++ b/net/sctp/sm_statefuns.c
@@ -3237,36 +3237,34 @@ static sctp_disposition_t sctp_sf_tabort_8_4_8(struct net *net,
struct sctp_chunk *abort;
packet = sctp_ootb_pkt_new(net, asoc, chunk);
+ if (!packet)
+ return SCTP_DISPOSITION_NOMEM;
- if (packet) {
- /* Make an ABORT. The T bit will be set if the asoc
- * is NULL.
- */
- abort = sctp_make_abort(asoc, chunk, 0);
- if (!abort) {
- sctp_ootb_pkt_free(packet);
- return SCTP_DISPOSITION_NOMEM;
- }
+ /* Make an ABORT. The T bit will be set if the asoc
+ * is NULL.
+ */
+ abort = sctp_make_abort(asoc, chunk, 0);
+ if (!abort) {
+ sctp_ootb_pkt_free(packet);
+ return SCTP_DISPOSITION_NOMEM;
+ }
- /* Reflect vtag if T-Bit is set */
- if (sctp_test_T_bit(abort))
- packet->vtag = ntohl(chunk->sctp_hdr->vtag);
+ /* Reflect vtag if T-Bit is set */
+ if (sctp_test_T_bit(abort))
+ packet->vtag = ntohl(chunk->sctp_hdr->vtag);
- /* Set the skb to the belonging sock for accounting. */
- abort->skb->sk = ep->base.sk;
+ /* Set the skb to the belonging sock for accounting. */
+ abort->skb->sk = ep->base.sk;
- sctp_packet_append_chunk(packet, abort);
+ sctp_packet_append_chunk(packet, abort);
- sctp_add_cmd_sf(commands, SCTP_CMD_SEND_PKT,
- SCTP_PACKET(packet));
-
- SCTP_INC_STATS(net, SCTP_MIB_OUTCTRLCHUNKS);
+ sctp_add_cmd_sf(commands, SCTP_CMD_SEND_PKT,
+ SCTP_PACKET(packet));
- sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
- return SCTP_DISPOSITION_CONSUME;
- }
+ SCTP_INC_STATS(net, SCTP_MIB_OUTCTRLCHUNKS);
- return SCTP_DISPOSITION_NOMEM;
+ sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
+ return SCTP_DISPOSITION_CONSUME;
}
/*
--
2.9.3
^ permalink raw reply related
* [PATCH RESEND net-next 2/5] sctp: reduce indent level in sctp_sf_shut_8_4_5
From: Marcelo Ricardo Leitner @ 2016-12-28 11:26 UTC (permalink / raw)
To: netdev; +Cc: linux-sctp, Vlad Yasevich, Neil Horman
In-Reply-To: <cover.1482924092.git.marcelo.leitner@gmail.com>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
---
net/sctp/sm_statefuns.c | 58 ++++++++++++++++++++++++-------------------------
1 file changed, 28 insertions(+), 30 deletions(-)
diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
index 32587b1f84e729221965e270607fea7ef93a7430..a95915ef9dbabad30f6171f77b26feab61752f36 100644
--- a/net/sctp/sm_statefuns.c
+++ b/net/sctp/sm_statefuns.c
@@ -3501,45 +3501,43 @@ static sctp_disposition_t sctp_sf_shut_8_4_5(struct net *net,
struct sctp_chunk *shut;
packet = sctp_ootb_pkt_new(net, asoc, chunk);
+ if (!packet)
+ return SCTP_DISPOSITION_NOMEM;
- if (packet) {
- /* Make an SHUTDOWN_COMPLETE.
- * The T bit will be set if the asoc is NULL.
- */
- shut = sctp_make_shutdown_complete(asoc, chunk);
- if (!shut) {
- sctp_ootb_pkt_free(packet);
- return SCTP_DISPOSITION_NOMEM;
- }
-
- /* Reflect vtag if T-Bit is set */
- if (sctp_test_T_bit(shut))
- packet->vtag = ntohl(chunk->sctp_hdr->vtag);
+ /* Make an SHUTDOWN_COMPLETE.
+ * The T bit will be set if the asoc is NULL.
+ */
+ shut = sctp_make_shutdown_complete(asoc, chunk);
+ if (!shut) {
+ sctp_ootb_pkt_free(packet);
+ return SCTP_DISPOSITION_NOMEM;
+ }
- /* Set the skb to the belonging sock for accounting. */
- shut->skb->sk = ep->base.sk;
+ /* Reflect vtag if T-Bit is set */
+ if (sctp_test_T_bit(shut))
+ packet->vtag = ntohl(chunk->sctp_hdr->vtag);
- sctp_packet_append_chunk(packet, shut);
+ /* Set the skb to the belonging sock for accounting. */
+ shut->skb->sk = ep->base.sk;
- sctp_add_cmd_sf(commands, SCTP_CMD_SEND_PKT,
- SCTP_PACKET(packet));
+ sctp_packet_append_chunk(packet, shut);
- SCTP_INC_STATS(net, SCTP_MIB_OUTCTRLCHUNKS);
+ sctp_add_cmd_sf(commands, SCTP_CMD_SEND_PKT,
+ SCTP_PACKET(packet));
- /* If the chunk length is invalid, we don't want to process
- * the reset of the packet.
- */
- if (!sctp_chunk_length_valid(chunk, sizeof(sctp_chunkhdr_t)))
- return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
+ SCTP_INC_STATS(net, SCTP_MIB_OUTCTRLCHUNKS);
- /* We need to discard the rest of the packet to prevent
- * potential bomming attacks from additional bundled chunks.
- * This is documented in SCTP Threats ID.
- */
+ /* If the chunk length is invalid, we don't want to process
+ * the reset of the packet.
+ */
+ if (!sctp_chunk_length_valid(chunk, sizeof(sctp_chunkhdr_t)))
return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
- }
- return SCTP_DISPOSITION_NOMEM;
+ /* We need to discard the rest of the packet to prevent
+ * potential bomming attacks from additional bundled chunks.
+ * This is documented in SCTP Threats ID.
+ */
+ return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
}
/*
--
2.9.3
^ permalink raw reply related
* [PATCH RESEND net-next 3/5] sctp: simplify addr copy
From: Marcelo Ricardo Leitner @ 2016-12-28 11:26 UTC (permalink / raw)
To: netdev; +Cc: linux-sctp, Vlad Yasevich, Neil Horman
In-Reply-To: <cover.1482924092.git.marcelo.leitner@gmail.com>
Make it a bit easier to read.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
---
net/sctp/ipv6.c | 16 +++++++---------
net/sctp/protocol.c | 18 +++++++-----------
2 files changed, 14 insertions(+), 20 deletions(-)
diff --git a/net/sctp/ipv6.c b/net/sctp/ipv6.c
index 176af3080a2b8f8ffc56b55f3ccb13a169e195fe..434167ac0c8e89078b0bac80f0a81da92be37e07 100644
--- a/net/sctp/ipv6.c
+++ b/net/sctp/ipv6.c
@@ -412,22 +412,20 @@ static void sctp_v6_copy_addrlist(struct list_head *addrlist,
static void sctp_v6_from_skb(union sctp_addr *addr, struct sk_buff *skb,
int is_saddr)
{
- __be16 *port;
- struct sctphdr *sh;
+ /* Always called on head skb, so this is safe */
+ struct sctphdr *sh = sctp_hdr(skb);
+ struct sockaddr_in6 *sa = &addr->v6;
- port = &addr->v6.sin6_port;
addr->v6.sin6_family = AF_INET6;
addr->v6.sin6_flowinfo = 0; /* FIXME */
addr->v6.sin6_scope_id = ((struct inet6_skb_parm *)skb->cb)->iif;
- /* Always called on head skb, so this is safe */
- sh = sctp_hdr(skb);
if (is_saddr) {
- *port = sh->source;
- addr->v6.sin6_addr = ipv6_hdr(skb)->saddr;
+ sa->sin6_port = sh->source;
+ sa->sin6_addr = ipv6_hdr(skb)->saddr;
} else {
- *port = sh->dest;
- addr->v6.sin6_addr = ipv6_hdr(skb)->daddr;
+ sa->sin6_port = sh->dest;
+ sa->sin6_addr = ipv6_hdr(skb)->daddr;
}
}
diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c
index ad40fa153878f354a367b5cefa2ccd4322463eb2..fe6c62607e99961b5b3bf524b6f6f46d4be3251d 100644
--- a/net/sctp/protocol.c
+++ b/net/sctp/protocol.c
@@ -237,23 +237,19 @@ int sctp_copy_local_addr_list(struct net *net, struct sctp_bind_addr *bp,
static void sctp_v4_from_skb(union sctp_addr *addr, struct sk_buff *skb,
int is_saddr)
{
- void *from;
- __be16 *port;
- struct sctphdr *sh;
+ /* Always called on head skb, so this is safe */
+ struct sctphdr *sh = sctp_hdr(skb);
+ struct sockaddr_in *sa = &addr->v4;
- port = &addr->v4.sin_port;
addr->v4.sin_family = AF_INET;
- /* Always called on head skb, so this is safe */
- sh = sctp_hdr(skb);
if (is_saddr) {
- *port = sh->source;
- from = &ip_hdr(skb)->saddr;
+ sa->sin_port = sh->source;
+ sa->sin_addr.s_addr = ip_hdr(skb)->saddr;
} else {
- *port = sh->dest;
- from = &ip_hdr(skb)->daddr;
+ sa->sin_port = sh->dest;
+ sa->sin_addr.s_addr = ip_hdr(skb)->daddr;
}
- memcpy(&addr->v4.sin_addr.s_addr, from, sizeof(struct in_addr));
}
/* Initialize an sctp_addr from a socket. */
--
2.9.3
^ permalink raw reply related
* [PATCH RESEND net-next 4/5] sctp: remove return value from sctp_packet_init/config
From: Marcelo Ricardo Leitner @ 2016-12-28 11:26 UTC (permalink / raw)
To: netdev; +Cc: linux-sctp, Vlad Yasevich, Neil Horman
In-Reply-To: <cover.1482924092.git.marcelo.leitner@gmail.com>
There is no reason to use this cascading. It doesn't add anything.
Let's remove it and simplify.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
---
include/net/sctp/structs.h | 7 +++----
net/sctp/output.c | 14 +++++---------
net/sctp/sm_statefuns.c | 5 +++--
3 files changed, 11 insertions(+), 15 deletions(-)
diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index 92daabdc007d94a544baa10d278d229e42e40eb3..87d56cc80a3c1d8549b97391245dacc3b6193618 100644
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -722,10 +722,9 @@ struct sctp_packet {
ipfragok:1; /* So let ip fragment this packet */
};
-struct sctp_packet *sctp_packet_init(struct sctp_packet *,
- struct sctp_transport *,
- __u16 sport, __u16 dport);
-struct sctp_packet *sctp_packet_config(struct sctp_packet *, __u32 vtag, int);
+void sctp_packet_init(struct sctp_packet *, struct sctp_transport *,
+ __u16 sport, __u16 dport);
+void sctp_packet_config(struct sctp_packet *, __u32 vtag, int);
sctp_xmit_t sctp_packet_transmit_chunk(struct sctp_packet *,
struct sctp_chunk *, int, gfp_t);
sctp_xmit_t sctp_packet_append_chunk(struct sctp_packet *,
diff --git a/net/sctp/output.c b/net/sctp/output.c
index f5320a87341e160d46b1160edf4c38b569e7e79b..07ab5062e541dc663b619a278742425a92002945 100644
--- a/net/sctp/output.c
+++ b/net/sctp/output.c
@@ -81,8 +81,8 @@ static void sctp_packet_reset(struct sctp_packet *packet)
/* Config a packet.
* This appears to be a followup set of initializations.
*/
-struct sctp_packet *sctp_packet_config(struct sctp_packet *packet,
- __u32 vtag, int ecn_capable)
+void sctp_packet_config(struct sctp_packet *packet, __u32 vtag,
+ int ecn_capable)
{
struct sctp_transport *tp = packet->transport;
struct sctp_association *asoc = tp->asoc;
@@ -123,14 +123,12 @@ struct sctp_packet *sctp_packet_config(struct sctp_packet *packet,
if (chunk)
sctp_packet_append_chunk(packet, chunk);
}
-
- return packet;
}
/* Initialize the packet structure. */
-struct sctp_packet *sctp_packet_init(struct sctp_packet *packet,
- struct sctp_transport *transport,
- __u16 sport, __u16 dport)
+void sctp_packet_init(struct sctp_packet *packet,
+ struct sctp_transport *transport,
+ __u16 sport, __u16 dport)
{
struct sctp_association *asoc = transport->asoc;
size_t overhead;
@@ -151,8 +149,6 @@ struct sctp_packet *sctp_packet_init(struct sctp_packet *packet,
packet->overhead = overhead;
sctp_packet_reset(packet);
packet->vtag = 0;
-
- return packet;
}
/* Free a packet. */
diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
index a95915ef9dbabad30f6171f77b26feab61752f36..9a223d5b2314ff166be0446462c33219b7eec1b9 100644
--- a/net/sctp/sm_statefuns.c
+++ b/net/sctp/sm_statefuns.c
@@ -6032,8 +6032,9 @@ static struct sctp_packet *sctp_ootb_pkt_new(struct net *net,
sctp_transport_route(transport, (union sctp_addr *)&chunk->dest,
sctp_sk(net->sctp.ctl_sock));
- packet = sctp_packet_init(&transport->packet, transport, sport, dport);
- packet = sctp_packet_config(packet, vtag, 0);
+ packet = &transport->packet;
+ sctp_packet_init(packet, transport, sport, dport);
+ sctp_packet_config(packet, vtag, 0);
return packet;
--
2.9.3
^ permalink raw reply related
* [PATCH RESEND net-next 5/5] sctp: sctp_chunk_length_valid should return bool
From: Marcelo Ricardo Leitner @ 2016-12-28 11:26 UTC (permalink / raw)
To: netdev; +Cc: linux-sctp, Vlad Yasevich, Neil Horman
In-Reply-To: <cover.1482924092.git.marcelo.leitner@gmail.com>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
---
net/sctp/sm_statefuns.c | 15 +++++++--------
1 file changed, 7 insertions(+), 8 deletions(-)
diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
index 9a223d5b2314ff166be0446462c33219b7eec1b9..3382ef254e7b41ae4723f2e72e5aca30d46a4a8e 100644
--- a/net/sctp/sm_statefuns.c
+++ b/net/sctp/sm_statefuns.c
@@ -160,23 +160,22 @@ static sctp_disposition_t __sctp_sf_do_9_1_abort(struct net *net,
/* Small helper function that checks if the chunk length
* is of the appropriate length. The 'required_length' argument
* is set to be the size of a specific chunk we are testing.
- * Return Values: 1 = Valid length
- * 0 = Invalid length
+ * Return Values: true = Valid length
+ * false = Invalid length
*
*/
-static inline int
-sctp_chunk_length_valid(struct sctp_chunk *chunk,
- __u16 required_length)
+static inline bool
+sctp_chunk_length_valid(struct sctp_chunk *chunk, __u16 required_length)
{
__u16 chunk_length = ntohs(chunk->chunk_hdr->length);
/* Previously already marked? */
if (unlikely(chunk->pdiscard))
- return 0;
+ return false;
if (unlikely(chunk_length < required_length))
- return 0;
+ return false;
- return 1;
+ return true;
}
/**********************************************************
--
2.9.3
^ permalink raw reply related
* Re: [PATCH v3] stmmac: enable rx queues
From: Joao Pinto @ 2016-12-28 11:53 UTC (permalink / raw)
To: Kishan Sandeep, Joao Pinto; +Cc: davem, netdev
In-Reply-To: <CAJ3s=NB-ic4OrXPVPpYyoYrEwfwCmrAB8ZNPorpyU+5Do_b2=g@mail.gmail.com>
Hello,
Às 11:50 AM de 12/28/2016, Kishan Sandeep escreveu:
> On Wed, Dec 28, 2016 at 4:45 PM, Joao Pinto <Joao.Pinto@synopsys.com> wrote:
>
>> When the hardware is synthesized with multiple queues, all queues are
>> disabled for default. This patch adds the rx queues configuration.
>> This patch was successfully tested in a Synopsys QoS Reference design.
>>
>> Signed-off-by: Joao Pinto <jpinto@synopsys.com>
>> ---
>> changes v2 -> v3 (Seraphin Bonnaffe):
>> - GMAC_RX_QUEUE_CLEAR macro simplified
>> changes v1 -> v2 (Niklas Cassel and Seraphin Bonnaffe):
>> - Instead of using number of DMA channels, lets use number of queues
>> - Create 2 flavors of RX queue enable Macros: AV and DCB (AV by default)
>> - Make sure that the RX queue related bits are cleared before setting
>> - Check if rx_queue_enable is available before executing
>>
>> drivers/net/ethernet/stmicro/stmmac/common.h | 5 +++++
>> drivers/net/ethernet/stmicro/stmmac/dwmac4.h | 8 ++++++++
>> drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c | 12 ++++++++++++
>> drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c | 5 +++++
>> drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 22
>> ++++++++++++++++++++++
>> 5 files changed, 52 insertions(+)
>>
>> diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h
>> b/drivers/net/ethernet/stmicro/stmmac/common.h
>> index b13a144..6c96291 100644
>> --- a/drivers/net/ethernet/stmicro/stmmac/common.h
>> +++ b/drivers/net/ethernet/stmicro/stmmac/common.h
>> @@ -323,6 +323,9 @@ struct dma_features {
>> /* TX and RX number of channels */
>> unsigned int number_rx_channel;
>> unsigned int number_tx_channel;
>> + /* TX and RX number of queues */
>> + unsigned int number_rx_queues;
>> + unsigned int number_tx_queues;
>>
>
> Looks like this variable is unused. Current driver is only supporting
> single Tx channel/queue. ?? Is current driver not supporting
> multichannel/multiqueue ?
stmmac has a good base for multichannel / multiple queue support, but is only
using 1 for now. My next task is to add complete multiple queues / channel to
stmmac amd that is why number of tx queues is being available.
Thanks!
>
> Thanks!
>
^ permalink raw reply
* Re: [PATCH v2] net: stmmac: bug fix to synchronize stmmac_open and stmmac_dvr_probe
From: Kishan Sandeep @ 2016-12-28 11:56 UTC (permalink / raw)
To: Kweh, Hock Leong
Cc: David Miller, f.fainelli@gmail.com, Joao.Pinto@synopsys.com,
peppe.cavallaro@st.com, seraphin.bonnaffe@st.com,
alexandre.torgue@gmail.com, manabian@gmail.com,
niklas.cassel@axis.com, johan@kernel.org, pavel@ucw.cz,
lars.persson@axis.com, netdev@vger.kernel.org,
linux-kernel@vger.kernel.org
In-Reply-To: <F54AEECA5E2B9541821D670476DAE19C5A916C0E@PGSMSX102.gar.corp.intel.com>
On Wed, Dec 28, 2016 at 7:10 AM, Kweh, Hock Leong
<hock.leong.kweh@intel.com> wrote:
>> -----Original Message-----
>> From: David Miller [mailto:davem@davemloft.net]
>> Sent: Wednesday, December 28, 2016 12:34 AM
>> To: Kweh, Hock Leong <hock.leong.kweh@intel.com>
>> Cc: Joao.Pinto@synopsys.com; peppe.cavallaro@st.com;
>> seraphin.bonnaffe@st.com; f.fainelli@gmail.com;
>> alexandre.torgue@gmail.com; manabian@gmail.com; niklas.cassel@axis.com;
>> johan@kernel.org; pavel@ucw.cz; Ong, Boon Leong
>> <boon.leong.ong@intel.com>; Voon, Weifeng <weifeng.voon@intel.com>;
>> lars.persson@axis.com; netdev@vger.kernel.org; linux-kernel@vger.kernel.org
>> Subject: Re: [PATCH v2] net: stmmac: bug fix to synchronize stmmac_open and
>> stmmac_dvr_probe
>>
>> From: "Kweh, Hock Leong" <hock.leong.kweh@intel.com>
>> Date: Tue, 27 Dec 2016 22:42:36 +0800
>>
>> > From: "Kweh, Hock Leong" <hock.leong.kweh@intel.com>
>>
>> You are not the author of this change, do not take credit for it.
>>
>> You have copied Florian's patch character by character, therefore
>> he is the author.
>>
>> You also didn't CC: the netdev mailing list properly.
>
> Noted & Thanks.
>
> Hi Florian, could you submit this fix from your side so that you are the author.
> I will help to test out.
>
> Thanks & Regards,
> Wilson
>
I think you can give *--author* for giving author name. Try git commit
-am "commit message" --author="Author_name <author_email>"
^ permalink raw reply
* [PATCH net-next] sctp: add pr_debug for tracking asocs not found
From: Marcelo Ricardo Leitner @ 2016-12-28 11:51 UTC (permalink / raw)
To: netdev; +Cc: linux-sctp, Vlad Yasevich, Neil Horman, Xin Long
This pr_debug may help identify why the system is generating some
Aborts. It's not something a sysadmin would be expected to use.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
---
net/sctp/input.c | 17 +++++++++++++++--
1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/net/sctp/input.c b/net/sctp/input.c
index 458e506ef84bae3c53c239d6cf89a9349faafb11..704ad19c1565fe5d8a8aad8e73be124f4c21e0f1 100644
--- a/net/sctp/input.c
+++ b/net/sctp/input.c
@@ -1229,13 +1229,26 @@ static struct sctp_association *__sctp_rcv_lookup(struct net *net,
struct sctp_association *asoc;
asoc = __sctp_lookup_association(net, laddr, paddr, transportp);
+ if (asoc)
+ goto out;
/* Further lookup for INIT/INIT-ACK packets.
* SCTP Implementors Guide, 2.18 Handling of address
* parameters within the INIT or INIT-ACK.
*/
- if (!asoc)
- asoc = __sctp_rcv_lookup_harder(net, skb, laddr, transportp);
+ asoc = __sctp_rcv_lookup_harder(net, skb, laddr, transportp);
+ if (asoc)
+ goto out;
+ if (paddr->sa.sa_family == AF_INET)
+ pr_debug("sctp: asoc not found for src:%pI4:%d dst:%pI4:%d\n",
+ &laddr->v4.sin_addr, ntohs(laddr->v4.sin_port),
+ &paddr->v4.sin_addr, ntohs(paddr->v4.sin_port));
+ else
+ pr_debug("sctp: asoc not found for src:%pI6:%d dst:%pI6:%d\n",
+ &laddr->v6.sin6_addr, ntohs(laddr->v6.sin6_port),
+ &paddr->v6.sin6_addr, ntohs(paddr->v6.sin6_port));
+
+out:
return asoc;
}
--
2.9.3
^ permalink raw reply related
* Re: [PATCH v3] stmmac: enable rx queues
From: Kishan Sandeep @ 2016-12-28 12:02 UTC (permalink / raw)
To: Joao Pinto; +Cc: David Miller, netdev
In-Reply-To: <089cbc8f-ed57-f5e5-51a7-c62e37998f6c@synopsys.com>
On Wed, Dec 28, 2016 at 5:23 PM, Joao Pinto <Joao.Pinto@synopsys.com> wrote:
> Hello,
>
> Às 11:50 AM de 12/28/2016, Kishan Sandeep escreveu:
>> On Wed, Dec 28, 2016 at 4:45 PM, Joao Pinto <Joao.Pinto@synopsys.com> wrote:
>>
>>> When the hardware is synthesized with multiple queues, all queues are
>>> disabled for default. This patch adds the rx queues configuration.
>>> This patch was successfully tested in a Synopsys QoS Reference design.
>>>
>>> Signed-off-by: Joao Pinto <jpinto@synopsys.com>
>>> ---
>>> changes v2 -> v3 (Seraphin Bonnaffe):
>>> - GMAC_RX_QUEUE_CLEAR macro simplified
>>> changes v1 -> v2 (Niklas Cassel and Seraphin Bonnaffe):
>>> - Instead of using number of DMA channels, lets use number of queues
>>> - Create 2 flavors of RX queue enable Macros: AV and DCB (AV by default)
>>> - Make sure that the RX queue related bits are cleared before setting
>>> - Check if rx_queue_enable is available before executing
>>>
>>> drivers/net/ethernet/stmicro/stmmac/common.h | 5 +++++
>>> drivers/net/ethernet/stmicro/stmmac/dwmac4.h | 8 ++++++++
>>> drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c | 12 ++++++++++++
>>> drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c | 5 +++++
>>> drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 22
>>> ++++++++++++++++++++++
>>> 5 files changed, 52 insertions(+)
>>>
>>> diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h
>>> b/drivers/net/ethernet/stmicro/stmmac/common.h
>>> index b13a144..6c96291 100644
>>> --- a/drivers/net/ethernet/stmicro/stmmac/common.h
>>> +++ b/drivers/net/ethernet/stmicro/stmmac/common.h
>>> @@ -323,6 +323,9 @@ struct dma_features {
>>> /* TX and RX number of channels */
>>> unsigned int number_rx_channel;
>>> unsigned int number_tx_channel;
>>> + /* TX and RX number of queues */
>>> + unsigned int number_rx_queues;
>>> + unsigned int number_tx_queues;
>>>
>>
>> Looks like this variable is unused. Current driver is only supporting
>> single Tx channel/queue. ?? Is current driver not supporting
>> multichannel/multiqueue ?
>
> stmmac has a good base for multichannel / multiple queue support, but is only
> using 1 for now. My next task is to add complete multiple queues / channel to
> stmmac amd that is why number of tx queues is being available.
>
> Thanks!
I believe *number_tx_queues* variable can be added at the time support
we are providing.
One query is that what is the algorithm that we are following
currently in the driver if we enable multi Rx queues/DMA channels ?
Thanks!
^ permalink raw reply
* Re: [PATCH v3] stmmac: enable rx queues
From: Joao Pinto @ 2016-12-28 12:05 UTC (permalink / raw)
To: Kishan Sandeep, Joao Pinto; +Cc: David Miller, netdev
In-Reply-To: <CAJ3s=NCGnzgfWF2T5d=OSD4qSPpSnP=E4ioAmLbdFFR23m4zoQ@mail.gmail.com>
Às 12:02 PM de 12/28/2016, Kishan Sandeep escreveu:
> On Wed, Dec 28, 2016 at 5:23 PM, Joao Pinto <Joao.Pinto@synopsys.com> wrote:
>> Hello,
>>
>> Às 11:50 AM de 12/28/2016, Kishan Sandeep escreveu:
>>> On Wed, Dec 28, 2016 at 4:45 PM, Joao Pinto <Joao.Pinto@synopsys.com> wrote:
>>>
>>>> When the hardware is synthesized with multiple queues, all queues are
>>>> disabled for default. This patch adds the rx queues configuration.
>>>> This patch was successfully tested in a Synopsys QoS Reference design.
>>>>
>>>> Signed-off-by: Joao Pinto <jpinto@synopsys.com>
>>>> ---
>>>> changes v2 -> v3 (Seraphin Bonnaffe):
>>>> - GMAC_RX_QUEUE_CLEAR macro simplified
>>>> changes v1 -> v2 (Niklas Cassel and Seraphin Bonnaffe):
>>>> - Instead of using number of DMA channels, lets use number of queues
>>>> - Create 2 flavors of RX queue enable Macros: AV and DCB (AV by default)
>>>> - Make sure that the RX queue related bits are cleared before setting
>>>> - Check if rx_queue_enable is available before executing
>>>>
>>>> drivers/net/ethernet/stmicro/stmmac/common.h | 5 +++++
>>>> drivers/net/ethernet/stmicro/stmmac/dwmac4.h | 8 ++++++++
>>>> drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c | 12 ++++++++++++
>>>> drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c | 5 +++++
>>>> drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 22
>>>> ++++++++++++++++++++++
>>>> 5 files changed, 52 insertions(+)
>>>>
>>>> diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h
>>>> b/drivers/net/ethernet/stmicro/stmmac/common.h
>>>> index b13a144..6c96291 100644
>>>> --- a/drivers/net/ethernet/stmicro/stmmac/common.h
>>>> +++ b/drivers/net/ethernet/stmicro/stmmac/common.h
>>>> @@ -323,6 +323,9 @@ struct dma_features {
>>>> /* TX and RX number of channels */
>>>> unsigned int number_rx_channel;
>>>> unsigned int number_tx_channel;
>>>> + /* TX and RX number of queues */
>>>> + unsigned int number_rx_queues;
>>>> + unsigned int number_tx_queues;
>>>>
>>>
>>> Looks like this variable is unused. Current driver is only supporting
>>> single Tx channel/queue. ?? Is current driver not supporting
>>> multichannel/multiqueue ?
>>
>> stmmac has a good base for multichannel / multiple queue support, but is only
>> using 1 for now. My next task is to add complete multiple queues / channel to
>> stmmac amd that is why number of tx queues is being available.
>>
>> Thanks!
>
> I believe *number_tx_queues* variable can be added at the time support
> we are providing.
>
> One query is that what is the algorithm that we are following
> currently in the driver if we enable multi Rx queues/DMA channels ?
In a synthesized core with single channel, multi dma/queues is not an issue
since it has a single queue and that's it. For a synthesized core with multiple
there are some algorithms that can be configured in the driver. I will include
them later. I can change this patch and enale rx queue 0 only, no problem.
>
> Thanks!
>
^ permalink raw reply
* Re: [PATCH v3] stmmac: enable rx queues
From: Joao Pinto @ 2016-12-28 12:26 UTC (permalink / raw)
To: Kishan Sandeep, Joao Pinto, netdev
In-Reply-To: <CAJ3s=NCs5j_iWkYTyNto6UaX-bmLjXVYu6DGKXQNfbafa5AVvg@mail.gmail.com>
Às 12:23 PM de 12/28/2016, Kishan Sandeep escreveu:
> On Dec 28, 2016 5:50 PM, "Joao Pinto" <Joao.Pinto@synopsys.com> wrote:
>>
>> Às 12:19 PM de 12/28/2016, Joao Pinto escreveu:
>>> Às 12:15 PM de 12/28/2016, Kishan Sandeep escreveu:
>>>> On Dec 28, 2016 5:36 PM, "Joao Pinto" <Joao.Pinto@synopsys.com> wrote:
>>>>>
>>>>> Às 12:02 PM de 12/28/2016, Kishan Sandeep escreveu:
>>>>>> On Wed, Dec 28, 2016 at 5:23 PM, Joao Pinto <Joao.Pinto@synopsys.com>
>>>> wrote:
>>>>>>> Hello,
>>>>>>>
>>>>>>> Às 11:50 AM de 12/28/2016, Kishan Sandeep escreveu:
>>>>>>>> On Wed, Dec 28, 2016 at 4:45 PM, Joao Pinto <
> Joao.Pinto@synopsys.com>
>>>> wrote:
>>>>>>>>
>>>>>>>>> When the hardware is synthesized with multiple queues, all queues
> are
>>>>>>>>> disabled for default. This patch adds the rx queues configuration.
>>>>>>>>> This patch was successfully tested in a Synopsys QoS Reference
>>>> design.
>>>>>>>>>
>>>>>>>>> Signed-off-by: Joao Pinto <jpinto@synopsys.com>
>>>>>>>>> ---
>>>>>>>>> changes v2 -> v3 (Seraphin Bonnaffe):
>>>>>>>>> - GMAC_RX_QUEUE_CLEAR macro simplified
>>>>>>>>> changes v1 -> v2 (Niklas Cassel and Seraphin Bonnaffe):
>>>>>>>>> - Instead of using number of DMA channels, lets use number of
> queues
>>>>>>>>> - Create 2 flavors of RX queue enable Macros: AV and DCB (AV by
>>>> default)
>>>>>>>>> - Make sure that the RX queue related bits are cleared before
> setting
>>>>>>>>> - Check if rx_queue_enable is available before executing
>>>>>>>>>
>>>>>>>>> drivers/net/ethernet/stmicro/stmmac/common.h | 5 +++++
>>>>>>>>> drivers/net/ethernet/stmicro/stmmac/dwmac4.h | 8 ++++++++
>>>>>>>>> drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c | 12
> ++++++++++++
>>>>>>>>> drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c | 5 +++++
>>>>>>>>> drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 22
>>>>>>>>> ++++++++++++++++++++++
>>>>>>>>> 5 files changed, 52 insertions(+)
>>>>>>>>>
>>>>>>>>> diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h
>>>>>>>>> b/drivers/net/ethernet/stmicro/stmmac/common.h
>>>>>>>>> index b13a144..6c96291 100644
>>>>>>>>> --- a/drivers/net/ethernet/stmicro/stmmac/common.h
>>>>>>>>> +++ b/drivers/net/ethernet/stmicro/stmmac/common.h
>>>>>>>>> @@ -323,6 +323,9 @@ struct dma_features {
>>>>>>>>> /* TX and RX number of channels */
>>>>>>>>> unsigned int number_rx_channel;
>>>>>>>>> unsigned int number_tx_channel;
>>>>>>>>> + /* TX and RX number of queues */
>>>>>>>>> + unsigned int number_rx_queues;
>>>>>>>>> + unsigned int number_tx_queues;
>>>>>>>>>
>>>>>>>>
>>>>>>>> Looks like this variable is unused. Current driver is only
> supporting
>>>>>>>> single Tx channel/queue. ?? Is current driver not supporting
>>>>>>>> multichannel/multiqueue ?
>>>>>>>
>>>>>>> stmmac has a good base for multichannel / multiple queue support,
> but
>>>> is only
>>>>>>> using 1 for now. My next task is to add complete multiple queues /
>>>> channel to
>>>>>>> stmmac amd that is why number of tx queues is being available.
>>>>>>>
>>>>>>> Thanks!
>>>>>>
>>>>>> I believe *number_tx_queues* variable can be added at the time
> support
>>>>>> we are providing.
>>>>>>
>>>>>> One query is that what is the algorithm that we are following
>>>>>> currently in the driver if we enable multi Rx queues/DMA channels ?
>>>>>
>>>>> In a synthesized core with single channel, multi dma/queues is not an
>>>> issue
>>>>> since it has a single queue and that's it. For a synthesized core with
>>>> multiple
>>>>> there are some algorithms that can be configured in the driver. I will
>>>> include
>>>>> them later. I can change this patch and enale rx queue 0 only, no
> problem.
>>>>
>>>> Generally there will be one to one mapping from the DMA channel to
>>>> corresponding MTL queue. If we enable multiple MTL RX queues, is there
> any
>>>> requirement to use those many number of DMA channels in RX path or
> single
>>>> DMA RX channel is sufficient ???
>>>
>>> I am using an ethernet QoS reference design synthesized with 4
> channels / 4
>>> queues. Since stmmac only uses one 1 queue / 1 dma channel. Of course
> the
>>> performance is not the same as using the 4 available ones. I would say
> the
>>> correct approach would be tp get the available rx/tx dma channel number
> and
>>> rx/tx queue number configured in the hardware and get the driver adapt
> and use
>>> as many as there are.
>>
>> Of course there should be always dma channels = queues.
>>
>>>
>>>>
>>>
> In that case here we need to enable multi DMA channels along with multiple
> RX channels also we may need to program the algorithm that needs to run on
> queues
I understand your concern. In order to do this step by step, I am going to
change the patch and enable RX queue 0 only, and then when I finish multiple
support I will add multi rx and tx. I will send a v4 in the afternoon. This way
it will be cleaner and less confusing.
Thanks.
^ permalink raw reply
* [PATCH net] net/sched: cls_flower: Fix missing addr_type in classify
From: Paul Blakey @ 2016-12-28 12:54 UTC (permalink / raw)
To: David S. Miller, netdev
Cc: Jiri Pirko, Hadar Hen Zion, Or Gerlitz, Roi Dayan, Paul Blakey
Since we now use a non zero mask on addr_type, we are matching on its
value (IPV4/IPV6). So before this fix, matching on enc_src_ip/enc_dst_ip
failed in SW/classify path since its value was zero.
This patch sets the proper value of addr_type for encapsulated packets.
Fixes: 970bfcd09791 ('net/sched: cls_flower: Use mask for addr_type')
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Hadar Hen Zion <hadarh@mellanox.com>
---
net/sched/cls_flower.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 333f8e2..970db7a 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -153,10 +153,14 @@ static int fl_classify(struct sk_buff *skb, const struct tcf_proto *tp,
switch (ip_tunnel_info_af(info)) {
case AF_INET:
+ skb_key.enc_control.addr_type =
+ FLOW_DISSECTOR_KEY_IPV4_ADDRS;
skb_key.enc_ipv4.src = key->u.ipv4.src;
skb_key.enc_ipv4.dst = key->u.ipv4.dst;
break;
case AF_INET6:
+ skb_key.enc_control.addr_type =
+ FLOW_DISSECTOR_KEY_IPV6_ADDRS;
skb_key.enc_ipv6.src = key->u.ipv6.src;
skb_key.enc_ipv6.dst = key->u.ipv6.dst;
break;
--
1.8.3.1
^ permalink raw reply related
* [PATCH v4] stmmac: enable rx queues
From: Joao Pinto @ 2016-12-28 12:57 UTC (permalink / raw)
To: davem; +Cc: netdev, sandeepkishan108, Joao Pinto
When the hardware is synthesized with multiple queues, all queues are
disabled for default. This patch adds the rx queues configuration.
This patch was successfully tested in a Synopsys QoS Reference design.
Signed-off-by: Joao Pinto <jpinto@synopsys.com>
---
changes v3 -> v4
- To avoid confusing users only rx queue 0 is being enabled
- Tx queues count were kept for future use when implementing multiple
queue / channel support
changes v2 -> v3 (Seraphin Bonnaffe):
- GMAC_RX_QUEUE_CLEAR macro simplified
changes v1 -> v2 (Niklas Cassel and Seraphin Bonnaffe):
- Instead of using number of DMA channels, lets use number of queues
- Create 2 flavors of RX queue enable Macros: AV and DCB (AV by default)
- Make sure that the RX queue related bits are cleared before setting
- Check if rx_queue_enable is available before executing
drivers/net/ethernet/stmicro/stmmac/common.h | 5 +++++
drivers/net/ethernet/stmicro/stmmac/dwmac4.h | 8 +++++++
drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c | 12 +++++++++++
drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c | 5 +++++
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 26 +++++++++++++++++++++++
5 files changed, 56 insertions(+)
diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h b/drivers/net/ethernet/stmicro/stmmac/common.h
index b13a144..6c96291 100644
--- a/drivers/net/ethernet/stmicro/stmmac/common.h
+++ b/drivers/net/ethernet/stmicro/stmmac/common.h
@@ -323,6 +323,9 @@ struct dma_features {
/* TX and RX number of channels */
unsigned int number_rx_channel;
unsigned int number_tx_channel;
+ /* TX and RX number of queues */
+ unsigned int number_rx_queues;
+ unsigned int number_tx_queues;
/* Alternate (enhanced) DESC mode */
unsigned int enh_desc;
};
@@ -454,6 +457,8 @@ struct stmmac_ops {
void (*core_init)(struct mac_device_info *hw, int mtu);
/* Enable and verify that the IPC module is supported */
int (*rx_ipc)(struct mac_device_info *hw);
+ /* Enable RX Queues */
+ void (*rx_queue_enable)(struct mac_device_info *hw, u32 queue);
/* Dump MAC registers */
void (*dump_regs)(struct mac_device_info *hw);
/* Handle extra events on specific interrupts hw dependent */
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4.h b/drivers/net/ethernet/stmicro/stmmac/dwmac4.h
index 3e8d4fe..b524598 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac4.h
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4.h
@@ -22,6 +22,7 @@
#define GMAC_HASH_TAB_32_63 0x00000014
#define GMAC_RX_FLOW_CTRL 0x00000090
#define GMAC_QX_TX_FLOW_CTRL(x) (0x70 + x * 4)
+#define GMAC_RXQ_CTRL0 0x000000a0
#define GMAC_INT_STATUS 0x000000b0
#define GMAC_INT_EN 0x000000b4
#define GMAC_PCS_BASE 0x000000e0
@@ -44,6 +45,11 @@
#define GMAC_MAX_PERFECT_ADDRESSES 128
+/* MAC RX Queue Enable */
+#define GMAC_RX_QUEUE_CLEAR(queue) ~(GENMASK(1, 0) << ((queue) * 2))
+#define GMAC_RX_AV_QUEUE_ENABLE(queue) BIT((queue) * 2)
+#define GMAC_RX_DCB_QUEUE_ENABLE(queue) BIT(((queue) * 2) + 1)
+
/* MAC Flow Control RX */
#define GMAC_RX_FLOW_CTRL_RFE BIT(0)
@@ -133,6 +139,8 @@ enum power_event {
/* MAC HW features2 bitmap */
#define GMAC_HW_FEAT_TXCHCNT GENMASK(21, 18)
#define GMAC_HW_FEAT_RXCHCNT GENMASK(15, 12)
+#define GMAC_HW_FEAT_TXQCNT GENMASK(9, 6)
+#define GMAC_HW_FEAT_RXQCNT GENMASK(3, 0)
/* MAC HW ADDR regs */
#define GMAC_HI_DCS GENMASK(18, 16)
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c b/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c
index eaed7cb..ecfbf57 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c
@@ -59,6 +59,17 @@ static void dwmac4_core_init(struct mac_device_info *hw, int mtu)
writel(value, ioaddr + GMAC_INT_EN);
}
+static void dwmac4_rx_queue_enable(struct mac_device_info *hw, u32 queue)
+{
+ void __iomem *ioaddr = hw->pcsr;
+ u32 value = readl(ioaddr + GMAC_RXQ_CTRL0);
+
+ value &= GMAC_RX_QUEUE_CLEAR(queue);
+ value |= GMAC_RX_AV_QUEUE_ENABLE(queue);
+
+ writel(value, ioaddr + GMAC_RXQ_CTRL0);
+}
+
static void dwmac4_dump_regs(struct mac_device_info *hw)
{
void __iomem *ioaddr = hw->pcsr;
@@ -392,6 +403,7 @@ static void dwmac4_debug(void __iomem *ioaddr, struct stmmac_extra_stats *x)
static const struct stmmac_ops dwmac4_ops = {
.core_init = dwmac4_core_init,
.rx_ipc = dwmac4_rx_ipc_enable,
+ .rx_queue_enable = dwmac4_rx_queue_enable,
.dump_regs = dwmac4_dump_regs,
.host_irq_status = dwmac4_irq_status,
.flow_ctrl = dwmac4_flow_ctrl,
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c b/drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c
index 8196ab5..377d1b4 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_dma.c
@@ -303,6 +303,11 @@ static void dwmac4_get_hw_feature(void __iomem *ioaddr,
((hw_cap & GMAC_HW_FEAT_RXCHCNT) >> 12) + 1;
dma_cap->number_tx_channel =
((hw_cap & GMAC_HW_FEAT_TXCHCNT) >> 18) + 1;
+ /* TX and RX number of queues */
+ dma_cap->number_rx_queues =
+ ((hw_cap & GMAC_HW_FEAT_RXQCNT) >> 0) + 1;
+ dma_cap->number_tx_queues =
+ ((hw_cap & GMAC_HW_FEAT_TXQCNT) >> 6) + 1;
/* IEEE 1588-2002 */
dma_cap->time_stamp = 0;
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index 3e40578..6318127 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -1271,6 +1271,28 @@ static void free_dma_desc_resources(struct stmmac_priv *priv)
}
/**
+ * stmmac_mac_enable_rx_queues - Enable MAC rx queues
+ * @priv: driver private structure
+ * Description: It is used for enabling the rx queues in the MAC
+ */
+static void stmmac_mac_enable_rx_queues(struct stmmac_priv *priv)
+{
+ int rx_count = priv->dma_cap.number_rx_queues;
+ int queue = 0;
+
+ /* If GMAC does not have multiple queues, then this is not necessary*/
+ if (rx_count == 1)
+ return;
+
+ /**
+ * If the core is synthesized with multiple rx queues / multiple
+ * dma channels, then rx queues will be disabled by default.
+ * For now only rx queue 0 is enabled.
+ */
+ priv->hw->mac->rx_queue_enable(priv->hw, queue);
+}
+
+/**
* stmmac_dma_operation_mode - HW DMA operation mode
* @priv: driver private structure
* Description: it is used for configuring the DMA operation mode register in
@@ -1691,6 +1713,10 @@ static int stmmac_hw_setup(struct net_device *dev, bool init_ptp)
/* Initialize the MAC Core */
priv->hw->mac->core_init(priv->hw, dev->mtu);
+ /* Initialize MAC RX Queues */
+ if (priv->hw->mac->rx_queue_enable)
+ stmmac_mac_enable_rx_queues(priv);
+
ret = priv->hw->mac->rx_ipc(priv->hw);
if (!ret) {
netdev_warn(priv->dev, "RX IPC Checksum Offload disabled\n");
--
2.9.3
^ permalink raw reply related
* [PATCH net 02/12] net/mlx5: Check FW limitations on log_max_qp before setting it
From: Saeed Mahameed @ 2016-12-28 12:58 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Noa Osherovich, Saeed Mahameed
In-Reply-To: <1482929922-32626-1-git-send-email-saeedm@mellanox.com>
From: Noa Osherovich <noaos@mellanox.com>
When setting HCA capabilities, set log_max_qp to be the minimum
between the selected profile's value and the HCA limitation.
Fixes: 938fe83c8dcb ('net/mlx5_core: New device capabilities...')
Signed-off-by: Noa Osherovich <noaos@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/main.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 54e5a78..23c12f1 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -503,6 +503,13 @@ static int handle_hca_cap(struct mlx5_core_dev *dev)
MLX5_SET(cmd_hca_cap, set_hca_cap, pkey_table_size,
to_fw_pkey_sz(dev, 128));
+ /* Check log_max_qp from HCA caps to set in current profile */
+ if (MLX5_CAP_GEN_MAX(dev, log_max_qp) < profile[prof_sel].log_max_qp) {
+ mlx5_core_warn(dev, "log_max_qp value in current profile is %d, changing it to HCA capability limit (%d)\n",
+ profile[prof_sel].log_max_qp,
+ MLX5_CAP_GEN_MAX(dev, log_max_qp));
+ profile[prof_sel].log_max_qp = MLX5_CAP_GEN_MAX(dev, log_max_qp);
+ }
if (prof->mask & MLX5_PROF_MASK_QP_SIZE)
MLX5_SET(cmd_hca_cap, set_hca_cap, log_max_qp,
prof->log_max_qp);
--
2.7.4
^ permalink raw reply related
* [PATCH net 01/12] net/mlx5: Disable RoCE on the e-switch management port under switchdev mode
From: Saeed Mahameed @ 2016-12-28 12:58 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Or Gerlitz, Saeed Mahameed
In-Reply-To: <1482929922-32626-1-git-send-email-saeedm@mellanox.com>
From: Or Gerlitz <ogerlitz@mellanox.com>
Under the switchdev/offloads mode, packets that don't match any
e-switch steering rule are sent towards the e-switch management
port. We use a NIC HW steering rule set per vport (uplink and VFs)
to make them be received into the host OS through the respective
vport representor netdevice.
Currnetly such missed RoCE packets will not get to this NIC steering
rule, and hence VF RoCE will not work over the slow path of the offloads
mode. This is b/c these packets will be matched by a steering rule added
by the firmware that serves RoCE traffic set on the PF NIC vport which
is also the e-switch management port under SRIOV.
Disabling RoCE on the e-switch management vport when we are in the offloads
mode, will signal to the firmware to remove their RoCE rule, and then the
missed RoCE packets will be matched by the representor NIC steering rule
as any other missed packets.
To achieve that, we disable RoCE on the PF vport. We do that by removing
(hot-unplugging) the IB device instance associated with the PF. This is
also required by our current model where the PF serves as the uplink
representor and hence only SW switching (TC, bridge, OVS) applications
and slow path vport mlx5e net-device should be running over that vport.
Fixes: c930a3ad7453 ('net/mlx5e: Add devlink based SRIOV mode changes')
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Hadar Hen Zion <hadarh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 466e161..03293ed 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -695,6 +695,12 @@ int esw_offloads_init(struct mlx5_eswitch *esw, int nvports)
if (err)
goto err_reps;
}
+
+ /* disable PF RoCE so missed packets don't go through RoCE steering */
+ mlx5_dev_list_lock();
+ mlx5_remove_dev_by_protocol(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
+ mlx5_dev_list_unlock();
+
return 0;
err_reps:
@@ -718,6 +724,11 @@ static int esw_offloads_stop(struct mlx5_eswitch *esw)
{
int err, err1, num_vfs = esw->dev->priv.sriov.num_vfs;
+ /* enable back PF RoCE */
+ mlx5_dev_list_lock();
+ mlx5_add_dev_by_protocol(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
+ mlx5_dev_list_unlock();
+
mlx5_eswitch_disable_sriov(esw);
err = mlx5_eswitch_enable_sriov(esw, num_vfs, SRIOV_LEGACY);
if (err) {
--
2.7.4
^ permalink raw reply related
* [PATCH net 07/12] net/mlx5: Prevent setting multicast macs for VFs
From: Saeed Mahameed @ 2016-12-28 12:58 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Mohamad Haj Yahia, Saeed Mahameed
In-Reply-To: <1482929922-32626-1-git-send-email-saeedm@mellanox.com>
From: Mohamad Haj Yahia <mohamad@mellanox.com>
Need to check that VF mac address entered by the admin user is either
zero or unicast mac.
Multicast mac addresses are prohibited.
Fixes: 77256579c6b4 ('net/mlx5: E-Switch, Introduce Vport administration functions')
Signed-off-by: Mohamad Haj Yahia <mohamad@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index d6807c3..f14d9c9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1860,7 +1860,7 @@ int mlx5_eswitch_set_vport_mac(struct mlx5_eswitch *esw,
if (!ESW_ALLOWED(esw))
return -EPERM;
- if (!LEGAL_VPORT(esw, vport))
+ if (!LEGAL_VPORT(esw, vport) || is_multicast_ether_addr(mac))
return -EINVAL;
mutex_lock(&esw->state_lock);
--
2.7.4
^ permalink raw reply related
* [PATCH net 03/12] net/mlx5: Cancel recovery work in remove flow
From: Saeed Mahameed @ 2016-12-28 12:58 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Daniel Jurgens, Saeed Mahameed
In-Reply-To: <1482929922-32626-1-git-send-email-saeedm@mellanox.com>
From: Daniel Jurgens <danielj@mellanox.com>
If there is pending delayed work for health recovery it must be canceled
if the device is being unloaded.
Fixes: 05ac2c0b7438 ("net/mlx5: Fix race between PCI error handlers and health work")
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/main.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 23c12f1..0b49739 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -1196,6 +1196,8 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
{
int err = 0;
+ mlx5_drain_health_wq(dev);
+
mutex_lock(&dev->intf_state_mutex);
if (test_bit(MLX5_INTERFACE_STATE_DOWN, &dev->intf_state)) {
dev_warn(&dev->pdev->dev, "%s: interface is down, NOP\n",
@@ -1358,10 +1360,9 @@ static pci_ers_result_t mlx5_pci_err_detected(struct pci_dev *pdev,
mlx5_enter_error_state(dev);
mlx5_unload_one(dev, priv, false);
- /* In case of kernel call save the pci state and drain health wq */
+ /* In case of kernel call save the pci state */
if (state) {
pci_save_state(pdev);
- mlx5_drain_health_wq(dev);
mlx5_pci_disable_device(dev);
}
--
2.7.4
^ permalink raw reply related
* [PATCH net 12/12] net/mlx5e: Disable netdev after close
From: Saeed Mahameed @ 2016-12-28 12:58 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Saeed Mahameed
In-Reply-To: <1482929922-32626-1-git-send-email-saeedm@mellanox.com>
Disable netdev should come after it was closed, although no harm of doing it
before -hence the MLX5E_STATE_DESTROYING bit- but it is more natural this way.
Fixes: 26e59d8077a3 ("net/mlx5e: Implement mlx5e interface attach/detach callbacks")
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Mohamad Haj Yahia <mohamad@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index cf270f6..1236b27 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -3947,10 +3947,6 @@ void mlx5e_detach_netdev(struct mlx5_core_dev *mdev, struct net_device *netdev)
const struct mlx5e_profile *profile = priv->profile;
set_bit(MLX5E_STATE_DESTROYING, &priv->state);
- if (profile->disable)
- profile->disable(priv);
-
- flush_workqueue(priv->wq);
rtnl_lock();
if (netif_running(netdev))
@@ -3958,6 +3954,10 @@ void mlx5e_detach_netdev(struct mlx5_core_dev *mdev, struct net_device *netdev)
netif_device_detach(netdev);
rtnl_unlock();
+ if (profile->disable)
+ profile->disable(priv);
+ flush_workqueue(priv->wq);
+
mlx5e_destroy_q_counter(priv);
profile->cleanup_rx(priv);
mlx5e_close_drop_rq(priv);
--
2.7.4
^ permalink raw reply related
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