* Re: [PATCH 0/2] Get rid of ndo_xmit_flush
From: Jesper Dangaard Brouer @ 2014-08-27 7:48 UTC (permalink / raw)
To: Alexander Duyck
Cc: David Miller, netdev, therbert, jhs, hannes, edumazet,
jeffrey.t.kirsher, rusty, dborkman, brouer, Jeremy Eder
In-Reply-To: <53FCB944.9060904@intel.com>
On Tue, 26 Aug 2014 09:43:48 -0700
Alexander Duyck <alexander.h.duyck@intel.com> wrote:
> On 08/26/2014 05:52 AM, Jesper Dangaard Brouer wrote:
> >
[...]
> >
> > Setting: (inc TX ring size)
> > sudo ethtool -G eth5 tx 1024
> > sudo ethtool -C eth5 rx-usecs 1 #(default setting)
> > Result pktgen:
> > * instant rx:1 tx:5745632 pps n:118 average: rx:1 tx:5748818 pps
> > (instant variation TX -0.096 ns (min:-0.293 max:0.897) RX 0.000 ns)
> >
[...]
> > Setting:
> > sudo ethtool -G eth5 tx 512
> > sudo ethtool -C eth5 rx-usecs 30
> > Result pktgen:
> > * instant rx:1 tx:5920856 pps n:114 average: rx:1 tx:5918350 pps
> > (instant variation TX 0.071 ns (min:-0.177 max:0.135) RX 0.000 ns)
> >
>
> My advice would be to disable all C states and P states (including
> turbo) if possible, and try using idle=poll. Any processor frequency
> and/or C state transitions will totally wreak havoc with trying to get
> reliable results out of any performance test.
Thanks for the advice.
Reading Jeremy Eder's blog post:
http://www.breakage.org/2012/11/14/processor-max_cstate-intel_idle-max_cstate-and-devcpu_dma_latency/
It seems the best method for disabling these C and P states, and
keeping all CPUs in C0/C1 state is doing:
# tuned-adm profile latency-performance
Accuracy of: "ethtool -C eth5 rx-usecs 30" got slightly little better, (see
below) latency variations below 1ns min:-0.209 ns and max:0.114 ns. Which is
very good, and should be good enough for us, to measure the upcoming
code changes.
Just increasing TX ring queue still have variations.
I'm also going to disable Hyper-threading in BIOS, even-though I'm only
using one CPU in these tests (I worry that some process running on a
sibling could disturb the accuracy).
Setting:
* sudo ethtool -C eth5 rx-usecs 30
Result:
* instant rx:1 tx:5603644 pps n:120 average: rx:0 tx:5603317 pps
(instant variation TX 0.010 ns (min:-0.132 max:0.114) RX 0.000 ns)
* instant rx:1 tx:5599820 pps n:300 average: rx:1 tx:5602382 pps
(instant variation TX -0.082 ns (min:-0.209 max:0.114) RX 0.000 ns)
Setting:
* sudo ethtool -G eth5 tx 1024
Result:
* instant rx:1 tx:5398208 pps n:300 average: rx:1 tx:5404893 pps
(instant variation TX -0.229 ns (min:-0.257 max:1.666) RX 0.000 ns)
--
Best regards,
Jesper Dangaard Brouer
MSc.CS, Sr. Network Kernel Developer at Red Hat
Author of http://www.iptv-analyzer.org
LinkedIn: http://www.linkedin.com/in/brouer
^ permalink raw reply
* Re: [PATCH 0/2] Get rid of ndo_xmit_flush
From: Jesper Dangaard Brouer @ 2014-08-27 8:37 UTC (permalink / raw)
To: Jesper Dangaard Brouer
Cc: Alexander Duyck, David Miller, netdev, therbert, jhs, hannes,
edumazet, jeffrey.t.kirsher, rusty, dborkman, Jeremy Eder
In-Reply-To: <20140827094859.65f4f0fc@redhat.com>
On Wed, 27 Aug 2014 09:48:59 +0200
Jesper Dangaard Brouer <brouer@redhat.com> wrote:
> Reading Jeremy Eder's blog post:
> http://www.breakage.org/2012/11/14/processor-max_cstate-intel_idle-max_cstate-and-devcpu_dma_latency/
>
> It seems the best method for disabling these C and P states, and
> keeping all CPUs in C0/C1 state is doing:
>
> # tuned-adm profile latency-performance
[...]
> I'm also going to disable Hyper-threading in BIOS, even-though I'm only
> using one CPU in these tests (I worry that some process running on a
> sibling could disturb the accuracy).
Disabling Hyper-Threading in BIOS helped the accuracy, now the latency
variation is below 0.1ns min:-0.031 and max:0.054 over 120 sec.
Still using:
# tuned-adm profile latency-performance
> Setting:
> * sudo ethtool -C eth5 rx-usecs 30
> Result:
> * instant rx:1 tx:5603644 pps n:120 average: rx:0 tx:5603317 pps
> (instant variation TX 0.010 ns (min:-0.132 max:0.114) RX 0.000 ns)
> * instant rx:1 tx:5599820 pps n:300 average: rx:1 tx:5602382 pps
> (instant variation TX -0.082 ns (min:-0.209 max:0.114) RX 0.000 ns)
With no HT:
* instant rx:1 tx:5597616 pps n:120 average: rx:1 tx:5596283 pps
(instant variation TX 0.043 ns (min:-0.031 max:0.054) RX 0.000 ns)
With this setup, I think we are ready to measure the effect of our API
changes with sufficient accuracy.
--
Best regards,
Jesper Dangaard Brouer
MSc.CS, Sr. Network Kernel Developer at Red Hat
Author of http://www.iptv-analyzer.org
LinkedIn: http://www.linkedin.com/in/brouer
^ permalink raw reply
* [PATCH (net.git) (v2)] stmmac: ptp: fix the reference clock
From: Giuseppe Cavallaro @ 2014-08-27 8:37 UTC (permalink / raw)
To: netdev; +Cc: Giuseppe Cavallaro
In-Reply-To: <1409033072-10508-1-git-send-email-peppe.cavallaro@st.com>
The PTP reference clock, used for setting the addend in the Timestamp Addend
Register, was erroneously hard-coded (as reported in the databook just as
example).
The patch removes the macro named: STMMAC_SYSCLOCK and allows to use a
reference clock (clk_ptp_ref_i) that can be passed from the platform.
If not passed, the main driver clock will be used as default; note that
this can be fine on some platforms.
Note that, prior this patch, using the old STMMAC_SYSCLOCK on some platforms,
as side effect, the ptp clock can move faster/slower than the system clock.
Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
---
v2: fixed the release function in case of the new clock is not provided by the
platform.
Documentation/devicetree/bindings/net/stmmac.txt | 4 ++++
drivers/net/ethernet/stmicro/stmmac/stmmac.h | 2 ++
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 20 ++++++++++++++++----
drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h | 2 --
4 files changed, 22 insertions(+), 6 deletions(-)
diff --git a/Documentation/devicetree/bindings/net/stmmac.txt b/Documentation/devicetree/bindings/net/stmmac.txt
index 9b03c57..e45ac3f 100644
--- a/Documentation/devicetree/bindings/net/stmmac.txt
+++ b/Documentation/devicetree/bindings/net/stmmac.txt
@@ -39,6 +39,10 @@ Optional properties:
further clocks may be specified in derived bindings.
- clock-names: One name for each entry in the clocks property, the
first one should be "stmmaceth".
+- clk_ptp_ref: this is the PTP reference clock; in case of the PTP is
+ available this clock is used for programming the Timestamp Addend Register.
+ If not passed then the system clock will be used and this is fine on some
+ platforms.
Examples:
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac.h b/drivers/net/ethernet/stmicro/stmmac/stmmac.h
index ca01035..128a0b7 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac.h
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac.h
@@ -105,6 +105,8 @@ struct stmmac_priv {
struct ptp_clock *ptp_clock;
struct ptp_clock_info ptp_clock_ops;
unsigned int default_addend;
+ struct clk *clk_ptp_ref;
+ unsigned int clk_ptp_rate;
u32 adv_ts;
int use_riwt;
int irq_wake;
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index 51a89d4..0365289 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -603,16 +603,16 @@ static int stmmac_hwtstamp_ioctl(struct net_device *dev, struct ifreq *ifr)
/* calculate default added value:
* formula is :
* addend = (2^32)/freq_div_ratio;
- * where, freq_div_ratio = STMMAC_SYSCLOCK/50MHz
- * hence, addend = ((2^32) * 50MHz)/STMMAC_SYSCLOCK;
- * NOTE: STMMAC_SYSCLOCK should be >= 50MHz to
+ * where, freq_div_ratio = clk_ptp_ref_i/50MHz
+ * hence, addend = ((2^32) * 50MHz)/clk_ptp_ref_i;
+ * NOTE: clk_ptp_ref_i should be >= 50MHz to
* achive 20ns accuracy.
*
* 2^x * y == (y << x), hence
* 2^32 * 50000000 ==> (50000000 << 32)
*/
temp = (u64) (50000000ULL << 32);
- priv->default_addend = div_u64(temp, STMMAC_SYSCLOCK);
+ priv->default_addend = div_u64(temp, priv->clk_ptp_rate);
priv->hw->ptp->config_addend(priv->ioaddr,
priv->default_addend);
@@ -638,6 +638,16 @@ static int stmmac_init_ptp(struct stmmac_priv *priv)
if (!(priv->dma_cap.time_stamp || priv->dma_cap.atime_stamp))
return -EOPNOTSUPP;
+ /* Fall-back to main clock in case of no PTP ref is passed */
+ priv->clk_ptp_ref = devm_clk_get(priv->device, "clk_ptp_ref");
+ if (IS_ERR(priv->clk_ptp_ref)) {
+ priv->clk_ptp_rate = clk_get_rate(priv->stmmac_clk);
+ priv->clk_ptp_ref = NULL;
+ } else {
+ clk_prepare_enable(priv->clk_ptp_ref);
+ priv->clk_ptp_rate = clk_get_rate(priv->clk_ptp_ref);
+ }
+
priv->adv_ts = 0;
if (priv->dma_cap.atime_stamp && priv->extend_desc)
priv->adv_ts = 1;
@@ -657,6 +667,8 @@ static int stmmac_init_ptp(struct stmmac_priv *priv)
static void stmmac_release_ptp(struct stmmac_priv *priv)
{
+ if (priv->clk_ptp_ref)
+ clk_disable_unprepare(priv->clk_ptp_ref);
stmmac_ptp_unregister(priv);
}
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h
index 3dbc047..4535df3 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h
@@ -25,8 +25,6 @@
#ifndef __STMMAC_PTP_H__
#define __STMMAC_PTP_H__
-#define STMMAC_SYSCLOCK 62500000
-
/* IEEE 1588 PTP register offsets */
#define PTP_TCR 0x0700 /* Timestamp Control Reg */
#define PTP_SSIR 0x0704 /* Sub-Second Increment Reg */
--
1.7.4.4
^ permalink raw reply related
* [PATCH RFC] lib/rhashtable: allow users to set the minimum shifts of shrinking
From: Ying Xue @ 2014-08-27 8:57 UTC (permalink / raw)
To: tgraf; +Cc: davem, eric.dumazet, netdev
Now the resizeable hash table size is allowed to shrink a too smaller
size - HASH_MIN_SIZE(4) although users initially specify a rather big
size when table is created. Especially when the number of objects
saved in the table keeps a small value in comparison with the initial
setting of table size during a quite long time, lots of actions of
expanding and shrinking are involved with objects being inserted or
removed from table. However, as synchronize_rcu() has to be called
during expanding and shrinking, these unnecessary actions would
seriously hit users' performance.
Therefore, we should permit users to set the minimum table size
through configuring the minimum of number of shifts when table is
created according to users specific requirement.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
---
include/linux/rhashtable.h | 2 ++
lib/rhashtable.c | 16 ++++++++++++----
2 files changed, 14 insertions(+), 4 deletions(-)
diff --git a/include/linux/rhashtable.h b/include/linux/rhashtable.h
index 36826c0..fb298e9d 100644
--- a/include/linux/rhashtable.h
+++ b/include/linux/rhashtable.h
@@ -44,6 +44,7 @@ struct rhashtable;
* @head_offset: Offset of rhash_head in struct to be hashed
* @hash_rnd: Seed to use while hashing
* @max_shift: Maximum number of shifts while expanding
+ * @min_shift: Minimum number of shifts while shrinking
* @hashfn: Function to hash key
* @obj_hashfn: Function to hash object
* @grow_decision: If defined, may return true if table should expand
@@ -57,6 +58,7 @@ struct rhashtable_params {
size_t head_offset;
u32 hash_rnd;
size_t max_shift;
+ size_t min_shift;
rht_hashfn_t hashfn;
rht_obj_hashfn_t obj_hashfn;
bool (*grow_decision)(const struct rhashtable *ht,
diff --git a/lib/rhashtable.c b/lib/rhashtable.c
index a2c7881..1466e2d 100644
--- a/lib/rhashtable.c
+++ b/lib/rhashtable.c
@@ -293,12 +293,15 @@ EXPORT_SYMBOL_GPL(rhashtable_expand);
int rhashtable_shrink(struct rhashtable *ht, gfp_t flags)
{
struct bucket_table *ntbl, *tbl = rht_dereference(ht->tbl, ht);
+ size_t min_shift = ilog2(HASH_MIN_SIZE);
struct rhash_head __rcu **pprev;
unsigned int i;
ASSERT_RHT_MUTEX(ht);
- if (tbl->size <= HASH_MIN_SIZE)
+ if (ht->p.min_shift)
+ min_shift = max(ht->p.min_shift, min_shift);
+ if (ht->shift <= min_shift)
return 0;
ntbl = bucket_table_alloc(tbl->size / 2, flags);
@@ -506,9 +509,14 @@ void *rhashtable_lookup_compare(const struct rhashtable *ht, u32 hash,
}
EXPORT_SYMBOL_GPL(rhashtable_lookup_compare);
-static size_t rounded_hashtable_size(unsigned int nelem)
+static size_t rounded_hashtable_size(struct rhashtable_params *params)
{
- return max(roundup_pow_of_two(nelem * 4 / 3), HASH_MIN_SIZE);
+ size_t size = HASH_MIN_SIZE;
+
+ if (params->min_shift)
+ size = max((1UL << params->min_shift), HASH_MIN_SIZE);
+
+ return max(roundup_pow_of_two(params->nelem_hint * 4 / 3), size);
}
/**
@@ -567,7 +575,7 @@ int rhashtable_init(struct rhashtable *ht, struct rhashtable_params *params)
return -EINVAL;
if (params->nelem_hint)
- size = rounded_hashtable_size(params->nelem_hint);
+ size = rounded_hashtable_size(params);
tbl = bucket_table_alloc(size, GFP_KERNEL);
if (tbl == NULL)
--
1.7.9.5
^ permalink raw reply related
* Re: [RFC] net: ipv4: drop unicast encapsulated in L2 multicast
From: Johannes Berg @ 2014-08-27 9:05 UTC (permalink / raw)
To: Hannes Frederic Sowa; +Cc: linux-wireless, netdev
In-Reply-To: <1409125114.11976.14.camel@localhost>
On Wed, 2014-08-27 at 09:38 +0200, Hannes Frederic Sowa wrote:
> > And if it's *not* in the IPv6 RFCs, how should we implement this?
>
> I haven't found anything, too. Should I bring this up with IETF?
I don't know if that's really useful? OTOH, there surely must have been
a reason for this to be in the IPv4 RFC, so maybe for that same reason
it should also be in the IPv6 RFC?
However, in our particular case, it's really meant only to close the
so-called "hole-196" vulnerability where rogue clients in your network
can abuse the GTK to do some attacks. Those attacks are also always
possible on non-managed ethernet segments, but those can be segregated
more easily by client than shared medium wireless.
This is only one building block for addressing the vulnerability. The
idea here was that in the wireless stack we already check
frame encrypted with GTK => must have multicast destination address
and in the IPv4 stack we can check
frame has multicast destination address => must have multicast/broadcast
IP addr
This would address this point.
The question now is, in the absence of such a latter required check (and
indeed, in the case of CLUSTERIP), how we implement such a check.
Perhaps a sysctl is needed after all?
johannes
^ permalink raw reply
* [PATCH net-next v2] net: add skb_get_tx_queue() helper
From: Daniel Borkmann @ 2014-08-27 9:11 UTC (permalink / raw)
To: davem; +Cc: therbert, netdev
Replace occurences of skb_get_queue_mapping() and follow-up
netdev_get_tx_queue() with an actual helper function.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
---
v1->v2:
- Lets keep skb_get_tx_queue() a one-liner
include/linux/netdevice.h | 6 ++++++
net/core/netpoll.c | 2 +-
net/core/pktgen.c | 4 +---
net/packet/af_packet.c | 4 +---
net/sched/sch_generic.c | 6 ++++--
5 files changed, 13 insertions(+), 9 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 039b237..dfbd149 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -1747,6 +1747,12 @@ struct netdev_queue *netdev_get_tx_queue(const struct net_device *dev,
return &dev->_tx[index];
}
+static inline struct netdev_queue *skb_get_tx_queue(const struct net_device *dev,
+ const struct sk_buff *skb)
+{
+ return netdev_get_tx_queue(dev, skb_get_queue_mapping(skb));
+}
+
static inline void netdev_for_each_tx_queue(struct net_device *dev,
void (*f)(struct net_device *,
struct netdev_queue *,
diff --git a/net/core/netpoll.c b/net/core/netpoll.c
index a5ad068..12b1df9 100644
--- a/net/core/netpoll.c
+++ b/net/core/netpoll.c
@@ -115,7 +115,7 @@ static void queue_process(struct work_struct *work)
continue;
}
- txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb));
+ txq = skb_get_tx_queue(dev, skb);
local_irq_save(flags);
HARD_TX_LOCK(dev, txq, smp_processor_id());
diff --git a/net/core/pktgen.c b/net/core/pktgen.c
index 83e2b4b..d81b540 100644
--- a/net/core/pktgen.c
+++ b/net/core/pktgen.c
@@ -3286,7 +3286,6 @@ static void pktgen_xmit(struct pktgen_dev *pkt_dev)
{
struct net_device *odev = pkt_dev->odev;
struct netdev_queue *txq;
- u16 queue_map;
int ret;
/* If device is offline, then don't send */
@@ -3324,8 +3323,7 @@ static void pktgen_xmit(struct pktgen_dev *pkt_dev)
if (pkt_dev->delay && pkt_dev->last_ok)
spin(pkt_dev, pkt_dev->next_tx);
- queue_map = skb_get_queue_mapping(pkt_dev->skb);
- txq = netdev_get_tx_queue(odev, queue_map);
+ txq = skb_get_tx_queue(odev, pkt_dev->skb);
local_bh_disable();
diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
index 0dfa990..b7a7f5a 100644
--- a/net/packet/af_packet.c
+++ b/net/packet/af_packet.c
@@ -243,7 +243,6 @@ static int packet_direct_xmit(struct sk_buff *skb)
netdev_features_t features;
struct netdev_queue *txq;
int ret = NETDEV_TX_BUSY;
- u16 queue_map;
if (unlikely(!netif_running(dev) ||
!netif_carrier_ok(dev)))
@@ -254,8 +253,7 @@ static int packet_direct_xmit(struct sk_buff *skb)
__skb_linearize(skb))
goto drop;
- queue_map = skb_get_queue_mapping(skb);
- txq = netdev_get_tx_queue(dev, queue_map);
+ txq = skb_get_tx_queue(dev, skb);
local_bh_disable();
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index fc04fe9..05b3f5d 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -63,7 +63,7 @@ static inline struct sk_buff *dequeue_skb(struct Qdisc *q)
if (unlikely(skb)) {
/* check the reason of requeuing without tx lock first */
- txq = netdev_get_tx_queue(txq->dev, skb_get_queue_mapping(skb));
+ txq = skb_get_tx_queue(txq->dev, skb);
if (!netif_xmit_frozen_or_stopped(txq)) {
q->gso_skb = NULL;
q->q.qlen--;
@@ -183,10 +183,12 @@ static inline int qdisc_restart(struct Qdisc *q)
skb = dequeue_skb(q);
if (unlikely(!skb))
return 0;
+
WARN_ON_ONCE(skb_dst_is_noref(skb));
+
root_lock = qdisc_lock(q);
dev = qdisc_dev(q);
- txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb));
+ txq = skb_get_tx_queue(dev, skb);
return sch_direct_xmit(skb, q, dev, txq, root_lock);
}
--
1.7.11.7
^ permalink raw reply related
* Re: [PATCH net-next] net: add skb_get_tx_queue() helper
From: Daniel Borkmann @ 2014-08-27 9:12 UTC (permalink / raw)
To: Tom Herbert; +Cc: David Miller, Linux Netdev List
In-Reply-To: <CA+mtBx855UwnZsRwn3ugLU35gzrL0=7WD9G6g43MgVST86oWhA@mail.gmail.com>
On 08/27/2014 01:51 AM, Tom Herbert wrote:
> On Tue, Aug 26, 2014 at 2:06 PM, Daniel Borkmann <dborkman@redhat.com> wrote:
>> Replace occurences of skb_get_queue_mapping() and follow-up
>> netdev_get_tx_queue() with an actual helper function.
>>
>> Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
>> ---
>> include/linux/netdevice.h | 8 ++++++++
>> net/core/netpoll.c | 2 +-
>> net/core/pktgen.c | 4 +---
>> net/packet/af_packet.c | 4 +---
>> net/sched/sch_generic.c | 6 ++++--
>> 5 files changed, 15 insertions(+), 9 deletions(-)
>>
>> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
>> index 039b237..200ee4f 100644
>> --- a/include/linux/netdevice.h
>> +++ b/include/linux/netdevice.h
>> @@ -1747,6 +1747,14 @@ struct netdev_queue *netdev_get_tx_queue(const struct net_device *dev,
>> return &dev->_tx[index];
>> }
>>
>> +static inline struct netdev_queue *skb_get_tx_queue(const struct net_device *dev,
>> + const struct sk_buff *skb)
>> +{
>> + u16 index = skb_get_queue_mapping(skb);
>> +
>> + return netdev_get_tx_queue(dev, index);
>> +}
>> +
> return netdev_get_tx_queue(dev, skb_get_queue_mapping(skb));
>
> if you want to be a little more concise.
That's fine by me too, lets keep it a one-liner.
^ permalink raw reply
* Re: [RFC] net: ipv4: drop unicast encapsulated in L2 multicast
From: Johannes Berg @ 2014-08-27 9:13 UTC (permalink / raw)
To: David Miller; +Cc: ja, linux-wireless, netdev
In-Reply-To: <20140822.105405.1982870131653082781.davem@davemloft.net>
On Fri, 2014-08-22 at 10:54 -0700, David Miller wrote:
> > Is this place better, after checking for RTN_BROADCAST?
> >
> > /* ARP link-layer broadcasts are acceptable here */
> > if ((skb->pkt_type == PACKET_BROADCAST ||
> > skb->pkt_type == PACKET_MULTICAST) &&
> > skb->protocol == htons(ETH_P_IP))
> > goto e_inval;
>
> Indeed, this would make ARP happier, but that still leaves open the
> issue of CLUSTERIP.
Unfortunately, I have no idea how to determine that CLUSTERIP is active
here? Do we need to tag frames, or would a sysctl work?
Or should we go back to the drawing board and not make this change in
the IP stack at all? But parsing all the IP layer in the wireless stack
is really quite ugly as well.
johannes
^ permalink raw reply
* [net-next 00/11][pull request] Intel Wired LAN Driver Updates 2014-08-27
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Jeff Kirsher, netdev, nhorman, sassmann
This series contains updates to i40e and i40evf.
Carolyn provides two patches, first changes the wording of the flow
director add/remove and asynchronous failure messages to include the
fd_id to try and add some way to track the operations on a given fd_id.
Second adds a check during handle_link_event for unqualified modules
when link is down and there is a module plugged in.
Anjali provides four patches to i40e/i40evf. First update flow director
messages so that a user can tell if a filter was added or deleted. Then
updates the ATR policy to not auto-disable ATR when we have errors in
programming. The disabling of ATR when we got programming errors was
buggy and was still adding new rules and causing continuous errors.
With this policy change, we flush instead when we see too many errors.
In addition she adds a flow director flush counter to ethtool to help
know how many times the interface had to flush and replay the flow
director filter table. Updates the driver to ignores a driver
perceived transmit hang if the number of descriptors pending is less
than 4, and instead log a stat when this situation happens. This is
because the queue progresses forward and the stack never experiences
a real hang in these situations.
Shannon provides three patches for i40e/i40evf, first enables the
l2tsel bit on receive queue contexts that are assigned to VFs so that
the VF can get the stripped VLAN tag. Then adds a max buffer size
parameter to the print helper to be sure the code knows when to stop.
Lastly, remove the complaint when removing the default MAC VLAN filter.
This was because old firmware had an incorrect MAC VLAN filter that
needed to be replaced at startup, and now newer firmware does not have
this problem. So now we only add the new filter if the removal
succeeded and no need to complain if the removal fails.
Ashish provides a change to vsi->num_queue_pairs to equal the number
that is configured by the VF. This limits the number of queues that
are enabled/disabled and fixes the mismatch case for when a VF
configures fewer queues than is allocated to it by the PF.
The following are changes since commit a3d1214688d5259a200414def4d38f1e4531febd:
neigh: document gc_thresh2
and are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
Anjali Singhai Jain (4):
i40e: Some FD message fixes
i40e: ATR policy change to flush the table to clean stale ATR rules
i40e: Add a FD flush counter to ethtool
i40e/i40evf: Ignore a driver perceived Tx hang if the number of desc
pending < 4
Ashish Shah (1):
i40e: set num_queue_pairs to num configured by VF
Carolyn Wyborny (2):
i40e: Update flow director error messages to reduce user confusion
i40e: Add checks and message for Qualified Module info
Catherine Sullivan (1):
i40e/i40evf: Bump i40e & i40evf version
Shannon Nelson (3):
i40e: Enable l2tsel bit for VLAN tag control
i40e/i40evf: add max buf len to aq debug print helper
i40e: quiet complaints when removing default MAC VLAN filter and make
set_mac reversible
drivers/net/ethernet/intel/i40e/i40e.h | 8 +
drivers/net/ethernet/intel/i40e/i40e_adminq.c | 8 +-
drivers/net/ethernet/intel/i40e/i40e_common.c | 8 +-
drivers/net/ethernet/intel/i40e/i40e_debugfs.c | 3 +
drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 15 ++
drivers/net/ethernet/intel/i40e/i40e_main.c | 216 +++++++++++++++++----
drivers/net/ethernet/intel/i40e/i40e_prototype.h | 6 +-
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 88 ++++++---
drivers/net/ethernet/intel/i40e/i40e_txrx.h | 1 +
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 14 +-
drivers/net/ethernet/intel/i40evf/i40e_adminq.c | 9 +-
drivers/net/ethernet/intel/i40evf/i40e_common.c | 8 +-
drivers/net/ethernet/intel/i40evf/i40e_prototype.h | 6 +-
drivers/net/ethernet/intel/i40evf/i40e_txrx.c | 6 +-
drivers/net/ethernet/intel/i40evf/i40e_txrx.h | 1 +
drivers/net/ethernet/intel/i40evf/i40evf_main.c | 2 +-
16 files changed, 301 insertions(+), 98 deletions(-)
--
1.9.3
^ permalink raw reply
* [net-next 01/11] i40e: Update flow director error messages to reduce user confusion
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Carolyn Wyborny, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Carolyn Wyborny <carolyn.wyborny@intel.com>
This patch changes the wording of the flow director add/remove and
asynchronous failure messages to include fd_id to try and add some
way to track the operations on a given fd_id. Its not perfect, but
its better than what we had as PCTYPE can apply to several different
filter requests.
This patch also removes a redundant message when filter
addition fails due to full condition.
Change-ID: Icf58b0603d4f162d9fc542f11a74866a907049f2
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Jim Young <jamesx.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 33 ++++++++++++++---------------
1 file changed, 16 insertions(+), 17 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index a51aa37..a6cc8d6 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -224,15 +224,14 @@ static int i40e_add_del_fdir_udpv4(struct i40e_vsi *vsi,
ret = i40e_program_fdir_filter(fd_data, raw_packet, pf, add);
if (ret) {
dev_info(&pf->pdev->dev,
- "Filter command send failed for PCTYPE %d (ret = %d)\n",
- fd_data->pctype, ret);
+ "PCTYPE:%d, Filter command send failed for fd_id:%d (ret = %d)\n",
+ fd_data->pctype, fd_data->fd_id, ret);
err = true;
} else {
dev_info(&pf->pdev->dev,
- "Filter OK for PCTYPE %d (ret = %d)\n",
- fd_data->pctype, ret);
+ "PCTYPE:%d, Filter send OK for fd_id:%d\n",
+ fd_data->pctype, fd_data->fd_id);
}
-
return err ? -EOPNOTSUPP : 0;
}
@@ -287,12 +286,13 @@ static int i40e_add_del_fdir_tcpv4(struct i40e_vsi *vsi,
if (ret) {
dev_info(&pf->pdev->dev,
- "Filter command send failed for PCTYPE %d (ret = %d)\n",
- fd_data->pctype, ret);
+ "PCTYPE:%d, Filter command send failed for fd_id:%d (ret = %d)\n",
+ fd_data->pctype, fd_data->fd_id, ret);
err = true;
} else {
- dev_info(&pf->pdev->dev, "Filter OK for PCTYPE %d (ret = %d)\n",
- fd_data->pctype, ret);
+ dev_info(&pf->pdev->dev,
+ "PCTYPE:%d, Filter send OK for fd_id:%d\n",
+ fd_data->pctype, fd_data->fd_id);
}
return err ? -EOPNOTSUPP : 0;
@@ -355,13 +355,13 @@ static int i40e_add_del_fdir_ipv4(struct i40e_vsi *vsi,
if (ret) {
dev_info(&pf->pdev->dev,
- "Filter command send failed for PCTYPE %d (ret = %d)\n",
- fd_data->pctype, ret);
+ "PCTYPE:%d, Filter command send failed for fd_id:%d (ret = %d)\n",
+ fd_data->pctype, fd_data->fd_id, ret);
err = true;
} else {
dev_info(&pf->pdev->dev,
- "Filter OK for PCTYPE %d (ret = %d)\n",
- fd_data->pctype, ret);
+ "PCTYPE:%d, Filter send OK for fd_id:%d\n",
+ fd_data->pctype, fd_data->fd_id);
}
}
@@ -443,8 +443,6 @@ static void i40e_fd_handle_status(struct i40e_ring *rx_ring,
I40E_RX_PROG_STATUS_DESC_QW1_ERROR_SHIFT;
if (error == (0x1 << I40E_RX_PROG_STATUS_DESC_FD_TBL_FULL_SHIFT)) {
- dev_warn(&pdev->dev, "ntuple filter loc = %d, could not be added\n",
- rx_desc->wb.qword0.hi_dword.fd_id);
/* filter programming failed most likely due to table full */
fcnt_prog = i40e_get_cur_guaranteed_fd_count(pf);
@@ -471,12 +469,13 @@ static void i40e_fd_handle_status(struct i40e_ring *rx_ring,
pf->flags |= I40E_FLAG_FDIR_REQUIRES_REINIT;
}
} else {
- dev_info(&pdev->dev, "FD filter programming error\n");
+ dev_info(&pdev->dev,
+ "FD filter programming failed due to incorrect filter parameters\n");
}
} else if (error ==
(0x1 << I40E_RX_PROG_STATUS_DESC_NO_FD_ENTRY_SHIFT)) {
if (I40E_DEBUG_FD & pf->hw.debug_mask)
- dev_info(&pdev->dev, "ntuple filter loc = %d, could not be removed\n",
+ dev_info(&pdev->dev, "ntuple filter fd_id = %d, could not be removed\n",
rx_desc->wb.qword0.hi_dword.fd_id);
}
}
--
1.9.3
^ permalink raw reply related
* [net-next 02/11] i40e: Some FD message fixes
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Anjali Singhai Jain, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Anjali Singhai Jain <anjali.singhai@intel.com>
Change the message that gets printed when adding/deleting a filter to
the SB, so that user can tell if a filter was added or deleted.
Print filter add failures only in case of SB filters. For ATR the
information is not useful to the user and hence suppress it unless in
higher debug mode.
Change-ID: I78d7a7a6ecfa82a38a582b0d7b4da038355e3735
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <jamesx.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 38 +++++++++++++++++++++--------
1 file changed, 28 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index a6cc8d6..267c0e6 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -228,9 +228,14 @@ static int i40e_add_del_fdir_udpv4(struct i40e_vsi *vsi,
fd_data->pctype, fd_data->fd_id, ret);
err = true;
} else {
- dev_info(&pf->pdev->dev,
- "PCTYPE:%d, Filter send OK for fd_id:%d\n",
- fd_data->pctype, fd_data->fd_id);
+ if (add)
+ dev_info(&pf->pdev->dev,
+ "Filter OK for PCTYPE %d loc = %d\n",
+ fd_data->pctype, fd_data->fd_id);
+ else
+ dev_info(&pf->pdev->dev,
+ "Filter deleted for PCTYPE %d loc = %d\n",
+ fd_data->pctype, fd_data->fd_id);
}
return err ? -EOPNOTSUPP : 0;
}
@@ -290,9 +295,13 @@ static int i40e_add_del_fdir_tcpv4(struct i40e_vsi *vsi,
fd_data->pctype, fd_data->fd_id, ret);
err = true;
} else {
- dev_info(&pf->pdev->dev,
- "PCTYPE:%d, Filter send OK for fd_id:%d\n",
- fd_data->pctype, fd_data->fd_id);
+ if (add)
+ dev_info(&pf->pdev->dev, "Filter OK for PCTYPE %d loc = %d)\n",
+ fd_data->pctype, fd_data->fd_id);
+ else
+ dev_info(&pf->pdev->dev,
+ "Filter deleted for PCTYPE %d loc = %d\n",
+ fd_data->pctype, fd_data->fd_id);
}
return err ? -EOPNOTSUPP : 0;
@@ -359,9 +368,14 @@ static int i40e_add_del_fdir_ipv4(struct i40e_vsi *vsi,
fd_data->pctype, fd_data->fd_id, ret);
err = true;
} else {
- dev_info(&pf->pdev->dev,
- "PCTYPE:%d, Filter send OK for fd_id:%d\n",
- fd_data->pctype, fd_data->fd_id);
+ if (add)
+ dev_info(&pf->pdev->dev,
+ "Filter OK for PCTYPE %d loc = %d\n",
+ fd_data->pctype, fd_data->fd_id);
+ else
+ dev_info(&pf->pdev->dev,
+ "Filter deleted for PCTYPE %d loc = %d\n",
+ fd_data->pctype, fd_data->fd_id);
}
}
@@ -443,6 +457,10 @@ static void i40e_fd_handle_status(struct i40e_ring *rx_ring,
I40E_RX_PROG_STATUS_DESC_QW1_ERROR_SHIFT;
if (error == (0x1 << I40E_RX_PROG_STATUS_DESC_FD_TBL_FULL_SHIFT)) {
+ if ((rx_desc->wb.qword0.hi_dword.fd_id != 0) ||
+ (I40E_DEBUG_FD & pf->hw.debug_mask))
+ dev_warn(&pdev->dev, "ntuple filter loc = %d, could not be added\n",
+ rx_desc->wb.qword0.hi_dword.fd_id);
/* filter programming failed most likely due to table full */
fcnt_prog = i40e_get_cur_guaranteed_fd_count(pf);
@@ -470,7 +488,7 @@ static void i40e_fd_handle_status(struct i40e_ring *rx_ring,
}
} else {
dev_info(&pdev->dev,
- "FD filter programming failed due to incorrect filter parameters\n");
+ "FD filter programming failed due to incorrect filter parameters\n");
}
} else if (error ==
(0x1 << I40E_RX_PROG_STATUS_DESC_NO_FD_ENTRY_SHIFT)) {
--
1.9.3
^ permalink raw reply related
* [net-next 06/11] i40e: set num_queue_pairs to num configured by VF
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Ashish Shah, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Ashish Shah <ashish.n.shah@intel.com>
Change vsi->num_queue_pairs to equal the number that are configured
by the VF. This, in turn, limits the number of queues that are
enable/disabled. This fixes the mismatched case for when a VF configures
fewer queues than is allocated to it by the PF.
Change other sections to use alloc_queue_pairs as warranted.
Change-ID: I0de1b55c9084e7be6acc818da8569f12128a82c2
Signed-off-by: Ashish Shah <ashish.n.shah@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 13 ++++++++-----
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 0309bf8..da0f005 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -73,7 +73,7 @@ static inline bool i40e_vc_isvalid_queue_id(struct i40e_vf *vf, u8 vsi_id,
{
struct i40e_pf *pf = vf->pf;
- return qid < pf->vsi[vsi_id]->num_queue_pairs;
+ return qid < pf->vsi[vsi_id]->alloc_queue_pairs;
}
/**
@@ -469,7 +469,7 @@ static void i40e_enable_vf_mappings(struct i40e_vf *vf)
wr32(hw, I40E_VPLAN_MAPENA(vf->vf_id), reg);
/* map PF queues to VF queues */
- for (j = 0; j < pf->vsi[vf->lan_vsi_index]->num_queue_pairs; j++) {
+ for (j = 0; j < pf->vsi[vf->lan_vsi_index]->alloc_queue_pairs; j++) {
u16 qid = i40e_vc_get_pf_queue_id(vf, vf->lan_vsi_index, j);
reg = (qid & I40E_VPLAN_QTABLE_QINDEX_MASK);
wr32(hw, I40E_VPLAN_QTABLE(total_queue_pairs, vf->vf_id), reg);
@@ -478,7 +478,7 @@ static void i40e_enable_vf_mappings(struct i40e_vf *vf)
/* map PF queues to VSI */
for (j = 0; j < 7; j++) {
- if (j * 2 >= pf->vsi[vf->lan_vsi_index]->num_queue_pairs) {
+ if (j * 2 >= pf->vsi[vf->lan_vsi_index]->alloc_queue_pairs) {
reg = 0x07FF07FF; /* unused */
} else {
u16 qid = i40e_vc_get_pf_queue_id(vf, vf->lan_vsi_index,
@@ -585,7 +585,7 @@ static int i40e_alloc_vf_res(struct i40e_vf *vf)
ret = i40e_alloc_vsi_res(vf, I40E_VSI_SRIOV);
if (ret)
goto error_alloc;
- total_queue_pairs += pf->vsi[vf->lan_vsi_index]->num_queue_pairs;
+ total_queue_pairs += pf->vsi[vf->lan_vsi_index]->alloc_queue_pairs;
set_bit(I40E_VIRTCHNL_VF_CAP_PRIVILEGE, &vf->vf_caps);
/* store the total qps number for the runtime
@@ -1124,7 +1124,7 @@ static int i40e_vc_get_vf_resources_msg(struct i40e_vf *vf)
vfres->vsi_res[i].vsi_id = vf->lan_vsi_index;
vfres->vsi_res[i].vsi_type = I40E_VSI_SRIOV;
vfres->vsi_res[i].num_queue_pairs =
- pf->vsi[vf->lan_vsi_index]->num_queue_pairs;
+ pf->vsi[vf->lan_vsi_index]->alloc_queue_pairs;
memcpy(vfres->vsi_res[i].default_mac_addr,
vf->default_lan_addr.addr, ETH_ALEN);
i++;
@@ -1210,6 +1210,7 @@ static int i40e_vc_config_queues_msg(struct i40e_vf *vf, u8 *msg, u16 msglen)
struct i40e_virtchnl_vsi_queue_config_info *qci =
(struct i40e_virtchnl_vsi_queue_config_info *)msg;
struct i40e_virtchnl_queue_pair_info *qpi;
+ struct i40e_pf *pf = vf->pf;
u16 vsi_id, vsi_queue_id;
i40e_status aq_ret = 0;
int i;
@@ -1243,6 +1244,8 @@ static int i40e_vc_config_queues_msg(struct i40e_vf *vf, u8 *msg, u16 msglen)
goto error_param;
}
}
+ /* set vsi num_queue_pairs in use to num configured by vf */
+ pf->vsi[vf->lan_vsi_index]->num_queue_pairs = qci->num_queue_pairs;
error_param:
/* send the response to the vf */
--
1.9.3
^ permalink raw reply related
* [net-next 03/11] i40e: ATR policy change to flush the table to clean stale ATR rules
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Anjali Singhai Jain, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Anjali Singhai Jain <anjali.singhai@intel.com>
Instead of disabling ATR when we get a programming error, we now
will wait it out to see if some room gets created by ATR rule deletion.
If we still have too many errors and ATR filter count did not change
much, its time to flush and replay. We no more auto-disable ATR when
we have errors in programming.
The disabling of ATR when we get programming error was buggy and
was still adding new rules and causing continuous errors. With this
policy change we flush instead when we see too many errors.
ATR is still disabled if we add a SB rule for TCP/IPv4 flow type,
more logic is added to re-enable it once all SB TCP/IPv4 rules are gone.
Change-ID: I77edcbeab9500c72a7e0bd7b5c5b113ced133a9c
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <jamesx.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e.h | 6 ++
drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 6 ++
drivers/net/ethernet/intel/i40e/i40e_main.c | 96 +++++++++++++++++++++++---
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 23 +++---
4 files changed, 111 insertions(+), 20 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 801da39..0c6a9b9 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -144,6 +144,7 @@ enum i40e_state_t {
__I40E_PTP_TX_IN_PROGRESS,
__I40E_BAD_EEPROM,
__I40E_DOWN_REQUESTED,
+ __I40E_FD_FLUSH_REQUESTED,
};
enum i40e_interrupt_policy {
@@ -250,6 +251,10 @@ struct i40e_pf {
u16 fdir_pf_active_filters;
u16 fd_sb_cnt_idx;
u16 fd_atr_cnt_idx;
+ unsigned long fd_flush_timestamp;
+ u32 fd_add_err;
+ u32 fd_atr_cnt;
+ u32 fd_tcp_rule;
#ifdef CONFIG_I40E_VXLAN
__be16 vxlan_ports[I40E_MAX_PF_UDP_OFFLOAD_PORTS];
@@ -608,6 +613,7 @@ int i40e_add_del_fdir(struct i40e_vsi *vsi,
void i40e_fdir_check_and_reenable(struct i40e_pf *pf);
int i40e_get_current_fd_count(struct i40e_pf *pf);
int i40e_get_cur_guaranteed_fd_count(struct i40e_pf *pf);
+int i40e_get_current_atr_cnt(struct i40e_pf *pf);
bool i40e_set_ntuple(struct i40e_pf *pf, netdev_features_t features);
void i40e_set_ethtool_ops(struct net_device *netdev);
struct i40e_mac_filter *i40e_add_filter(struct i40e_vsi *vsi,
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index e8ba747..7204e29 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -1977,6 +1977,9 @@ static int i40e_del_fdir_entry(struct i40e_vsi *vsi,
struct i40e_pf *pf = vsi->back;
int ret = 0;
+ if (test_bit(__I40E_FD_FLUSH_REQUESTED, &pf->state))
+ return -EBUSY;
+
ret = i40e_update_ethtool_fdir_entry(vsi, NULL, fsp->location, cmd);
i40e_fdir_check_and_reenable(pf);
@@ -2010,6 +2013,9 @@ static int i40e_add_fdir_ethtool(struct i40e_vsi *vsi,
if (pf->auto_disable_flags & I40E_FLAG_FD_SB_ENABLED)
return -ENOSPC;
+ if (test_bit(__I40E_FD_FLUSH_REQUESTED, &pf->state))
+ return -EBUSY;
+
fsp = (struct ethtool_rx_flow_spec *)&cmd->fs;
if (fsp->location >= (pf->hw.func_caps.fd_filters_best_effort +
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index eddec6b..cab9c9e 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -4483,8 +4483,16 @@ static int i40e_up_complete(struct i40e_vsi *vsi)
}
/* replay FDIR SB filters */
- if (vsi->type == I40E_VSI_FDIR)
+ if (vsi->type == I40E_VSI_FDIR) {
+ /* reset fd counters */
+ pf->fd_add_err = pf->fd_atr_cnt = 0;
+ if (pf->fd_tcp_rule > 0) {
+ pf->flags &= ~I40E_FLAG_FD_ATR_ENABLED;
+ dev_info(&pf->pdev->dev, "Forcing ATR off, sideband rules for TCP/IPv4 exist\n");
+ pf->fd_tcp_rule = 0;
+ }
i40e_fdir_filter_restore(vsi);
+ }
i40e_service_event_schedule(pf);
return 0;
@@ -5125,6 +5133,7 @@ int i40e_get_current_fd_count(struct i40e_pf *pf)
I40E_PFQF_FDSTAT_BEST_CNT_SHIFT);
return fcnt_prog;
}
+
/**
* i40e_fdir_check_and_reenable - Function to reenabe FD ATR or SB if disabled
* @pf: board private structure
@@ -5133,15 +5142,17 @@ void i40e_fdir_check_and_reenable(struct i40e_pf *pf)
{
u32 fcnt_prog, fcnt_avail;
+ if (test_bit(__I40E_FD_FLUSH_REQUESTED, &pf->state))
+ return;
+
/* Check if, FD SB or ATR was auto disabled and if there is enough room
* to re-enable
*/
- if ((pf->flags & I40E_FLAG_FD_ATR_ENABLED) &&
- (pf->flags & I40E_FLAG_FD_SB_ENABLED))
- return;
fcnt_prog = i40e_get_cur_guaranteed_fd_count(pf);
fcnt_avail = pf->fdir_pf_filter_count;
- if (fcnt_prog < (fcnt_avail - I40E_FDIR_BUFFER_HEAD_ROOM)) {
+ if ((fcnt_prog < (fcnt_avail - I40E_FDIR_BUFFER_HEAD_ROOM)) ||
+ (pf->fd_add_err == 0) ||
+ (i40e_get_current_atr_cnt(pf) < pf->fd_atr_cnt)) {
if ((pf->flags & I40E_FLAG_FD_SB_ENABLED) &&
(pf->auto_disable_flags & I40E_FLAG_FD_SB_ENABLED)) {
pf->auto_disable_flags &= ~I40E_FLAG_FD_SB_ENABLED;
@@ -5158,23 +5169,83 @@ void i40e_fdir_check_and_reenable(struct i40e_pf *pf)
}
}
+#define I40E_MIN_FD_FLUSH_INTERVAL 10
+/**
+ * i40e_fdir_flush_and_replay - Function to flush all FD filters and replay SB
+ * @pf: board private structure
+ **/
+static void i40e_fdir_flush_and_replay(struct i40e_pf *pf)
+{
+ int flush_wait_retry = 50;
+ int reg;
+
+ if (time_after(jiffies, pf->fd_flush_timestamp +
+ (I40E_MIN_FD_FLUSH_INTERVAL * HZ))) {
+ set_bit(__I40E_FD_FLUSH_REQUESTED, &pf->state);
+ pf->fd_flush_timestamp = jiffies;
+ pf->auto_disable_flags |= I40E_FLAG_FD_SB_ENABLED;
+ pf->flags &= ~I40E_FLAG_FD_ATR_ENABLED;
+ /* flush all filters */
+ wr32(&pf->hw, I40E_PFQF_CTL_1,
+ I40E_PFQF_CTL_1_CLEARFDTABLE_MASK);
+ i40e_flush(&pf->hw);
+ pf->fd_add_err = 0;
+ do {
+ /* Check FD flush status every 5-6msec */
+ usleep_range(5000, 6000);
+ reg = rd32(&pf->hw, I40E_PFQF_CTL_1);
+ if (!(reg & I40E_PFQF_CTL_1_CLEARFDTABLE_MASK))
+ break;
+ } while (flush_wait_retry--);
+ if (reg & I40E_PFQF_CTL_1_CLEARFDTABLE_MASK) {
+ dev_warn(&pf->pdev->dev, "FD table did not flush, needs more time\n");
+ } else {
+ /* replay sideband filters */
+ i40e_fdir_filter_restore(pf->vsi[pf->lan_vsi]);
+
+ pf->flags |= I40E_FLAG_FD_ATR_ENABLED;
+ pf->auto_disable_flags &= ~I40E_FLAG_FD_ATR_ENABLED;
+ pf->auto_disable_flags &= ~I40E_FLAG_FD_SB_ENABLED;
+ clear_bit(__I40E_FD_FLUSH_REQUESTED, &pf->state);
+ dev_info(&pf->pdev->dev, "FD Filter table flushed and FD-SB replayed.\n");
+ }
+ }
+}
+
+/**
+ * i40e_get_current_atr_count - Get the count of total FD ATR filters programmed
+ * @pf: board private structure
+ **/
+int i40e_get_current_atr_cnt(struct i40e_pf *pf)
+{
+ return i40e_get_current_fd_count(pf) - pf->fdir_pf_active_filters;
+}
+
+/* We can see up to 256 filter programming desc in transit if the filters are
+ * being applied really fast; before we see the first
+ * filter miss error on Rx queue 0. Accumulating enough error messages before
+ * reacting will make sure we don't cause flush too often.
+ */
+#define I40E_MAX_FD_PROGRAM_ERROR 256
+
/**
* i40e_fdir_reinit_subtask - Worker thread to reinit FDIR filter table
* @pf: board private structure
**/
static void i40e_fdir_reinit_subtask(struct i40e_pf *pf)
{
- if (!(pf->flags & I40E_FLAG_FDIR_REQUIRES_REINIT))
- return;
/* if interface is down do nothing */
if (test_bit(__I40E_DOWN, &pf->state))
return;
+
+ if ((pf->fd_add_err >= I40E_MAX_FD_PROGRAM_ERROR) &&
+ (i40e_get_current_atr_cnt(pf) >= pf->fd_atr_cnt) &&
+ (i40e_get_current_atr_cnt(pf) > pf->fdir_pf_filter_count))
+ i40e_fdir_flush_and_replay(pf);
+
i40e_fdir_check_and_reenable(pf);
- if ((pf->flags & I40E_FLAG_FD_ATR_ENABLED) &&
- (pf->flags & I40E_FLAG_FD_SB_ENABLED))
- pf->flags &= ~I40E_FLAG_FDIR_REQUIRES_REINIT;
}
/**
@@ -7086,6 +7157,11 @@ bool i40e_set_ntuple(struct i40e_pf *pf, netdev_features_t features)
}
pf->flags &= ~I40E_FLAG_FD_SB_ENABLED;
pf->auto_disable_flags &= ~I40E_FLAG_FD_SB_ENABLED;
+ /* reset fd counters */
+ pf->fd_add_err = pf->fd_atr_cnt = pf->fd_tcp_rule = 0;
+ pf->fdir_pf_active_filters = 0;
+ pf->flags |= I40E_FLAG_FD_ATR_ENABLED;
+ dev_info(&pf->pdev->dev, "ATR re-enabled.\n");
/* if ATR was auto disabled it can be re-enabled. */
if ((pf->flags & I40E_FLAG_FD_ATR_ENABLED) &&
(pf->auto_disable_flags & I40E_FLAG_FD_ATR_ENABLED))
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 267c0e6..366624a 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -280,10 +280,18 @@ static int i40e_add_del_fdir_tcpv4(struct i40e_vsi *vsi,
tcp->source = fd_data->src_port;
if (add) {
+ pf->fd_tcp_rule++;
if (pf->flags & I40E_FLAG_FD_ATR_ENABLED) {
dev_info(&pf->pdev->dev, "Forcing ATR off, sideband rules for TCP/IPv4 flow being applied\n");
pf->flags &= ~I40E_FLAG_FD_ATR_ENABLED;
}
+ } else {
+ pf->fd_tcp_rule = (pf->fd_tcp_rule > 0) ?
+ (pf->fd_tcp_rule - 1) : 0;
+ if (pf->fd_tcp_rule == 0) {
+ pf->flags |= I40E_FLAG_FD_ATR_ENABLED;
+ dev_info(&pf->pdev->dev, "ATR re-enabled due to no sideband TCP/IPv4 rules\n");
+ }
}
fd_data->pctype = I40E_FILTER_PCTYPE_NONF_IPV4_TCP;
@@ -462,6 +470,10 @@ static void i40e_fd_handle_status(struct i40e_ring *rx_ring,
dev_warn(&pdev->dev, "ntuple filter loc = %d, could not be added\n",
rx_desc->wb.qword0.hi_dword.fd_id);
+ pf->fd_add_err++;
+ /* store the current atr filter count */
+ pf->fd_atr_cnt = i40e_get_current_atr_cnt(pf);
+
/* filter programming failed most likely due to table full */
fcnt_prog = i40e_get_cur_guaranteed_fd_count(pf);
fcnt_avail = pf->fdir_pf_filter_count;
@@ -470,21 +482,12 @@ static void i40e_fd_handle_status(struct i40e_ring *rx_ring,
* FD ATR/SB and then re-enable it when there is room.
*/
if (fcnt_prog >= (fcnt_avail - I40E_FDIR_BUFFER_FULL_MARGIN)) {
- /* Turn off ATR first */
- if ((pf->flags & I40E_FLAG_FD_ATR_ENABLED) &&
+ if ((pf->flags & I40E_FLAG_FD_SB_ENABLED) &&
!(pf->auto_disable_flags &
- I40E_FLAG_FD_ATR_ENABLED)) {
- dev_warn(&pdev->dev, "FD filter space full, ATR for further flows will be turned off\n");
- pf->auto_disable_flags |=
- I40E_FLAG_FD_ATR_ENABLED;
- pf->flags |= I40E_FLAG_FDIR_REQUIRES_REINIT;
- } else if ((pf->flags & I40E_FLAG_FD_SB_ENABLED) &&
- !(pf->auto_disable_flags &
I40E_FLAG_FD_SB_ENABLED)) {
dev_warn(&pdev->dev, "FD filter space full, new ntuple rules will not be added\n");
pf->auto_disable_flags |=
I40E_FLAG_FD_SB_ENABLED;
- pf->flags |= I40E_FLAG_FDIR_REQUIRES_REINIT;
}
} else {
dev_info(&pdev->dev,
--
1.9.3
^ permalink raw reply related
* [net-next 05/11] i40e: Enable l2tsel bit for VLAN tag control
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Shannon Nelson, netdev, nhorman, sassmann, Ashish Shah,
Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Shannon Nelson <shannon.nelson@intel.com>
Enable the l2tsel bit on Rx queue contexts that are assigned to VFs so
that the VF can get the stripped VLAN tag.
Change-ID: I7d9bc56238a9ea9baf5e8a97e69b9e27ebb9d169
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Ashish Shah <ashish.n.shah@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 3ac6a0d..0309bf8 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -350,6 +350,7 @@ static int i40e_config_vsi_rx_queue(struct i40e_vf *vf, u16 vsi_idx,
rx_ctx.lrxqthresh = 2;
rx_ctx.crcstrip = 1;
rx_ctx.prefena = 1;
+ rx_ctx.l2tsel = 1;
/* clear the context in the HMC */
ret = i40e_clear_lan_rx_queue_context(hw, pf_queue_id);
--
1.9.3
^ permalink raw reply related
* [net-next 04/11] i40e: Add a FD flush counter to ethtool
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Anjali Singhai Jain, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Anjali Singhai Jain <anjali.singhai@intel.com>
This helps know how many times the interface had to flush and replay FD
filter table, which gives an indication on how often we are getting FD
table full situation.
Also check on certain pf states before proceeding to add or delete
filters since we can't add or delete filters if we are in those states.
Change-ID: I97f5bbbea7146833ea61af0e08ea794fccba1780
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <jamesx.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e.h | 1 +
drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 9 +++++++++
drivers/net/ethernet/intel/i40e/i40e_main.c | 1 +
3 files changed, 11 insertions(+)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 0c6a9b9..8526950 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -252,6 +252,7 @@ struct i40e_pf {
u16 fd_sb_cnt_idx;
u16 fd_atr_cnt_idx;
unsigned long fd_flush_timestamp;
+ u32 fd_flush_cnt;
u32 fd_add_err;
u32 fd_atr_cnt;
u32 fd_tcp_rule;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index 7204e29..571d527 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -145,6 +145,7 @@ static struct i40e_stats i40e_gstrings_stats[] = {
I40E_PF_STAT("rx_jabber", stats.rx_jabber),
I40E_PF_STAT("VF_admin_queue_requests", vf_aq_requests),
I40E_PF_STAT("rx_hwtstamp_cleared", rx_hwtstamp_cleared),
+ I40E_PF_STAT("fdir_flush_cnt", fd_flush_cnt),
I40E_PF_STAT("fdir_atr_match", stats.fd_atr_match),
I40E_PF_STAT("fdir_sb_match", stats.fd_sb_match),
@@ -1977,6 +1978,10 @@ static int i40e_del_fdir_entry(struct i40e_vsi *vsi,
struct i40e_pf *pf = vsi->back;
int ret = 0;
+ if (test_bit(__I40E_RESET_RECOVERY_PENDING, &pf->state) ||
+ test_bit(__I40E_RESET_INTR_RECEIVED, &pf->state))
+ return -EBUSY;
+
if (test_bit(__I40E_FD_FLUSH_REQUESTED, &pf->state))
return -EBUSY;
@@ -2013,6 +2018,10 @@ static int i40e_add_fdir_ethtool(struct i40e_vsi *vsi,
if (pf->auto_disable_flags & I40E_FLAG_FD_SB_ENABLED)
return -ENOSPC;
+ if (test_bit(__I40E_RESET_RECOVERY_PENDING, &pf->state) ||
+ test_bit(__I40E_RESET_INTR_RECEIVED, &pf->state))
+ return -EBUSY;
+
if (test_bit(__I40E_FD_FLUSH_REQUESTED, &pf->state))
return -EBUSY;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index cab9c9e..712a23b 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -5189,6 +5189,7 @@ static void i40e_fdir_flush_and_replay(struct i40e_pf *pf)
wr32(&pf->hw, I40E_PFQF_CTL_1,
I40E_PFQF_CTL_1_CLEARFDTABLE_MASK);
i40e_flush(&pf->hw);
+ pf->fd_flush_cnt++;
pf->fd_add_err = 0;
do {
/* Check FD flush status every 5-6msec */
--
1.9.3
^ permalink raw reply related
* [net-next 08/11] i40e/i40evf: add max buf len to aq debug print helper
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Shannon Nelson, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Shannon Nelson <shannon.nelson@intel.com>
There is at least one case in the Firmware API where the response to a
command changes the buffer size field in the AQ descriptor to a larger
number than what the request's buffer size started as. This is in addition
to setting an error flag and is in order to tell the requester how much
larger a buffer is required for the answer. We need to be sure not to
use that number when dumping the contents of the data buffer because it
can send us into the weeds and generate an invalid pointer exception.
This patch adds a max buffer size parameter to the print helper to be
sure the code knows when to stop.
Change-ID: Ib84f7ed72140fe9d600086d8f2002fc5d8753092
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Tested-by: Jim Young <jamesx.m.young@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_adminq.c | 8 +++++---
drivers/net/ethernet/intel/i40e/i40e_common.c | 8 ++++++--
drivers/net/ethernet/intel/i40e/i40e_prototype.h | 6 ++----
drivers/net/ethernet/intel/i40evf/i40e_adminq.c | 9 ++++++---
drivers/net/ethernet/intel/i40evf/i40e_common.c | 8 ++++++--
drivers/net/ethernet/intel/i40evf/i40e_prototype.h | 6 ++----
6 files changed, 27 insertions(+), 18 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_adminq.c b/drivers/net/ethernet/intel/i40e/i40e_adminq.c
index b29c157..72f5d25 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_adminq.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_adminq.c
@@ -840,7 +840,8 @@ i40e_status i40e_asq_send_command(struct i40e_hw *hw,
/* bump the tail */
i40e_debug(hw, I40E_DEBUG_AQ_MESSAGE, "AQTX: desc and buffer:\n");
- i40e_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc_on_ring, buff);
+ i40e_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc_on_ring,
+ buff, buff_size);
(hw->aq.asq.next_to_use)++;
if (hw->aq.asq.next_to_use == hw->aq.asq.count)
hw->aq.asq.next_to_use = 0;
@@ -891,7 +892,7 @@ i40e_status i40e_asq_send_command(struct i40e_hw *hw,
i40e_debug(hw, I40E_DEBUG_AQ_MESSAGE,
"AQTX: desc and buffer writeback:\n");
- i40e_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc, buff);
+ i40e_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc, buff, buff_size);
/* update the error if time out occurred */
if ((!cmd_completed) &&
@@ -987,7 +988,8 @@ i40e_status i40e_clean_arq_element(struct i40e_hw *hw,
e->msg_size);
i40e_debug(hw, I40E_DEBUG_AQ_MESSAGE, "AQRX: desc and buffer:\n");
- i40e_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc, e->msg_buf);
+ i40e_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc, e->msg_buf,
+ hw->aq.arq_buf_size);
/* Restore the original datalen and buffer address in the desc,
* FW updates datalen to indicate the event message
diff --git a/drivers/net/ethernet/intel/i40e/i40e_common.c b/drivers/net/ethernet/intel/i40e/i40e_common.c
index df43e7c..01874c0 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_common.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_common.c
@@ -75,13 +75,15 @@ static i40e_status i40e_set_mac_type(struct i40e_hw *hw)
* @mask: debug mask
* @desc: pointer to admin queue descriptor
* @buffer: pointer to command buffer
+ * @buf_len: max length of buffer
*
* Dumps debug log about adminq command with descriptor contents.
**/
void i40e_debug_aq(struct i40e_hw *hw, enum i40e_debug_mask mask, void *desc,
- void *buffer)
+ void *buffer, u16 buf_len)
{
struct i40e_aq_desc *aq_desc = (struct i40e_aq_desc *)desc;
+ u16 len = le16_to_cpu(aq_desc->datalen);
u8 *aq_buffer = (u8 *)buffer;
u32 data[4];
u32 i = 0;
@@ -105,7 +107,9 @@ void i40e_debug_aq(struct i40e_hw *hw, enum i40e_debug_mask mask, void *desc,
if ((buffer != NULL) && (aq_desc->datalen != 0)) {
memset(data, 0, sizeof(data));
i40e_debug(hw, mask, "AQ CMD Buffer:\n");
- for (i = 0; i < le16_to_cpu(aq_desc->datalen); i++) {
+ if (buf_len < len)
+ len = buf_len;
+ for (i = 0; i < len; i++) {
data[((i % 16) / 4)] |=
((u32)aq_buffer[i]) << (8 * (i % 4));
if ((i % 16) == 15) {
diff --git a/drivers/net/ethernet/intel/i40e/i40e_prototype.h b/drivers/net/ethernet/intel/i40e/i40e_prototype.h
index 949a9a0..0988b5c 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_prototype.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_prototype.h
@@ -52,10 +52,8 @@ i40e_status i40e_asq_send_command(struct i40e_hw *hw,
struct i40e_asq_cmd_details *cmd_details);
/* debug function for adminq */
-void i40e_debug_aq(struct i40e_hw *hw,
- enum i40e_debug_mask mask,
- void *desc,
- void *buffer);
+void i40e_debug_aq(struct i40e_hw *hw, enum i40e_debug_mask mask,
+ void *desc, void *buffer, u16 buf_len);
void i40e_idle_aq(struct i40e_hw *hw);
bool i40e_check_asq_alive(struct i40e_hw *hw);
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_adminq.c b/drivers/net/ethernet/intel/i40evf/i40e_adminq.c
index 0030060..f206be9 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_adminq.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_adminq.c
@@ -788,7 +788,8 @@ i40e_status i40evf_asq_send_command(struct i40e_hw *hw,
/* bump the tail */
i40e_debug(hw, I40E_DEBUG_AQ_MESSAGE, "AQTX: desc and buffer:\n");
- i40evf_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc_on_ring, buff);
+ i40evf_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc_on_ring,
+ buff, buff_size);
(hw->aq.asq.next_to_use)++;
if (hw->aq.asq.next_to_use == hw->aq.asq.count)
hw->aq.asq.next_to_use = 0;
@@ -842,7 +843,8 @@ i40e_status i40evf_asq_send_command(struct i40e_hw *hw,
i40e_debug(hw, I40E_DEBUG_AQ_MESSAGE,
"AQTX: desc and buffer writeback:\n");
- i40evf_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc, buff);
+ i40evf_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc, buff,
+ buff_size);
/* update the error if time out occurred */
if ((!cmd_completed) &&
@@ -938,7 +940,8 @@ i40e_status i40evf_clean_arq_element(struct i40e_hw *hw,
hw->aq.nvm_busy = false;
i40e_debug(hw, I40E_DEBUG_AQ_MESSAGE, "AQRX: desc and buffer:\n");
- i40evf_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc, e->msg_buf);
+ i40evf_debug_aq(hw, I40E_DEBUG_AQ_COMMAND, (void *)desc, e->msg_buf,
+ hw->aq.arq_buf_size);
/* Restore the original datalen and buffer address in the desc,
* FW updates datalen to indicate the event message
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_common.c b/drivers/net/ethernet/intel/i40evf/i40e_common.c
index 4ea90bf..9525605 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_common.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_common.c
@@ -75,13 +75,15 @@ i40e_status i40e_set_mac_type(struct i40e_hw *hw)
* @mask: debug mask
* @desc: pointer to admin queue descriptor
* @buffer: pointer to command buffer
+ * @buf_len: max length of buffer
*
* Dumps debug log about adminq command with descriptor contents.
**/
void i40evf_debug_aq(struct i40e_hw *hw, enum i40e_debug_mask mask, void *desc,
- void *buffer)
+ void *buffer, u16 buf_len)
{
struct i40e_aq_desc *aq_desc = (struct i40e_aq_desc *)desc;
+ u16 len = le16_to_cpu(aq_desc->datalen);
u8 *aq_buffer = (u8 *)buffer;
u32 data[4];
u32 i = 0;
@@ -105,7 +107,9 @@ void i40evf_debug_aq(struct i40e_hw *hw, enum i40e_debug_mask mask, void *desc,
if ((buffer != NULL) && (aq_desc->datalen != 0)) {
memset(data, 0, sizeof(data));
i40e_debug(hw, mask, "AQ CMD Buffer:\n");
- for (i = 0; i < le16_to_cpu(aq_desc->datalen); i++) {
+ if (buf_len < len)
+ len = buf_len;
+ for (i = 0; i < len; i++) {
data[((i % 16) / 4)] |=
((u32)aq_buffer[i]) << (8 * (i % 4));
if ((i % 16) == 15) {
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_prototype.h b/drivers/net/ethernet/intel/i40evf/i40e_prototype.h
index 849edcc..9173834 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_prototype.h
+++ b/drivers/net/ethernet/intel/i40evf/i40e_prototype.h
@@ -53,10 +53,8 @@ i40e_status i40evf_asq_send_command(struct i40e_hw *hw,
bool i40evf_asq_done(struct i40e_hw *hw);
/* debug function for adminq */
-void i40evf_debug_aq(struct i40e_hw *hw,
- enum i40e_debug_mask mask,
- void *desc,
- void *buffer);
+void i40evf_debug_aq(struct i40e_hw *hw, enum i40e_debug_mask mask,
+ void *desc, void *buffer, u16 buf_len);
void i40e_idle_aq(struct i40e_hw *hw);
void i40evf_resume_aq(struct i40e_hw *hw);
--
1.9.3
^ permalink raw reply related
* [net-next 07/11] i40e: Add checks and message for Qualified Module info
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Carolyn Wyborny, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Carolyn Wyborny <carolyn.wyborny@intel.com>
This patch adds a check during handle_link_event for unqualified
module when link is down and there is a module plugged. If found,
print a message.
Change-ID: Ibd8666d77d3044c2a3dd4d762d3ae9ac6e18e943
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Jim Young <jamesx.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_main.c | 14 ++++++++++++++
1 file changed, 14 insertions(+)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 712a23b..172f756 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -4480,6 +4480,13 @@ static int i40e_up_complete(struct i40e_vsi *vsi)
netif_carrier_on(vsi->netdev);
} else if (vsi->netdev) {
i40e_print_link_message(vsi, false);
+ /* need to check for qualified module here*/
+ if ((pf->hw.phy.link_info.link_info &
+ I40E_AQ_MEDIA_AVAILABLE) &&
+ (!(pf->hw.phy.link_info.an_info &
+ I40E_AQ_QUALIFIED_MODULE)))
+ netdev_err(vsi->netdev,
+ "the driver failed to link because an unqualified module was detected.");
}
/* replay FDIR SB filters */
@@ -5492,6 +5499,13 @@ static void i40e_handle_link_event(struct i40e_pf *pf,
memcpy(&pf->hw.phy.link_info_old, hw_link_info,
sizeof(pf->hw.phy.link_info_old));
+ /* check for unqualified module, if link is down */
+ if ((status->link_info & I40E_AQ_MEDIA_AVAILABLE) &&
+ (!(status->an_info & I40E_AQ_QUALIFIED_MODULE)) &&
+ (!(status->link_info & I40E_AQ_LINK_UP)))
+ dev_err(&pf->pdev->dev,
+ "The driver failed to link because an unqualified module was detected.\n");
+
/* update link status */
hw_link_info->phy_type = (enum i40e_aq_phy_type)status->phy_type;
hw_link_info->link_speed = (enum i40e_aq_link_speed)status->link_speed;
--
1.9.3
^ permalink raw reply related
* [net-next 11/11] i40e/i40evf: Bump i40e & i40evf version
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Catherine Sullivan, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Catherine Sullivan <catherine.sullivan@intel.com>
Bump versions for i40e to 1.0.4 and i40evf to 1.0.1.
Change-ID: I960c04da2c91bdf1d02f8e5011e68c34a634122d
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-By: Jim Young <jamesx.m.young@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_main.c | 6 +++---
drivers/net/ethernet/intel/i40evf/i40evf_main.c | 2 +-
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 3216aa5..bd192b8 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -37,9 +37,9 @@ static const char i40e_driver_string[] =
#define DRV_KERN "-k"
-#define DRV_VERSION_MAJOR 0
-#define DRV_VERSION_MINOR 4
-#define DRV_VERSION_BUILD 21
+#define DRV_VERSION_MAJOR 1
+#define DRV_VERSION_MINOR 0
+#define DRV_VERSION_BUILD 4
#define DRV_VERSION __stringify(DRV_VERSION_MAJOR) "." \
__stringify(DRV_VERSION_MINOR) "." \
__stringify(DRV_VERSION_BUILD) DRV_KERN
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 38429fa..8c80925 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -36,7 +36,7 @@ char i40evf_driver_name[] = "i40evf";
static const char i40evf_driver_string[] =
"Intel(R) XL710/X710 Virtual Function Network Driver";
-#define DRV_VERSION "0.9.40"
+#define DRV_VERSION "1.0.1"
const char i40evf_driver_version[] = DRV_VERSION;
static const char i40evf_copyright[] =
"Copyright (c) 2013 - 2014 Intel Corporation.";
--
1.9.3
^ permalink raw reply related
* [net-next 09/11] i40e: quiet complaints when removing default MAC VLAN filter and make set_mac reversible
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Shannon Nelson, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Shannon Nelson <shannon.nelson@intel.com>
Older firmware has an incorrect MAC VLAN filter that needs to be replaced
at startup, and now newer firmware doesn't have this problem. With this
change we no longer complain if the remove fails, and we only add the
new filter if the remove succeeded.
Setting a new LAA worked the first time, but didn't work well in successive
operations, including returning to the HW default address. This simplifies
the code that was trying to be too smart.
Lastly, this pulls the hardware default mac address out into separate
handling code and keeps the broadcast filtering from getting munged.
Change-ID: I1f54b002def04ffef2546febb9a4044385452f85
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Tested-by: Jim Young <jamesx.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_main.c | 99 ++++++++++++++++++++---------
1 file changed, 70 insertions(+), 29 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 172f756..3216aa5 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -1239,8 +1239,11 @@ struct i40e_mac_filter *i40e_put_mac_in_vlan(struct i40e_vsi *vsi, u8 *macaddr,
* i40e_rm_default_mac_filter - Remove the default MAC filter set by NVM
* @vsi: the PF Main VSI - inappropriate for any other VSI
* @macaddr: the MAC address
+ *
+ * Some older firmware configurations set up a default promiscuous VLAN
+ * filter that needs to be removed.
**/
-static void i40e_rm_default_mac_filter(struct i40e_vsi *vsi, u8 *macaddr)
+static int i40e_rm_default_mac_filter(struct i40e_vsi *vsi, u8 *macaddr)
{
struct i40e_aqc_remove_macvlan_element_data element;
struct i40e_pf *pf = vsi->back;
@@ -1248,15 +1251,18 @@ static void i40e_rm_default_mac_filter(struct i40e_vsi *vsi, u8 *macaddr)
/* Only appropriate for the PF main VSI */
if (vsi->type != I40E_VSI_MAIN)
- return;
+ return -EINVAL;
+ memset(&element, 0, sizeof(element));
ether_addr_copy(element.mac_addr, macaddr);
element.vlan_tag = 0;
element.flags = I40E_AQC_MACVLAN_DEL_PERFECT_MATCH |
I40E_AQC_MACVLAN_DEL_IGNORE_VLAN;
aq_ret = i40e_aq_remove_macvlan(&pf->hw, vsi->seid, &element, 1, NULL);
if (aq_ret)
- dev_err(&pf->pdev->dev, "Could not remove default MAC-VLAN\n");
+ return -ENOENT;
+
+ return 0;
}
/**
@@ -1385,18 +1391,30 @@ static int i40e_set_mac(struct net_device *netdev, void *p)
{
struct i40e_netdev_priv *np = netdev_priv(netdev);
struct i40e_vsi *vsi = np->vsi;
+ struct i40e_pf *pf = vsi->back;
+ struct i40e_hw *hw = &pf->hw;
struct sockaddr *addr = p;
struct i40e_mac_filter *f;
if (!is_valid_ether_addr(addr->sa_data))
return -EADDRNOTAVAIL;
- netdev_info(netdev, "set mac address=%pM\n", addr->sa_data);
+ if (ether_addr_equal(netdev->dev_addr, addr->sa_data)) {
+ netdev_info(netdev, "already using mac address %pM\n",
+ addr->sa_data);
+ return 0;
+ }
if (test_bit(__I40E_DOWN, &vsi->back->state) ||
test_bit(__I40E_RESET_RECOVERY_PENDING, &vsi->back->state))
return -EADDRNOTAVAIL;
+ if (ether_addr_equal(hw->mac.addr, addr->sa_data))
+ netdev_info(netdev, "returning to hw mac address %pM\n",
+ hw->mac.addr);
+ else
+ netdev_info(netdev, "set new mac address %pM\n", addr->sa_data);
+
if (vsi->type == I40E_VSI_MAIN) {
i40e_status ret;
ret = i40e_aq_mac_address_write(&vsi->back->hw,
@@ -1410,25 +1428,34 @@ static int i40e_set_mac(struct net_device *netdev, void *p)
}
}
- f = i40e_find_mac(vsi, addr->sa_data, false, true);
- if (!f) {
- /* In order to be sure to not drop any packets, add the
- * new address first then delete the old one.
- */
- f = i40e_add_filter(vsi, addr->sa_data, I40E_VLAN_ANY,
- false, false);
- if (!f)
- return -ENOMEM;
+ if (ether_addr_equal(netdev->dev_addr, hw->mac.addr)) {
+ struct i40e_aqc_remove_macvlan_element_data element;
- i40e_sync_vsi_filters(vsi);
+ memset(&element, 0, sizeof(element));
+ ether_addr_copy(element.mac_addr, netdev->dev_addr);
+ element.flags = I40E_AQC_MACVLAN_DEL_PERFECT_MATCH;
+ i40e_aq_remove_macvlan(&pf->hw, vsi->seid, &element, 1, NULL);
+ } else {
i40e_del_filter(vsi, netdev->dev_addr, I40E_VLAN_ANY,
false, false);
- i40e_sync_vsi_filters(vsi);
}
- f->is_laa = true;
- if (!ether_addr_equal(netdev->dev_addr, addr->sa_data))
- ether_addr_copy(netdev->dev_addr, addr->sa_data);
+ if (ether_addr_equal(addr->sa_data, hw->mac.addr)) {
+ struct i40e_aqc_add_macvlan_element_data element;
+
+ memset(&element, 0, sizeof(element));
+ ether_addr_copy(element.mac_addr, hw->mac.addr);
+ element.flags = cpu_to_le16(I40E_AQC_MACVLAN_ADD_PERFECT_MATCH);
+ i40e_aq_add_macvlan(&pf->hw, vsi->seid, &element, 1, NULL);
+ } else {
+ f = i40e_add_filter(vsi, addr->sa_data, I40E_VLAN_ANY,
+ false, false);
+ if (f)
+ f->is_laa = true;
+ }
+
+ i40e_sync_vsi_filters(vsi);
+ ether_addr_copy(netdev->dev_addr, addr->sa_data);
return 0;
}
@@ -1796,9 +1823,8 @@ int i40e_sync_vsi_filters(struct i40e_vsi *vsi)
kfree(add_list);
add_list = NULL;
- if (add_happened && (!aq_ret)) {
- /* do nothing */;
- } else if (add_happened && (aq_ret)) {
+ if (add_happened && aq_ret &&
+ pf->hw.aq.asq_last_status != I40E_AQ_RC_EINVAL) {
dev_info(&pf->pdev->dev,
"add filter failed, err %d, aq_err %d\n",
aq_ret, pf->hw.aq.asq_last_status);
@@ -7512,14 +7538,14 @@ static int i40e_config_netdev(struct i40e_vsi *vsi)
if (vsi->type == I40E_VSI_MAIN) {
SET_NETDEV_DEV(netdev, &pf->pdev->dev);
ether_addr_copy(mac_addr, hw->mac.perm_addr);
- /* The following two steps are necessary to prevent reception
- * of tagged packets - by default the NVM loads a MAC-VLAN
- * filter that will accept any tagged packet. This is to
- * prevent that during normal operations until a specific
- * VLAN tag filter has been set.
+ /* The following steps are necessary to prevent reception
+ * of tagged packets - some older NVM configurations load a
+ * default a MAC-VLAN filter that accepts any tagged packet
+ * which must be replaced by a normal filter.
*/
- i40e_rm_default_mac_filter(vsi, mac_addr);
- i40e_add_filter(vsi, mac_addr, I40E_VLAN_ANY, false, true);
+ if (!i40e_rm_default_mac_filter(vsi, mac_addr))
+ i40e_add_filter(vsi, mac_addr,
+ I40E_VLAN_ANY, false, true);
} else {
/* relate the VSI_VMDQ name to the VSI_MAIN name */
snprintf(netdev->name, IFNAMSIZ, "%sv%%d",
@@ -7735,7 +7761,22 @@ static int i40e_add_vsi(struct i40e_vsi *vsi)
f_count++;
if (f->is_laa && vsi->type == I40E_VSI_MAIN) {
- i40e_aq_mac_address_write(&vsi->back->hw,
+ struct i40e_aqc_remove_macvlan_element_data element;
+
+ memset(&element, 0, sizeof(element));
+ ether_addr_copy(element.mac_addr, f->macaddr);
+ element.flags = I40E_AQC_MACVLAN_DEL_PERFECT_MATCH;
+ ret = i40e_aq_remove_macvlan(hw, vsi->seid,
+ &element, 1, NULL);
+ if (ret) {
+ /* some older FW has a different default */
+ element.flags |=
+ I40E_AQC_MACVLAN_DEL_IGNORE_VLAN;
+ i40e_aq_remove_macvlan(hw, vsi->seid,
+ &element, 1, NULL);
+ }
+
+ i40e_aq_mac_address_write(hw,
I40E_AQC_WRITE_TYPE_LAA_WOL,
f->macaddr, NULL);
}
--
1.9.3
^ permalink raw reply related
* [net-next 10/11] i40e/i40evf: Ignore a driver perceived Tx hang if the number of desc pending < 4
From: Jeff Kirsher @ 2014-08-27 9:26 UTC (permalink / raw)
To: davem; +Cc: Anjali Singhai Jain, netdev, nhorman, sassmann, Jeff Kirsher
In-Reply-To: <1409131606-15011-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Anjali Singhai Jain <anjali.singhai@intel.com>
We are seeing situations where the driver sees a hang with less than 4
desc pending, if the driver chooses to ignore it the queue progresses
forward and the stack never experiences a real hang.
With this patch we will log a stat when this situation happens
"tx_sluggish" will increment and we can see some more details
at a higher debug level. Other than that we will ignore this
particular case of Tx hang.
Change-ID: I7d1d1666d990e2b12f4f6bed0d17d22e1b6410d5
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e.h | 1 +
drivers/net/ethernet/intel/i40e/i40e_debugfs.c | 3 +++
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 10 +++++++++-
drivers/net/ethernet/intel/i40e/i40e_txrx.h | 1 +
drivers/net/ethernet/intel/i40evf/i40e_txrx.c | 6 ++++--
drivers/net/ethernet/intel/i40evf/i40e_txrx.h | 1 +
6 files changed, 19 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 8526950..4e97ba1 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -316,6 +316,7 @@ struct i40e_pf {
u32 tx_timeout_count;
u32 tx_timeout_recovery_level;
unsigned long tx_timeout_last_recovery;
+ u32 tx_sluggish_count;
u32 hw_csum_rx_error;
u32 led_status;
u16 corer_count; /* Core reset count */
diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
index 5a0cabe..7067f4b 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
@@ -1356,6 +1356,9 @@ static ssize_t i40e_dbg_command_write(struct file *filp,
"emp reset count: %d\n", pf->empr_count);
dev_info(&pf->pdev->dev,
"pf reset count: %d\n", pf->pfr_count);
+ dev_info(&pf->pdev->dev,
+ "pf tx sluggish count: %d\n",
+ pf->tx_sluggish_count);
} else if (strncmp(&cmd_buf[5], "port", 4) == 0) {
struct i40e_aqc_query_port_ets_config_resp *bw_data;
struct i40e_dcbx_config *cfg =
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 366624a..4bf49d2 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -607,6 +607,7 @@ static u32 i40e_get_tx_pending(struct i40e_ring *ring)
static bool i40e_check_tx_hang(struct i40e_ring *tx_ring)
{
u32 tx_pending = i40e_get_tx_pending(tx_ring);
+ struct i40e_pf *pf = tx_ring->vsi->back;
bool ret = false;
clear_check_for_tx_hang(tx_ring);
@@ -623,10 +624,17 @@ static bool i40e_check_tx_hang(struct i40e_ring *tx_ring)
* pending but without time to complete it yet.
*/
if ((tx_ring->tx_stats.tx_done_old == tx_ring->stats.packets) &&
- tx_pending) {
+ (tx_pending >= I40E_MIN_DESC_PENDING)) {
/* make sure it is true for two checks in a row */
ret = test_and_set_bit(__I40E_HANG_CHECK_ARMED,
&tx_ring->state);
+ } else if ((tx_ring->tx_stats.tx_done_old == tx_ring->stats.packets) &&
+ (tx_pending < I40E_MIN_DESC_PENDING) &&
+ (tx_pending > 0)) {
+ if (I40E_DEBUG_FLOW & pf->hw.debug_mask)
+ dev_info(tx_ring->dev, "HW needs some more descs to do a cacheline flush. tx_pending %d, queue %d",
+ tx_pending, tx_ring->queue_index);
+ pf->tx_sluggish_count++;
} else {
/* update completed stats and disarm the hang check */
tx_ring->tx_stats.tx_done_old = tx_ring->stats.packets;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.h b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
index 73f4fa4..d7a625a 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
@@ -121,6 +121,7 @@ enum i40e_dyn_idx_t {
/* Tx Descriptors needed, worst case */
#define TXD_USE_COUNT(S) DIV_ROUND_UP((S), I40E_MAX_DATA_PER_TXD)
#define DESC_NEEDED (MAX_SKB_FRAGS + 4)
+#define I40E_MIN_DESC_PENDING 4
#define I40E_TX_FLAGS_CSUM (u32)(1)
#define I40E_TX_FLAGS_HW_VLAN (u32)(1 << 1)
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index 79bf96c..64b0891 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -163,11 +163,13 @@ static bool i40e_check_tx_hang(struct i40e_ring *tx_ring)
* pending but without time to complete it yet.
*/
if ((tx_ring->tx_stats.tx_done_old == tx_ring->stats.packets) &&
- tx_pending) {
+ (tx_pending >= I40E_MIN_DESC_PENDING)) {
/* make sure it is true for two checks in a row */
ret = test_and_set_bit(__I40E_HANG_CHECK_ARMED,
&tx_ring->state);
- } else {
+ } else if (!(tx_ring->tx_stats.tx_done_old == tx_ring->stats.packets) ||
+ !(tx_pending < I40E_MIN_DESC_PENDING) ||
+ !(tx_pending > 0)) {
/* update completed stats and disarm the hang check */
tx_ring->tx_stats.tx_done_old = tx_ring->stats.packets;
clear_bit(__I40E_HANG_CHECK_ARMED, &tx_ring->state);
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
index 8bc6858..f6dcf9d 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
@@ -121,6 +121,7 @@ enum i40e_dyn_idx_t {
/* Tx Descriptors needed, worst case */
#define TXD_USE_COUNT(S) DIV_ROUND_UP((S), I40E_MAX_DATA_PER_TXD)
#define DESC_NEEDED (MAX_SKB_FRAGS + 4)
+#define I40E_MIN_DESC_PENDING 4
#define I40E_TX_FLAGS_CSUM (u32)(1)
#define I40E_TX_FLAGS_HW_VLAN (u32)(1 << 1)
--
1.9.3
^ permalink raw reply related
* [PATCH (net.git)] stmmac: fix dma api misuse
From: Giuseppe Cavallaro @ 2014-08-27 9:27 UTC (permalink / raw)
To: netdev; +Cc: Giuseppe Cavallaro, Andre Draszik, Hans de Goede
In-Reply-To: <53F6E8E8.1030900@st.com>
Enabling DMA_API_DEBUG, warnings are reported at runtime
because the device driver frees DMA memory with wrong functions
and it does not call dma_mapping_error after mapping dma memory.
The first problem is fixed by of introducing a flag that helps us
keeping track which mapping technique was used, so that we can use
the right API for unmap.
This approach was inspired by the e1000 driver, which uses a similar
technique.
Signed-off-by: Andre Draszik <andre.draszik@st.com>
Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Reviewed-by: Denis Kirjanov <kda@linux-powerpc.org>
Cc: Hans de Goede <hdegoede@redhat.com>
---
drivers/net/ethernet/stmicro/stmmac/chain_mode.c | 14 +++-
drivers/net/ethernet/stmicro/stmmac/common.h | 2 +-
drivers/net/ethernet/stmicro/stmmac/ring_mode.c | 15 +++-
drivers/net/ethernet/stmicro/stmmac/stmmac.h | 7 ++-
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 68 +++++++++++++++-----
5 files changed, 79 insertions(+), 27 deletions(-)
diff --git a/drivers/net/ethernet/stmicro/stmmac/chain_mode.c b/drivers/net/ethernet/stmicro/stmmac/chain_mode.c
index c553f6b..cf28dab 100644
--- a/drivers/net/ethernet/stmicro/stmmac/chain_mode.c
+++ b/drivers/net/ethernet/stmicro/stmmac/chain_mode.c
@@ -28,7 +28,7 @@
#include "stmmac.h"
-static unsigned int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
+static int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
{
struct stmmac_priv *priv = (struct stmmac_priv *)p;
unsigned int txsize = priv->dma_tx_size;
@@ -47,7 +47,9 @@ static unsigned int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
desc->des2 = dma_map_single(priv->device, skb->data,
bmax, DMA_TO_DEVICE);
- priv->tx_skbuff_dma[entry] = desc->des2;
+ if (dma_mapping_error(priv->device, desc->des2))
+ return -1;
+ priv->tx_skbuff_dma[entry].buf = desc->des2;
priv->hw->desc->prepare_tx_desc(desc, 1, bmax, csum, STMMAC_CHAIN_MODE);
while (len != 0) {
@@ -59,7 +61,9 @@ static unsigned int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
desc->des2 = dma_map_single(priv->device,
(skb->data + bmax * i),
bmax, DMA_TO_DEVICE);
- priv->tx_skbuff_dma[entry] = desc->des2;
+ if (dma_mapping_error(priv->device, desc->des2))
+ return -1;
+ priv->tx_skbuff_dma[entry].buf = desc->des2;
priv->hw->desc->prepare_tx_desc(desc, 0, bmax, csum,
STMMAC_CHAIN_MODE);
priv->hw->desc->set_tx_owner(desc);
@@ -69,7 +73,9 @@ static unsigned int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
desc->des2 = dma_map_single(priv->device,
(skb->data + bmax * i), len,
DMA_TO_DEVICE);
- priv->tx_skbuff_dma[entry] = desc->des2;
+ if (dma_mapping_error(priv->device, desc->des2))
+ return -1;
+ priv->tx_skbuff_dma[entry].buf = desc->des2;
priv->hw->desc->prepare_tx_desc(desc, 0, len, csum,
STMMAC_CHAIN_MODE);
priv->hw->desc->set_tx_owner(desc);
diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h b/drivers/net/ethernet/stmicro/stmmac/common.h
index de507c3..bd54238 100644
--- a/drivers/net/ethernet/stmicro/stmmac/common.h
+++ b/drivers/net/ethernet/stmicro/stmmac/common.h
@@ -425,7 +425,7 @@ struct stmmac_mode_ops {
void (*init) (void *des, dma_addr_t phy_addr, unsigned int size,
unsigned int extend_desc);
unsigned int (*is_jumbo_frm) (int len, int ehn_desc);
- unsigned int (*jumbo_frm) (void *priv, struct sk_buff *skb, int csum);
+ int (*jumbo_frm)(void *priv, struct sk_buff *skb, int csum);
int (*set_16kib_bfsize)(int mtu);
void (*init_desc3)(struct dma_desc *p);
void (*refill_desc3) (void *priv, struct dma_desc *p);
diff --git a/drivers/net/ethernet/stmicro/stmmac/ring_mode.c b/drivers/net/ethernet/stmicro/stmmac/ring_mode.c
index 650a4be..5dd50c6 100644
--- a/drivers/net/ethernet/stmicro/stmmac/ring_mode.c
+++ b/drivers/net/ethernet/stmicro/stmmac/ring_mode.c
@@ -28,7 +28,7 @@
#include "stmmac.h"
-static unsigned int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
+static int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
{
struct stmmac_priv *priv = (struct stmmac_priv *)p;
unsigned int txsize = priv->dma_tx_size;
@@ -53,7 +53,10 @@ static unsigned int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
desc->des2 = dma_map_single(priv->device, skb->data,
bmax, DMA_TO_DEVICE);
- priv->tx_skbuff_dma[entry] = desc->des2;
+ if (dma_mapping_error(priv->device, desc->des2))
+ return -1;
+
+ priv->tx_skbuff_dma[entry].buf = desc->des2;
desc->des3 = desc->des2 + BUF_SIZE_4KiB;
priv->hw->desc->prepare_tx_desc(desc, 1, bmax, csum,
STMMAC_RING_MODE);
@@ -68,7 +71,9 @@ static unsigned int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
desc->des2 = dma_map_single(priv->device, skb->data + bmax,
len, DMA_TO_DEVICE);
- priv->tx_skbuff_dma[entry] = desc->des2;
+ if (dma_mapping_error(priv->device, desc->des2))
+ return -1;
+ priv->tx_skbuff_dma[entry].buf = desc->des2;
desc->des3 = desc->des2 + BUF_SIZE_4KiB;
priv->hw->desc->prepare_tx_desc(desc, 0, len, csum,
STMMAC_RING_MODE);
@@ -77,7 +82,9 @@ static unsigned int stmmac_jumbo_frm(void *p, struct sk_buff *skb, int csum)
} else {
desc->des2 = dma_map_single(priv->device, skb->data,
nopaged_len, DMA_TO_DEVICE);
- priv->tx_skbuff_dma[entry] = desc->des2;
+ if (dma_mapping_error(priv->device, desc->des2))
+ return -1;
+ priv->tx_skbuff_dma[entry].buf = desc->des2;
desc->des3 = desc->des2 + BUF_SIZE_4KiB;
priv->hw->desc->prepare_tx_desc(desc, 1, nopaged_len, csum,
STMMAC_RING_MODE);
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac.h b/drivers/net/ethernet/stmicro/stmmac/stmmac.h
index ca01035..b69c156 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac.h
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac.h
@@ -34,6 +34,11 @@
#include <linux/ptp_clock_kernel.h>
#include <linux/reset.h>
+struct stmmac_tx_info {
+ dma_addr_t buf;
+ bool map_as_page;
+};
+
struct stmmac_priv {
/* Frequently used values are kept adjacent for cache effect */
struct dma_extended_desc *dma_etx ____cacheline_aligned_in_smp;
@@ -45,7 +50,7 @@ struct stmmac_priv {
u32 tx_count_frames;
u32 tx_coal_frames;
u32 tx_coal_timer;
- dma_addr_t *tx_skbuff_dma;
+ struct stmmac_tx_info *tx_skbuff_dma;
dma_addr_t dma_tx_phy;
int tx_coalesce;
int hwts_tx_en;
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index 51a89d4..56ad0dc 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -1061,7 +1061,8 @@ static int init_dma_desc_rings(struct net_device *dev)
else
p = priv->dma_tx + i;
p->des2 = 0;
- priv->tx_skbuff_dma[i] = 0;
+ priv->tx_skbuff_dma[i].buf = 0;
+ priv->tx_skbuff_dma[i].map_as_page = false;
priv->tx_skbuff[i] = NULL;
}
@@ -1100,17 +1101,24 @@ static void dma_free_tx_skbufs(struct stmmac_priv *priv)
else
p = priv->dma_tx + i;
- if (priv->tx_skbuff_dma[i]) {
- dma_unmap_single(priv->device,
- priv->tx_skbuff_dma[i],
- priv->hw->desc->get_tx_len(p),
- DMA_TO_DEVICE);
- priv->tx_skbuff_dma[i] = 0;
+ if (priv->tx_skbuff_dma[i].buf) {
+ if (priv->tx_skbuff_dma[i].map_as_page)
+ dma_unmap_page(priv->device,
+ priv->tx_skbuff_dma[i].buf,
+ priv->hw->desc->get_tx_len(p),
+ DMA_TO_DEVICE);
+ else
+ dma_unmap_single(priv->device,
+ priv->tx_skbuff_dma[i].buf,
+ priv->hw->desc->get_tx_len(p),
+ DMA_TO_DEVICE);
}
if (priv->tx_skbuff[i] != NULL) {
dev_kfree_skb_any(priv->tx_skbuff[i]);
priv->tx_skbuff[i] = NULL;
+ priv->tx_skbuff_dma[i].buf = 0;
+ priv->tx_skbuff_dma[i].map_as_page = false;
}
}
}
@@ -1131,7 +1139,8 @@ static int alloc_dma_desc_resources(struct stmmac_priv *priv)
if (!priv->rx_skbuff)
goto err_rx_skbuff;
- priv->tx_skbuff_dma = kmalloc_array(txsize, sizeof(dma_addr_t),
+ priv->tx_skbuff_dma = kmalloc_array(txsize,
+ sizeof(*priv->tx_skbuff_dma),
GFP_KERNEL);
if (!priv->tx_skbuff_dma)
goto err_tx_skbuff_dma;
@@ -1293,12 +1302,19 @@ static void stmmac_tx_clean(struct stmmac_priv *priv)
pr_debug("%s: curr %d, dirty %d\n", __func__,
priv->cur_tx, priv->dirty_tx);
- if (likely(priv->tx_skbuff_dma[entry])) {
- dma_unmap_single(priv->device,
- priv->tx_skbuff_dma[entry],
- priv->hw->desc->get_tx_len(p),
- DMA_TO_DEVICE);
- priv->tx_skbuff_dma[entry] = 0;
+ if (likely(priv->tx_skbuff_dma[entry].buf)) {
+ if (priv->tx_skbuff_dma[entry].map_as_page)
+ dma_unmap_page(priv->device,
+ priv->tx_skbuff_dma[entry].buf,
+ priv->hw->desc->get_tx_len(p),
+ DMA_TO_DEVICE);
+ else
+ dma_unmap_single(priv->device,
+ priv->tx_skbuff_dma[entry].buf,
+ priv->hw->desc->get_tx_len(p),
+ DMA_TO_DEVICE);
+ priv->tx_skbuff_dma[entry].buf = 0;
+ priv->tx_skbuff_dma[entry].map_as_page = false;
}
priv->hw->mode->clean_desc3(priv, p);
@@ -1893,12 +1909,16 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
if (likely(!is_jumbo)) {
desc->des2 = dma_map_single(priv->device, skb->data,
nopaged_len, DMA_TO_DEVICE);
- priv->tx_skbuff_dma[entry] = desc->des2;
+ if (dma_mapping_error(priv->device, desc->des2))
+ goto dma_map_err;
+ priv->tx_skbuff_dma[entry].buf = desc->des2;
priv->hw->desc->prepare_tx_desc(desc, 1, nopaged_len,
csum_insertion, priv->mode);
} else {
desc = first;
entry = priv->hw->mode->jumbo_frm(priv, skb, csum_insertion);
+ if (unlikely(entry < 0))
+ goto dma_map_err;
}
for (i = 0; i < nfrags; i++) {
@@ -1914,7 +1934,11 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
desc->des2 = skb_frag_dma_map(priv->device, frag, 0, len,
DMA_TO_DEVICE);
- priv->tx_skbuff_dma[entry] = desc->des2;
+ if (dma_mapping_error(priv->device, desc->des2))
+ goto dma_map_err; /* should reuse desc w/o issues */
+
+ priv->tx_skbuff_dma[entry].buf = desc->des2;
+ priv->tx_skbuff_dma[entry].map_as_page = true;
priv->hw->desc->prepare_tx_desc(desc, 0, len, csum_insertion,
priv->mode);
wmb();
@@ -1981,7 +2005,12 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
priv->hw->dma->enable_dma_transmission(priv->ioaddr);
spin_unlock(&priv->tx_lock);
+ return NETDEV_TX_OK;
+dma_map_err:
+ dev_err(priv->device, "Tx dma map failed\n");
+ dev_kfree_skb(skb);
+ priv->dev->stats.tx_dropped++;
return NETDEV_TX_OK;
}
@@ -2034,7 +2063,12 @@ static inline void stmmac_rx_refill(struct stmmac_priv *priv)
priv->rx_skbuff_dma[entry] =
dma_map_single(priv->device, skb->data, bfsize,
DMA_FROM_DEVICE);
-
+ if (dma_mapping_error(priv->device,
+ priv->rx_skbuff_dma[entry])) {
+ dev_err(priv->device, "Rx dma map failed\n");
+ dev_kfree_skb(skb);
+ break;
+ }
p->des2 = priv->rx_skbuff_dma[entry];
priv->hw->mode->refill_desc3(priv, p);
--
1.7.4.4
^ permalink raw reply related
* Re: [hyperv] BUG at drivers/hv/channel.c:462 while changing MTU
From: Sitsofe Wheeler @ 2014-08-27 9:33 UTC (permalink / raw)
To: KY Srinivasan
Cc: Greg Kroah-Hartman, Jason Wang, Haiyang Zhang,
linux-kernel@vger.kernel.org, Daniel Borkmann,
netdev@vger.kernel.org, devel@linuxdriverproject.org,
David S. Miller, Dan Carpenter
In-Reply-To: <d282e41ecdb54a2ab3012a3471f649f3@BY2PR0301MB0711.namprd03.prod.outlook.com>
On Mon, Aug 25, 2014 at 09:48:31PM +0000, KY Srinivasan wrote:
>
> > There is also a case of the BUG_ON at line 460 being hit (from
> > https://lkml.org/lkml/2014/8/19/708 ):
> >
> > 457 ret = vmbus_post_msg(msg,
> > 458 sizeof(struct vmbus_channel_gpadl_teardown));
> > 459
> > 460 BUG_ON(ret != 0);
>
> I will take care of all BUG_ON() instances in channel.c
Running a kernel patched with "Drivers: hv: vmbus: Eliminate calls to
BUG_ON()" boots but repeatedly changing the MTU with the following script now
results in a GPF in netvsc_open:
dev=eth1; while true; do ifconfig eth1 down; ifconfig eth1 mtu 1500; ifconfig eth1 up; ifconfig eth1 mtu 9000; done
[ 58.031328] hv_netvsc vmbus_0_15: net device safe to remove
[ 58.083975] hv_netvsc: hv_netvsc channel opened successfully
[ 58.973625] hv_netvsc vmbus_0_15: Send section size: 6144, Section count:2560
[ 59.030456] hv_netvsc vmbus_0_15: Device MAC 00:15:5d:6f:02:a5 link state up
[ 59.142723] hv_netvsc vmbus_0_15: net device safe to remove
[ 59.184701] hv_netvsc: hv_netvsc channel opened successfully
[ 59.867690] hv_netvsc vmbus_0_15: Send section size: 6144, Section count:2560
[ 59.919317] hv_netvsc vmbus_0_15: Device MAC 00:15:5d:6f:02:a5 link state up
[ 59.988054] hv_netvsc vmbus_0_15 eth1: unable to teardown receive buffer's gpadl
[ 60.038752] hv_netvsc vmbus_0_15: net device safe to remove
[ 60.070163] hv_vmbus: Close failed: close post msg return is 4
[ 60.121381] hv_vmbus: Close failed: close post msg return is 4
[ 60.157974] hv_vmbus: Close failed: close post msg return is 4
[ 60.190205] hv_vmbus: Close failed: close post msg return is 4
[ 60.232070] hv_vmbus: Close failed: close post msg return is 4
[ 65.275820] hv_netvsc vmbus_0_15 eth1: unable to open channel: -110
[ 65.325009] general protection fault: 0000 [#1] SMP
[ 65.356804] CPU: 7 PID: 852 Comm: ifconfig Not tainted 3.17.0-rc1.x86_64-dirty #127
[ 65.356804] Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS 090006 05/23/2012
[ 65.356804] task: ffff8800f268b9f0 ti: ffff8800eed38000 task.ti: ffff8800eed38000
[ 65.356804] RIP: 0010:[<ffffffff814e9fff>] [<ffffffff814e9fff>] rndis_filter_open+0x1f/0x60
[ 65.356804] RSP: 0018:ffff8800eed3bd28 EFLAGS: 00010246
[ 65.356804] RAX: 0000000000000000 RBX: 6b6b6b6b6b6b6b6b RCX: 0000000000000006
[ 65.356804] RDX: 0000000000000006 RSI: ffff8800f268c130 RDI: ffff8801fbb8d480
[ 65.356804] RBP: ffff8800eed3bd30 R08: 0000000000000000 R09: 0000000000000000
[ 65.356804] R10: 0000000000000001 R11: 0000000000000001 R12: ffff8801fbb8d480
[ 65.356804] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000001
[ 65.356804] FS: 00007fdfdbf94740(0000) GS:ffff880207ce0000(0000) knlGS:0000000000000000
[ 65.356804] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 65.356804] CR2: 00007fdfdc1b51c4 CR3: 00000000ec53b000 CR4: 00000000000406e0
[ 65.356804] Stack:
[ 65.356804] ffff8800f1021160 ffff8800eed3bd58 ffffffff814e6515 ffff8800f1021160
[ 65.356804] ffffffff8188f980 0000000000000000 ffff8800eed3bd80 ffffffff815d0998
[ 65.356804] ffff8800f1021160 ffff8800f1021160 0000000000001043 ffff8800eed3bdb8
[ 65.356804] Call Trace:
[ 65.356804] [<ffffffff814e6515>] netvsc_open+0x25/0xb0
[ 65.356804] [<ffffffff815d0998>] __dev_open+0x98/0x110
[ 65.356804] [<ffffffff815d0c99>] __dev_change_flags+0xb9/0x160
[ 65.356804] [<ffffffff815d0d69>] dev_change_flags+0x29/0x60
[ 65.356804] [<ffffffff81641b6b>] devinet_ioctl+0x31b/0x6f0
[ 65.356804] [<ffffffff81642c4d>] inet_ioctl+0x6d/0xa0
[ 65.356804] [<ffffffff815b27c0>] sock_ioctl+0x1e0/0x210
[ 65.356804] [<ffffffff811d4dc0>] do_vfs_ioctl+0x4d0/0x510
[ 65.356804] [<ffffffff816a2d55>] ? sysret_check+0x22/0x5d
[ 65.356804] [<ffffffff810b976d>] ? trace_hardirqs_on_caller+0x17d/0x210
[ 65.356804] [<ffffffff811d4e53>] SyS_ioctl+0x53/0x90
[ 65.356804] [<ffffffff816a2d29>] system_call_fastpath+0x16/0x1b
[ 65.356804] Code: 41 5e 41 5f 5d c3 66 0f 1f 44 00 00 66 66 66 66 90 48 8b 87 20 01 00 00 48 85 c0 74 2f 55 48 89 e5 53 48 8b 98 40 02 00 00 31 c0 <83> 7b 08 02 75 2b be 0d 00 00 00 48 89 df e8 9e f9 ff ff 85 c0
[ 65.356804] RIP [<ffffffff814e9fff>] rndis_filter_open+0x1f/0x60
[ 65.356804] RSP <ffff8800eed3bd28>
[ 66.789324] ---[ end trace 1b6075f9340eb5bc ]---
--
Sitsofe | http://sucs.org/~sits/
^ permalink raw reply
* [PATCH net-next] xfrm: remove useless hash_resize_mutex locks
From: Ying Xue @ 2014-08-27 9:49 UTC (permalink / raw)
To: steffen.klassert, davem; +Cc: netdev
In xfrm_policy.c, hash_resize_mutex is defined as a local variable
and only used in xfrm_hash_resize() which is declared as a work
handler of xfrm.policy_hash_work. But when the xfrm.policy_hash_work
work is put in the global workqueue(system_wq) with schedule_work(),
the work will be really inserted in the global workqueue if it was
not already queued, otherwise, it is still left in the same position
on the the global workqueue. This means the xfrm_hash_resize() work
handler is only executed once at any time no matter how many times
its work is scheduled, that is, xfrm_hash_resize() is not called
concurrently at all, so hash_resize_mutex is redundant for us.
Additionally hash_resize_mutex defined in xfrm_state.c can be removed
as the same reason.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Acked-by: David S. Miller <davem@davemloft.net>
---
Just resend the patch after RFC flag is removed from below
version:
http://patchwork.ozlabs.org/patch/369818/
net/xfrm/xfrm_policy.c | 5 -----
net/xfrm/xfrm_state.c | 13 +++----------
2 files changed, 3 insertions(+), 15 deletions(-)
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index beeed60..b559a90 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -510,14 +510,11 @@ void xfrm_spd_getinfo(struct net *net, struct xfrmk_spdinfo *si)
}
EXPORT_SYMBOL(xfrm_spd_getinfo);
-static DEFINE_MUTEX(hash_resize_mutex);
static void xfrm_hash_resize(struct work_struct *work)
{
struct net *net = container_of(work, struct net, xfrm.policy_hash_work);
int dir, total;
- mutex_lock(&hash_resize_mutex);
-
total = 0;
for (dir = 0; dir < XFRM_POLICY_MAX * 2; dir++) {
if (xfrm_bydst_should_resize(net, dir, &total))
@@ -525,8 +522,6 @@ static void xfrm_hash_resize(struct work_struct *work)
}
if (xfrm_byidx_should_resize(net, total))
xfrm_byidx_resize(net, total);
-
- mutex_unlock(&hash_resize_mutex);
}
/* Generate new index... KAME seems to generate them ordered by cost
diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index 0ab5413..de971b6 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -97,8 +97,6 @@ static unsigned long xfrm_hash_new_size(unsigned int state_hmask)
return ((state_hmask + 1) << 1) * sizeof(struct hlist_head);
}
-static DEFINE_MUTEX(hash_resize_mutex);
-
static void xfrm_hash_resize(struct work_struct *work)
{
struct net *net = container_of(work, struct net, xfrm.state_hash_work);
@@ -107,22 +105,20 @@ static void xfrm_hash_resize(struct work_struct *work)
unsigned int nhashmask, ohashmask;
int i;
- mutex_lock(&hash_resize_mutex);
-
nsize = xfrm_hash_new_size(net->xfrm.state_hmask);
ndst = xfrm_hash_alloc(nsize);
if (!ndst)
- goto out_unlock;
+ return;
nsrc = xfrm_hash_alloc(nsize);
if (!nsrc) {
xfrm_hash_free(ndst, nsize);
- goto out_unlock;
+ return;
}
nspi = xfrm_hash_alloc(nsize);
if (!nspi) {
xfrm_hash_free(ndst, nsize);
xfrm_hash_free(nsrc, nsize);
- goto out_unlock;
+ return;
}
spin_lock_bh(&net->xfrm.xfrm_state_lock);
@@ -148,9 +144,6 @@ static void xfrm_hash_resize(struct work_struct *work)
xfrm_hash_free(odst, osize);
xfrm_hash_free(osrc, osize);
xfrm_hash_free(ospi, osize);
-
-out_unlock:
- mutex_unlock(&hash_resize_mutex);
}
static DEFINE_SPINLOCK(xfrm_state_afinfo_lock);
--
1.7.9.5
^ permalink raw reply related
* Re: [RFC] net: ipv4: drop unicast encapsulated in L2 multicast
From: Hannes Frederic Sowa @ 2014-08-27 9:53 UTC (permalink / raw)
To: Johannes Berg; +Cc: linux-wireless, netdev
In-Reply-To: <1409130313.2505.3.camel@jlt4.sipsolutions.net>
On Mi, 2014-08-27 at 11:05 +0200, Johannes Berg wrote:
> On Wed, 2014-08-27 at 09:38 +0200, Hannes Frederic Sowa wrote:
>
> > > And if it's *not* in the IPv6 RFCs, how should we implement this?
> >
> > I haven't found anything, too. Should I bring this up with IETF?
>
> I don't know if that's really useful? OTOH, there surely must have been
> a reason for this to be in the IPv4 RFC, so maybe for that same reason
> it should also be in the IPv6 RFC?
Either it is an oversight, but RFC6085 3) tries to at least clarify the
multicast destination with LL unicast address. So there must have been
people trying to enfore a relationship between LL address and IPv6 one.
I think it would be OK to drop it by default in case we don't break any
other assumptions in the stack (e.g. CLUSTERIP).
> However, in our particular case, it's really meant only to close the
> so-called "hole-196" vulnerability where rogue clients in your network
> can abuse the GTK to do some attacks. Those attacks are also always
> possible on non-managed ethernet segments, but those can be segregated
> more easily by client than shared medium wireless.
>
> This is only one building block for addressing the vulnerability. The
> idea here was that in the wireless stack we already check
>
> frame encrypted with GTK => must have multicast destination address
>
> and in the IPv4 stack we can check
>
> frame has multicast destination address => must have multicast/broadcast
> IP addr
>
> This would address this point.
Yes, I was in the room when we discussed this. ;) To me at that time it
seemed like an easy and good solution.
> The question now is, in the absence of such a latter required check (and
> indeed, in the case of CLUSTERIP), how we implement such a check.
> Perhaps a sysctl is needed after all?
Yeah, unfortunate situation.
One could add those IP addresses as broadcast addresses (/32) to the
routing table, so the brd_input jump would be taken.
But this would still break users of CLUSTERIP until they install those
routes. :(
Bye,
Hannes
^ permalink raw reply
* Re: BCM4313 & brcmsmac & 3.12: only semi-working?
From: Michael Tokarev @ 2014-08-27 10:02 UTC (permalink / raw)
To: Arend van Spriel; +Cc: brcm80211-dev-list, linux-wireless, netdev
In-Reply-To: <53FCFE1A.3060508@broadcom.com>
27.08.2014 01:37, Arend van Spriel wrote:
> On 08/26/14 18:15, Michael Tokarev wrote:
[]
> Well, sorry about that. I did see the other messages fly by and noticed you were using the wl driver so assumed you were fine with that.
That's past already. I had several issues with wl driver,
and current issue is that even the latest (Aug-2014) version
of wl driver doesn't work with current kernel. So I can't
really even compare wl and brcmsmac, -- in kernels < 3.16
brcmsmac does not work, but wl can't be compiled for 3.16,
and using different kernels for comparison is a bit wrong
because there may be differences in other areas.
> Admittedly the brcmsmac got very little attention as all our resources were put on brcmfmac.
That happens. :)
[]
> Ok. Let's put frustration aside and make an effort. So could you make a trace using trace-cmd utility. The log can get quite big. The brcmsmac driver needs to be built with CONFIG_BRCM_TRACING enabled. Please execute the following commands (assuming you use ubuntu with network-manager):
>
> $ sudo stop network-manager
> $ sudo insmod brcmfmac.ko
> $ sudo trace-cmd record -e brcmsmac:*
>
> In another terminal:
>
> $ sudo start network-manager
>
> The trace-cmd must be stopped using ctrl-c.
Okay. This turned out to be not so simple.
My initial attempt indicated that brcmsmac in 3.16 does
not work at all. This isn't actually true - subsequent
attempts shows that it works. I was ready to conclude
the problem is fixed (after transferring several gigs
of data over wifi, with tracing enabled or disabled,
after fresh boot or after reboot from wl-enabled kernel,
etc - it all worked.
Until I hit the same stall as I described initially, the
same which happened numerous times with kernel 3.12 ($subj).
After several mins of transferring it stalled. But this
time (unlike with 3.12), it continued after about 30 secs.
So, while my initial test of 3.16 indicated the prob is still
here, at the same (or even worse) state, I can't really
reproduce it, at least in a reliable way. There's something
wrong still, but at least current version is significantly
more useful than before (in a hope it wont stall at the
very wrong moment exactly ;).
There's one more difference between brcmsmac and wl -- with
wl, I see significantly better speed, -- it is about 5MB/sec,
while with brcmsmac it jumps between 2.0..4.5MB/sec (with
58..65Mbps connection rate in both cases). Here's a typical
iwconfig output for brcmsmac version:
wlan0 IEEE 802.11bgn ESSID:"mjt"
Mode:Managed Frequency:2.412 GHz Access Point: 64:70:02:29:D9:30
Bit Rate=65 Mb/s Tx-Power=19 dBm
Retry short limit:7 RTS thr:off Fragment thr:off
Power Management:off
Link Quality=58/70 Signal level=-61 dBm
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:56355 Invalid misc:472 Missed beacon:0
I'll keep trying/testing various cases, in attempt to
understand what's going on. For now, I can't provide the
requested traces (it wont be very useful, I guess).
BTW, are there other things not implemented in brcmsmac?
I see the module reminds about power management, what
does it mean? Anything else missing?
Thank you!
/mjt
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox