* Re: hitting lockdep warning as of too early VF probe with 3.9-rc1
From: Or Gerlitz @ 2013-03-06 20:54 UTC (permalink / raw)
To: Ming Lei
Cc: Or Gerlitz, Greg Kroah-Hartman, David Miller, Roland Dreier,
netdev, Yan Burman, Jack Morgenstein, Liran Liss
In-Reply-To: <CACVXFVM2MLR6HxYyvKRnqddqvRbHvzL59+_NxqFek7YN7nDhQQ@mail.gmail.com>
On Wed, Mar 6, 2013 at 4:43 AM, Ming Lei <ming.lei@canonical.com> wrote:
> You are adding one new PCI device inside another PCI device's probe(),
> so the new device will be probed, since PCI probe() is scheduled by
> work_on_cpu, then cause flush_work() called inside worker function,
> which might be a real deadlock.
So if I understand correct, you recommend to somehow avoid this nested probing?
> I am wondering why this commit can cause the problem, since the PCI
> device will be probed with its driver if there is one driver for it. There is no
> any limit on when the driver should be loaded into system, either before
> device is added or after.
FWIW to undertstanding the issue - the same driver (mlx4_core) is used
by the PF and VF, so the VF driver is already loaded at the time its
been added as new PCI device.
> From driver core view, looks no wrong things are found.
So this got me confused, you pointed on possible deadlock, are you
saying the deadlock wouldn't be the result of how the driver code is
going nor the commited we bisected?
Or.
^ permalink raw reply
* Re: [PATCH] 8139too: send NETDEV_CHANGE manually when autoneg is disabled
From: David Miller @ 2013-03-06 20:53 UTC (permalink / raw)
To: bhutchings
Cc: vfalico, netdev, wfp5p, jasowang, junchangwang, greearb, ivecera
In-Reply-To: <1362601372.3126.23.camel@bwh-desktop.uk.solarflarecom.com>
From: Ben Hutchings <bhutchings@solarflare.com>
Date: Wed, 6 Mar 2013 20:22:52 +0000
> On Wed, 2013-03-06 at 20:06 +0100, Veaceslav Falico wrote:
>> When setting autoneg off (with any additional parameters, like
>> speed/duplex), 8139too doesn't do an interface reset, and thus doesn't
>> notify anyone that its speed/duplex might have changed (bonding and bridge
>> will not see the speed changes, per example).
>>
>> Verify if we've force_media and send notification manually, so that the
>> listeners have a chance to see the changes. It's quite ugly, however I
>> don't see anything better.
>
> Isn't this really a bug in mii_check_media()? It shouldn't shortcut the
> calls to netif_carrier_{off,on}() just because mii->force_media is set.
I think mii_check_media() is responsible for handling this too.
^ permalink raw reply
* [PATCH 3/3 net-next] tcp: implement RFC5682 F-RTO
From: Yuchung Cheng @ 2013-03-06 20:52 UTC (permalink / raw)
To: davem, ncardwell, edumazet, nanditad; +Cc: ilpo.jarvinen, netdev, Yuchung Cheng
In-Reply-To: <1362603130-16501-1-git-send-email-ycheng@google.com>
This patch implements F-RTO (foward RTO recovery):
When the first retransmission after timeout is acknowledged, F-RTO
sends new data instead of old data. If the next ACK acknowledges
some never-retransmitted data, then the timeout was spurious and the
congestion state is reverted. Otherwise if the next ACK selectively
acknowledges the new data, then the timeout was genuine and the
loss recovery continues. This idea applies to recurring timeouts
as well. While F-RTO sends different data during timeout recovery,
it does not (and should not) change the congestion control.
The implementaion follows the three steps of SACK enhanced algorithm
(section 3) in RFC5682. Step 1 is in tcp_enter_loss(). Step 2 and
3 are in tcp_process_loss(). The basic version is not supported
because SACK enhanced version also works for non-SACK connections.
The new implementation is functionally in parity with the old F-RTO
implementation except the one case where it increases undo events:
In addition to the RFC algorithm, a spurious timeout may be detected
without sending data in step 2, as long as the SACK confirms not
all the original data are dropped. When this happens, the sender
will undo the cwnd and perhaps enter fast recovery instead. This
additional check increases the F-RTO undo events by 5x compared
to the prior implementation on Google Web servers, since the sender
often does not have new data to send for HTTP.
Note F-RTO may detect spurious timeout before Eifel with timestamps
does so.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
---
Documentation/networking/ip-sysctl.txt | 18 +++------
include/linux/tcp.h | 3 +-
net/ipv4/tcp_input.c | 74 ++++++++++++++++++++++++++++------
3 files changed, 69 insertions(+), 26 deletions(-)
diff --git a/Documentation/networking/ip-sysctl.txt b/Documentation/networking/ip-sysctl.txt
index df03fcb..7dcc8bf 100644
--- a/Documentation/networking/ip-sysctl.txt
+++ b/Documentation/networking/ip-sysctl.txt
@@ -229,19 +229,13 @@ tcp_fin_timeout - INTEGER
Default: 60 seconds
tcp_frto - INTEGER
- Enables Forward RTO-Recovery (F-RTO) defined in RFC4138.
+ Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
F-RTO is an enhanced recovery algorithm for TCP retransmission
- timeouts. It is particularly beneficial in wireless environments
- where packet loss is typically due to random radio interference
- rather than intermediate router congestion. F-RTO is sender-side
- only modification. Therefore it does not require any support from
- the peer.
-
- If set to 1, basic version is enabled. 2 enables SACK enhanced
- F-RTO if flow uses SACK. The basic version can be used also when
- SACK is in use though scenario(s) with it exists where F-RTO
- interacts badly with the packet counting of the SACK enabled TCP
- flow.
+ timeouts. It is particularly beneficial in networks where the
+ RTT fluctuates (e.g., wireless). F-RTO is sender-side only
+ modification. It does not require any support from the peer.
+
+ By default it's enabled with a non-zero value. 0 disables F-RTO.
tcp_keepalive_time - INTEGER
How often TCP sends out keepalive messages when keepalive is enabled.
diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 7319831..d7d6ea0 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -196,7 +196,8 @@ struct tcp_sock {
u8 nonagle : 4,/* Disable Nagle algorithm? */
thin_lto : 1,/* Use linear timeouts for thin streams */
thin_dupack : 1,/* Fast retransmit on first dupack */
- repair : 1;
+ repair : 1,
+ frto : 1;/* F-RTO (RFC5682) activated in CA_Loss */
u8 repair_queue;
u8 do_early_retrans:1,/* Enable RFC5827 early-retransmit */
early_retrans_delayed:1, /* Delayed ER timer installed */
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 07a05e0..a56657f 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -107,6 +107,7 @@ int sysctl_tcp_early_retrans __read_mostly = 2;
#define FLAG_DATA_SACKED 0x20 /* New SACK. */
#define FLAG_ECE 0x40 /* ECE in this ACK */
#define FLAG_SLOWPATH 0x100 /* Do not skip RFC checks for window update.*/
+#define FLAG_ORIG_SACK_ACKED 0x200 /* Never retransmitted data are (s)acked */
#define FLAG_SND_UNA_ADVANCED 0x400 /* Snd_una was changed (!= FLAG_DATA_ACKED) */
#define FLAG_DSACKING_ACK 0x800 /* SACK blocks contained D-SACK info */
#define FLAG_SACK_RENEGING 0x2000 /* snd_una advanced to a sacked seq */
@@ -1155,6 +1156,8 @@ static u8 tcp_sacktag_one(struct sock *sk,
tcp_highest_sack_seq(tp)))
state->reord = min(fack_count,
state->reord);
+ if (!after(end_seq, tp->high_seq))
+ state->flag |= FLAG_ORIG_SACK_ACKED;
}
if (sacked & TCPCB_LOST) {
@@ -1835,10 +1838,13 @@ void tcp_enter_loss(struct sock *sk, int how)
const struct inet_connection_sock *icsk = inet_csk(sk);
struct tcp_sock *tp = tcp_sk(sk);
struct sk_buff *skb;
+ bool new_recovery = false;
/* Reduce ssthresh if it has not yet been made inside this window. */
- if (icsk->icsk_ca_state <= TCP_CA_Disorder || tp->snd_una == tp->high_seq ||
+ if (icsk->icsk_ca_state <= TCP_CA_Disorder ||
+ !after(tp->high_seq, tp->snd_una) ||
(icsk->icsk_ca_state == TCP_CA_Loss && !icsk->icsk_retransmits)) {
+ new_recovery = true;
tp->prior_ssthresh = tcp_current_ssthresh(sk);
tp->snd_ssthresh = icsk->icsk_ca_ops->ssthresh(sk);
tcp_ca_event(sk, CA_EVENT_LOSS);
@@ -1883,6 +1889,15 @@ void tcp_enter_loss(struct sock *sk, int how)
tcp_set_ca_state(sk, TCP_CA_Loss);
tp->high_seq = tp->snd_nxt;
TCP_ECN_queue_cwr(tp);
+
+ /* F-RTO RFC5682 sec 3.1 step 1: retransmit SND.UNA if no previous
+ * loss recovery is underway except recurring timeout(s) on
+ * the same SND.UNA (sec 3.2). Disable F-RTO on path MTU probing or
+ * other abnormal timeouts (e.g., sack reneging).
+ */
+ tp->frto = sysctl_tcp_frto &&
+ (new_recovery || icsk->icsk_retransmits) &&
+ !inet_csk(sk)->icsk_mtup.probe_size && !how;
}
/* If ACK arrived pointing to a remembered SACK, it means that our
@@ -2425,12 +2440,12 @@ static int tcp_try_undo_partial(struct sock *sk, int acked)
return failed;
}
-/* Undo during loss recovery after partial ACK. */
-static bool tcp_try_undo_loss(struct sock *sk)
+/* Undo during loss recovery after partial ACK or using F-RTO. */
+static bool tcp_try_undo_loss(struct sock *sk, bool frto_undo)
{
struct tcp_sock *tp = tcp_sk(sk);
- if (tcp_may_undo(tp)) {
+ if (frto_undo || tcp_may_undo(tp)) {
struct sk_buff *skb;
tcp_for_write_queue(skb, sk) {
if (skb == tcp_send_head(sk))
@@ -2444,9 +2459,12 @@ static bool tcp_try_undo_loss(struct sock *sk)
tp->lost_out = 0;
tcp_undo_cwr(sk, true);
NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPLOSSUNDO);
+ if (frto_undo)
+ NET_INC_STATS_BH(sock_net(sk),
+ LINUX_MIB_TCPSPURIOUSRTOS);
inet_csk(sk)->icsk_retransmits = 0;
tp->undo_marker = 0;
- if (tcp_is_sack(tp))
+ if (frto_undo || tcp_is_sack(tp))
tcp_set_ca_state(sk, TCP_CA_Open);
return true;
}
@@ -2665,24 +2683,52 @@ static void tcp_enter_recovery(struct sock *sk, bool ece_ack)
/* Process an ACK in CA_Loss state. Move to CA_Open if lost data are
* recovered or spurious. Otherwise retransmits more on partial ACKs.
*/
-static void tcp_process_loss(struct sock *sk, int flag)
+static void tcp_process_loss(struct sock *sk, int flag, bool is_dupack)
{
struct inet_connection_sock *icsk = inet_csk(sk);
struct tcp_sock *tp = tcp_sk(sk);
+ bool recovered = !before(tp->snd_una, tp->high_seq);
- if (!before(tp->snd_una, tp->high_seq)) {
+ if (tp->frto) { /* F-RTO RFC5682 sec 3.1 (sack enhanced version). */
+ if (flag & FLAG_ORIG_SACK_ACKED) {
+ /* Step 3.b. A timeout is spurious if not all data are
+ * lost, i.e., never-retransmitted data are (s)acked.
+ */
+ tcp_try_undo_loss(sk, true);
+ return;
+ }
+ if (after(tp->snd_nxt, tp->high_seq) &&
+ (flag & FLAG_DATA_SACKED || is_dupack)) {
+ tp->frto = 0; /* Loss was real: 2nd part of step 3.a */
+ } else if (flag & FLAG_SND_UNA_ADVANCED && !recovered) {
+ tp->high_seq = tp->snd_nxt;
+ __tcp_push_pending_frames(sk, tcp_current_mss(sk),
+ TCP_NAGLE_OFF);
+ if (after(tp->snd_nxt, tp->high_seq))
+ return; /* Step 2.b */
+ tp->frto = 0;
+ }
+ }
+
+ if (recovered) {
+ /* F-RTO RFC5682 sec 3.1 step 2.a and 1st part of step 3.a */
icsk->icsk_retransmits = 0;
tcp_try_undo_recovery(sk);
return;
}
-
if (flag & FLAG_DATA_ACKED)
icsk->icsk_retransmits = 0;
- if (tcp_is_reno(tp) && flag & FLAG_SND_UNA_ADVANCED)
- tcp_reset_reno_sack(tp);
- if (tcp_try_undo_loss(sk))
+ if (tcp_is_reno(tp)) {
+ /* A Reno DUPACK means new data in F-RTO step 2.b above are
+ * delivered. Lower inflight to clock out (re)tranmissions.
+ */
+ if (after(tp->snd_nxt, tp->high_seq) && is_dupack)
+ tcp_add_reno_sack(sk);
+ else if (flag & FLAG_SND_UNA_ADVANCED)
+ tcp_reset_reno_sack(tp);
+ }
+ if (tcp_try_undo_loss(sk, false))
return;
- tcp_moderate_cwnd(tp);
tcp_xmit_retransmit_queue(sk);
}
@@ -2762,7 +2808,7 @@ static void tcp_fastretrans_alert(struct sock *sk, int pkts_acked,
newly_acked_sacked = pkts_acked + tp->sacked_out - prior_sacked;
break;
case TCP_CA_Loss:
- tcp_process_loss(sk, flag);
+ tcp_process_loss(sk, flag, is_dupack);
if (icsk->icsk_ca_state != TCP_CA_Open)
return;
/* Fall through to processing in Open state. */
@@ -3000,6 +3046,8 @@ static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets,
}
if (!(sacked & TCPCB_SACKED_ACKED))
reord = min(pkts_acked, reord);
+ if (!after(scb->end_seq, tp->high_seq))
+ flag |= FLAG_ORIG_SACK_ACKED;
}
if (sacked & TCPCB_SACKED_ACKED)
--
1.8.1.3
^ permalink raw reply related
* [PATCH 1/3 net-next] tcp: refactor F-RTO
From: Yuchung Cheng @ 2013-03-06 20:52 UTC (permalink / raw)
To: davem, ncardwell, edumazet, nanditad; +Cc: ilpo.jarvinen, netdev, Yuchung Cheng
The patch series refactor the F-RTO feature (RFC4138/5682).
This is to simplify the loss recovery processing. Existing F-RTO
was developed during the experimental stage (RFC4138) and has
many experimental features. It takes a separate code path from
the traditional timeout processing by overloading CA_Disorder
instead of using CA_Loss state. This complicates CA_Disorder state
handling because it's also used for handling dubious ACKs and undos.
While the algorithm in the RFC does not change the congestion control,
the implementation intercepts congestion control in various places
(e.g., frto_cwnd in tcp_ack()).
The new code implements newer F-RTO RFC5682 using CA_Loss processing
path. F-RTO becomes a small extension in the timeout processing
and interfaces with congestion control and Eifel undo modules.
It lets congestion control (module) determines how many to send
independently. F-RTO only chooses what to send in order to detect
spurious retranmission. If timeout is found spurious it invokes
existing Eifel undo algorithms like DSACK or TCP timestamp based
detection.
The first patch removes all F-RTO code except the sysctl_tcp_frto is
left for the new implementation. Since CA_EVENT_FRTO is removed, TCP
westwood now computes ssthresh on regular timeout CA_EVENT_LOSS event.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
---
Documentation/networking/ip-sysctl.txt | 17 --
include/linux/tcp.h | 6 +-
include/net/tcp.h | 4 -
net/ipv4/sysctl_net_ipv4.c | 7 -
net/ipv4/tcp_input.c | 375 +--------------------------------
net/ipv4/tcp_minisocks.c | 3 -
net/ipv4/tcp_output.c | 11 +-
net/ipv4/tcp_timer.c | 6 +-
net/ipv4/tcp_westwood.c | 2 +-
9 files changed, 10 insertions(+), 421 deletions(-)
diff --git a/Documentation/networking/ip-sysctl.txt b/Documentation/networking/ip-sysctl.txt
index dc2dc87..df03fcb 100644
--- a/Documentation/networking/ip-sysctl.txt
+++ b/Documentation/networking/ip-sysctl.txt
@@ -243,23 +243,6 @@ tcp_frto - INTEGER
interacts badly with the packet counting of the SACK enabled TCP
flow.
-tcp_frto_response - INTEGER
- When F-RTO has detected that a TCP retransmission timeout was
- spurious (i.e, the timeout would have been avoided had TCP set a
- longer retransmission timeout), TCP has several options what to do
- next. Possible values are:
- 0 Rate halving based; a smooth and conservative response,
- results in halved cwnd and ssthresh after one RTT
- 1 Very conservative response; not recommended because even
- though being valid, it interacts poorly with the rest of
- Linux TCP, halves cwnd and ssthresh immediately
- 2 Aggressive response; undoes congestion control measures
- that are now known to be unnecessary (ignoring the
- possibility of a lost retransmission that would require
- TCP to be more cautious), cwnd and ssthresh are restored
- to the values prior timeout
- Default: 0 (rate halving based)
-
tcp_keepalive_time - INTEGER
How often TCP sends out keepalive messages when keepalive is enabled.
Default: 2hours.
diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index f28408c..7319831 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -191,14 +191,12 @@ struct tcp_sock {
u32 window_clamp; /* Maximal window to advertise */
u32 rcv_ssthresh; /* Current window clamp */
- u32 frto_highmark; /* snd_nxt when RTO occurred */
u16 advmss; /* Advertised MSS */
- u8 frto_counter; /* Number of new acks after RTO */
+ u8 unused;
u8 nonagle : 4,/* Disable Nagle algorithm? */
thin_lto : 1,/* Use linear timeouts for thin streams */
thin_dupack : 1,/* Fast retransmit on first dupack */
- repair : 1,
- unused : 1;
+ repair : 1;
u8 repair_queue;
u8 do_early_retrans:1,/* Enable RFC5827 early-retransmit */
early_retrans_delayed:1, /* Delayed ER timer installed */
diff --git a/include/net/tcp.h b/include/net/tcp.h
index cf0694d..f537e23 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -273,7 +273,6 @@ extern int sysctl_tcp_app_win;
extern int sysctl_tcp_adv_win_scale;
extern int sysctl_tcp_tw_reuse;
extern int sysctl_tcp_frto;
-extern int sysctl_tcp_frto_response;
extern int sysctl_tcp_low_latency;
extern int sysctl_tcp_dma_copybreak;
extern int sysctl_tcp_nometrics_save;
@@ -425,8 +424,6 @@ extern struct sock * tcp_check_req(struct sock *sk,struct sk_buff *skb,
bool fastopen);
extern int tcp_child_process(struct sock *parent, struct sock *child,
struct sk_buff *skb);
-extern bool tcp_use_frto(struct sock *sk);
-extern void tcp_enter_frto(struct sock *sk);
extern void tcp_enter_loss(struct sock *sk, int how);
extern void tcp_clear_retrans(struct tcp_sock *tp);
extern void tcp_update_metrics(struct sock *sk);
@@ -756,7 +753,6 @@ enum tcp_ca_event {
CA_EVENT_TX_START, /* first transmit when no packets in flight */
CA_EVENT_CWND_RESTART, /* congestion window restart */
CA_EVENT_COMPLETE_CWR, /* end of congestion recovery */
- CA_EVENT_FRTO, /* fast recovery timeout */
CA_EVENT_LOSS, /* loss timeout */
CA_EVENT_FAST_ACK, /* in sequence ack */
CA_EVENT_SLOW_ACK, /* other ack */
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 960fd29..bfb8fda 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -592,13 +592,6 @@ static struct ctl_table ipv4_table[] = {
.proc_handler = proc_dointvec
},
{
- .procname = "tcp_frto_response",
- .data = &sysctl_tcp_frto_response,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = proc_dointvec
- },
- {
.procname = "tcp_low_latency",
.data = &sysctl_tcp_low_latency,
.maxlen = sizeof(int),
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 0d9bdac..9933b6b 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -93,7 +93,6 @@ int sysctl_tcp_stdurg __read_mostly;
int sysctl_tcp_rfc1337 __read_mostly;
int sysctl_tcp_max_orphans __read_mostly = NR_FILE;
int sysctl_tcp_frto __read_mostly = 2;
-int sysctl_tcp_frto_response __read_mostly;
int sysctl_tcp_thin_dupack __read_mostly;
@@ -108,17 +107,14 @@ int sysctl_tcp_early_retrans __read_mostly = 2;
#define FLAG_DATA_SACKED 0x20 /* New SACK. */
#define FLAG_ECE 0x40 /* ECE in this ACK */
#define FLAG_SLOWPATH 0x100 /* Do not skip RFC checks for window update.*/
-#define FLAG_ONLY_ORIG_SACKED 0x200 /* SACKs only non-rexmit sent before RTO */
#define FLAG_SND_UNA_ADVANCED 0x400 /* Snd_una was changed (!= FLAG_DATA_ACKED) */
#define FLAG_DSACKING_ACK 0x800 /* SACK blocks contained D-SACK info */
-#define FLAG_NONHEAD_RETRANS_ACKED 0x1000 /* Non-head rexmitted data was ACKed */
#define FLAG_SACK_RENEGING 0x2000 /* snd_una advanced to a sacked seq */
#define FLAG_ACKED (FLAG_DATA_ACKED|FLAG_SYN_ACKED)
#define FLAG_NOT_DUP (FLAG_DATA|FLAG_WIN_UPDATE|FLAG_ACKED)
#define FLAG_CA_ALERT (FLAG_DATA_SACKED|FLAG_ECE)
#define FLAG_FORWARD_PROGRESS (FLAG_ACKED|FLAG_DATA_SACKED)
-#define FLAG_ANY_PROGRESS (FLAG_FORWARD_PROGRESS|FLAG_SND_UNA_ADVANCED)
#define TCP_REMNANT (TCP_FLAG_FIN|TCP_FLAG_URG|TCP_FLAG_SYN|TCP_FLAG_PSH)
#define TCP_HP_BITS (~(TCP_RESERVED_BITS|TCP_FLAG_PSH))
@@ -1159,10 +1155,6 @@ static u8 tcp_sacktag_one(struct sock *sk,
tcp_highest_sack_seq(tp)))
state->reord = min(fack_count,
state->reord);
-
- /* SACK enhanced F-RTO (RFC4138; Appendix B) */
- if (!after(end_seq, tp->frto_highmark))
- state->flag |= FLAG_ONLY_ORIG_SACKED;
}
if (sacked & TCPCB_LOST) {
@@ -1555,7 +1547,6 @@ static int
tcp_sacktag_write_queue(struct sock *sk, const struct sk_buff *ack_skb,
u32 prior_snd_una)
{
- const struct inet_connection_sock *icsk = inet_csk(sk);
struct tcp_sock *tp = tcp_sk(sk);
const unsigned char *ptr = (skb_transport_header(ack_skb) +
TCP_SKB_CB(ack_skb)->sacked);
@@ -1728,12 +1719,6 @@ walk:
start_seq, end_seq, dup_sack);
advance_sp:
- /* SACK enhanced FRTO (RFC4138, Appendix B): Clearing correct
- * due to in-order walk
- */
- if (after(end_seq, tp->frto_highmark))
- state.flag &= ~FLAG_ONLY_ORIG_SACKED;
-
i++;
}
@@ -1750,8 +1735,7 @@ advance_sp:
tcp_verify_left_out(tp);
if ((state.reord < tp->fackets_out) &&
- ((icsk->icsk_ca_state != TCP_CA_Loss) || tp->undo_marker) &&
- (!tp->frto_highmark || after(tp->snd_una, tp->frto_highmark)))
+ ((inet_csk(sk)->icsk_ca_state != TCP_CA_Loss) || tp->undo_marker))
tcp_update_reordering(sk, tp->fackets_out - state.reord, 0);
out:
@@ -1825,197 +1809,6 @@ static inline void tcp_reset_reno_sack(struct tcp_sock *tp)
tp->sacked_out = 0;
}
-static int tcp_is_sackfrto(const struct tcp_sock *tp)
-{
- return (sysctl_tcp_frto == 0x2) && !tcp_is_reno(tp);
-}
-
-/* F-RTO can only be used if TCP has never retransmitted anything other than
- * head (SACK enhanced variant from Appendix B of RFC4138 is more robust here)
- */
-bool tcp_use_frto(struct sock *sk)
-{
- const struct tcp_sock *tp = tcp_sk(sk);
- const struct inet_connection_sock *icsk = inet_csk(sk);
- struct sk_buff *skb;
-
- if (!sysctl_tcp_frto)
- return false;
-
- /* MTU probe and F-RTO won't really play nicely along currently */
- if (icsk->icsk_mtup.probe_size)
- return false;
-
- if (tcp_is_sackfrto(tp))
- return true;
-
- /* Avoid expensive walking of rexmit queue if possible */
- if (tp->retrans_out > 1)
- return false;
-
- skb = tcp_write_queue_head(sk);
- if (tcp_skb_is_last(sk, skb))
- return true;
- skb = tcp_write_queue_next(sk, skb); /* Skips head */
- tcp_for_write_queue_from(skb, sk) {
- if (skb == tcp_send_head(sk))
- break;
- if (TCP_SKB_CB(skb)->sacked & TCPCB_RETRANS)
- return false;
- /* Short-circuit when first non-SACKed skb has been checked */
- if (!(TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED))
- break;
- }
- return true;
-}
-
-/* RTO occurred, but do not yet enter Loss state. Instead, defer RTO
- * recovery a bit and use heuristics in tcp_process_frto() to detect if
- * the RTO was spurious. Only clear SACKED_RETRANS of the head here to
- * keep retrans_out counting accurate (with SACK F-RTO, other than head
- * may still have that bit set); TCPCB_LOST and remaining SACKED_RETRANS
- * bits are handled if the Loss state is really to be entered (in
- * tcp_enter_frto_loss).
- *
- * Do like tcp_enter_loss() would; when RTO expires the second time it
- * does:
- * "Reduce ssthresh if it has not yet been made inside this window."
- */
-void tcp_enter_frto(struct sock *sk)
-{
- const struct inet_connection_sock *icsk = inet_csk(sk);
- struct tcp_sock *tp = tcp_sk(sk);
- struct sk_buff *skb;
-
- if ((!tp->frto_counter && icsk->icsk_ca_state <= TCP_CA_Disorder) ||
- tp->snd_una == tp->high_seq ||
- ((icsk->icsk_ca_state == TCP_CA_Loss || tp->frto_counter) &&
- !icsk->icsk_retransmits)) {
- tp->prior_ssthresh = tcp_current_ssthresh(sk);
- /* Our state is too optimistic in ssthresh() call because cwnd
- * is not reduced until tcp_enter_frto_loss() when previous F-RTO
- * recovery has not yet completed. Pattern would be this: RTO,
- * Cumulative ACK, RTO (2xRTO for the same segment does not end
- * up here twice).
- * RFC4138 should be more specific on what to do, even though
- * RTO is quite unlikely to occur after the first Cumulative ACK
- * due to back-off and complexity of triggering events ...
- */
- if (tp->frto_counter) {
- u32 stored_cwnd;
- stored_cwnd = tp->snd_cwnd;
- tp->snd_cwnd = 2;
- tp->snd_ssthresh = icsk->icsk_ca_ops->ssthresh(sk);
- tp->snd_cwnd = stored_cwnd;
- } else {
- tp->snd_ssthresh = icsk->icsk_ca_ops->ssthresh(sk);
- }
- /* ... in theory, cong.control module could do "any tricks" in
- * ssthresh(), which means that ca_state, lost bits and lost_out
- * counter would have to be faked before the call occurs. We
- * consider that too expensive, unlikely and hacky, so modules
- * using these in ssthresh() must deal these incompatibility
- * issues if they receives CA_EVENT_FRTO and frto_counter != 0
- */
- tcp_ca_event(sk, CA_EVENT_FRTO);
- }
-
- tp->undo_marker = tp->snd_una;
- tp->undo_retrans = 0;
-
- skb = tcp_write_queue_head(sk);
- if (TCP_SKB_CB(skb)->sacked & TCPCB_RETRANS)
- tp->undo_marker = 0;
- if (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_RETRANS) {
- TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS;
- tp->retrans_out -= tcp_skb_pcount(skb);
- }
- tcp_verify_left_out(tp);
-
- /* Too bad if TCP was application limited */
- tp->snd_cwnd = min(tp->snd_cwnd, tcp_packets_in_flight(tp) + 1);
-
- /* Earlier loss recovery underway (see RFC4138; Appendix B).
- * The last condition is necessary at least in tp->frto_counter case.
- */
- if (tcp_is_sackfrto(tp) && (tp->frto_counter ||
- ((1 << icsk->icsk_ca_state) & (TCPF_CA_Recovery|TCPF_CA_Loss))) &&
- after(tp->high_seq, tp->snd_una)) {
- tp->frto_highmark = tp->high_seq;
- } else {
- tp->frto_highmark = tp->snd_nxt;
- }
- tcp_set_ca_state(sk, TCP_CA_Disorder);
- tp->high_seq = tp->snd_nxt;
- tp->frto_counter = 1;
-}
-
-/* Enter Loss state after F-RTO was applied. Dupack arrived after RTO,
- * which indicates that we should follow the traditional RTO recovery,
- * i.e. mark everything lost and do go-back-N retransmission.
- */
-static void tcp_enter_frto_loss(struct sock *sk, int allowed_segments, int flag)
-{
- struct tcp_sock *tp = tcp_sk(sk);
- struct sk_buff *skb;
-
- tp->lost_out = 0;
- tp->retrans_out = 0;
- if (tcp_is_reno(tp))
- tcp_reset_reno_sack(tp);
-
- tcp_for_write_queue(skb, sk) {
- if (skb == tcp_send_head(sk))
- break;
-
- TCP_SKB_CB(skb)->sacked &= ~TCPCB_LOST;
- /*
- * Count the retransmission made on RTO correctly (only when
- * waiting for the first ACK and did not get it)...
- */
- if ((tp->frto_counter == 1) && !(flag & FLAG_DATA_ACKED)) {
- /* For some reason this R-bit might get cleared? */
- if (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_RETRANS)
- tp->retrans_out += tcp_skb_pcount(skb);
- /* ...enter this if branch just for the first segment */
- flag |= FLAG_DATA_ACKED;
- } else {
- if (TCP_SKB_CB(skb)->sacked & TCPCB_RETRANS)
- tp->undo_marker = 0;
- TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS;
- }
-
- /* Marking forward transmissions that were made after RTO lost
- * can cause unnecessary retransmissions in some scenarios,
- * SACK blocks will mitigate that in some but not in all cases.
- * We used to not mark them but it was causing break-ups with
- * receivers that do only in-order receival.
- *
- * TODO: we could detect presence of such receiver and select
- * different behavior per flow.
- */
- if (!(TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED)) {
- TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
- tp->lost_out += tcp_skb_pcount(skb);
- tp->retransmit_high = TCP_SKB_CB(skb)->end_seq;
- }
- }
- tcp_verify_left_out(tp);
-
- tp->snd_cwnd = tcp_packets_in_flight(tp) + allowed_segments;
- tp->snd_cwnd_cnt = 0;
- tp->snd_cwnd_stamp = tcp_time_stamp;
- tp->frto_counter = 0;
-
- tp->reordering = min_t(unsigned int, tp->reordering,
- sysctl_tcp_reordering);
- tcp_set_ca_state(sk, TCP_CA_Loss);
- tp->high_seq = tp->snd_nxt;
- TCP_ECN_queue_cwr(tp);
-
- tcp_clear_all_retrans_hints(tp);
-}
-
static void tcp_clear_retrans_partial(struct tcp_sock *tp)
{
tp->retrans_out = 0;
@@ -2090,8 +1883,6 @@ void tcp_enter_loss(struct sock *sk, int how)
tcp_set_ca_state(sk, TCP_CA_Loss);
tp->high_seq = tp->snd_nxt;
TCP_ECN_queue_cwr(tp);
- /* Abort F-RTO algorithm if one is in progress */
- tp->frto_counter = 0;
}
/* If ACK arrived pointing to a remembered SACK, it means that our
@@ -2274,10 +2065,6 @@ static bool tcp_time_to_recover(struct sock *sk, int flag)
struct tcp_sock *tp = tcp_sk(sk);
__u32 packets_out;
- /* Do not perform any recovery during F-RTO algorithm */
- if (tp->frto_counter)
- return false;
-
/* Trick#1: The loss is proven. */
if (tp->lost_out)
return true;
@@ -2758,7 +2545,7 @@ static void tcp_try_to_open(struct sock *sk, int flag, int newly_acked_sacked)
tcp_verify_left_out(tp);
- if (!tp->frto_counter && !tcp_any_retrans_done(sk))
+ if (!tcp_any_retrans_done(sk))
tp->retrans_stamp = 0;
if (flag & FLAG_ECE)
@@ -3195,8 +2982,6 @@ static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets,
flag |= FLAG_RETRANS_DATA_ACKED;
ca_seq_rtt = -1;
seq_rtt = -1;
- if ((flag & FLAG_DATA_ACKED) || (acked_pcount > 1))
- flag |= FLAG_NONHEAD_RETRANS_ACKED;
} else {
ca_seq_rtt = now - scb->when;
last_ackt = skb->tstamp;
@@ -3405,150 +3190,6 @@ static int tcp_ack_update_window(struct sock *sk, const struct sk_buff *skb, u32
return flag;
}
-/* A very conservative spurious RTO response algorithm: reduce cwnd and
- * continue in congestion avoidance.
- */
-static void tcp_conservative_spur_to_response(struct tcp_sock *tp)
-{
- tp->snd_cwnd = min(tp->snd_cwnd, tp->snd_ssthresh);
- tp->snd_cwnd_cnt = 0;
- TCP_ECN_queue_cwr(tp);
- tcp_moderate_cwnd(tp);
-}
-
-/* A conservative spurious RTO response algorithm: reduce cwnd using
- * PRR and continue in congestion avoidance.
- */
-static void tcp_cwr_spur_to_response(struct sock *sk)
-{
- tcp_enter_cwr(sk, 0);
-}
-
-static void tcp_undo_spur_to_response(struct sock *sk, int flag)
-{
- if (flag & FLAG_ECE)
- tcp_cwr_spur_to_response(sk);
- else
- tcp_undo_cwr(sk, true);
-}
-
-/* F-RTO spurious RTO detection algorithm (RFC4138)
- *
- * F-RTO affects during two new ACKs following RTO (well, almost, see inline
- * comments). State (ACK number) is kept in frto_counter. When ACK advances
- * window (but not to or beyond highest sequence sent before RTO):
- * On First ACK, send two new segments out.
- * On Second ACK, RTO was likely spurious. Do spurious response (response
- * algorithm is not part of the F-RTO detection algorithm
- * given in RFC4138 but can be selected separately).
- * Otherwise (basically on duplicate ACK), RTO was (likely) caused by a loss
- * and TCP falls back to conventional RTO recovery. F-RTO allows overriding
- * of Nagle, this is done using frto_counter states 2 and 3, when a new data
- * segment of any size sent during F-RTO, state 2 is upgraded to 3.
- *
- * Rationale: if the RTO was spurious, new ACKs should arrive from the
- * original window even after we transmit two new data segments.
- *
- * SACK version:
- * on first step, wait until first cumulative ACK arrives, then move to
- * the second step. In second step, the next ACK decides.
- *
- * F-RTO is implemented (mainly) in four functions:
- * - tcp_use_frto() is used to determine if TCP is can use F-RTO
- * - tcp_enter_frto() prepares TCP state on RTO if F-RTO is used, it is
- * called when tcp_use_frto() showed green light
- * - tcp_process_frto() handles incoming ACKs during F-RTO algorithm
- * - tcp_enter_frto_loss() is called if there is not enough evidence
- * to prove that the RTO is indeed spurious. It transfers the control
- * from F-RTO to the conventional RTO recovery
- */
-static bool tcp_process_frto(struct sock *sk, int flag)
-{
- struct tcp_sock *tp = tcp_sk(sk);
-
- tcp_verify_left_out(tp);
-
- /* Duplicate the behavior from Loss state (fastretrans_alert) */
- if (flag & FLAG_DATA_ACKED)
- inet_csk(sk)->icsk_retransmits = 0;
-
- if ((flag & FLAG_NONHEAD_RETRANS_ACKED) ||
- ((tp->frto_counter >= 2) && (flag & FLAG_RETRANS_DATA_ACKED)))
- tp->undo_marker = 0;
-
- if (!before(tp->snd_una, tp->frto_highmark)) {
- tcp_enter_frto_loss(sk, (tp->frto_counter == 1 ? 2 : 3), flag);
- return true;
- }
-
- if (!tcp_is_sackfrto(tp)) {
- /* RFC4138 shortcoming in step 2; should also have case c):
- * ACK isn't duplicate nor advances window, e.g., opposite dir
- * data, winupdate
- */
- if (!(flag & FLAG_ANY_PROGRESS) && (flag & FLAG_NOT_DUP))
- return true;
-
- if (!(flag & FLAG_DATA_ACKED)) {
- tcp_enter_frto_loss(sk, (tp->frto_counter == 1 ? 0 : 3),
- flag);
- return true;
- }
- } else {
- if (!(flag & FLAG_DATA_ACKED) && (tp->frto_counter == 1)) {
- if (!tcp_packets_in_flight(tp)) {
- tcp_enter_frto_loss(sk, 2, flag);
- return true;
- }
-
- /* Prevent sending of new data. */
- tp->snd_cwnd = min(tp->snd_cwnd,
- tcp_packets_in_flight(tp));
- return true;
- }
-
- if ((tp->frto_counter >= 2) &&
- (!(flag & FLAG_FORWARD_PROGRESS) ||
- ((flag & FLAG_DATA_SACKED) &&
- !(flag & FLAG_ONLY_ORIG_SACKED)))) {
- /* RFC4138 shortcoming (see comment above) */
- if (!(flag & FLAG_FORWARD_PROGRESS) &&
- (flag & FLAG_NOT_DUP))
- return true;
-
- tcp_enter_frto_loss(sk, 3, flag);
- return true;
- }
- }
-
- if (tp->frto_counter == 1) {
- /* tcp_may_send_now needs to see updated state */
- tp->snd_cwnd = tcp_packets_in_flight(tp) + 2;
- tp->frto_counter = 2;
-
- if (!tcp_may_send_now(sk))
- tcp_enter_frto_loss(sk, 2, flag);
-
- return true;
- } else {
- switch (sysctl_tcp_frto_response) {
- case 2:
- tcp_undo_spur_to_response(sk, flag);
- break;
- case 1:
- tcp_conservative_spur_to_response(tp);
- break;
- default:
- tcp_cwr_spur_to_response(sk);
- break;
- }
- tp->frto_counter = 0;
- tp->undo_marker = 0;
- NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPSPURIOUSRTOS);
- }
- return false;
-}
-
/* RFC 5961 7 [ACK Throttling] */
static void tcp_send_challenge_ack(struct sock *sk)
{
@@ -3581,7 +3222,6 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag)
int prior_packets;
int prior_sacked = tp->sacked_out;
int pkts_acked = 0;
- bool frto_cwnd = false;
/* If the ack is older than previous acks
* then we can probably ignore it.
@@ -3654,22 +3294,15 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag)
pkts_acked = prior_packets - tp->packets_out;
- if (tp->frto_counter)
- frto_cwnd = tcp_process_frto(sk, flag);
- /* Guarantee sacktag reordering detection against wrap-arounds */
- if (before(tp->frto_highmark, tp->snd_una))
- tp->frto_highmark = 0;
-
if (tcp_ack_is_dubious(sk, flag)) {
/* Advance CWND, if state allows this. */
- if ((flag & FLAG_DATA_ACKED) && !frto_cwnd &&
- tcp_may_raise_cwnd(sk, flag))
+ if ((flag & FLAG_DATA_ACKED) && tcp_may_raise_cwnd(sk, flag))
tcp_cong_avoid(sk, ack, prior_in_flight);
is_dupack = !(flag & (FLAG_SND_UNA_ADVANCED | FLAG_NOT_DUP));
tcp_fastretrans_alert(sk, pkts_acked, prior_sacked,
is_dupack, flag);
} else {
- if ((flag & FLAG_DATA_ACKED) && !frto_cwnd)
+ if (flag & FLAG_DATA_ACKED)
tcp_cong_avoid(sk, ack, prior_in_flight);
}
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index b83a49c..344f185 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -449,9 +449,6 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req,
newtp->snd_cwnd = TCP_INIT_CWND;
newtp->snd_cwnd_cnt = 0;
- newtp->frto_counter = 0;
- newtp->frto_highmark = 0;
-
if (newicsk->icsk_ca_ops != &tcp_init_congestion_ops &&
!try_module_get(newicsk->icsk_ca_ops->owner))
newicsk->icsk_ca_ops = &tcp_init_congestion_ops;
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index e2b4461..5756ada 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -80,10 +80,6 @@ static void tcp_event_new_data_sent(struct sock *sk, const struct sk_buff *skb)
tcp_advance_send_head(sk, skb);
tp->snd_nxt = TCP_SKB_CB(skb)->end_seq;
- /* Don't override Nagle indefinitely with F-RTO */
- if (tp->frto_counter == 2)
- tp->frto_counter = 3;
-
tp->packets_out += tcp_skb_pcount(skb);
if (!prior_packets || tp->early_retrans_delayed)
tcp_rearm_rto(sk);
@@ -1633,11 +1629,8 @@ static inline bool tcp_nagle_test(const struct tcp_sock *tp, const struct sk_buf
if (nonagle & TCP_NAGLE_PUSH)
return true;
- /* Don't use the nagle rule for urgent data (or for the final FIN).
- * Nagle can be ignored during F-RTO too (see RFC4138).
- */
- if (tcp_urg_mode(tp) || (tp->frto_counter == 2) ||
- (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN))
+ /* Don't use the nagle rule for urgent data (or for the final FIN). */
+ if (tcp_urg_mode(tp) || (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN))
return true;
if (!tcp_nagle_check(tp, skb, cur_mss, nonagle))
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index b78aac3..a846ca1 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -418,11 +418,7 @@ void tcp_retransmit_timer(struct sock *sk)
NET_INC_STATS_BH(sock_net(sk), mib_idx);
}
- if (tcp_use_frto(sk)) {
- tcp_enter_frto(sk);
- } else {
- tcp_enter_loss(sk, 0);
- }
+ tcp_enter_loss(sk, 0);
if (tcp_retransmit_skb(sk, tcp_write_queue_head(sk)) > 0) {
/* Retransmission failed because of local congestion,
diff --git a/net/ipv4/tcp_westwood.c b/net/ipv4/tcp_westwood.c
index 1b91bf4..76a1e23 100644
--- a/net/ipv4/tcp_westwood.c
+++ b/net/ipv4/tcp_westwood.c
@@ -236,7 +236,7 @@ static void tcp_westwood_event(struct sock *sk, enum tcp_ca_event event)
tp->snd_cwnd = tp->snd_ssthresh = tcp_westwood_bw_rttmin(sk);
break;
- case CA_EVENT_FRTO:
+ case CA_EVENT_LOSS:
tp->snd_ssthresh = tcp_westwood_bw_rttmin(sk);
/* Update RTT_min when next ack arrives */
w->reset_rtt_min = 1;
--
1.8.1.3
^ permalink raw reply related
* [PATCH 2/3 net-next] tcp: refactor CA_Loss state processing
From: Yuchung Cheng @ 2013-03-06 20:52 UTC (permalink / raw)
To: davem, ncardwell, edumazet, nanditad; +Cc: ilpo.jarvinen, netdev, Yuchung Cheng
In-Reply-To: <1362603130-16501-1-git-send-email-ycheng@google.com>
Consolidate all of TCP CA_Loss state processing in
tcp_fastretrans_alert() into a new function called tcp_process_loss().
This is to prepare the new F-RTO implementation in the next patch.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
---
net/ipv4/tcp_input.c | 42 ++++++++++++++++++++++++++----------------
1 file changed, 26 insertions(+), 16 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 9933b6b..07a05e0 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -2662,6 +2662,30 @@ static void tcp_enter_recovery(struct sock *sk, bool ece_ack)
tcp_set_ca_state(sk, TCP_CA_Recovery);
}
+/* Process an ACK in CA_Loss state. Move to CA_Open if lost data are
+ * recovered or spurious. Otherwise retransmits more on partial ACKs.
+ */
+static void tcp_process_loss(struct sock *sk, int flag)
+{
+ struct inet_connection_sock *icsk = inet_csk(sk);
+ struct tcp_sock *tp = tcp_sk(sk);
+
+ if (!before(tp->snd_una, tp->high_seq)) {
+ icsk->icsk_retransmits = 0;
+ tcp_try_undo_recovery(sk);
+ return;
+ }
+
+ if (flag & FLAG_DATA_ACKED)
+ icsk->icsk_retransmits = 0;
+ if (tcp_is_reno(tp) && flag & FLAG_SND_UNA_ADVANCED)
+ tcp_reset_reno_sack(tp);
+ if (tcp_try_undo_loss(sk))
+ return;
+ tcp_moderate_cwnd(tp);
+ tcp_xmit_retransmit_queue(sk);
+}
+
/* Process an event, which can update packets-in-flight not trivially.
* Main goal of this function is to calculate new estimate for left_out,
* taking into account both packets sitting in receiver's buffer and
@@ -2708,12 +2732,6 @@ static void tcp_fastretrans_alert(struct sock *sk, int pkts_acked,
tp->retrans_stamp = 0;
} else if (!before(tp->snd_una, tp->high_seq)) {
switch (icsk->icsk_ca_state) {
- case TCP_CA_Loss:
- icsk->icsk_retransmits = 0;
- if (tcp_try_undo_recovery(sk))
- return;
- break;
-
case TCP_CA_CWR:
/* CWR is to be held something *above* high_seq
* is ACKed for CWR bit to reach receiver. */
@@ -2744,18 +2762,10 @@ static void tcp_fastretrans_alert(struct sock *sk, int pkts_acked,
newly_acked_sacked = pkts_acked + tp->sacked_out - prior_sacked;
break;
case TCP_CA_Loss:
- if (flag & FLAG_DATA_ACKED)
- icsk->icsk_retransmits = 0;
- if (tcp_is_reno(tp) && flag & FLAG_SND_UNA_ADVANCED)
- tcp_reset_reno_sack(tp);
- if (!tcp_try_undo_loss(sk)) {
- tcp_moderate_cwnd(tp);
- tcp_xmit_retransmit_queue(sk);
- return;
- }
+ tcp_process_loss(sk, flag);
if (icsk->icsk_ca_state != TCP_CA_Open)
return;
- /* Loss is undone; fall through to processing in Open state. */
+ /* Fall through to processing in Open state. */
default:
if (tcp_is_reno(tp)) {
if (flag & FLAG_SND_UNA_ADVANCED)
--
1.8.1.3
^ permalink raw reply related
* Re: destroy_conntrack GPF in 3.7.9
From: David Miller @ 2013-03-06 20:46 UTC (permalink / raw)
To: eric.dumazet; +Cc: davej, netdev, kernel-team
In-Reply-To: <1362588847.15793.180.camel@edumazet-glaptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Wed, 06 Mar 2013 08:54:07 -0800
> tun driver lacks a nf_reset(skb) call
>
> I would try :
>
> diff --git a/drivers/net/tun.c b/drivers/net/tun.c
> index 2c6a22e..b7c457a 100644
> --- a/drivers/net/tun.c
> +++ b/drivers/net/tun.c
> @@ -747,6 +747,8 @@ static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
> goto drop;
> skb_orphan(skb);
>
> + nf_reset(skb);
> +
We just recently fixed a similar bug in the VXLAN driver too.
Eric please submit the TUN fix formally, thanks.
commit 88c4c066c6b4db26dc4909ee94e6bf377e8e8e81
Author: Zang MingJie <zealot0630@gmail.com>
Date: Mon Mar 4 06:07:34 2013 +0000
reset nf before xmit vxlan encapsulated packet
We should reset nf settings bond to the skb as ipip/ipgre do.
If not, the conntrack/nat info bond to the origin packet may continually
redirect the packet to vxlan interface causing a routing loop.
this is the scenario:
VETP VXLAN Gateway
/----\ /---------------\
| | | |
| vx+--+vx --NAT-> eth0+--> Internet
| | | |
\----/ \---------------/
when there are any packet coming from internet to the vetp, there will be lots
of garbage packets coming out the gateway's vxlan interface, but none actually
sent to the physical interface, because they are redirected back to the vxlan
interface in the postrouting chain of NAT rule, and dmesg complains:
Mar 1 21:52:53 debian kernel: [ 8802.997699] Dead loop on virtual device vxlan0, fix it urgently!
Mar 1 21:52:54 debian kernel: [ 8804.004907] Dead loop on virtual device vxlan0, fix it urgently!
Mar 1 21:52:55 debian kernel: [ 8805.012189] Dead loop on virtual device vxlan0, fix it urgently!
Mar 1 21:52:56 debian kernel: [ 8806.020593] Dead loop on virtual device vxlan0, fix it urgently!
the patch should fix the problem
Signed-off-by: Zang MingJie <zealot0630@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index f10e58a..c3e3d29 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -961,6 +961,8 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
iph->ttl = ttl ? : ip4_dst_hoplimit(&rt->dst);
tunnel_ip_select_ident(skb, old_iph, &rt->dst);
+ nf_reset(skb);
+
vxlan_set_owner(dev, skb);
/* See iptunnel_xmit() */
^ permalink raw reply related
* Re: [PATCH net-next] netconf: add the handler to dump entries
From: David Miller @ 2013-03-06 20:42 UTC (permalink / raw)
To: nicolas.dichtel; +Cc: netdev
In-Reply-To: <1362562926-4181-1-git-send-email-nicolas.dichtel@6wind.com>
From: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Date: Wed, 6 Mar 2013 10:42:06 +0100
> It's useful to be able to get the initial state of all entries. The patch adds
> the support for IPv4 and IPv6.
>
> Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Applied.
^ permalink raw reply
* Re: [PATCH net-next] r6040: check MDIO register busy waiting result
From: David Miller @ 2013-03-06 20:42 UTC (permalink / raw)
To: florian; +Cc: netdev
In-Reply-To: <1362566492-28176-1-git-send-email-florian@openwrt.org>
From: Florian Fainelli <florian@openwrt.org>
Date: Wed, 6 Mar 2013 11:41:32 +0100
> We are currently busy waiting for MDIO registers to complete their
> operation but we did not propagate the result back to the caller.
> Update r6040_phy_{read,write} to report the busy waiting result
> accordingly.
>
> Signed-off-by: Florian Fainelli <florian@openwrt.org>
Applied.
^ permalink raw reply
* Re: [PATCH net-next] htb: add HTB_DIRECT_QLEN attribute
From: David Miller @ 2013-03-06 20:42 UTC (permalink / raw)
To: eric.dumazet; +Cc: netdev, jhs
In-Reply-To: <1362588561.15793.178.camel@edumazet-glaptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Wed, 06 Mar 2013 08:49:21 -0800
> From: Eric Dumazet <edumazet@google.com>
>
> HTB uses an internal pfifo queue, which limit is not reported
> to userland tools (tc), and value inherited from device tx_queue_len
> at setup time.
>
> Introduce TCA_HTB_DIRECT_QLEN attribute to allow finer control.
>
> Remove two obsolete pr_err() calls as well.
>
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Applied.
^ permalink raw reply
* Re: [PATCH net-next] vmxnet3: prevent div-by-zero panic when ring resizing uninitialized dev
From: David Miller @ 2013-03-06 20:38 UTC (permalink / raw)
To: bhavesh; +Cc: netdev, linux-kernel, pv-drivers, sbhatewara
In-Reply-To: <1362591984-16587-1-git-send-email-bhavesh@vmware.com>
From: Bhavesh Davda <bhavesh@vmware.com>
Date: Wed, 6 Mar 2013 09:46:24 -0800
> Linux is free to call ethtool ops as soon as a netdev exists when probe
> finishes. However, we only allocate vmxnet3 tx/rx queues and initialize the
> rx_buf_per_pkt field in struct vmxnet3_adapter when the interface is
> opened (UP).
>
> Signed-off-by: Bhavesh Davda <bhavesh@vmware.com>
> Signed-off-by: Shreyas N Bhatewara <sbhatewara@vmware.com>
> ---
> drivers/net/vmxnet3/vmxnet3_drv.c | 3 ++-
> drivers/net/vmxnet3/vmxnet3_ethtool.c | 6 ++++++
> drivers/net/vmxnet3/vmxnet3_int.h | 4 ++--
> 3 files changed, 10 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c b/drivers/net/vmxnet3/vmxnet3_drv.c
> index 4aad350..bfc5898 100644
> --- a/drivers/net/vmxnet3/vmxnet3_drv.c
> +++ b/drivers/net/vmxnet3/vmxnet3_drv.c
> @@ -2922,7 +2922,7 @@ vmxnet3_probe_device(struct pci_dev *pdev,
> (int)num_online_cpus());
> else
> #endif
> - num_rx_queues = 1;
> + num_rx_queues = 1;
Why are you un-indenting this else branch? Please don't do that.
^ permalink raw reply
* Re: [PATCH v2 net-next 0/4] Provide default fdb operation to allow mac filter setting
From: Or Gerlitz @ 2013-03-06 20:37 UTC (permalink / raw)
To: David Miller; +Cc: vyasevic, netdev, john.r.fastabend, Amir Vadai, Yan Burman
In-Reply-To: <20130301.162612.2115492006742815404.davem@davemloft.net>
On Fri, Mar 1, 2013 at 11:26 PM, David Miller <davem@davemloft.net> wrote:
> These changes look great, please resubmit them once I open net-next back up.
yep, once this resubmission is done, we'd like to test the patches
before actually acknowledging them for net-next
Or.
^ permalink raw reply
* Re: pull request: wireless 2013-03-06
From: David Miller @ 2013-03-06 20:37 UTC (permalink / raw)
To: linville; +Cc: linux-wireless, netdev, linux-kernel
In-Reply-To: <20130306192912.GA17452@tuxdriver.com>
From: "John W. Linville" <linville@tuxdriver.com>
Date: Wed, 6 Mar 2013 14:29:14 -0500
> As Johannes says, it is a bit large. But I hope it is still early
> enough in the cycle to make that worthwhile.
It's quite sizable, but I guess better during -rc1 than -rc6 or -rc7
:-)
Pulled, thanks John.
^ permalink raw reply
* Re: [PATCH] 8139too: send NETDEV_CHANGE manually when autoneg is disabled
From: Ben Hutchings @ 2013-03-06 20:22 UTC (permalink / raw)
To: Veaceslav Falico; +Cc: netdev, wfp5p, jasowang, junchangwang, greearb, ivecera
In-Reply-To: <1362596784-19443-1-git-send-email-vfalico@redhat.com>
On Wed, 2013-03-06 at 20:06 +0100, Veaceslav Falico wrote:
> When setting autoneg off (with any additional parameters, like
> speed/duplex), 8139too doesn't do an interface reset, and thus doesn't
> notify anyone that its speed/duplex might have changed (bonding and bridge
> will not see the speed changes, per example).
>
> Verify if we've force_media and send notification manually, so that the
> listeners have a chance to see the changes. It's quite ugly, however I
> don't see anything better.
Isn't this really a bug in mii_check_media()? It shouldn't shortcut the
calls to netif_carrier_{off,on}() just because mii->force_media is set.
Ben.
> Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
> ---
> drivers/net/ethernet/realtek/8139too.c | 5 +++++
> 1 files changed, 5 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/net/ethernet/realtek/8139too.c b/drivers/net/ethernet/realtek/8139too.c
> index 1276ac7..96ee18c 100644
> --- a/drivers/net/ethernet/realtek/8139too.c
> +++ b/drivers/net/ethernet/realtek/8139too.c
> @@ -2393,6 +2393,11 @@ static int rtl8139_set_settings(struct net_device *dev, struct ethtool_cmd *cmd)
> spin_lock_irq(&tp->lock);
> rc = mii_ethtool_sset(&tp->mii, cmd);
> spin_unlock_irq(&tp->lock);
> + /*
> + * we don't restart on autoneg off, so notify manually
> + */
> + if (tp->mii.force_media)
> + call_netdevice_notifiers(NETDEV_CHANGE, dev);
> return rc;
> }
>
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [PATCH net-next 06/12] ipvs: optimize dst usage for real server
From: David Miller @ 2013-03-06 20:18 UTC (permalink / raw)
To: ja; +Cc: horms, lvs-devel, netdev
In-Reply-To: <1362559342-18784-7-git-send-email-ja@ssi.bg>
From: Julian Anastasov <ja@ssi.bg>
Date: Wed, 6 Mar 2013 10:42:16 +0200
> Currently when forwarding requests to real servers
> we use dst_lock and atomic operations when cloning the
> dst_cache value. As the dst_cache value does not change
> most of the time it is better to use RCU and to lock
> dst_lock only when we need to replace the obsoleted dst.
> For this to work we keep dst_cache in new structure protected
> by RCU. For packets to remote real servers we will use noref
> version of dst_cache, it will be valid while we are in RCU
> read-side critical section because now dst_release for replaced
> dsts will be invoked after the grace period. NAT-ed packets
> via loopback that are not sent but are passed to local stack
> with NF_ACCEPT need a dst clone (skb_dst_force).
>
> Signed-off-by: Julian Anastasov <ja@ssi.bg>
It think it's much cleaner to have the callers of route lookups do
skb_set_dst_noref().
Then you don't need any new interfaces, and you therefore don't need
to expose bits of the noref implementation and semantics at all.
^ permalink raw reply
* Re: [PATCH net-next] bond: add support to read speed and duplex via ethtool
From: Ben Hutchings @ 2013-03-06 20:17 UTC (permalink / raw)
To: Andy Gospodarek; +Cc: netdev
In-Reply-To: <1362595173-11442-1-git-send-email-andy@greyhouse.net>
On Wed, 2013-03-06 at 13:39 -0500, Andy Gospodarek wrote:
> This patch adds support for the get_settings ethtool op to the bonding
> driver. This was motivated by users who wanted to get the speed of the
> bond and compare that against throughput to understand utilization.
> The behavior before this patch was added was problematic when computing
> line utilization after trying to get link-speed and throughput via SNMP.
>
> The general plan for computing link-speed was as follows:
>
> Mode Formula
> ---- -------
> active-backup speed of current active slave
> broadcast speed of first slave with known speed
> all other modes aggregate speed of all slaves with known speed
>
> Output from ethtool looks like this for a round-robin bond:
>
> Settings for bond0:
> Supported ports: [ ]
> Supported link modes: Not reported
> Supported pause frame use: No
> Supports auto-negotiation: No
> Advertised link modes: Not reported
> Advertised pause frame use: No
> Advertised auto-negotiation: No
> Speed: 11000Mb/s
> Duplex: Full
> Port: Twisted Pair
I think port should be PORT_OTHER. Or perhaps you could ask the current
slave, if using active-backup. (Tricky since you can't do that while
holding the bond rwlock.)
> PHYAD: 0
I don't think we have a properly defined 'not applicable' value for
phy_address, though I've used 0xff in sfc.
> Transceiver: internal
> Auto-negotiation: off
> MDI-X: Unknown
> Link detected: yes
>
> I tested this and verified it works as expected. A test was also done
> on a version backported to an older kernel and it worked well there.
>
> Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
> ---
> drivers/net/bonding/bond_main.c | 47 +++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 47 insertions(+)
>
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index 7bd068a..6e70ff0 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -4224,6 +4224,52 @@ void bond_set_mode_ops(struct bonding *bond, int mode)
> }
> }
>
> +static int bond_ethtool_get_settings(struct net_device *bond_dev,
> + struct ethtool_cmd *ecmd)
> +{
> + struct bonding *bond = netdev_priv(bond_dev);
> + struct slave *slave;
> + int i;
> + unsigned long speed = 0;
> +
> + ecmd->speed = SPEED_UNKNOWN;
[...]
You must use ethtool_cmd_speed_set() instead of assigning just
ecmd->speed.
(bonding appears to be the first driver that could report a value >
65535.)
Ben.
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [PATCH net-next] bond: add support to read speed and duplex via ethtool
From: Andy Gospodarek @ 2013-03-06 20:01 UTC (permalink / raw)
To: Jay Vosburgh; +Cc: netdev
In-Reply-To: <22416.1362597912@death.nxdomain>
On Wed, Mar 06, 2013 at 11:25:12AM -0800, Jay Vosburgh wrote:
> Andy Gospodarek <andy@greyhouse.net> wrote:
>
> >This patch adds support for the get_settings ethtool op to the bonding
> >driver. This was motivated by users who wanted to get the speed of the
> >bond and compare that against throughput to understand utilization.
> >The behavior before this patch was added was problematic when computing
> >line utilization after trying to get link-speed and throughput via SNMP.
> >
> >The general plan for computing link-speed was as follows:
> >
> >Mode Formula
> >---- -------
> >active-backup speed of current active slave
> >broadcast speed of first slave with known speed
> >all other modes aggregate speed of all slaves with known speed
>
> I'll just point out that the balance-tlb mode is asymmetric; it
> uses all slaves for transmission, but only one slave for reception.
> Ethtool only has a single speed for both directions, so this is probably
> the best choice, but it should still be noted.
Thanks for pointing that out. I have a feeling there will be a v2, so
I'll try and update the changelog to reflect that. For the record, this
same limitation exists when connecting to most switches and using
round-robin, so I didn't feel the need to differentiate possibly
asymmetric speeds.
> >Output from ethtool looks like this for a round-robin bond:
> >
> >Settings for bond0:
> > Supported ports: [ ]
> > Supported link modes: Not reported
> > Supported pause frame use: No
> > Supports auto-negotiation: No
> > Advertised link modes: Not reported
> > Advertised pause frame use: No
> > Advertised auto-negotiation: No
> > Speed: 11000Mb/s
> > Duplex: Full
> > Port: Twisted Pair
> > PHYAD: 0
> > Transceiver: internal
> > Auto-negotiation: off
> > MDI-X: Unknown
> > Link detected: yes
> >
> >I tested this and verified it works as expected. A test was also done
> >on a version backported to an older kernel and it worked well there.
> >
> >Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
> >---
> > drivers/net/bonding/bond_main.c | 47 +++++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 47 insertions(+)
> >
> >diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> >index 7bd068a..6e70ff0 100644
> >--- a/drivers/net/bonding/bond_main.c
> >+++ b/drivers/net/bonding/bond_main.c
> >@@ -4224,6 +4224,52 @@ void bond_set_mode_ops(struct bonding *bond, int mode)
> > }
> > }
> >
> >+static int bond_ethtool_get_settings(struct net_device *bond_dev,
> >+ struct ethtool_cmd *ecmd)
> >+{
> >+ struct bonding *bond = netdev_priv(bond_dev);
> >+ struct slave *slave;
> >+ int i;
> >+ unsigned long speed = 0;
> >+
> >+ ecmd->speed = SPEED_UNKNOWN;
> >+ ecmd->duplex = DUPLEX_UNKNOWN;
> >+
> >+ read_lock(&bond->lock);
> >+ switch (bond->params.mode) {
> >+ case BOND_MODE_ACTIVEBACKUP:
> >+ read_lock(&bond->curr_slave_lock);
> >+ if (bond->curr_active_slave &&
> >+ bond->curr_active_slave->speed != SPEED_UNKNOWN) {
> >+ ecmd->speed = bond->curr_active_slave->speed;
> >+ ecmd->duplex = bond->curr_active_slave->duplex;
> >+ }
> >+ read_unlock(&bond->curr_slave_lock);
> >+ break;
> >+ case BOND_MODE_BROADCAST:
> >+ bond_for_each_slave(bond, slave, i) {
> >+ if (slave->speed != SPEED_UNKNOWN) {
> >+ ecmd->speed = slave->speed;
> >+ ecmd->duplex = slave->duplex;
> >+ break;
> >+ }
> >+ }
> >+ break;
>
> Does anybody really use broadcast mode? Not that I'm saying
> this is incorrect, I'm just wondering in general.
>
I don't imagine they do, but wanted to add something for it since it
would not reallyu fall into the default case well.
> >+ default:
> >+ bond_for_each_slave(bond, slave, i) {
> >+ if (slave->speed != SPEED_UNKNOWN) {
> >+ speed += slave->speed;
> >+ }
> >+ if (ecmd->duplex == DUPLEX_UNKNOWN &&
> >+ slave->duplex != DUPLEX_UNKNOWN)
> >+ ecmd->duplex = slave->duplex;
>
> Should the calculations skip slaves that are not BOND_LINK_UP?
> If the ARP monitor is running, some slaves may be carrier up (and have
> slave->speed set), but are not actually in use by the bond, at least for
> transmission.
>
That would be fine with me. If you would like I can add that for a v2.
It would produce a more honest estimate of what the maximum throughput
would be at that point in time.
> -J
>
> >+ }
> >+ ecmd->speed = speed;
> >+ }
> >+ read_unlock(&bond->lock);
> >+ return 0;
> >+}
> >+
> > static void bond_ethtool_get_drvinfo(struct net_device *bond_dev,
> > struct ethtool_drvinfo *drvinfo)
> > {
> >@@ -4235,6 +4281,7 @@ static void bond_ethtool_get_drvinfo(struct net_device *bond_dev,
> >
> > static const struct ethtool_ops bond_ethtool_ops = {
> > .get_drvinfo = bond_ethtool_get_drvinfo,
> >+ .get_settings = bond_ethtool_get_settings,
> > .get_link = ethtool_op_get_link,
> > };
> >
> >--
> >1.7.11.7
>
> ---
> -Jay Vosburgh, IBM Linux Technology Center, fubar@us.ibm.com
>
> --
> 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
^ permalink raw reply
* Re: [PATCH net-next 2/2] bnx2x: use the default NAPI weight
From: David Miller @ 2013-03-06 19:59 UTC (permalink / raw)
To: eric.dumazet; +Cc: netdev, eilong, jhs, herbert
In-Reply-To: <1362553398.15793.168.camel@edumazet-glaptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Tue, 05 Mar 2013 23:03:18 -0800
> On Tue, 2013-03-05 at 23:37 -0500, David Miller wrote:
>
>> Thanks for the explanation.
>>
>> Since you haven't completely resolved the issues you were running into
>> I'll target this to net-next for now.
>
> Thanks David
>
> An other issue is the spin_trylock() attempted in net_tx_action()
>
> It seems we can miss a qdisc_run(), and have to wait the following
> NET_TX softirq(s) to send more data. NET_RX being interleaved, we can
> have to wait a long time (not mentioning other softirq handlers like
> RCU ...)
>
> I might be too tired right now, but cant see the reason of the trylock.
>
> qdisc lock is already BH safe, so we should do a spinlock
...
> @@ -3201,22 +3201,11 @@ static void net_tx_action(struct softirq_action *h)
> head = head->next_sched;
>
> root_lock = qdisc_lock(q);
> - if (spin_trylock(root_lock)) {
> - smp_mb__before_clear_bit();
> - clear_bit(__QDISC_STATE_SCHED,
> - &q->state);
> - qdisc_run(q);
> - spin_unlock(root_lock);
I think this trylock is intentional, but not to deal with BH safeness,
but rather to allow another cpu already processing the qdisc to
continue doing so.
I think this is what Jamal's amazing flash animations back at netconf
in Toronto were all about :-)
Herbert Xu and Jamal have touched upon this issue several times in
the past.
^ permalink raw reply
* Re: [PATCH] be2net: use CSR-BAR SEMAPHORE reg for BE2/BE3
From: David Miller @ 2013-03-06 19:57 UTC (permalink / raw)
To: sathya.perla; +Cc: netdev
In-Reply-To: <eb9ec39f-6c12-4f1d-ae31-9f1e9ba76848@CMEXHTCAS1.ad.emulex.com>
From: Sathya Perla <sathya.perla@emulex.com>
Date: Wed, 6 Mar 2013 13:53:20 +0530
> The SLIPORT_SEMAPHORE register shadowed in the
> config-space may not reflect the correct POST stage after
> an EEH reset in BE2/3; it may return FW_READY state even though
> FW is not ready. This causes the driver to prematurely
> poll the FW mailbox and fail.
>
> For BE2/3 use the CSR-BAR/0xac instead.
>
> Reported-by: Gavin Shan <shangw@linux.vnet.ibm.com>
> Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Applied, thanks.
^ permalink raw reply
* Re: [V2 PATCH] net: docs: document multiqueue tuntap API
From: David Miller @ 2013-03-06 19:56 UTC (permalink / raw)
To: jasowang; +Cc: maxk, rob, linux-doc, netdev, linux-kernel, mst, hkchu
In-Reply-To: <1362546626-29119-1-git-send-email-jasowang@redhat.com>
From: Jason Wang <jasowang@redhat.com>
Date: Wed, 6 Mar 2013 13:10:26 +0800
> Signed-off-by: Jason Wang <jasowang@redhat.com>
Applied, thanks Jason.
^ permalink raw reply
* Re: [patch net-next,repost 0/2] team: introduce random mode
From: David Miller @ 2013-03-06 19:55 UTC (permalink / raw)
To: jiri; +Cc: netdev, edumazet, fbl
In-Reply-To: <1362569473-4310-1-git-send-email-jiri@resnulli.us>
From: Jiri Pirko <jiri@resnulli.us>
Date: Wed, 6 Mar 2013 12:31:11 +0100
> Jiri Pirko (2):
> team: introduce two default team_modeop functions and use them in
> modes
> team: introduce random mode
Both applied, thanks Jiri.
^ permalink raw reply
* Re: [PATCH net-next] bond: add support to read speed and duplex via ethtool
From: Andy Gospodarek @ 2013-03-06 19:48 UTC (permalink / raw)
To: Veaceslav Falico; +Cc: netdev
In-Reply-To: <CAE3RKN=d2WmY9Sf-fOuWuTJk+O9_OxZeKPTfjMwWS3iQh2B7ig@mail.gmail.com>
On Wed, Mar 06, 2013 at 08:13:06PM +0100, Veaceslav Falico wrote:
> On Wed, Mar 6, 2013 at 7:39 PM, Andy Gospodarek <andy@greyhouse.net> wrote:
> > This patch adds support for the get_settings ethtool op to the bonding
> > driver. This was motivated by users who wanted to get the speed of the
> > bond and compare that against throughput to understand utilization.
> > The behavior before this patch was added was problematic when computing
> > line utilization after trying to get link-speed and throughput via SNMP.
> >
> > The general plan for computing link-speed was as follows:
> >
> > Mode Formula
> > ---- -------
> > active-backup speed of current active slave
> > broadcast speed of first slave with known speed
> > all other modes aggregate speed of all slaves with known speed
> >
> > Output from ethtool looks like this for a round-robin bond:
> >
> > Settings for bond0:
> > Supported ports: [ ]
> > Supported link modes: Not reported
> > Supported pause frame use: No
> > Supports auto-negotiation: No
> > Advertised link modes: Not reported
> > Advertised pause frame use: No
> > Advertised auto-negotiation: No
> > Speed: 11000Mb/s
> > Duplex: Full
> > Port: Twisted Pair
> > PHYAD: 0
> > Transceiver: internal
> > Auto-negotiation: off
> > MDI-X: Unknown
> > Link detected: yes
> >
> > I tested this and verified it works as expected. A test was also done
> > on a version backported to an older kernel and it worked well there.
> >
> > Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
> > ---
> > drivers/net/bonding/bond_main.c | 47 +++++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 47 insertions(+)
> >
> > diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> > index 7bd068a..6e70ff0 100644
> > --- a/drivers/net/bonding/bond_main.c
> > +++ b/drivers/net/bonding/bond_main.c
> > @@ -4224,6 +4224,52 @@ void bond_set_mode_ops(struct bonding *bond, int mode)
> > }
> > }
> >
> > +static int bond_ethtool_get_settings(struct net_device *bond_dev,
> > + struct ethtool_cmd *ecmd)
> > +{
> > + struct bonding *bond = netdev_priv(bond_dev);
> > + struct slave *slave;
> > + int i;
> > + unsigned long speed = 0;
> > +
> > + ecmd->speed = SPEED_UNKNOWN;
> > + ecmd->duplex = DUPLEX_UNKNOWN;
> > +
> > + read_lock(&bond->lock);
> > + switch (bond->params.mode) {
> > + case BOND_MODE_ACTIVEBACKUP:
> > + read_lock(&bond->curr_slave_lock);
> > + if (bond->curr_active_slave &&
> > + bond->curr_active_slave->speed != SPEED_UNKNOWN) {
> > + ecmd->speed = bond->curr_active_slave->speed;
> > + ecmd->duplex = bond->curr_active_slave->duplex;
> > + }
> > + read_unlock(&bond->curr_slave_lock);
> > + break;
> > + case BOND_MODE_BROADCAST:
> > + bond_for_each_slave(bond, slave, i) {
> > + if (slave->speed != SPEED_UNKNOWN) {
> > + ecmd->speed = slave->speed;
> > + ecmd->duplex = slave->duplex;
>
> Maybe it should be 'the smallest speed (and then duplex)' here? In
> broadcast we send all packets
> through all the slaves, so the effective speed/duplex would be the
> speed/duplex of the slowest slave.
My read of the transmit side of bonding is that broadcast mode will
send frames to all drivers via dev_queue_xmit without checking return
codes, so frames on faster interfaces will be send while those on slower
ones will be dropped in the driver, dropped in hardware or reqeued.
If sending UDP traffic frames would still go out at the max rate of one
of the interfaces. This means there is a chance that a mixed bond (a
10Gbps interface with a 1Gbps interface) could both send at 100%
capacity even though one was an order of magnitude faster than another.
I considered setting to either the max speed or minimum speed, but
decided against it as I really could not decide which was better. It
would work just fine to use broadcast mode with different speeds, but I
decided that reporting only the speed of the first interface that was up
would provide an additional way for admins to notice there might be a
problem in their configuration.
I'm happy to do either if others would like to see that and we can come
up with a good reason to use max over minimum.
>
> > + break;
> > + }
> > + }
> > + break;
> > + default:
> > + bond_for_each_slave(bond, slave, i) {
> > + if (slave->speed != SPEED_UNKNOWN) {
> > + speed += slave->speed;
> > + }
> > + if (ecmd->duplex == DUPLEX_UNKNOWN &&
> > + slave->duplex != DUPLEX_UNKNOWN)
> > + ecmd->duplex = slave->duplex;
> > + }
> > + ecmd->speed = speed;
> > + }
> > + read_unlock(&bond->lock);
> > + return 0;
> > +}
> > +
> > static void bond_ethtool_get_drvinfo(struct net_device *bond_dev,
> > struct ethtool_drvinfo *drvinfo)
> > {
> > @@ -4235,6 +4281,7 @@ static void bond_ethtool_get_drvinfo(struct net_device *bond_dev,
> >
> > static const struct ethtool_ops bond_ethtool_ops = {
> > .get_drvinfo = bond_ethtool_get_drvinfo,
> > + .get_settings = bond_ethtool_get_settings,
> > .get_link = ethtool_op_get_link,
> > };
> >
> > --
> > 1.7.11.7
> >
> > --
> > 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
>
>
>
> --
> Best regards,
> Veaceslav Falico
> --
> 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
^ permalink raw reply
* Re: Pull request: sfc 2013-03-06
From: David Miller @ 2013-03-06 19:51 UTC (permalink / raw)
To: bhutchings; +Cc: linux-net-drivers, scrum-linux, netdev
In-Reply-To: <1362597261.3126.11.camel@bwh-desktop.uk.solarflarecom.com>
From: Ben Hutchings <bhutchings@solarflare.com>
Date: Wed, 6 Mar 2013 19:14:21 +0000
> The following changes since commit aab2b4bf224ef8358d262f95b568b8ad0cecf0a0:
>
> tcp: fix double-counted receiver RTT when leaving receiver fast path (2013-03-04 14:12:07 -0500)
>
> are available in the git repository at:
> git://git.kernel.org/pub/scm/linux/kernel/git/bwh/sfc.git sfc-3.9
>
> Fix regressions introduced by the last set of fixes (sorry):
>
> 1. Potential deadlock when disabling TX queues.
> 2. RX was broken on architectures other than x86 and powerpc.
>
> I still expect to send one more bug fix for 3.9, but as it sometimes
> takes days to reproduce the bug it's going to take a couple of weeks of
> testing to be confident that it's really fixed.
Pulled, thanks Ben.
^ permalink raw reply
* Re: Softirqs without captured packets
From: David Miller @ 2013-03-06 19:40 UTC (permalink / raw)
To: javierdo1; +Cc: eric.dumazet, netdev
In-Reply-To: <CALZVapniE32NfuOjmfDd7DmVHxGLmMTunJu2CxBHRsyoE1Kqaw@mail.gmail.com>
From: Javier Domingo <javierdo1@gmail.com>
Date: Wed, 6 Mar 2013 18:20:35 +0100
> Another question, that is a little related, in tg3, it deactivates
> napi, when it has a TX_RECOVERY_PENDING, what does that flag mean?
It means that we are resetting the device.
^ permalink raw reply
* pull request: wireless 2013-03-06
From: John W. Linville @ 2013-03-06 19:29 UTC (permalink / raw)
To: davem; +Cc: linux-wireless, netdev, linux-kernel
[-- Attachment #1: Type: text/plain, Size: 53016 bytes --]
Dave,
This time just passing along a big batch of fixes from Johannes...
For the mac80211 bits:
"Here I have fixes from Ben Greear for stray work items when deleting
interfaces, another idle handling fix from Felix, a fix from Marco ro a
mesh PS buffering crash and I have a fix for the VHT MCS calculation in
association request frames and more nl80211 feature advertising removal
as well as a workaround to increase the dump size if the SKB overhead is
too large. For 3.10 I already have a complete fix queued, but that also
requires (simple) userspace changes."
And for the iwlwifi bits:
"The patches from Dor fix a bunch of calibration issues in the new MVM
driver, and Emmanuel has a number of fixes there as well. Also, we
decided to disable 8k A-MSDU by default, so that's in there. My own
patches are addressing an issue we found with the new devices but that
seems to also exist on older ones, the DMA writeback the devices do can
be delayed and cause issues. The fix is unfortunately relatively large
and depends on two other changes (to not be hugely conflicting), but I
think it's still worth it at this point."
As Johannes says, it is a bit large. But I hope it is still early
enough in the cycle to make that worthwhile.
Please let me know if there are problems!
John
---
The following changes since commit 66d29cbc59433ba538922a9e958495156b31b83b:
benet: Wait f/w POST until timeout (2013-03-06 02:47:06 -0500)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless.git for-davem
for you to fetch changes up to 32cdd592b723fc88ecca699e550197cd48bb4ad6:
Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless into for-davem (2013-03-06 10:21:17 -0500)
----------------------------------------------------------------
Ben Greear (2):
mac80211: stop timers before canceling work items
mac80211: Fix crash due to un-canceled work-items
Dor Shaish (3):
iwlwifi: mvm: Set valid TX antennas value before calib request
iwlwifi: mvm: Take the phy_cfg from the TLV value
iwlwifi: mvm: Remove overriding calibrations for the 7000 family
Emmanuel Grumbach (5):
iwlwifi: disable 8K A-MSDU by default
iwlwifi: mvm: restart the NIC of the cmd queue gets full
iwlwifi: mvm: ignore STOP_AGG when restarting
iwlwifi: mvm: update the rssi calculation
iwlwifi: mvm: don't warn on normal BAR sending
Felix Fietkau (1):
mac80211: another fix for idle handling in monitor mode
Johannes Berg (7):
nl80211: remove channel width and extended capa advertising
mac80211: really fix monitor mode channel reporting
iwlwifi: don't map complete commands bidirectionally
iwlwifi: rename IWL_MAX_CMD_TFDS to IWL_MAX_CMD_TBS_PER_TFD
iwlwifi: use coherent DMA memory for command header
nl80211: increase wiphy dump size dynamically
mac80211: fix VHT MCS calculation
John W. Linville (3):
Merge branch 'for-john' of git://git.kernel.org/.../jberg/mac80211
Merge branch 'for-john' of git://git.kernel.org/.../iwlwifi/iwlwifi-fixes
Merge branch 'master' of git://git.kernel.org/.../linville/wireless into for-davem
Marco Porsch (1):
mac80211: fix oops on mesh PS broadcast forwarding
drivers/net/wireless/iwlwifi/dvm/sta.c | 2 +-
drivers/net/wireless/iwlwifi/iwl-devtrace.h | 2 +-
drivers/net/wireless/iwlwifi/iwl-drv.c | 3 +-
drivers/net/wireless/iwlwifi/iwl-modparams.h | 2 +-
drivers/net/wireless/iwlwifi/iwl-trans.h | 20 +-
drivers/net/wireless/iwlwifi/mvm/fw-api.h | 18 +-
drivers/net/wireless/iwlwifi/mvm/fw.c | 133 +-------------
drivers/net/wireless/iwlwifi/mvm/mvm.h | 3 +-
drivers/net/wireless/iwlwifi/mvm/ops.c | 18 +-
drivers/net/wireless/iwlwifi/mvm/rx.c | 37 ++--
drivers/net/wireless/iwlwifi/mvm/sta.c | 10 +
drivers/net/wireless/iwlwifi/mvm/tx.c | 6 +-
drivers/net/wireless/iwlwifi/pcie/internal.h | 34 +++-
drivers/net/wireless/iwlwifi/pcie/rx.c | 14 +-
drivers/net/wireless/iwlwifi/pcie/tx.c | 266 +++++++++++++--------------
net/mac80211/cfg.c | 21 ++-
net/mac80211/iface.c | 6 +
net/mac80211/mlme.c | 28 ++-
net/mac80211/tx.c | 3 +-
net/wireless/core.c | 3 +-
net/wireless/nl80211.c | 51 +++--
21 files changed, 299 insertions(+), 381 deletions(-)
diff --git a/drivers/net/wireless/iwlwifi/dvm/sta.c b/drivers/net/wireless/iwlwifi/dvm/sta.c
index 94ef338..b775769 100644
--- a/drivers/net/wireless/iwlwifi/dvm/sta.c
+++ b/drivers/net/wireless/iwlwifi/dvm/sta.c
@@ -151,7 +151,7 @@ int iwl_send_add_sta(struct iwl_priv *priv,
sta_id, sta->sta.addr, flags & CMD_ASYNC ? "a" : "");
if (!(flags & CMD_ASYNC)) {
- cmd.flags |= CMD_WANT_SKB | CMD_WANT_HCMD;
+ cmd.flags |= CMD_WANT_SKB;
might_sleep();
}
diff --git a/drivers/net/wireless/iwlwifi/iwl-devtrace.h b/drivers/net/wireless/iwlwifi/iwl-devtrace.h
index 10f0179..81aa91f 100644
--- a/drivers/net/wireless/iwlwifi/iwl-devtrace.h
+++ b/drivers/net/wireless/iwlwifi/iwl-devtrace.h
@@ -363,7 +363,7 @@ TRACE_EVENT(iwlwifi_dev_hcmd,
__entry->flags = cmd->flags;
memcpy(__get_dynamic_array(hcmd), hdr, sizeof(*hdr));
- for (i = 0; i < IWL_MAX_CMD_TFDS; i++) {
+ for (i = 0; i < IWL_MAX_CMD_TBS_PER_TFD; i++) {
if (!cmd->len[i])
continue;
memcpy((u8 *)__get_dynamic_array(hcmd) + offset,
diff --git a/drivers/net/wireless/iwlwifi/iwl-drv.c b/drivers/net/wireless/iwlwifi/iwl-drv.c
index 6f228bb..fbfd2d1 100644
--- a/drivers/net/wireless/iwlwifi/iwl-drv.c
+++ b/drivers/net/wireless/iwlwifi/iwl-drv.c
@@ -1102,7 +1102,6 @@ void iwl_drv_stop(struct iwl_drv *drv)
/* shared module parameters */
struct iwl_mod_params iwlwifi_mod_params = {
- .amsdu_size_8K = 1,
.restart_fw = 1,
.plcp_check = true,
.bt_coex_active = true,
@@ -1207,7 +1206,7 @@ MODULE_PARM_DESC(11n_disable,
"disable 11n functionality, bitmap: 1: full, 2: agg TX, 4: agg RX");
module_param_named(amsdu_size_8K, iwlwifi_mod_params.amsdu_size_8K,
int, S_IRUGO);
-MODULE_PARM_DESC(amsdu_size_8K, "enable 8K amsdu size");
+MODULE_PARM_DESC(amsdu_size_8K, "enable 8K amsdu size (default 0)");
module_param_named(fw_restart, iwlwifi_mod_params.restart_fw, int, S_IRUGO);
MODULE_PARM_DESC(fw_restart, "restart firmware in case of error");
diff --git a/drivers/net/wireless/iwlwifi/iwl-modparams.h b/drivers/net/wireless/iwlwifi/iwl-modparams.h
index e5e3a79..2c2a729 100644
--- a/drivers/net/wireless/iwlwifi/iwl-modparams.h
+++ b/drivers/net/wireless/iwlwifi/iwl-modparams.h
@@ -91,7 +91,7 @@ enum iwl_power_level {
* @sw_crypto: using hardware encryption, default = 0
* @disable_11n: disable 11n capabilities, default = 0,
* use IWL_DISABLE_HT_* constants
- * @amsdu_size_8K: enable 8K amsdu size, default = 1
+ * @amsdu_size_8K: enable 8K amsdu size, default = 0
* @restart_fw: restart firmware, default = 1
* @plcp_check: enable plcp health check, default = true
* @wd_disable: enable stuck queue check, default = 0
diff --git a/drivers/net/wireless/iwlwifi/iwl-trans.h b/drivers/net/wireless/iwlwifi/iwl-trans.h
index 8c7bec6..0cac2b7 100644
--- a/drivers/net/wireless/iwlwifi/iwl-trans.h
+++ b/drivers/net/wireless/iwlwifi/iwl-trans.h
@@ -186,19 +186,13 @@ struct iwl_rx_packet {
* @CMD_ASYNC: Return right away and don't want for the response
* @CMD_WANT_SKB: valid only with CMD_SYNC. The caller needs the buffer of the
* response. The caller needs to call iwl_free_resp when done.
- * @CMD_WANT_HCMD: The caller needs to get the HCMD that was sent in the
- * response handler. Chunks flagged by %IWL_HCMD_DFL_NOCOPY won't be
- * copied. The pointer passed to the response handler is in the transport
- * ownership and don't need to be freed by the op_mode. This also means
- * that the pointer is invalidated after the op_mode's handler returns.
* @CMD_ON_DEMAND: This command is sent by the test mode pipe.
*/
enum CMD_MODE {
CMD_SYNC = 0,
CMD_ASYNC = BIT(0),
CMD_WANT_SKB = BIT(1),
- CMD_WANT_HCMD = BIT(2),
- CMD_ON_DEMAND = BIT(3),
+ CMD_ON_DEMAND = BIT(2),
};
#define DEF_CMD_PAYLOAD_SIZE 320
@@ -217,7 +211,11 @@ struct iwl_device_cmd {
#define TFD_MAX_PAYLOAD_SIZE (sizeof(struct iwl_device_cmd))
-#define IWL_MAX_CMD_TFDS 2
+/*
+ * number of transfer buffers (fragments) per transmit frame descriptor;
+ * this is just the driver's idea, the hardware supports 20
+ */
+#define IWL_MAX_CMD_TBS_PER_TFD 2
/**
* struct iwl_hcmd_dataflag - flag for each one of the chunks of the command
@@ -254,15 +252,15 @@ enum iwl_hcmd_dataflag {
* @id: id of the host command
*/
struct iwl_host_cmd {
- const void *data[IWL_MAX_CMD_TFDS];
+ const void *data[IWL_MAX_CMD_TBS_PER_TFD];
struct iwl_rx_packet *resp_pkt;
unsigned long _rx_page_addr;
u32 _rx_page_order;
int handler_status;
u32 flags;
- u16 len[IWL_MAX_CMD_TFDS];
- u8 dataflags[IWL_MAX_CMD_TFDS];
+ u16 len[IWL_MAX_CMD_TBS_PER_TFD];
+ u8 dataflags[IWL_MAX_CMD_TBS_PER_TFD];
u8 id;
};
diff --git a/drivers/net/wireless/iwlwifi/mvm/fw-api.h b/drivers/net/wireless/iwlwifi/mvm/fw-api.h
index 23eebda..2adb61f 100644
--- a/drivers/net/wireless/iwlwifi/mvm/fw-api.h
+++ b/drivers/net/wireless/iwlwifi/mvm/fw-api.h
@@ -762,18 +762,20 @@ struct iwl_phy_context_cmd {
#define IWL_RX_INFO_PHY_CNT 8
#define IWL_RX_INFO_AGC_IDX 1
#define IWL_RX_INFO_RSSI_AB_IDX 2
-#define IWL_RX_INFO_RSSI_C_IDX 3
-#define IWL_OFDM_AGC_DB_MSK 0xfe00
-#define IWL_OFDM_AGC_DB_POS 9
+#define IWL_OFDM_AGC_A_MSK 0x0000007f
+#define IWL_OFDM_AGC_A_POS 0
+#define IWL_OFDM_AGC_B_MSK 0x00003f80
+#define IWL_OFDM_AGC_B_POS 7
+#define IWL_OFDM_AGC_CODE_MSK 0x3fe00000
+#define IWL_OFDM_AGC_CODE_POS 20
#define IWL_OFDM_RSSI_INBAND_A_MSK 0x00ff
-#define IWL_OFDM_RSSI_ALLBAND_A_MSK 0xff00
#define IWL_OFDM_RSSI_A_POS 0
+#define IWL_OFDM_RSSI_ALLBAND_A_MSK 0xff00
+#define IWL_OFDM_RSSI_ALLBAND_A_POS 8
#define IWL_OFDM_RSSI_INBAND_B_MSK 0xff0000
-#define IWL_OFDM_RSSI_ALLBAND_B_MSK 0xff000000
#define IWL_OFDM_RSSI_B_POS 16
-#define IWL_OFDM_RSSI_INBAND_C_MSK 0x00ff
-#define IWL_OFDM_RSSI_ALLBAND_C_MSK 0xff00
-#define IWL_OFDM_RSSI_C_POS 0
+#define IWL_OFDM_RSSI_ALLBAND_B_MSK 0xff000000
+#define IWL_OFDM_RSSI_ALLBAND_B_POS 24
/**
* struct iwl_rx_phy_info - phy info
diff --git a/drivers/net/wireless/iwlwifi/mvm/fw.c b/drivers/net/wireless/iwlwifi/mvm/fw.c
index d3d959d..500f818 100644
--- a/drivers/net/wireless/iwlwifi/mvm/fw.c
+++ b/drivers/net/wireless/iwlwifi/mvm/fw.c
@@ -79,17 +79,8 @@
#define UCODE_VALID_OK cpu_to_le32(0x1)
/* Default calibration values for WkP - set to INIT image w/o running */
-static const u8 wkp_calib_values_bb_filter[] = { 0xbf, 0x00, 0x5f, 0x00, 0x2f,
- 0x00, 0x18, 0x00 };
-static const u8 wkp_calib_values_rx_dc[] = { 0x7f, 0x7f, 0x7f, 0x7f, 0x7f,
- 0x7f, 0x7f, 0x7f };
-static const u8 wkp_calib_values_tx_lo[] = { 0x00, 0x00, 0x00, 0x00 };
-static const u8 wkp_calib_values_tx_iq[] = { 0xff, 0x00, 0xff, 0x00, 0x00,
- 0x00 };
-static const u8 wkp_calib_values_rx_iq[] = { 0xff, 0x00, 0x00, 0x00 };
static const u8 wkp_calib_values_rx_iq_skew[] = { 0x00, 0x00, 0x01, 0x00 };
static const u8 wkp_calib_values_tx_iq_skew[] = { 0x01, 0x00, 0x00, 0x00 };
-static const u8 wkp_calib_values_xtal[] = { 0xd2, 0xd2 };
struct iwl_calib_default_data {
u16 size;
@@ -99,12 +90,7 @@ struct iwl_calib_default_data {
#define CALIB_SIZE_N_DATA(_buf) {.size = sizeof(_buf), .data = &_buf}
static const struct iwl_calib_default_data wkp_calib_default_data[12] = {
- [5] = CALIB_SIZE_N_DATA(wkp_calib_values_rx_dc),
- [6] = CALIB_SIZE_N_DATA(wkp_calib_values_bb_filter),
- [7] = CALIB_SIZE_N_DATA(wkp_calib_values_tx_lo),
- [8] = CALIB_SIZE_N_DATA(wkp_calib_values_tx_iq),
[9] = CALIB_SIZE_N_DATA(wkp_calib_values_tx_iq_skew),
- [10] = CALIB_SIZE_N_DATA(wkp_calib_values_rx_iq),
[11] = CALIB_SIZE_N_DATA(wkp_calib_values_rx_iq_skew),
};
@@ -241,20 +227,6 @@ static int iwl_mvm_load_ucode_wait_alive(struct iwl_mvm *mvm,
return 0;
}
-#define IWL_HW_REV_ID_RAINBOW 0x2
-#define IWL_PROJ_TYPE_LHP 0x5
-
-static u32 iwl_mvm_build_phy_cfg(struct iwl_mvm *mvm)
-{
- struct iwl_nvm_data *data = mvm->nvm_data;
- /* Temp calls to static definitions, will be changed to CSR calls */
- u8 hw_rev_id = IWL_HW_REV_ID_RAINBOW;
- u8 project_type = IWL_PROJ_TYPE_LHP;
-
- return data->radio_cfg_dash | (data->radio_cfg_step << 2) |
- (hw_rev_id << 4) | ((project_type & 0x7f) << 6) |
- (data->valid_tx_ant << 16) | (data->valid_rx_ant << 20);
-}
static int iwl_send_phy_cfg_cmd(struct iwl_mvm *mvm)
{
@@ -262,7 +234,7 @@ static int iwl_send_phy_cfg_cmd(struct iwl_mvm *mvm)
enum iwl_ucode_type ucode_type = mvm->cur_ucode;
/* Set parameters */
- phy_cfg_cmd.phy_cfg = cpu_to_le32(iwl_mvm_build_phy_cfg(mvm));
+ phy_cfg_cmd.phy_cfg = cpu_to_le32(mvm->fw->phy_config);
phy_cfg_cmd.calib_control.event_trigger =
mvm->fw->default_calib[ucode_type].event_trigger;
phy_cfg_cmd.calib_control.flow_trigger =
@@ -275,103 +247,6 @@ static int iwl_send_phy_cfg_cmd(struct iwl_mvm *mvm)
sizeof(phy_cfg_cmd), &phy_cfg_cmd);
}
-/* Starting with the new PHY DB implementation - New calibs are enabled */
-/* Value - 0x405e7 */
-#define IWL_CALIB_DEFAULT_FLOW_INIT (IWL_CALIB_CFG_XTAL_IDX |\
- IWL_CALIB_CFG_TEMPERATURE_IDX |\
- IWL_CALIB_CFG_VOLTAGE_READ_IDX |\
- IWL_CALIB_CFG_DC_IDX |\
- IWL_CALIB_CFG_BB_FILTER_IDX |\
- IWL_CALIB_CFG_LO_LEAKAGE_IDX |\
- IWL_CALIB_CFG_TX_IQ_IDX |\
- IWL_CALIB_CFG_RX_IQ_IDX |\
- IWL_CALIB_CFG_AGC_IDX)
-
-#define IWL_CALIB_DEFAULT_EVENT_INIT 0x0
-
-/* Value 0x41567 */
-#define IWL_CALIB_DEFAULT_FLOW_RUN (IWL_CALIB_CFG_XTAL_IDX |\
- IWL_CALIB_CFG_TEMPERATURE_IDX |\
- IWL_CALIB_CFG_VOLTAGE_READ_IDX |\
- IWL_CALIB_CFG_BB_FILTER_IDX |\
- IWL_CALIB_CFG_DC_IDX |\
- IWL_CALIB_CFG_TX_IQ_IDX |\
- IWL_CALIB_CFG_RX_IQ_IDX |\
- IWL_CALIB_CFG_SENSITIVITY_IDX |\
- IWL_CALIB_CFG_AGC_IDX)
-
-#define IWL_CALIB_DEFAULT_EVENT_RUN (IWL_CALIB_CFG_XTAL_IDX |\
- IWL_CALIB_CFG_TEMPERATURE_IDX |\
- IWL_CALIB_CFG_VOLTAGE_READ_IDX |\
- IWL_CALIB_CFG_TX_PWR_IDX |\
- IWL_CALIB_CFG_DC_IDX |\
- IWL_CALIB_CFG_TX_IQ_IDX |\
- IWL_CALIB_CFG_SENSITIVITY_IDX)
-
-/*
- * Sets the calibrations trigger values that will be sent to the FW for runtime
- * and init calibrations.
- * The ones given in the FW TLV are not correct.
- */
-static void iwl_set_default_calib_trigger(struct iwl_mvm *mvm)
-{
- struct iwl_tlv_calib_ctrl default_calib;
-
- /*
- * WkP FW TLV calib bits are wrong, overwrite them.
- * This defines the dynamic calibrations which are implemented in the
- * uCode both for init(flow) calculation and event driven calibs.
- */
-
- /* Init Image */
- default_calib.event_trigger = cpu_to_le32(IWL_CALIB_DEFAULT_EVENT_INIT);
- default_calib.flow_trigger = cpu_to_le32(IWL_CALIB_DEFAULT_FLOW_INIT);
-
- if (default_calib.event_trigger !=
- mvm->fw->default_calib[IWL_UCODE_INIT].event_trigger)
- IWL_ERR(mvm,
- "Updating the event calib for INIT image: 0x%x -> 0x%x\n",
- mvm->fw->default_calib[IWL_UCODE_INIT].event_trigger,
- default_calib.event_trigger);
- if (default_calib.flow_trigger !=
- mvm->fw->default_calib[IWL_UCODE_INIT].flow_trigger)
- IWL_ERR(mvm,
- "Updating the flow calib for INIT image: 0x%x -> 0x%x\n",
- mvm->fw->default_calib[IWL_UCODE_INIT].flow_trigger,
- default_calib.flow_trigger);
-
- memcpy((void *)&mvm->fw->default_calib[IWL_UCODE_INIT],
- &default_calib, sizeof(struct iwl_tlv_calib_ctrl));
- IWL_ERR(mvm,
- "Setting uCode init calibrations event 0x%x, trigger 0x%x\n",
- default_calib.event_trigger,
- default_calib.flow_trigger);
-
- /* Run time image */
- default_calib.event_trigger = cpu_to_le32(IWL_CALIB_DEFAULT_EVENT_RUN);
- default_calib.flow_trigger = cpu_to_le32(IWL_CALIB_DEFAULT_FLOW_RUN);
-
- if (default_calib.event_trigger !=
- mvm->fw->default_calib[IWL_UCODE_REGULAR].event_trigger)
- IWL_ERR(mvm,
- "Updating the event calib for RT image: 0x%x -> 0x%x\n",
- mvm->fw->default_calib[IWL_UCODE_REGULAR].event_trigger,
- default_calib.event_trigger);
- if (default_calib.flow_trigger !=
- mvm->fw->default_calib[IWL_UCODE_REGULAR].flow_trigger)
- IWL_ERR(mvm,
- "Updating the flow calib for RT image: 0x%x -> 0x%x\n",
- mvm->fw->default_calib[IWL_UCODE_REGULAR].flow_trigger,
- default_calib.flow_trigger);
-
- memcpy((void *)&mvm->fw->default_calib[IWL_UCODE_REGULAR],
- &default_calib, sizeof(struct iwl_tlv_calib_ctrl));
- IWL_ERR(mvm,
- "Setting uCode runtime calibs event 0x%x, trigger 0x%x\n",
- default_calib.event_trigger,
- default_calib.flow_trigger);
-}
-
static int iwl_set_default_calibrations(struct iwl_mvm *mvm)
{
u8 cmd_raw[16]; /* holds the variable size commands */
@@ -446,8 +321,10 @@ int iwl_run_init_mvm_ucode(struct iwl_mvm *mvm, bool read_nvm)
ret = iwl_nvm_check_version(mvm->nvm_data, mvm->trans);
WARN_ON(ret);
- /* Override the calibrations from TLV and the const of fw */
- iwl_set_default_calib_trigger(mvm);
+ /* Send TX valid antennas before triggering calibrations */
+ ret = iwl_send_tx_ant_cfg(mvm, mvm->nvm_data->valid_tx_ant);
+ if (ret)
+ goto error;
/* WkP doesn't have all calibrations, need to set default values */
if (mvm->cfg->device_family == IWL_DEVICE_FAMILY_7000) {
diff --git a/drivers/net/wireless/iwlwifi/mvm/mvm.h b/drivers/net/wireless/iwlwifi/mvm/mvm.h
index 537711b..bdae700 100644
--- a/drivers/net/wireless/iwlwifi/mvm/mvm.h
+++ b/drivers/net/wireless/iwlwifi/mvm/mvm.h
@@ -80,7 +80,8 @@
#define IWL_INVALID_MAC80211_QUEUE 0xff
#define IWL_MVM_MAX_ADDRESSES 2
-#define IWL_RSSI_OFFSET 44
+/* RSSI offset for WkP */
+#define IWL_RSSI_OFFSET 50
enum iwl_mvm_tx_fifo {
IWL_MVM_TX_FIFO_BK = 0,
diff --git a/drivers/net/wireless/iwlwifi/mvm/ops.c b/drivers/net/wireless/iwlwifi/mvm/ops.c
index aa59adf..d0f9c1e 100644
--- a/drivers/net/wireless/iwlwifi/mvm/ops.c
+++ b/drivers/net/wireless/iwlwifi/mvm/ops.c
@@ -624,12 +624,8 @@ static void iwl_mvm_free_skb(struct iwl_op_mode *op_mode, struct sk_buff *skb)
ieee80211_free_txskb(mvm->hw, skb);
}
-static void iwl_mvm_nic_error(struct iwl_op_mode *op_mode)
+static void iwl_mvm_nic_restart(struct iwl_mvm *mvm)
{
- struct iwl_mvm *mvm = IWL_OP_MODE_GET_MVM(op_mode);
-
- iwl_mvm_dump_nic_error_log(mvm);
-
iwl_abort_notification_waits(&mvm->notif_wait);
/*
@@ -663,9 +659,21 @@ static void iwl_mvm_nic_error(struct iwl_op_mode *op_mode)
}
}
+static void iwl_mvm_nic_error(struct iwl_op_mode *op_mode)
+{
+ struct iwl_mvm *mvm = IWL_OP_MODE_GET_MVM(op_mode);
+
+ iwl_mvm_dump_nic_error_log(mvm);
+
+ iwl_mvm_nic_restart(mvm);
+}
+
static void iwl_mvm_cmd_queue_full(struct iwl_op_mode *op_mode)
{
+ struct iwl_mvm *mvm = IWL_OP_MODE_GET_MVM(op_mode);
+
WARN_ON(1);
+ iwl_mvm_nic_restart(mvm);
}
static const struct iwl_op_mode_ops iwl_mvm_ops = {
diff --git a/drivers/net/wireless/iwlwifi/mvm/rx.c b/drivers/net/wireless/iwlwifi/mvm/rx.c
index 3f40ab0..b0b190d 100644
--- a/drivers/net/wireless/iwlwifi/mvm/rx.c
+++ b/drivers/net/wireless/iwlwifi/mvm/rx.c
@@ -131,33 +131,42 @@ static void iwl_mvm_pass_packet_to_mac80211(struct iwl_mvm *mvm,
static int iwl_mvm_calc_rssi(struct iwl_mvm *mvm,
struct iwl_rx_phy_info *phy_info)
{
- u32 rssi_a, rssi_b, rssi_c, max_rssi, agc_db;
+ int rssi_a, rssi_b, rssi_a_dbm, rssi_b_dbm, max_rssi_dbm;
+ int rssi_all_band_a, rssi_all_band_b;
+ u32 agc_a, agc_b, max_agc;
u32 val;
- /* Find max rssi among 3 possible receivers.
+ /* Find max rssi among 2 possible receivers.
* These values are measured by the Digital Signal Processor (DSP).
* They should stay fairly constant even as the signal strength varies,
* if the radio's Automatic Gain Control (AGC) is working right.
* AGC value (see below) will provide the "interesting" info.
*/
+ val = le32_to_cpu(phy_info->non_cfg_phy[IWL_RX_INFO_AGC_IDX]);
+ agc_a = (val & IWL_OFDM_AGC_A_MSK) >> IWL_OFDM_AGC_A_POS;
+ agc_b = (val & IWL_OFDM_AGC_B_MSK) >> IWL_OFDM_AGC_B_POS;
+ max_agc = max_t(u32, agc_a, agc_b);
+
val = le32_to_cpu(phy_info->non_cfg_phy[IWL_RX_INFO_RSSI_AB_IDX]);
rssi_a = (val & IWL_OFDM_RSSI_INBAND_A_MSK) >> IWL_OFDM_RSSI_A_POS;
rssi_b = (val & IWL_OFDM_RSSI_INBAND_B_MSK) >> IWL_OFDM_RSSI_B_POS;
- val = le32_to_cpu(phy_info->non_cfg_phy[IWL_RX_INFO_RSSI_C_IDX]);
- rssi_c = (val & IWL_OFDM_RSSI_INBAND_C_MSK) >> IWL_OFDM_RSSI_C_POS;
-
- val = le32_to_cpu(phy_info->non_cfg_phy[IWL_RX_INFO_AGC_IDX]);
- agc_db = (val & IWL_OFDM_AGC_DB_MSK) >> IWL_OFDM_AGC_DB_POS;
+ rssi_all_band_a = (val & IWL_OFDM_RSSI_ALLBAND_A_MSK) >>
+ IWL_OFDM_RSSI_ALLBAND_A_POS;
+ rssi_all_band_b = (val & IWL_OFDM_RSSI_ALLBAND_B_MSK) >>
+ IWL_OFDM_RSSI_ALLBAND_B_POS;
- max_rssi = max_t(u32, rssi_a, rssi_b);
- max_rssi = max_t(u32, max_rssi, rssi_c);
+ /*
+ * dBm = rssi dB - agc dB - constant.
+ * Higher AGC (higher radio gain) means lower signal.
+ */
+ rssi_a_dbm = rssi_a - IWL_RSSI_OFFSET - agc_a;
+ rssi_b_dbm = rssi_b - IWL_RSSI_OFFSET - agc_b;
+ max_rssi_dbm = max_t(int, rssi_a_dbm, rssi_b_dbm);
- IWL_DEBUG_STATS(mvm, "Rssi In A %d B %d C %d Max %d AGC dB %d\n",
- rssi_a, rssi_b, rssi_c, max_rssi, agc_db);
+ IWL_DEBUG_STATS(mvm, "Rssi In A %d B %d Max %d AGCA %d AGCB %d\n",
+ rssi_a_dbm, rssi_b_dbm, max_rssi_dbm, agc_a, agc_b);
- /* dBm = max_rssi dB - agc dB - constant.
- * Higher AGC (higher radio gain) means lower signal. */
- return max_rssi - agc_db - IWL_RSSI_OFFSET;
+ return max_rssi_dbm;
}
/*
diff --git a/drivers/net/wireless/iwlwifi/mvm/sta.c b/drivers/net/wireless/iwlwifi/mvm/sta.c
index 861a7f9..274f44e 100644
--- a/drivers/net/wireless/iwlwifi/mvm/sta.c
+++ b/drivers/net/wireless/iwlwifi/mvm/sta.c
@@ -770,6 +770,16 @@ int iwl_mvm_sta_tx_agg_stop(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
u16 txq_id;
int err;
+
+ /*
+ * If mac80211 is cleaning its state, then say that we finished since
+ * our state has been cleared anyway.
+ */
+ if (test_bit(IWL_MVM_STATUS_IN_HW_RESTART, &mvm->status)) {
+ ieee80211_stop_tx_ba_cb_irqsafe(vif, sta->addr, tid);
+ return 0;
+ }
+
spin_lock_bh(&mvmsta->lock);
txq_id = tid_data->txq_id;
diff --git a/drivers/net/wireless/iwlwifi/mvm/tx.c b/drivers/net/wireless/iwlwifi/mvm/tx.c
index 6b67ce3..6645efe 100644
--- a/drivers/net/wireless/iwlwifi/mvm/tx.c
+++ b/drivers/net/wireless/iwlwifi/mvm/tx.c
@@ -607,12 +607,8 @@ static void iwl_mvm_rx_tx_cmd_single(struct iwl_mvm *mvm,
/* Single frame failure in an AMPDU queue => send BAR */
if (txq_id >= IWL_FIRST_AMPDU_QUEUE &&
- !(info->flags & IEEE80211_TX_STAT_ACK)) {
- /* there must be only one skb in the skb_list */
- WARN_ON_ONCE(skb_freed > 1 ||
- !skb_queue_empty(&skbs));
+ !(info->flags & IEEE80211_TX_STAT_ACK))
info->flags |= IEEE80211_TX_STAT_AMPDU_NO_BACK;
- }
/* W/A FW bug: seq_ctl is wrong when the queue is flushed */
if (status == TX_STATUS_FAIL_FIFO_FLUSHED) {
diff --git a/drivers/net/wireless/iwlwifi/pcie/internal.h b/drivers/net/wireless/iwlwifi/pcie/internal.h
index 3d62e80..148843e 100644
--- a/drivers/net/wireless/iwlwifi/pcie/internal.h
+++ b/drivers/net/wireless/iwlwifi/pcie/internal.h
@@ -137,10 +137,6 @@ static inline int iwl_queue_dec_wrap(int index, int n_bd)
struct iwl_cmd_meta {
/* only for SYNC commands, iff the reply skb is wanted */
struct iwl_host_cmd *source;
-
- DEFINE_DMA_UNMAP_ADDR(mapping);
- DEFINE_DMA_UNMAP_LEN(len);
-
u32 flags;
};
@@ -185,25 +181,36 @@ struct iwl_queue {
/*
* The FH will write back to the first TB only, so we need
* to copy some data into the buffer regardless of whether
- * it should be mapped or not. This indicates how much to
- * copy, even for HCMDs it must be big enough to fit the
- * DRAM scratch from the TX cmd, at least 16 bytes.
+ * it should be mapped or not. This indicates how big the
+ * first TB must be to include the scratch buffer. Since
+ * the scratch is 4 bytes at offset 12, it's 16 now. If we
+ * make it bigger then allocations will be bigger and copy
+ * slower, so that's probably not useful.
*/
-#define IWL_HCMD_MIN_COPY_SIZE 16
+#define IWL_HCMD_SCRATCHBUF_SIZE 16
struct iwl_pcie_txq_entry {
struct iwl_device_cmd *cmd;
- struct iwl_device_cmd *copy_cmd;
struct sk_buff *skb;
/* buffer to free after command completes */
const void *free_buf;
struct iwl_cmd_meta meta;
};
+struct iwl_pcie_txq_scratch_buf {
+ struct iwl_cmd_header hdr;
+ u8 buf[8];
+ __le32 scratch;
+};
+
/**
* struct iwl_txq - Tx Queue for DMA
* @q: generic Rx/Tx queue descriptor
* @tfds: transmit frame descriptors (DMA memory)
+ * @scratchbufs: start of command headers, including scratch buffers, for
+ * the writeback -- this is DMA memory and an array holding one buffer
+ * for each command on the queue
+ * @scratchbufs_dma: DMA address for the scratchbufs start
* @entries: transmit entries (driver state)
* @lock: queue lock
* @stuck_timer: timer that fires if queue gets stuck
@@ -217,6 +224,8 @@ struct iwl_pcie_txq_entry {
struct iwl_txq {
struct iwl_queue q;
struct iwl_tfd *tfds;
+ struct iwl_pcie_txq_scratch_buf *scratchbufs;
+ dma_addr_t scratchbufs_dma;
struct iwl_pcie_txq_entry *entries;
spinlock_t lock;
struct timer_list stuck_timer;
@@ -225,6 +234,13 @@ struct iwl_txq {
u8 active;
};
+static inline dma_addr_t
+iwl_pcie_get_scratchbuf_dma(struct iwl_txq *txq, int idx)
+{
+ return txq->scratchbufs_dma +
+ sizeof(struct iwl_pcie_txq_scratch_buf) * idx;
+}
+
/**
* struct iwl_trans_pcie - PCIe transport specific data
* @rxq: all the RX queue data
diff --git a/drivers/net/wireless/iwlwifi/pcie/rx.c b/drivers/net/wireless/iwlwifi/pcie/rx.c
index b0ae06d..567e67a 100644
--- a/drivers/net/wireless/iwlwifi/pcie/rx.c
+++ b/drivers/net/wireless/iwlwifi/pcie/rx.c
@@ -637,22 +637,14 @@ static void iwl_pcie_rx_handle_rb(struct iwl_trans *trans,
index = SEQ_TO_INDEX(sequence);
cmd_index = get_cmd_index(&txq->q, index);
- if (reclaim) {
- struct iwl_pcie_txq_entry *ent;
- ent = &txq->entries[cmd_index];
- cmd = ent->copy_cmd;
- WARN_ON_ONCE(!cmd && ent->meta.flags & CMD_WANT_HCMD);
- } else {
+ if (reclaim)
+ cmd = txq->entries[cmd_index].cmd;
+ else
cmd = NULL;
- }
err = iwl_op_mode_rx(trans->op_mode, &rxcb, cmd);
if (reclaim) {
- /* The original command isn't needed any more */
- kfree(txq->entries[cmd_index].copy_cmd);
- txq->entries[cmd_index].copy_cmd = NULL;
- /* nor is the duplicated part of the command */
kfree(txq->entries[cmd_index].free_buf);
txq->entries[cmd_index].free_buf = NULL;
}
diff --git a/drivers/net/wireless/iwlwifi/pcie/tx.c b/drivers/net/wireless/iwlwifi/pcie/tx.c
index 8b625a7..8595c16 100644
--- a/drivers/net/wireless/iwlwifi/pcie/tx.c
+++ b/drivers/net/wireless/iwlwifi/pcie/tx.c
@@ -191,12 +191,9 @@ static void iwl_pcie_txq_stuck_timer(unsigned long data)
}
for (i = q->read_ptr; i != q->write_ptr;
- i = iwl_queue_inc_wrap(i, q->n_bd)) {
- struct iwl_tx_cmd *tx_cmd =
- (struct iwl_tx_cmd *)txq->entries[i].cmd->payload;
+ i = iwl_queue_inc_wrap(i, q->n_bd))
IWL_ERR(trans, "scratch %d = 0x%08x\n", i,
- get_unaligned_le32(&tx_cmd->scratch));
- }
+ le32_to_cpu(txq->scratchbufs[i].scratch));
iwl_op_mode_nic_error(trans->op_mode);
}
@@ -367,8 +364,8 @@ static inline u8 iwl_pcie_tfd_get_num_tbs(struct iwl_tfd *tfd)
}
static void iwl_pcie_tfd_unmap(struct iwl_trans *trans,
- struct iwl_cmd_meta *meta, struct iwl_tfd *tfd,
- enum dma_data_direction dma_dir)
+ struct iwl_cmd_meta *meta,
+ struct iwl_tfd *tfd)
{
int i;
int num_tbs;
@@ -382,17 +379,12 @@ static void iwl_pcie_tfd_unmap(struct iwl_trans *trans,
return;
}
- /* Unmap tx_cmd */
- if (num_tbs)
- dma_unmap_single(trans->dev,
- dma_unmap_addr(meta, mapping),
- dma_unmap_len(meta, len),
- DMA_BIDIRECTIONAL);
+ /* first TB is never freed - it's the scratchbuf data */
- /* Unmap chunks, if any. */
for (i = 1; i < num_tbs; i++)
dma_unmap_single(trans->dev, iwl_pcie_tfd_tb_get_addr(tfd, i),
- iwl_pcie_tfd_tb_get_len(tfd, i), dma_dir);
+ iwl_pcie_tfd_tb_get_len(tfd, i),
+ DMA_TO_DEVICE);
tfd->num_tbs = 0;
}
@@ -406,8 +398,7 @@ static void iwl_pcie_tfd_unmap(struct iwl_trans *trans,
* Does NOT advance any TFD circular buffer read/write indexes
* Does NOT free the TFD itself (which is within circular buffer)
*/
-static void iwl_pcie_txq_free_tfd(struct iwl_trans *trans, struct iwl_txq *txq,
- enum dma_data_direction dma_dir)
+static void iwl_pcie_txq_free_tfd(struct iwl_trans *trans, struct iwl_txq *txq)
{
struct iwl_tfd *tfd_tmp = txq->tfds;
@@ -418,8 +409,7 @@ static void iwl_pcie_txq_free_tfd(struct iwl_trans *trans, struct iwl_txq *txq,
lockdep_assert_held(&txq->lock);
/* We have only q->n_window txq->entries, but we use q->n_bd tfds */
- iwl_pcie_tfd_unmap(trans, &txq->entries[idx].meta, &tfd_tmp[rd_ptr],
- dma_dir);
+ iwl_pcie_tfd_unmap(trans, &txq->entries[idx].meta, &tfd_tmp[rd_ptr]);
/* free SKB */
if (txq->entries) {
@@ -479,6 +469,7 @@ static int iwl_pcie_txq_alloc(struct iwl_trans *trans,
{
struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
size_t tfd_sz = sizeof(struct iwl_tfd) * TFD_QUEUE_SIZE_MAX;
+ size_t scratchbuf_sz;
int i;
if (WARN_ON(txq->entries || txq->tfds))
@@ -514,9 +505,25 @@ static int iwl_pcie_txq_alloc(struct iwl_trans *trans,
IWL_ERR(trans, "dma_alloc_coherent(%zd) failed\n", tfd_sz);
goto error;
}
+
+ BUILD_BUG_ON(IWL_HCMD_SCRATCHBUF_SIZE != sizeof(*txq->scratchbufs));
+ BUILD_BUG_ON(offsetof(struct iwl_pcie_txq_scratch_buf, scratch) !=
+ sizeof(struct iwl_cmd_header) +
+ offsetof(struct iwl_tx_cmd, scratch));
+
+ scratchbuf_sz = sizeof(*txq->scratchbufs) * slots_num;
+
+ txq->scratchbufs = dma_alloc_coherent(trans->dev, scratchbuf_sz,
+ &txq->scratchbufs_dma,
+ GFP_KERNEL);
+ if (!txq->scratchbufs)
+ goto err_free_tfds;
+
txq->q.id = txq_id;
return 0;
+err_free_tfds:
+ dma_free_coherent(trans->dev, tfd_sz, txq->tfds, txq->q.dma_addr);
error:
if (txq->entries && txq_id == trans_pcie->cmd_queue)
for (i = 0; i < slots_num; i++)
@@ -565,22 +572,13 @@ static void iwl_pcie_txq_unmap(struct iwl_trans *trans, int txq_id)
struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
struct iwl_txq *txq = &trans_pcie->txq[txq_id];
struct iwl_queue *q = &txq->q;
- enum dma_data_direction dma_dir;
if (!q->n_bd)
return;
- /* In the command queue, all the TBs are mapped as BIDI
- * so unmap them as such.
- */
- if (txq_id == trans_pcie->cmd_queue)
- dma_dir = DMA_BIDIRECTIONAL;
- else
- dma_dir = DMA_TO_DEVICE;
-
spin_lock_bh(&txq->lock);
while (q->write_ptr != q->read_ptr) {
- iwl_pcie_txq_free_tfd(trans, txq, dma_dir);
+ iwl_pcie_txq_free_tfd(trans, txq);
q->read_ptr = iwl_queue_inc_wrap(q->read_ptr, q->n_bd);
}
spin_unlock_bh(&txq->lock);
@@ -610,7 +608,6 @@ static void iwl_pcie_txq_free(struct iwl_trans *trans, int txq_id)
if (txq_id == trans_pcie->cmd_queue)
for (i = 0; i < txq->q.n_window; i++) {
kfree(txq->entries[i].cmd);
- kfree(txq->entries[i].copy_cmd);
kfree(txq->entries[i].free_buf);
}
@@ -619,6 +616,10 @@ static void iwl_pcie_txq_free(struct iwl_trans *trans, int txq_id)
dma_free_coherent(dev, sizeof(struct iwl_tfd) *
txq->q.n_bd, txq->tfds, txq->q.dma_addr);
txq->q.dma_addr = 0;
+
+ dma_free_coherent(dev,
+ sizeof(*txq->scratchbufs) * txq->q.n_window,
+ txq->scratchbufs, txq->scratchbufs_dma);
}
kfree(txq->entries);
@@ -962,7 +963,7 @@ void iwl_trans_pcie_reclaim(struct iwl_trans *trans, int txq_id, int ssn,
iwl_pcie_txq_inval_byte_cnt_tbl(trans, txq);
- iwl_pcie_txq_free_tfd(trans, txq, DMA_TO_DEVICE);
+ iwl_pcie_txq_free_tfd(trans, txq);
}
iwl_pcie_txq_progress(trans_pcie, txq);
@@ -1152,29 +1153,29 @@ static int iwl_pcie_enqueue_hcmd(struct iwl_trans *trans,
void *dup_buf = NULL;
dma_addr_t phys_addr;
int idx;
- u16 copy_size, cmd_size, dma_size;
+ u16 copy_size, cmd_size, scratch_size;
bool had_nocopy = false;
int i;
u32 cmd_pos;
- const u8 *cmddata[IWL_MAX_CMD_TFDS];
- u16 cmdlen[IWL_MAX_CMD_TFDS];
+ const u8 *cmddata[IWL_MAX_CMD_TBS_PER_TFD];
+ u16 cmdlen[IWL_MAX_CMD_TBS_PER_TFD];
copy_size = sizeof(out_cmd->hdr);
cmd_size = sizeof(out_cmd->hdr);
/* need one for the header if the first is NOCOPY */
- BUILD_BUG_ON(IWL_MAX_CMD_TFDS > IWL_NUM_OF_TBS - 1);
+ BUILD_BUG_ON(IWL_MAX_CMD_TBS_PER_TFD > IWL_NUM_OF_TBS - 1);
- for (i = 0; i < IWL_MAX_CMD_TFDS; i++) {
+ for (i = 0; i < IWL_MAX_CMD_TBS_PER_TFD; i++) {
cmddata[i] = cmd->data[i];
cmdlen[i] = cmd->len[i];
if (!cmd->len[i])
continue;
- /* need at least IWL_HCMD_MIN_COPY_SIZE copied */
- if (copy_size < IWL_HCMD_MIN_COPY_SIZE) {
- int copy = IWL_HCMD_MIN_COPY_SIZE - copy_size;
+ /* need at least IWL_HCMD_SCRATCHBUF_SIZE copied */
+ if (copy_size < IWL_HCMD_SCRATCHBUF_SIZE) {
+ int copy = IWL_HCMD_SCRATCHBUF_SIZE - copy_size;
if (copy > cmdlen[i])
copy = cmdlen[i];
@@ -1260,15 +1261,15 @@ static int iwl_pcie_enqueue_hcmd(struct iwl_trans *trans,
/* and copy the data that needs to be copied */
cmd_pos = offsetof(struct iwl_device_cmd, payload);
copy_size = sizeof(out_cmd->hdr);
- for (i = 0; i < IWL_MAX_CMD_TFDS; i++) {
+ for (i = 0; i < IWL_MAX_CMD_TBS_PER_TFD; i++) {
int copy = 0;
if (!cmd->len)
continue;
- /* need at least IWL_HCMD_MIN_COPY_SIZE copied */
- if (copy_size < IWL_HCMD_MIN_COPY_SIZE) {
- copy = IWL_HCMD_MIN_COPY_SIZE - copy_size;
+ /* need at least IWL_HCMD_SCRATCHBUF_SIZE copied */
+ if (copy_size < IWL_HCMD_SCRATCHBUF_SIZE) {
+ copy = IWL_HCMD_SCRATCHBUF_SIZE - copy_size;
if (copy > cmd->len[i])
copy = cmd->len[i];
@@ -1286,50 +1287,38 @@ static int iwl_pcie_enqueue_hcmd(struct iwl_trans *trans,
}
}
- WARN_ON_ONCE(txq->entries[idx].copy_cmd);
-
- /*
- * since out_cmd will be the source address of the FH, it will write
- * the retry count there. So when the user needs to receivce the HCMD
- * that corresponds to the response in the response handler, it needs
- * to set CMD_WANT_HCMD.
- */
- if (cmd->flags & CMD_WANT_HCMD) {
- txq->entries[idx].copy_cmd =
- kmemdup(out_cmd, cmd_pos, GFP_ATOMIC);
- if (unlikely(!txq->entries[idx].copy_cmd)) {
- idx = -ENOMEM;
- goto out;
- }
- }
-
IWL_DEBUG_HC(trans,
"Sending command %s (#%x), seq: 0x%04X, %d bytes at %d[%d]:%d\n",
get_cmd_string(trans_pcie, out_cmd->hdr.cmd),
out_cmd->hdr.cmd, le16_to_cpu(out_cmd->hdr.sequence),
cmd_size, q->write_ptr, idx, trans_pcie->cmd_queue);
- /*
- * If the entire command is smaller than IWL_HCMD_MIN_COPY_SIZE, we must
- * still map at least that many bytes for the hardware to write back to.
- * We have enough space, so that's not a problem.
- */
- dma_size = max_t(u16, copy_size, IWL_HCMD_MIN_COPY_SIZE);
+ /* start the TFD with the scratchbuf */
+ scratch_size = min_t(int, copy_size, IWL_HCMD_SCRATCHBUF_SIZE);
+ memcpy(&txq->scratchbufs[q->write_ptr], &out_cmd->hdr, scratch_size);
+ iwl_pcie_txq_build_tfd(trans, txq,
+ iwl_pcie_get_scratchbuf_dma(txq, q->write_ptr),
+ scratch_size, 1);
+
+ /* map first command fragment, if any remains */
+ if (copy_size > scratch_size) {
+ phys_addr = dma_map_single(trans->dev,
+ ((u8 *)&out_cmd->hdr) + scratch_size,
+ copy_size - scratch_size,
+ DMA_TO_DEVICE);
+ if (dma_mapping_error(trans->dev, phys_addr)) {
+ iwl_pcie_tfd_unmap(trans, out_meta,
+ &txq->tfds[q->write_ptr]);
+ idx = -ENOMEM;
+ goto out;
+ }
- phys_addr = dma_map_single(trans->dev, &out_cmd->hdr, dma_size,
- DMA_BIDIRECTIONAL);
- if (unlikely(dma_mapping_error(trans->dev, phys_addr))) {
- idx = -ENOMEM;
- goto out;
+ iwl_pcie_txq_build_tfd(trans, txq, phys_addr,
+ copy_size - scratch_size, 0);
}
- dma_unmap_addr_set(out_meta, mapping, phys_addr);
- dma_unmap_len_set(out_meta, len, dma_size);
-
- iwl_pcie_txq_build_tfd(trans, txq, phys_addr, copy_size, 1);
-
/* map the remaining (adjusted) nocopy/dup fragments */
- for (i = 0; i < IWL_MAX_CMD_TFDS; i++) {
+ for (i = 0; i < IWL_MAX_CMD_TBS_PER_TFD; i++) {
const void *data = cmddata[i];
if (!cmdlen[i])
@@ -1340,11 +1329,10 @@ static int iwl_pcie_enqueue_hcmd(struct iwl_trans *trans,
if (cmd->dataflags[i] & IWL_HCMD_DFL_DUP)
data = dup_buf;
phys_addr = dma_map_single(trans->dev, (void *)data,
- cmdlen[i], DMA_BIDIRECTIONAL);
+ cmdlen[i], DMA_TO_DEVICE);
if (dma_mapping_error(trans->dev, phys_addr)) {
iwl_pcie_tfd_unmap(trans, out_meta,
- &txq->tfds[q->write_ptr],
- DMA_BIDIRECTIONAL);
+ &txq->tfds[q->write_ptr]);
idx = -ENOMEM;
goto out;
}
@@ -1418,7 +1406,7 @@ void iwl_pcie_hcmd_complete(struct iwl_trans *trans,
cmd = txq->entries[cmd_index].cmd;
meta = &txq->entries[cmd_index].meta;
- iwl_pcie_tfd_unmap(trans, meta, &txq->tfds[index], DMA_BIDIRECTIONAL);
+ iwl_pcie_tfd_unmap(trans, meta, &txq->tfds[index]);
/* Input error checking is done when commands are added to queue. */
if (meta->flags & CMD_WANT_SKB) {
@@ -1597,10 +1585,9 @@ int iwl_trans_pcie_tx(struct iwl_trans *trans, struct sk_buff *skb,
struct iwl_cmd_meta *out_meta;
struct iwl_txq *txq;
struct iwl_queue *q;
- dma_addr_t phys_addr = 0;
- dma_addr_t txcmd_phys;
- dma_addr_t scratch_phys;
- u16 len, firstlen, secondlen;
+ dma_addr_t tb0_phys, tb1_phys, scratch_phys;
+ void *tb1_addr;
+ u16 len, tb1_len, tb2_len;
u8 wait_write_ptr = 0;
__le16 fc = hdr->frame_control;
u8 hdr_len = ieee80211_hdrlen(fc);
@@ -1638,85 +1625,80 @@ int iwl_trans_pcie_tx(struct iwl_trans *trans, struct sk_buff *skb,
cpu_to_le16((u16)(QUEUE_TO_SEQ(txq_id) |
INDEX_TO_SEQ(q->write_ptr)));
+ tb0_phys = iwl_pcie_get_scratchbuf_dma(txq, q->write_ptr);
+ scratch_phys = tb0_phys + sizeof(struct iwl_cmd_header) +
+ offsetof(struct iwl_tx_cmd, scratch);
+
+ tx_cmd->dram_lsb_ptr = cpu_to_le32(scratch_phys);
+ tx_cmd->dram_msb_ptr = iwl_get_dma_hi_addr(scratch_phys);
+
/* Set up first empty entry in queue's array of Tx/cmd buffers */
out_meta = &txq->entries[q->write_ptr].meta;
/*
- * Use the first empty entry in this queue's command buffer array
- * to contain the Tx command and MAC header concatenated together
- * (payload data will be in another buffer).
- * Size of this varies, due to varying MAC header length.
- * If end is not dword aligned, we'll have 2 extra bytes at the end
- * of the MAC header (device reads on dword boundaries).
- * We'll tell device about this padding later.
+ * The second TB (tb1) points to the remainder of the TX command
+ * and the 802.11 header - dword aligned size
+ * (This calculation modifies the TX command, so do it before the
+ * setup of the first TB)
*/
- len = sizeof(struct iwl_tx_cmd) +
- sizeof(struct iwl_cmd_header) + hdr_len;
- firstlen = (len + 3) & ~3;
+ len = sizeof(struct iwl_tx_cmd) + sizeof(struct iwl_cmd_header) +
+ hdr_len - IWL_HCMD_SCRATCHBUF_SIZE;
+ tb1_len = (len + 3) & ~3;
/* Tell NIC about any 2-byte padding after MAC header */
- if (firstlen != len)
+ if (tb1_len != len)
tx_cmd->tx_flags |= TX_CMD_FLG_MH_PAD_MSK;
- /* Physical address of this Tx command's header (not MAC header!),
- * within command buffer array. */
- txcmd_phys = dma_map_single(trans->dev,
- &dev_cmd->hdr, firstlen,
- DMA_BIDIRECTIONAL);
- if (unlikely(dma_mapping_error(trans->dev, txcmd_phys)))
- goto out_err;
- dma_unmap_addr_set(out_meta, mapping, txcmd_phys);
- dma_unmap_len_set(out_meta, len, firstlen);
+ /* The first TB points to the scratchbuf data - min_copy bytes */
+ memcpy(&txq->scratchbufs[q->write_ptr], &dev_cmd->hdr,
+ IWL_HCMD_SCRATCHBUF_SIZE);
+ iwl_pcie_txq_build_tfd(trans, txq, tb0_phys,
+ IWL_HCMD_SCRATCHBUF_SIZE, 1);
- if (!ieee80211_has_morefrags(fc)) {
- txq->need_update = 1;
- } else {
- wait_write_ptr = 1;
- txq->need_update = 0;
- }
+ /* there must be data left over for TB1 or this code must be changed */
+ BUILD_BUG_ON(sizeof(struct iwl_tx_cmd) < IWL_HCMD_SCRATCHBUF_SIZE);
- /* Set up TFD's 2nd entry to point directly to remainder of skb,
- * if any (802.11 null frames have no payload). */
- secondlen = skb->len - hdr_len;
- if (secondlen > 0) {
- phys_addr = dma_map_single(trans->dev, skb->data + hdr_len,
- secondlen, DMA_TO_DEVICE);
- if (unlikely(dma_mapping_error(trans->dev, phys_addr))) {
- dma_unmap_single(trans->dev,
- dma_unmap_addr(out_meta, mapping),
- dma_unmap_len(out_meta, len),
- DMA_BIDIRECTIONAL);
+ /* map the data for TB1 */
+ tb1_addr = ((u8 *)&dev_cmd->hdr) + IWL_HCMD_SCRATCHBUF_SIZE;
+ tb1_phys = dma_map_single(trans->dev, tb1_addr, tb1_len, DMA_TO_DEVICE);
+ if (unlikely(dma_mapping_error(trans->dev, tb1_phys)))
+ goto out_err;
+ iwl_pcie_txq_build_tfd(trans, txq, tb1_phys, tb1_len, 0);
+
+ /*
+ * Set up TFD's third entry to point directly to remainder
+ * of skb, if any (802.11 null frames have no payload).
+ */
+ tb2_len = skb->len - hdr_len;
+ if (tb2_len > 0) {
+ dma_addr_t tb2_phys = dma_map_single(trans->dev,
+ skb->data + hdr_len,
+ tb2_len, DMA_TO_DEVICE);
+ if (unlikely(dma_mapping_error(trans->dev, tb2_phys))) {
+ iwl_pcie_tfd_unmap(trans, out_meta,
+ &txq->tfds[q->write_ptr]);
goto out_err;
}
+ iwl_pcie_txq_build_tfd(trans, txq, tb2_phys, tb2_len, 0);
}
- /* Attach buffers to TFD */
- iwl_pcie_txq_build_tfd(trans, txq, txcmd_phys, firstlen, 1);
- if (secondlen > 0)
- iwl_pcie_txq_build_tfd(trans, txq, phys_addr, secondlen, 0);
-
- scratch_phys = txcmd_phys + sizeof(struct iwl_cmd_header) +
- offsetof(struct iwl_tx_cmd, scratch);
-
- /* take back ownership of DMA buffer to enable update */
- dma_sync_single_for_cpu(trans->dev, txcmd_phys, firstlen,
- DMA_BIDIRECTIONAL);
- tx_cmd->dram_lsb_ptr = cpu_to_le32(scratch_phys);
- tx_cmd->dram_msb_ptr = iwl_get_dma_hi_addr(scratch_phys);
-
/* Set up entry for this TFD in Tx byte-count array */
iwl_pcie_txq_update_byte_cnt_tbl(trans, txq, le16_to_cpu(tx_cmd->len));
- dma_sync_single_for_device(trans->dev, txcmd_phys, firstlen,
- DMA_BIDIRECTIONAL);
-
trace_iwlwifi_dev_tx(trans->dev, skb,
&txq->tfds[txq->q.write_ptr],
sizeof(struct iwl_tfd),
- &dev_cmd->hdr, firstlen,
- skb->data + hdr_len, secondlen);
+ &dev_cmd->hdr, IWL_HCMD_SCRATCHBUF_SIZE + tb1_len,
+ skb->data + hdr_len, tb2_len);
trace_iwlwifi_dev_tx_data(trans->dev, skb,
- skb->data + hdr_len, secondlen);
+ skb->data + hdr_len, tb2_len);
+
+ if (!ieee80211_has_morefrags(fc)) {
+ txq->need_update = 1;
+ } else {
+ wait_write_ptr = 1;
+ txq->need_update = 0;
+ }
/* start timer if queue currently empty */
if (txq->need_update && q->read_ptr == q->write_ptr &&
diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c
index 808f5fc..fb30681 100644
--- a/net/mac80211/cfg.c
+++ b/net/mac80211/cfg.c
@@ -3290,14 +3290,19 @@ static int ieee80211_cfg_get_channel(struct wiphy *wiphy,
int ret = -ENODATA;
rcu_read_lock();
- if (local->use_chanctx) {
- chanctx_conf = rcu_dereference(sdata->vif.chanctx_conf);
- if (chanctx_conf) {
- *chandef = chanctx_conf->def;
- ret = 0;
- }
- } else if (local->open_count == local->monitors) {
- *chandef = local->monitor_chandef;
+ chanctx_conf = rcu_dereference(sdata->vif.chanctx_conf);
+ if (chanctx_conf) {
+ *chandef = chanctx_conf->def;
+ ret = 0;
+ } else if (local->open_count > 0 &&
+ local->open_count == local->monitors &&
+ sdata->vif.type == NL80211_IFTYPE_MONITOR) {
+ if (local->use_chanctx)
+ *chandef = local->monitor_chandef;
+ else
+ cfg80211_chandef_create(chandef,
+ local->_oper_channel,
+ local->_oper_channel_type);
ret = 0;
}
rcu_read_unlock();
diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c
index 640afab..baaa860 100644
--- a/net/mac80211/iface.c
+++ b/net/mac80211/iface.c
@@ -541,6 +541,9 @@ int ieee80211_do_open(struct wireless_dev *wdev, bool coming_up)
ieee80211_adjust_monitor_flags(sdata, 1);
ieee80211_configure_filter(local);
+ mutex_lock(&local->mtx);
+ ieee80211_recalc_idle(local);
+ mutex_unlock(&local->mtx);
netif_carrier_on(dev);
break;
@@ -812,6 +815,9 @@ static void ieee80211_do_stop(struct ieee80211_sub_if_data *sdata,
ieee80211_adjust_monitor_flags(sdata, -1);
ieee80211_configure_filter(local);
+ mutex_lock(&local->mtx);
+ ieee80211_recalc_idle(local);
+ mutex_unlock(&local->mtx);
break;
case NL80211_IFTYPE_P2P_DEVICE:
/* relies on synchronize_rcu() below */
diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c
index 9f6464f..1415774 100644
--- a/net/mac80211/mlme.c
+++ b/net/mac80211/mlme.c
@@ -647,6 +647,9 @@ static void ieee80211_add_vht_ie(struct ieee80211_sub_if_data *sdata,
our_mcs = (le16_to_cpu(vht_cap.vht_mcs.rx_mcs_map) &
mask) >> shift;
+ if (our_mcs == IEEE80211_VHT_MCS_NOT_SUPPORTED)
+ continue;
+
switch (ap_mcs) {
default:
if (our_mcs <= ap_mcs)
@@ -3503,6 +3506,14 @@ void ieee80211_sta_quiesce(struct ieee80211_sub_if_data *sdata)
struct ieee80211_if_managed *ifmgd = &sdata->u.mgd;
/*
+ * Stop timers before deleting work items, as timers
+ * could race and re-add the work-items. They will be
+ * re-established on connection.
+ */
+ del_timer_sync(&ifmgd->conn_mon_timer);
+ del_timer_sync(&ifmgd->bcn_mon_timer);
+
+ /*
* we need to use atomic bitops for the running bits
* only because both timers might fire at the same
* time -- the code here is properly synchronised.
@@ -3516,13 +3527,9 @@ void ieee80211_sta_quiesce(struct ieee80211_sub_if_data *sdata)
if (del_timer_sync(&ifmgd->timer))
set_bit(TMR_RUNNING_TIMER, &ifmgd->timers_running);
- cancel_work_sync(&ifmgd->chswitch_work);
if (del_timer_sync(&ifmgd->chswitch_timer))
set_bit(TMR_RUNNING_CHANSW, &ifmgd->timers_running);
-
- /* these will just be re-established on connection */
- del_timer_sync(&ifmgd->conn_mon_timer);
- del_timer_sync(&ifmgd->bcn_mon_timer);
+ cancel_work_sync(&ifmgd->chswitch_work);
}
void ieee80211_sta_restart(struct ieee80211_sub_if_data *sdata)
@@ -4315,6 +4322,17 @@ void ieee80211_mgd_stop(struct ieee80211_sub_if_data *sdata)
{
struct ieee80211_if_managed *ifmgd = &sdata->u.mgd;
+ /*
+ * Make sure some work items will not run after this,
+ * they will not do anything but might not have been
+ * cancelled when disconnecting.
+ */
+ cancel_work_sync(&ifmgd->monitor_work);
+ cancel_work_sync(&ifmgd->beacon_connection_loss_work);
+ cancel_work_sync(&ifmgd->request_smps_work);
+ cancel_work_sync(&ifmgd->csa_connection_drop_work);
+ cancel_work_sync(&ifmgd->chswitch_work);
+
mutex_lock(&ifmgd->mtx);
if (ifmgd->assoc_data)
ieee80211_destroy_assoc_data(sdata, false);
diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c
index ce78d11..8914d2d 100644
--- a/net/mac80211/tx.c
+++ b/net/mac80211/tx.c
@@ -2745,7 +2745,8 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw,
cpu_to_le16(IEEE80211_FCTL_MOREDATA);
}
- sdata = IEEE80211_DEV_TO_SUB_IF(skb->dev);
+ if (sdata->vif.type == NL80211_IFTYPE_AP_VLAN)
+ sdata = IEEE80211_DEV_TO_SUB_IF(skb->dev);
if (!ieee80211_tx_prepare(sdata, &tx, skb))
break;
dev_kfree_skb_any(skb);
diff --git a/net/wireless/core.c b/net/wireless/core.c
index 5ffff03..ea4155f 100644
--- a/net/wireless/core.c
+++ b/net/wireless/core.c
@@ -367,8 +367,7 @@ struct wiphy *wiphy_new(const struct cfg80211_ops *ops, int sizeof_priv)
rdev->wiphy.rts_threshold = (u32) -1;
rdev->wiphy.coverage_class = 0;
- rdev->wiphy.features = NL80211_FEATURE_SCAN_FLUSH |
- NL80211_FEATURE_ADVERTISE_CHAN_LIMITS;
+ rdev->wiphy.features = NL80211_FEATURE_SCAN_FLUSH;
return &rdev->wiphy;
}
diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c
index e652d05..d44ab21 100644
--- a/net/wireless/nl80211.c
+++ b/net/wireless/nl80211.c
@@ -557,18 +557,6 @@ static int nl80211_msg_put_channel(struct sk_buff *msg,
if ((chan->flags & IEEE80211_CHAN_RADAR) &&
nla_put_flag(msg, NL80211_FREQUENCY_ATTR_RADAR))
goto nla_put_failure;
- if ((chan->flags & IEEE80211_CHAN_NO_HT40MINUS) &&
- nla_put_flag(msg, NL80211_FREQUENCY_ATTR_NO_HT40_MINUS))
- goto nla_put_failure;
- if ((chan->flags & IEEE80211_CHAN_NO_HT40PLUS) &&
- nla_put_flag(msg, NL80211_FREQUENCY_ATTR_NO_HT40_PLUS))
- goto nla_put_failure;
- if ((chan->flags & IEEE80211_CHAN_NO_80MHZ) &&
- nla_put_flag(msg, NL80211_FREQUENCY_ATTR_NO_80MHZ))
- goto nla_put_failure;
- if ((chan->flags & IEEE80211_CHAN_NO_160MHZ) &&
- nla_put_flag(msg, NL80211_FREQUENCY_ATTR_NO_160MHZ))
- goto nla_put_failure;
if (nla_put_u32(msg, NL80211_FREQUENCY_ATTR_MAX_TX_POWER,
DBM_TO_MBM(chan->max_power)))
@@ -1310,15 +1298,6 @@ static int nl80211_send_wiphy(struct sk_buff *msg, u32 portid, u32 seq, int flag
dev->wiphy.max_acl_mac_addrs))
goto nla_put_failure;
- if (dev->wiphy.extended_capabilities &&
- (nla_put(msg, NL80211_ATTR_EXT_CAPA,
- dev->wiphy.extended_capabilities_len,
- dev->wiphy.extended_capabilities) ||
- nla_put(msg, NL80211_ATTR_EXT_CAPA_MASK,
- dev->wiphy.extended_capabilities_len,
- dev->wiphy.extended_capabilities_mask)))
- goto nla_put_failure;
-
return genlmsg_end(msg, hdr);
nla_put_failure:
@@ -1328,7 +1307,7 @@ static int nl80211_send_wiphy(struct sk_buff *msg, u32 portid, u32 seq, int flag
static int nl80211_dump_wiphy(struct sk_buff *skb, struct netlink_callback *cb)
{
- int idx = 0;
+ int idx = 0, ret;
int start = cb->args[0];
struct cfg80211_registered_device *dev;
@@ -1338,9 +1317,29 @@ static int nl80211_dump_wiphy(struct sk_buff *skb, struct netlink_callback *cb)
continue;
if (++idx <= start)
continue;
- if (nl80211_send_wiphy(skb, NETLINK_CB(cb->skb).portid,
- cb->nlh->nlmsg_seq, NLM_F_MULTI,
- dev) < 0) {
+ ret = nl80211_send_wiphy(skb, NETLINK_CB(cb->skb).portid,
+ cb->nlh->nlmsg_seq, NLM_F_MULTI,
+ dev);
+ if (ret < 0) {
+ /*
+ * If sending the wiphy data didn't fit (ENOBUFS or
+ * EMSGSIZE returned), this SKB is still empty (so
+ * it's not too big because another wiphy dataset is
+ * already in the skb) and we've not tried to adjust
+ * the dump allocation yet ... then adjust the alloc
+ * size to be bigger, and return 1 but with the empty
+ * skb. This results in an empty message being RX'ed
+ * in userspace, but that is ignored.
+ *
+ * We can then retry with the larger buffer.
+ */
+ if ((ret == -ENOBUFS || ret == -EMSGSIZE) &&
+ !skb->len &&
+ cb->min_dump_alloc < 4096) {
+ cb->min_dump_alloc = 4096;
+ mutex_unlock(&cfg80211_mutex);
+ return 1;
+ }
idx--;
break;
}
@@ -1357,7 +1356,7 @@ static int nl80211_get_wiphy(struct sk_buff *skb, struct genl_info *info)
struct sk_buff *msg;
struct cfg80211_registered_device *dev = info->user_ptr[0];
- msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
+ msg = nlmsg_new(4096, GFP_KERNEL);
if (!msg)
return -ENOMEM;
--
John W. Linville Someday the world will need a hero, and you
linville@tuxdriver.com might be all we have. Be ready.
[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply related
* [PATCH net 2/2] sfc: Correct efx_rx_buffer::page_offset when EFX_PAGE_IP_ALIGN != 0
From: Ben Hutchings @ 2013-03-06 19:29 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers, scrum-linux
In-Reply-To: <1362597261.3126.11.camel@bwh-desktop.uk.solarflarecom.com>
RX DMA buffers start at an offset of EFX_PAGE_IP_ALIGN bytes from the
start of a cache line. This offset obviously needs to be included in
the virtual address, but this was missed in commit b590ace09d51
('sfc: Fix efx_rx_buf_offset() in the presence of swiotlb') since
EFX_PAGE_IP_ALIGN is equal to 0 on both x86 and powerpc.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/rx.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/ethernet/sfc/rx.c b/drivers/net/ethernet/sfc/rx.c
index 879ff58..bb579a6 100644
--- a/drivers/net/ethernet/sfc/rx.c
+++ b/drivers/net/ethernet/sfc/rx.c
@@ -215,7 +215,7 @@ static int efx_init_rx_buffers_page(struct efx_rx_queue *rx_queue)
rx_buf = efx_rx_buffer(rx_queue, index);
rx_buf->dma_addr = dma_addr + EFX_PAGE_IP_ALIGN;
rx_buf->u.page = page;
- rx_buf->page_offset = page_offset;
+ rx_buf->page_offset = page_offset + EFX_PAGE_IP_ALIGN;
rx_buf->len = efx->rx_buffer_len - EFX_PAGE_IP_ALIGN;
rx_buf->flags = EFX_RX_BUF_PAGE;
++rx_queue->added_count;
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ 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