* Re: [PATCH net-next 3/5] rhashtable: use future table size to make expansion decision
From: Thomas Graf @ 2015-01-06 10:06 UTC (permalink / raw)
To: Ying Xue
Cc: davem, netdev, jon.maloy, Paul.Gortmaker, erik.hugne,
tipc-discussion
In-Reply-To: <54ABB147.6020904@windriver.com>
On 01/06/15 at 05:56pm, Ying Xue wrote:
> Thank you for above nice explanation. Regarding my understanding, as
> rhashtable_expand() and rhashtable_shrink() are always under the
> protection of "ht->mutex", the "future_tbl" and "tbl" absolutely point
> to the same bucket array once rhashtable_expand() or rhashtable_shrink()
> returns. Therefore, if rht_deferred_worker() takes the "ht->mutex" lock,
> the both "future_tbl" and "tbl" should point to the same bucket array.
> So the change made in the patch is useless for us, right?
Correct.
> But as you pointed in above patch, there is a bug in
> rhashtable_wakeup_worker(). As long as ht->tbl == ht->future_tbl, we
> should wake up the work. OK, I will drop the patch and fix the error in
> patch #2.
Awesome, thanks!
^ permalink raw reply
* Re: [PATCH net-next 3/5] rhashtable: use future table size to make expansion decision
From: Ying Xue @ 2015-01-06 9:56 UTC (permalink / raw)
To: Thomas Graf; +Cc: jon.maloy, netdev, Paul.Gortmaker, tipc-discussion, davem
In-Reply-To: <20150106093557.GC12468@casper.infradead.org>
On 01/06/2015 05:35 PM, Thomas Graf wrote:
> On 01/06/15 at 03:23pm, Ying Xue wrote:
>> Should use future table size instead of old table size to decide
>> whether hash table is worth being expanded.
>>
>> Signed-off-by: Ying Xue <ying.xue@windriver.com>
>> Cc: Thomas Graf <tgraf@suug.ch>
>> ---
>> lib/rhashtable.c | 5 +++--
>> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> Apologies as you were probably mislead by the bug as commented
> on in the previous patch.
>
> I don't think we need this. future_tbl only points to a different
> table until the old table entries are linked from the new table.
> The condition in the resize check where meant to exclude this
> phase so we would newer get to the deferred worker while relinking
> is happening.
>
>
Thank you for above nice explanation. Regarding my understanding, as
rhashtable_expand() and rhashtable_shrink() are always under the
protection of "ht->mutex", the "future_tbl" and "tbl" absolutely point
to the same bucket array once rhashtable_expand() or rhashtable_shrink()
returns. Therefore, if rht_deferred_worker() takes the "ht->mutex" lock,
the both "future_tbl" and "tbl" should point to the same bucket array.
So the change made in the patch is useless for us, right?
But as you pointed in above patch, there is a bug in
rhashtable_wakeup_worker(). As long as ht->tbl == ht->future_tbl, we
should wake up the work. OK, I will drop the patch and fix the error in
patch #2.
In all, thank you for quickly reviewing the series, and I would deliver
the next version soon in which your all comments would be resolved.
Please wait a moment.
Regards,
Ying
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
^ permalink raw reply
* Re: [PATCH v4 net-next 2/2 tuntap: Increase the number of queues in tun.
From: Michael S. Tsirkin @ 2015-01-06 9:49 UTC (permalink / raw)
To: Pankaj Gupta
Cc: linux-kernel, netdev, davem, jasowang, dgibson, vfalico, edumazet,
vyasevic, hkchu, wuzhy, xemul, therbert, bhutchings, xii, stephen,
jiri, sergei.shtylyov
In-Reply-To: <1420522756-15127-3-git-send-email-pagupta@redhat.com>
On Tue, Jan 06, 2015 at 11:09:16AM +0530, Pankaj Gupta wrote:
> Networking under kvm works best if we allocate a per-vCPU RX and TX
> queue in a virtual NIC. This requires a per-vCPU queue on the host side.
>
> It is now safe to increase the maximum number of queues.
> Preceding patch: 'net: allow large number of rx queues'
> made sure this won't cause failures due to high order memory
> allocations. Increase it to 256: this is the max number of vCPUs
> KVM supports.
>
> Size of tun_struct changes from 8512 to 10496 after this patch. This keeps
> pages allocated for tun_struct before and after the patch to 3.
>
> Signed-off-by: Pankaj Gupta <pagupta@redhat.com>
> Reviewed-by: David Gibson <dgibson@redhat.com>
> ---
> drivers/net/tun.c | 9 +++++----
> 1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/tun.c b/drivers/net/tun.c
> index e3fa65a..a19dc5f8 100644
> --- a/drivers/net/tun.c
> +++ b/drivers/net/tun.c
> @@ -113,10 +113,11 @@ struct tap_filter {
> unsigned char addr[FLT_EXACT_COUNT][ETH_ALEN];
> };
>
> -/* DEFAULT_MAX_NUM_RSS_QUEUES were chosen to let the rx/tx queues allocated for
> - * the netdevice to be fit in one page. So we can make sure the success of
> - * memory allocation. TODO: increase the limit. */
> -#define MAX_TAP_QUEUES DEFAULT_MAX_NUM_RSS_QUEUES
> +/* MAX_TAP_QUEUES 256 is chosen to allow rx/tx queues to be equal
> + * to max number of vCPUS in guest.
VCPUs I think.
> Also, we are making sure here
> + * queue memory allocation do not fail.
What does this mean? How are we making sure?
I would drop this phrase really.
> + */
> +#define MAX_TAP_QUEUES 256
> #define MAX_TAP_FLOWS 4096
>
> #define TUN_FLOW_EXPIRE (3 * HZ)
> --
> 1.8.3.1
^ permalink raw reply
* Re: [PATCH net-next 5/5] tipc: convert tipc reference table to use generic rhashtable
From: Thomas Graf @ 2015-01-06 9:42 UTC (permalink / raw)
To: Ying Xue
Cc: davem, netdev, jon.maloy, Paul.Gortmaker, erik.hugne,
tipc-discussion
In-Reply-To: <1420529003-22244-6-git-send-email-ying.xue@windriver.com>
On 01/06/15 at 03:23pm, Ying Xue wrote:
> As tipc reference table is statically allocated, its memory size
> requested on stack initialization stage is quite big even if the
> maximum port number is just restricted to 8191 currently, however,
> the number already becomes insufficient in practice. But if the
> maximum ports is allowed to its theory value - 2^32, its consumed
> memory size will reach a ridiculously unacceptable value. Apart from
> this, heavy tipc users spend a considerable amount of time in
> tipc_sk_get() due to the read-lock on ref_table_lock.
>
> If tipc reference table is converted with generic rhashtable, above
> mentioned both disadvantages would be resolved respectively: making
> use of the new resizable hash table can avoid locking on the lookup;
> smaller memory size is required at initial stage, for example, 256
> hash bucket slots are requested at the beginning phase instead of
> allocating the entire 8191 slots in old mode. The hash table will
> grow if entries exceeds 75% of table size up to a total table size
> of 1M, and it will automatically shrink if usage falls below 30%,
> but the minimum table size is allowed down to 256.
>
> Also converts ref_table_lock to a separate mutex to protect hash table
> mutations on write side. Lastly defers the release of the socket
> reference using call_rcu() to allow using an RCU read-side protected
> call to rhashtable_lookup().
>
> Signed-off-by: Ying Xue <ying.xue@windriver.com>
> Acked-by: Jon Maloy <jon.maloy@ericsson.com>
> Acked-by: Erik Hugne <erik.hugne@ericsson.com>
> Cc: Thomas Graf <tgraf@suug.ch>
The rhashtable usage looks good to me. I assume the TIPC maintainers
have reviewed the other pieces.
Acked-by: Thomas Graf <tgraf@suug.ch>
^ permalink raw reply
* [PATCH net] r8152: support ndo_features_check
From: Hayes Wang @ 2015-01-06 9:41 UTC (permalink / raw)
To: netdev; +Cc: nic_swsd, linux-kernel, linux-usb, Hayes Wang
Support ndo_features_check to avoid:
- the transport offset is more than the hw limitation when using hw checksum.
- the skb->len of a GSO packet is more than the limitation.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
---
drivers/net/usb/r8152.c | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c
index 2d1c77e..57ec23e 100644
--- a/drivers/net/usb/r8152.c
+++ b/drivers/net/usb/r8152.c
@@ -1897,6 +1897,22 @@ static void _rtl8152_set_rx_mode(struct net_device *netdev)
netif_wake_queue(netdev);
}
+static netdev_features_t
+rtl8152_features_check(struct sk_buff *skb, struct net_device *dev,
+ netdev_features_t features)
+{
+ u32 mss = skb_shinfo(skb)->gso_size;
+ int max_offset = mss ? GTTCPHO_MAX : TCPHO_MAX;
+ int offset = skb_transport_offset(skb);
+
+ if ((mss || skb->ip_summed == CHECKSUM_PARTIAL) && offset > max_offset)
+ features &= ~(NETIF_F_ALL_CSUM | NETIF_F_GSO_MASK);
+ else if ((skb->len + sizeof(struct tx_desc)) > agg_buf_sz)
+ features &= ~NETIF_F_GSO_MASK;
+
+ return features;
+}
+
static netdev_tx_t rtl8152_start_xmit(struct sk_buff *skb,
struct net_device *netdev)
{
@@ -3706,6 +3722,7 @@ static const struct net_device_ops rtl8152_netdev_ops = {
.ndo_set_mac_address = rtl8152_set_mac_address,
.ndo_change_mtu = rtl8152_change_mtu,
.ndo_validate_addr = eth_validate_addr,
+ .ndo_features_check = rtl8152_features_check,
};
static void r8152b_get_version(struct r8152 *tp)
--
2.1.0
^ permalink raw reply related
* Re: [PATCH net-next 4/5] rhashtable: involve rhashtable_lookup_insert routine
From: Thomas Graf @ 2015-01-06 9:41 UTC (permalink / raw)
To: Ying Xue
Cc: davem, netdev, jon.maloy, Paul.Gortmaker, erik.hugne,
tipc-discussion
In-Reply-To: <1420529003-22244-5-git-send-email-ying.xue@windriver.com>
On 01/06/15 at 03:23pm, Ying Xue wrote:
> Involve a new function called rhashtable_lookup_insert() which makes
> lookup and insertion atomic under bucket lock protection, helping us
> avoid to introduce an extra lock when we search and insert an object
> into hash table.
>
> Signed-off-by: Ying Xue <ying.xue@windriver.com>
> Signed-off-by: Thomas Graf <tgraf@suug.ch>
LGTM now, thanks!
Acked-by: Thomas Graf <tgraf@suug.ch>
^ permalink raw reply
* Re: [PATCH net-next 3/5] rhashtable: use future table size to make expansion decision
From: Thomas Graf @ 2015-01-06 9:35 UTC (permalink / raw)
To: Ying Xue
Cc: davem, netdev, jon.maloy, Paul.Gortmaker, erik.hugne,
tipc-discussion
In-Reply-To: <1420529003-22244-4-git-send-email-ying.xue@windriver.com>
On 01/06/15 at 03:23pm, Ying Xue wrote:
> Should use future table size instead of old table size to decide
> whether hash table is worth being expanded.
>
> Signed-off-by: Ying Xue <ying.xue@windriver.com>
> Cc: Thomas Graf <tgraf@suug.ch>
> ---
> lib/rhashtable.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
Apologies as you were probably mislead by the bug as commented
on in the previous patch.
I don't think we need this. future_tbl only points to a different
table until the old table entries are linked from the new table.
The condition in the resize check where meant to exclude this
phase so we would newer get to the deferred worker while relinking
is happening.
^ permalink raw reply
* Re: [PATCH net] gso: do GSO for local skb with size bigger than MTU
From: Fan Du @ 2015-01-06 9:34 UTC (permalink / raw)
To: Jesse Gross
Cc: Du, Fan, Thomas Graf, davem@davemloft.net, Michael S. Tsirkin,
Jason Wang, netdev@vger.kernel.org, fw@strlen.de,
dev@openvswitch.org, pshelar@nicira.com
In-Reply-To: <CAEP_g=-R1FrcA0sTJJjQhypRtVCwoRZ+LieKaSJxqA-HACZqEw@mail.gmail.com>
On 2015/1/6 1:58, Jesse Gross wrote:
> On Mon, Jan 5, 2015 at 1:02 AM, Fan Du <fengyuleidian0615@gmail.com> wrote:
>> 于 2014年12月03日 10:31, Du, Fan 写道:
>>
>>>
>>>> -----Original Message-----
>>>> From: Thomas Graf [mailto:tgr@infradead.org] On Behalf Of Thomas Graf
>>>> Sent: Wednesday, December 3, 2014 1:42 AM
>>>> To: Michael S. Tsirkin
>>>> Cc: Du, Fan; 'Jason Wang'; netdev@vger.kernel.org; davem@davemloft.net;
>>>> fw@strlen.de; dev@openvswitch.org; jesse@nicira.com; pshelar@nicira.com
>>>> Subject: Re: [PATCH net] gso: do GSO for local skb with size bigger than
>>>> MTU
>>>>
>>>> On 12/02/14 at 07:34pm, Michael S. Tsirkin wrote:
>>>>> On Tue, Dec 02, 2014 at 05:09:27PM +0000, Thomas Graf wrote:
>>>>>> On 12/02/14 at 01:48pm, Flavio Leitner wrote:
>>>>>>> What about containers or any other virtualization environment that
>>>>>>> doesn't use Virtio?
>>>>>>
>>>>>> The host can dictate the MTU in that case for both veth or OVS
>>>>>> internal which would be primary container plumbing techniques.
>>>>>
>>>>> It typically can't do this easily for VMs with emulated devices:
>>>>> real ethernet uses a fixed MTU.
>>>>>
>>>>> IMHO it's confusing to suggest MTU as a fix for this bug, it's an
>>>>> unrelated optimization.
>>>>> ICMP_DEST_UNREACH/ICMP_FRAG_NEEDED is the right fix here.
>>>>
>>>> PMTU discovery only resolves the issue if an actual IP stack is running
>>>> inside the
>>>> VM. This may not be the case at all.
>>> ^^^^^^^^^^^^^^^^^^^^^^^^^^^
>>>
>>> Some thoughts here:
>>>
>>> Think otherwise, this is indeed what host stack should forge a
>>> ICMP_DEST_UNREACH/ICMP_FRAG_NEEDED
>>> message with _inner_ skb network and transport header, do whatever type of
>>> encapsulation,
>>> and thereafter push such packet upward to Guest/Container, which make them
>>> feel, the intermediate node
>>> or the peer send such message. PMTU should be expected to work correct.
>>> And such behavior should be shared by all other encapsulation tech if they
>>> are also suffered.
>>
>> Hi David, Jesse and Thomas
>>
>> As discussed in here:
>> https://www.marc.info/?l=linux-netdev&m=141764712631150&w=4 and
>> quotes from Jesse:
>> My proposal would be something like this:
>> * For L2, reduce the VM MTU to the lowest common denominator on the
>> segment.
>> * For L3, use path MTU discovery or fragment inner packet (i.e.
>> normal routing behavior).
>> * As a last resort (such as if using an old version of virtio in the
>> guest), fragment the tunnel packet.
>>
>>
>> For L2, it's a administrative action
>> For L3, PMTU approach looks better, because once the sender is alerted the
>> reduced MTU,
>> packet size after encapsulation will not exceed physical MTU, so no
>> additional fragments
>> efforts needed.
>> For "As a last resort... fragment the tunnel packet", the original patch:
>> https://www.marc.info/?l=linux-netdev&m=141715655024090&w=4 did the job, but
>> seems it's
>> not welcomed.
> This needs to be properly integrated into IP processing if it is to
> work correctly.
Do you mean the original patch in this thread? yes, it works correctly
in my cloud env. If you has any other concerns, please let me know. :)
> One of the reasons for only doing path MTU discovery
> for L3 is that it operates seamlessly as part of normal operation -
> there is no need to forge addresses or potentially generate ICMP when
> on an L2 network. However, this ignores the IP handling that is going
> on (note that in OVS it is possible for L3 to be implemented as a set
> of flows coming from a controller).
>
> It also should not be VXLAN specific or duplicate VXLAN encapsulation
> code. As this is happening before encapsulation, the generated ICMP
> does not need to be encapsulated either if it is created in the right
> location.
Yes, I agree. GRE share the same issue from the code flow.
Pushing back ICMP msg back without encapsulation without circulating down
to physical device is possible. The "right location" as far as I know
could only be in ovs_vport_send. In addition this probably requires wrapper
route looking up operation for GRE/VXLAN, after get the under layer
device MTU
from the routing information, then calculate reduced MTU becomes feasible.
^ permalink raw reply
* Re: [PATCH net-next 2/5] rhashtable: introduce rhashtable_wakeup_worker helper function
From: Thomas Graf @ 2015-01-06 9:29 UTC (permalink / raw)
To: Ying Xue
Cc: davem, netdev, jon.maloy, Paul.Gortmaker, erik.hugne,
tipc-discussion
In-Reply-To: <1420529003-22244-3-git-send-email-ying.xue@windriver.com>
On 01/06/15 at 03:23pm, Ying Xue wrote:
> Introduce rhashtable_wakeup_worker() helper function to reduce
> duplicated code where to wake up worker.
>
> Signed-off-by: Ying Xue <ying.xue@windriver.com>
> Cc: Thomas Graf <tgraf@suug.ch>
> ---
> lib/rhashtable.c | 20 +++++++++++++-------
> 1 file changed, 13 insertions(+), 7 deletions(-)
>
> diff --git a/lib/rhashtable.c b/lib/rhashtable.c
> index f2fdd7a..6eda22f 100644
> --- a/lib/rhashtable.c
> +++ b/lib/rhashtable.c
> @@ -492,6 +492,17 @@ static void rht_deferred_worker(struct work_struct *work)
> mutex_unlock(&ht->mutex);
> }
>
> +static void rhashtable_wakeup_worker(struct rhashtable *ht)
> +{
> + struct bucket_table *tbl = rht_dereference_rcu(ht->tbl, ht);
> + struct bucket_table *new_tbl = rht_dereference_rcu(ht->future_tbl, ht);
> +
> + if (ht->tbl != ht->future_tbl &&
I just noticed that this is a bug in the original code. It should be
if (ht->tbl == ht->future_tbl &&
Do you want to fix this in this series?
> + ((ht->p.grow_decision && ht->p.grow_decision(ht, new_tbl->size)) ||
> + (ht->p.shrink_decision && ht->p.shrink_decision(ht, tbl->size))))
Indent to inner brakcet
> + schedule_delayed_work(&ht->run_work, 0);
> +}
> +
> /**
> * rhashtable_insert - insert object into hash hash table
> * @ht: hash table
> @@ -533,9 +544,7 @@ void rhashtable_insert(struct rhashtable *ht, struct rhash_head *obj)
> atomic_inc(&ht->nelems);
>
> /* Only grow the table if no resizing is currently in progress. */
This comment should be moved to the function as well.
^ permalink raw reply
* Re: [PATCH net-next 1/5] rhashtable: optimize rhashtable_lookup routine
From: Thomas Graf @ 2015-01-06 9:13 UTC (permalink / raw)
To: Ying Xue
Cc: davem, netdev, jon.maloy, Paul.Gortmaker, erik.hugne,
tipc-discussion
In-Reply-To: <1420529003-22244-2-git-send-email-ying.xue@windriver.com>
On 01/06/15 at 03:23pm, Ying Xue wrote:
> Define an internal compare function and relevant compare argument,
> and then make use of rhashtable_lookup_compare() to lookup key in
> hash table, reducing duplicated code between rhashtable_lookup()
> and rhashtable_lookup_compare().
>
> Signed-off-by: Ying Xue <ying.xue@windriver.com>
> Cc: Thomas Graf <tgraf@suug.ch>
Acked-by: Thomas Graf <tgraf@suug.ch>
^ permalink raw reply
* [net 2/3] i40e: Fix Rx checksum error counter
From: Jeff Kirsher @ 2015-01-06 8:44 UTC (permalink / raw)
To: davem
Cc: Anjali Singhai, netdev, nhorman, sassmann, jogreene, Greg Rose,
Jeff Kirsher
In-Reply-To: <1420533864-13125-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Anjali Singhai <anjali.singhai@intel.com>
The Rx port checksum error counter was incrementing incorrectly with
UDP encapsulated tunneled traffic. This patch fixes the problem so that
the port_rx_csum counter will show accurate statistics.
Signed-off-by: Anjali Singhai <anjali.singhai@intel.com>
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 24 +++++++++++++-----------
1 file changed, 13 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 9995598..2f19ea5 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -1314,9 +1314,7 @@ static inline void i40e_rx_checksum(struct i40e_vsi *vsi,
* so the total length of IPv4 header is IHL*4 bytes
* The UDP_0 bit *may* bet set if the *inner* header is UDP
*/
- if (ipv4_tunnel &&
- (decoded.inner_prot != I40E_RX_PTYPE_INNER_PROT_UDP) &&
- !(rx_status & (1 << I40E_RX_DESC_STATUS_UDP_0_SHIFT))) {
+ if (ipv4_tunnel) {
skb->transport_header = skb->mac_header +
sizeof(struct ethhdr) +
(ip_hdr(skb)->ihl * 4);
@@ -1326,15 +1324,19 @@ static inline void i40e_rx_checksum(struct i40e_vsi *vsi,
skb->protocol == htons(ETH_P_8021AD))
? VLAN_HLEN : 0;
- rx_udp_csum = udp_csum(skb);
- iph = ip_hdr(skb);
- csum = csum_tcpudp_magic(
- iph->saddr, iph->daddr,
- (skb->len - skb_transport_offset(skb)),
- IPPROTO_UDP, rx_udp_csum);
+ if ((ip_hdr(skb)->protocol == IPPROTO_UDP) &&
+ (udp_hdr(skb)->check != 0)) {
+ rx_udp_csum = udp_csum(skb);
+ iph = ip_hdr(skb);
+ csum = csum_tcpudp_magic(
+ iph->saddr, iph->daddr,
+ (skb->len - skb_transport_offset(skb)),
+ IPPROTO_UDP, rx_udp_csum);
- if (udp_hdr(skb)->check != csum)
- goto checksum_fail;
+ if (udp_hdr(skb)->check != csum)
+ goto checksum_fail;
+
+ } /* else its GRE and so no outer UDP header */
}
skb->ip_summed = CHECKSUM_UNNECESSARY;
--
1.9.3
^ permalink raw reply related
* [net 3/3] i40e: Fix bug with TCP over IPv6 over VXLAN
From: Jeff Kirsher @ 2015-01-06 8:44 UTC (permalink / raw)
To: davem
Cc: Anjali Singhai, netdev, nhorman, sassmann, jogreene, Greg Rose,
Jeff Kirsher
In-Reply-To: <1420533864-13125-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Anjali Singhai <anjali.singhai@intel.com>
The driver was examining the outer protocol layer to set the inner protocol
layer checksum offload. In the case of TCP over IPV6 over an IPv4 based
VXLAN the inner checksum offloads would be set to look for IPv4/UDP instead
of IPv6/TCP. This code fixes that so that the driver will look at the
proper layer for encapsulation offload settings.
Signed-off-by: Anjali Singhai <anjali.singhai@intel.com>
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 24 +++++++++++-------------
1 file changed, 11 insertions(+), 13 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 2f19ea5..35696ab 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -1866,17 +1866,16 @@ static int i40e_tso(struct i40e_ring *tx_ring, struct sk_buff *skb,
if (err < 0)
return err;
- if (protocol == htons(ETH_P_IP)) {
- iph = skb->encapsulation ? inner_ip_hdr(skb) : ip_hdr(skb);
+ iph = skb->encapsulation ? inner_ip_hdr(skb) : ip_hdr(skb);
+ ipv6h = skb->encapsulation ? inner_ipv6_hdr(skb) : ipv6_hdr(skb);
+
+ if (iph->version == 4) {
tcph = skb->encapsulation ? inner_tcp_hdr(skb) : tcp_hdr(skb);
iph->tot_len = 0;
iph->check = 0;
tcph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
0, IPPROTO_TCP, 0);
- } else if (skb_is_gso_v6(skb)) {
-
- ipv6h = skb->encapsulation ? inner_ipv6_hdr(skb)
- : ipv6_hdr(skb);
+ } else if (ipv6h->version == 6) {
tcph = skb->encapsulation ? inner_tcp_hdr(skb) : tcp_hdr(skb);
ipv6h->payload_len = 0;
tcph->check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
@@ -1972,13 +1971,9 @@ static void i40e_tx_enable_csum(struct sk_buff *skb, u32 tx_flags,
I40E_TX_CTX_EXT_IP_IPV4_NO_CSUM;
}
} else if (tx_flags & I40E_TX_FLAGS_IPV6) {
- if (tx_flags & I40E_TX_FLAGS_TSO) {
- *cd_tunneling |= I40E_TX_CTX_EXT_IP_IPV6;
+ *cd_tunneling |= I40E_TX_CTX_EXT_IP_IPV6;
+ if (tx_flags & I40E_TX_FLAGS_TSO)
ip_hdr(skb)->check = 0;
- } else {
- *cd_tunneling |=
- I40E_TX_CTX_EXT_IP_IPV4_NO_CSUM;
- }
}
/* Now set the ctx descriptor fields */
@@ -1988,7 +1983,10 @@ static void i40e_tx_enable_csum(struct sk_buff *skb, u32 tx_flags,
((skb_inner_network_offset(skb) -
skb_transport_offset(skb)) >> 1) <<
I40E_TXD_CTX_QW0_NATLEN_SHIFT;
-
+ if (this_ip_hdr->version == 6) {
+ tx_flags &= ~I40E_TX_FLAGS_IPV4;
+ tx_flags |= I40E_TX_FLAGS_IPV6;
+ }
} else {
network_hdr_len = skb_network_header_len(skb);
this_ip_hdr = ip_hdr(skb);
--
1.9.3
^ permalink raw reply related
* [net 1/3] i40e: fix un-necessary Tx hangs
From: Jeff Kirsher @ 2015-01-06 8:44 UTC (permalink / raw)
To: davem; +Cc: Jesse Brandeburg, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <1420533864-13125-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
When the driver was polling with interrupts disabled the hardware
will occasionally not write back descriptors. This patch causes
the driver to detect this situation and force an interrupt to
fire which will flush the stuck descriptor. Does not conflict
with NAPI because if we are already polling the napi_schedule is
ignored. Additionally the extra interrupts are rate limited, so
don't cause a burden to the CPU.
Change-ID: Iba4616d2a71288672a5f08e4512e2704b97335e8
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 35 ++++++++++++++++++++++++-----
1 file changed, 29 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 04b4414..9995598 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -658,6 +658,8 @@ static inline u32 i40e_get_head(struct i40e_ring *tx_ring)
return le32_to_cpu(*(volatile __le32 *)head);
}
+#define WB_STRIDE 0x3
+
/**
* i40e_clean_tx_irq - Reclaim resources after transmit completes
* @tx_ring: tx ring to clean
@@ -759,6 +761,25 @@ static bool i40e_clean_tx_irq(struct i40e_ring *tx_ring, int budget)
tx_ring->q_vector->tx.total_bytes += total_bytes;
tx_ring->q_vector->tx.total_packets += total_packets;
+ /* check to see if there are any non-cache aligned descriptors
+ * waiting to be written back, and kick the hardware to force
+ * them to be written back in case of napi polling
+ */
+ if (budget &&
+ !((i & WB_STRIDE) == WB_STRIDE) &&
+ !test_bit(__I40E_DOWN, &tx_ring->vsi->state) &&
+ (I40E_DESC_UNUSED(tx_ring) != tx_ring->count)) {
+ u32 val = I40E_PFINT_DYN_CTLN_INTENA_MASK |
+ I40E_PFINT_DYN_CTLN_SWINT_TRIG_MASK |
+ I40E_PFINT_DYN_CTLN_SW_ITR_INDX_ENA_MASK;
+ /* allow 00 to be written to the index */
+
+ wr32(&tx_ring->vsi->back->hw,
+ I40E_PFINT_DYN_CTLN(tx_ring->q_vector->v_idx +
+ tx_ring->vsi->base_vector - 1),
+ val);
+ }
+
if (check_for_tx_hang(tx_ring) && i40e_check_tx_hang(tx_ring)) {
/* schedule immediate reset if we believe we hung */
dev_info(tx_ring->dev, "Detected Tx Unit Hang\n"
@@ -777,13 +798,16 @@ static bool i40e_clean_tx_irq(struct i40e_ring *tx_ring, int budget)
netif_stop_subqueue(tx_ring->netdev, tx_ring->queue_index);
dev_info(tx_ring->dev,
- "tx hang detected on queue %d, resetting adapter\n",
+ "tx hang detected on queue %d, reset requested\n",
tx_ring->queue_index);
- tx_ring->netdev->netdev_ops->ndo_tx_timeout(tx_ring->netdev);
+ /* do not fire the reset immediately, wait for the stack to
+ * decide we are truly stuck, also prevents every queue from
+ * simultaneously requesting a reset
+ */
- /* the adapter is about to reset, no point in enabling stuff */
- return true;
+ /* the adapter is about to reset, no point in enabling polling */
+ budget = 1;
}
netdev_tx_completed_queue(netdev_get_tx_queue(tx_ring->netdev,
@@ -806,7 +830,7 @@ static bool i40e_clean_tx_irq(struct i40e_ring *tx_ring, int budget)
}
}
- return budget > 0;
+ return !!budget;
}
/**
@@ -2198,7 +2222,6 @@ static void i40e_tx_map(struct i40e_ring *tx_ring, struct sk_buff *skb,
/* Place RS bit on last descriptor of any packet that spans across the
* 4th descriptor (WB_STRIDE aka 0x3) in a 64B cacheline.
*/
-#define WB_STRIDE 0x3
if (((i & WB_STRIDE) != WB_STRIDE) &&
(first <= &tx_ring->tx_bi[i]) &&
(first >= &tx_ring->tx_bi[i & ~WB_STRIDE])) {
--
1.9.3
^ permalink raw reply related
* [net 0/3][pull request] Intel Wired LAN Driver Updates 2015-01-06
From: Jeff Kirsher @ 2015-01-06 8:44 UTC (permalink / raw)
To: davem; +Cc: Jeff Kirsher, netdev, nhorman, sassmann, jogreene
This series contains fixes to i40e only.
Jesse provides a fix for when the driver was polling with interrupts
disabled the hardware would occasionally not write back descriptors.
His fix causes the driver to detect this situation and force an interrupt
to fire which will flush the stuck descriptor.
Anjali provides a couple of fixes, the first corrects an issue where
the receive port checksum error counter was incrementing incorrectly with
UDP encapsulated tunneled traffic. The second fix resolves an issue where
the driver was examining the outer protocol layer to set the inner protocol
layer checksum offload. In the case of TCP over IPv6 over an IPv4 based
VXLAN, the inner checksum offloads would be set to look for IPv4/UDP
instead of IPv6/TCP, so fixed the issue so that the driver will look at
the proper layer for encapsulation offload settings.
The following are changes since commit 7ce67a38f799d1fb332f672b117efbadedaa5352:
net: ethernet: cpsw: fix hangs with interrupts
and are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net master
Anjali Singhai (2):
i40e: Fix Rx checksum error counter
i40e: Fix bug with TCP over IPv6 over VXLAN
Jesse Brandeburg (1):
i40e: fix un-necessary Tx hangs
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 83 ++++++++++++++++++-----------
1 file changed, 53 insertions(+), 30 deletions(-)
--
1.9.3
^ permalink raw reply
* Re: [PATCH] Revert "ipw2200: select CFG80211_WEXT"
From: Johannes Berg @ 2015-01-06 8:23 UTC (permalink / raw)
To: Paul Bolle
Cc: Arend van Spriel, Linus Torvalds, Marcel Holtmann,
Stanislav Yakovlev, Kalle Valo, Jiri Kosina, linux-wireless,
Network Development, Linux Kernel Mailing List
In-Reply-To: <1420495519.14308.29.camel@x220>
On Mon, 2015-01-05 at 23:05 +0100, Paul Bolle wrote:
> On Mon, 2015-01-05 at 19:57 +0100, Johannes Berg wrote:
> > Multiple other groups of ioctls could be converted in similar patches,
> > until at the end you can completely remove ipw_wx_handlers and rely
> > entirely on cfg80211's wext compatibility.
> >
> > So far the theory - in practice nobody cared enough to start working on
> > any of these drivers, let alone actually has the hardware today.
>
> So my suggestion to make ipw2200 no longer use cfg80211_wext_giwname()
> would actually be backwards. What's actually needed, in theory, is to
> use more of what's provided under CFG80211_WEXT (and, I guess, less of
> what's provided under WIRELESS_EXT). Did I get that right?
Yes, though I'd argue there are multiple levels of truth here.
Yours is the theoretical, hopefully-forward-looking one where we still
expect the driver to actually be modified to take advantage of the new
frameworks (which is independent of wext support towards userspace). In
that scenario, yes, it should use more until it uses all, and then it
can stop concerning itself with wext (which would be a win because
driver/wext interaction is always finicky) (*)
Then there's the other level that you were looking at earlier - simply
removing all of this again from this driver because nobody is going to
work on it. That'd actually make sense and shrink the driver footprint
(no need to load cfg80211 for using almost nothing of it) but this is no
longer feasible since it would again break userspace API - as I said you
can today discover the presence of this device/driver with nl80211 -
that would be broken by removing cfg80211 from here.
And then there's the practical third version that's likely to happen -
i.e. nothing :-)
johannes
(*) FWIW, if done to all drivers it would allow shrinking cfg80211-wext
by a lot because all the EXPORT_SYMBOL would no longer be needed - but
come to think of it we can fix that differently.
^ permalink raw reply
* Re: [PATCH] Revert "ipw2200: select CFG80211_WEXT"
From: Johannes Berg @ 2015-01-06 8:19 UTC (permalink / raw)
To: Arend van Spriel
Cc: Paul Bolle, Linus Torvalds, Marcel Holtmann, Stanislav Yakovlev,
Kalle Valo, Jiri Kosina, linux-wireless, Network Development,
Linux Kernel Mailing List
In-Reply-To: <54AB0C75.1090204-dY08KVG/lbpWk0Htik3J/w@public.gmane.org>
On Mon, 2015-01-05 at 23:13 +0100, Arend van Spriel wrote:
> On 01/05/15 23:05, Paul Bolle wrote:
> > On Mon, 2015-01-05 at 19:57 +0100, Johannes Berg wrote:
> >> Multiple other groups of ioctls could be converted in similar patches,
> >> until at the end you can completely remove ipw_wx_handlers and rely
> >> entirely on cfg80211's wext compatibility.
> >>
> >> So far the theory - in practice nobody cared enough to start working on
> >> any of these drivers, let alone actually has the hardware today.
> >
> > So my suggestion to make ipw2200 no longer use cfg80211_wext_giwname()
> > would actually be backwards. What's actually needed, in theory, is to
> > use more of what's provided under CFG80211_WEXT (and, I guess, less of
> > what's provided under WIRELESS_EXT). Did I get that right?
>
> Yes, but as Johannes indicated it needs consideration what to group in
> the patches.
Oh, that's not strictly necessary - that was just so it would actually
work after each single patch. If you wanted to, you could do it all in a
single huge patch as well :-) I was explaining though why we did the
cfg80211-wext code the way it is - which enables making smaller changes
that don't break the driver.
johannes
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH] Revert "ipw2200: select CFG80211_WEXT"
From: Kalle Valo @ 2015-01-06 8:03 UTC (permalink / raw)
To: Paul Bolle
Cc: Stanislav Yakovlev, Jiri Kosina, Linus Torvalds,
linux-wireless-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <1420297188.2397.3.camel-IWqWACnzNjzz+pZb47iToQ@public.gmane.org>
Paul Bolle <pebolle-IWqWACnzNjzz+pZb47iToQ@public.gmane.org> writes:
> This reverts commit dddd60220f41775e634258efd1b54c6fa81ce706.
>
> The raison d'être of commit dddd60220f41 ("ipw2200: select
> CFG80211_WEXT") was reverted in commit 2d36e008739e ("Revert "cfg80211:
> make WEXT compatibility unselectable""). So revert this commit too.
>
> Signed-off-by: Paul Bolle <pebolle-IWqWACnzNjzz+pZb47iToQ@public.gmane.org>
Based on the discussion I have dropped this patch from my queue.
--
Kalle Valo
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH] GMAC: fix simple_return.cocci warnings
From: David Miller @ 2015-01-06 7:52 UTC (permalink / raw)
To: roger.chen
Cc: joe, fengguang.wu, kbuild-all, peppe.cavallaro, netdev,
linux-kernel
In-Reply-To: <54AB8B23.3080303@rock-chips.com>
From: Roger <roger.chen@rock-chips.com>
Date: Tue, 06 Jan 2015 15:13:39 +0800
> What should I do now?
Nothing, I'm simply not applying this patch.
^ permalink raw reply
* Re: [net-next PATCH v1 08/11] net: rocker: add get flow API operation
From: Scott Feldman @ 2015-01-06 7:40 UTC (permalink / raw)
To: John Fastabend
Cc: Thomas Graf, Jiří Pírko, Jamal Hadi Salim,
simon.horman@netronome.com, Netdev, David S. Miller,
Andy Gospodarek
In-Reply-To: <20141231194852.31070.72727.stgit@nitbit.x32>
On Wed, Dec 31, 2014 at 11:48 AM, John Fastabend
<john.fastabend@gmail.com> wrote:
> Add operations to get flows. I wouldn't mind cleaning this code
> up a bit but my first attempt to do this used macros which shortered
> the code up but when I was done I decided it just made the code
> unreadable and unmaintainable.
>
> I might think about it a bit more but this implementation albeit
> a bit long and repeatative is easier to understand IMO.
Dang, you put a lot of work into this one.
Something doesn't feel right though. In this case, rocker driver just
happened to have cached all the flow/group stuff in hash tables in
software, so you don't need to query thru to the device to extract the
if_flow info. What doesn't feel right is all the work need in the
driver. For each and every driver. get_flows needs to go above
driver, somehow.
Seems the caller of if_flow already knows the flows pushed down with
add_flows/del_flows, and with the err handling can't mess it up.
Is one use-case for get_flows to recover from a fatal OS/driver crash,
and to rely on hardware to recover flow set? In this rocker example,
that's not going to work because driver didn't get thru to device to
get_flows. I think I'd like to know more about the use-cases of
get_flows.
-scott
^ permalink raw reply
* Re: [net-next PATCH v1 05/11] net: rocker: add set flow rules
From: Scott Feldman @ 2015-01-06 7:23 UTC (permalink / raw)
To: John Fastabend
Cc: Thomas Graf, Jiří Pírko, Jamal Hadi Salim,
simon.horman@netronome.com, Netdev, David S. Miller,
Andy Gospodarek
In-Reply-To: <20141231194735.31070.55480.stgit@nitbit.x32>
On Wed, Dec 31, 2014 at 11:47 AM, John Fastabend
<john.fastabend@gmail.com> wrote:
> Implement set flow operations for existing rocker tables.
>
> Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
> ---
> drivers/net/ethernet/rocker/rocker.c | 517 +++++++++++++++++++++++++
> drivers/net/ethernet/rocker/rocker_pipeline.h | 3
> 2 files changed, 519 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/rocker/rocker.c b/drivers/net/ethernet/rocker/rocker.c
> index 4c6787a..c40c58d 100644
> --- a/drivers/net/ethernet/rocker/rocker.c
> +++ b/drivers/net/ethernet/rocker/rocker.c
> @@ -3806,6 +3806,520 @@ static struct net_flow_hdr_node **rocker_get_hgraph(struct net_device *d)
> {
> return rocker_header_nodes;
> }
> +
> +static int is_valid_net_flow_action_arg(struct net_flow_action *a, int id)
> +{
> + struct net_flow_action_arg *args = a->args;
> + int i;
> +
> + for (i = 0; args[i].type != NET_FLOW_ACTION_ARG_TYPE_NULL; i++) {
> + if (a->args[i].type == NET_FLOW_ACTION_ARG_TYPE_NULL ||
> + args[i].type != a->args[i].type)
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int is_valid_net_flow_action(struct net_flow_action *a, int *actions)
> +{
> + int i;
> +
> + for (i = 0; actions[i]; i++) {
> + if (actions[i] == a->uid)
> + return is_valid_net_flow_action_arg(a, a->uid);
> + }
> + return -EINVAL;
> +}
> +
> +static int is_valid_net_flow_match(struct net_flow_field_ref *f,
> + struct net_flow_field_ref *fields)
> +{
> + int i;
> +
> + for (i = 0; fields[i].header; i++) {
> + if (f->header == fields[i].header &&
> + f->field == fields[i].field)
> + return 0;
> + }
> +
> + return -EINVAL;
> +}
> +
> +int is_valid_net_flow(struct net_flow_table *table, struct net_flow_flow *flow)
> +{
> + struct net_flow_field_ref *fields = table->matches;
> + int *actions = table->actions;
> + int i, err;
> +
> + for (i = 0; flow->actions[i].uid; i++) {
> + err = is_valid_net_flow_action(&flow->actions[i], actions);
> + if (err)
> + return -EINVAL;
> + }
> +
> + for (i = 0; flow->matches[i].header; i++) {
> + err = is_valid_net_flow_match(&flow->matches[i], fields);
> + if (err)
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
All the above doesn't look rocker-specific...up-level?
> +
> +static u32 rocker_goto_value(u32 id)
> +{
> + switch (id) {
> + case ROCKER_FLOW_TABLE_ID_INGRESS_PORT:
> + return ROCKER_OF_DPA_TABLE_ID_INGRESS_PORT;
> + case ROCKER_FLOW_TABLE_ID_VLAN:
> + return ROCKER_OF_DPA_TABLE_ID_VLAN;
> + case ROCKER_FLOW_TABLE_ID_TERMINATION_MAC:
> + return ROCKER_OF_DPA_TABLE_ID_TERMINATION_MAC;
> + case ROCKER_FLOW_TABLE_ID_UNICAST_ROUTING:
> + return ROCKER_OF_DPA_TABLE_ID_UNICAST_ROUTING;
> + case ROCKER_FLOW_TABLE_ID_MULTICAST_ROUTING:
> + return ROCKER_OF_DPA_TABLE_ID_MULTICAST_ROUTING;
> + case ROCKER_FLOW_TABLE_ID_BRIDGING:
> + return ROCKER_OF_DPA_TABLE_ID_BRIDGING;
> + case ROCKER_FLOW_TABLE_ID_ACL_POLICY:
> + return ROCKER_OF_DPA_TABLE_ID_ACL_POLICY;
> + default:
> + return 0;
> + }
> +}
Could the OF-DPA table IDs be used in the flow table defs? I think I
remember your answer was no because OF-DPA uses INGRESS_PORT ID == 0,
and 0 is a special value for if_flow tables. Bummer.
> +
> +static int rocker_flow_set_ig_port(struct net_device *dev,
> + struct net_flow_flow *flow)
> +{
> + struct rocker_port *rocker_port = netdev_priv(dev);
> + enum rocker_of_dpa_table_id goto_tbl;
> + u32 in_lport_mask = 0xffff0000;
> + u32 in_lport = 0;
why initialize these two?
> + int err, flags = 0;
> +
> + err = is_valid_net_flow(&ingress_port_table, flow);
> + if (err)
> + return err;
> +
> + /* ingress port table only supports one field/mask/action this
> + * simplifies the key construction and we can assume the values
> + * are the correct types/mask/action by valid check above. The
> + * user could pass multiple match/actions in a message with the
> + * same field multiple times currently the valid test does not
> + * catch this and we just use the first specified.
> + */
> + in_lport = flow->matches[0].value_u32;
> + in_lport_mask = flow->matches[0].mask_u32;
> + goto_tbl = rocker_goto_value(flow->actions[0].args[0].value_u16);
> +
> + err = rocker_flow_tbl_ig_port(rocker_port, flags,
> + in_lport, in_lport_mask,
> + goto_tbl);
> + return err;
> +}
> +
> +static int rocker_flow_set_vlan(struct net_device *dev,
> + struct net_flow_flow *flow)
> +{
> + enum rocker_of_dpa_table_id goto_tbl;
> + struct rocker_port *rocker_port = netdev_priv(dev);
rocker style thing: put rocker_port decl first (sorry for being so pedantic).
> + int i, err = 0, flags = 0;
> + u32 in_lport;
> + __be16 vlan_id, vlan_id_mask, new_vlan_id;
> + bool untagged, have_in_lport = false;
> +
> + err = is_valid_net_flow(&vlan_table, flow);
> + if (err)
> + return err;
> +
> + goto_tbl = ROCKER_OF_DPA_TABLE_ID_TERMINATION_MAC;
> +
> + /* If user does not specify vid match default to any */
> + vlan_id = 1;
htons()?
Not sure. Rocker convention is vlan_id is network-order, but some
places you'll see vid and that's host-order.
> + vlan_id_mask = 0;
> +
> + for (i = 0; flow->matches && flow->matches[i].instance; i++) {
> + switch (flow->matches[i].instance) {
> + case HEADER_INSTANCE_IN_LPORT:
> + in_lport = flow->matches[i].value_u32;
> + have_in_lport = true;
> + break;
> + case HEADER_INSTANCE_VLAN_OUTER:
> + if (flow->matches[i].field != HEADER_VLAN_VID)
> + break;
> +
> + vlan_id = htons(flow->matches[i].value_u16);
> + vlan_id_mask = htons(flow->matches[i].mask_u16);
> + break;
> + default:
> + return -EINVAL;
> + }
> + }
> +
> + /* If user does not specify a new vlan id use default vlan id */
> + new_vlan_id = rocker_port_vid_to_vlan(rocker_port, vlan_id, &untagged);
> +
> + for (i = 0; flow->actions && flow->actions[i].uid; i++) {
> + struct net_flow_action_arg *arg = &flow->actions[i].args[0];
> +
> + switch (flow->actions[i].uid) {
> + case ACTION_SET_GOTO_TABLE:
> + goto_tbl = rocker_goto_value(arg->value_u16);
> + break;
> + case ACTION_SET_VLAN_ID:
> + new_vlan_id = htons(arg->value_u16);
> + if (new_vlan_id)
> + untagged = false;
> + break;
> + }
> + }
> +
> + if (!have_in_lport)
> + return -EINVAL;
This can be moved up, before second for loop
> +
> + err = rocker_flow_tbl_vlan(rocker_port, flags, in_lport,
> + vlan_id, vlan_id_mask, goto_tbl,
> + untagged, new_vlan_id);
> + return err;
> +}
> +
> +static int rocker_flow_set_term_mac(struct net_device *dev,
> + struct net_flow_flow *flow)
> +{
> + struct rocker_port *rocker_port = netdev_priv(dev);
> + __be16 vlan_id, vlan_id_mask, ethtype = 0;
> + const u8 *eth_dst, *eth_dst_mask;
> + u32 in_lport, in_lport_mask;
> + int i, err = 0, flags = 0;
> + bool copy_to_cpu;
> +
> + eth_dst = NULL;
> + eth_dst_mask = NULL;
> +
Needed?
> + err = is_valid_net_flow(&term_mac_table, flow);
> + if (err)
> + return err;
> +
> + /* If user does not specify vid match default to any */
> + vlan_id = rocker_port->internal_vlan_id;
> + vlan_id_mask = 0;
> +
> + /* If user does not specify in_lport match default to any */
> + in_lport = rocker_port->lport;
> + in_lport_mask = 0;
> +
> + /* If user does not specify a mac address match any */
> + eth_dst = rocker_port->dev->dev_addr;
> + eth_dst_mask = zero_mac;
> +
> + for (i = 0; flow->matches && flow->matches[i].instance; i++) {
> + switch (flow->matches[i].instance) {
> + case HEADER_INSTANCE_IN_LPORT:
> + in_lport = flow->matches[i].value_u32;
> + in_lport_mask = flow->matches[i].mask_u32;
> + break;
> + case HEADER_INSTANCE_VLAN_OUTER:
> + if (flow->matches[i].field != HEADER_VLAN_VID)
> + break;
> +
> + vlan_id = htons(flow->matches[i].value_u16);
> + vlan_id_mask = htons(flow->matches[i].mask_u16);
> + break;
> + case HEADER_INSTANCE_ETHERNET:
> + switch (flow->matches[i].field) {
> + case HEADER_ETHERNET_DST_MAC:
> + eth_dst = (u8 *)&flow->matches[i].value_u64;
> + eth_dst_mask = (u8 *)&flow->matches[i].mask_u64;
> + break;
> + case HEADER_ETHERNET_ETHERTYPE:
> + ethtype = htons(flow->matches[i].value_u16);
> + break;
> + default:
> + return -EINVAL;
> + }
> + break;
> + default:
> + return -EINVAL;
> + }
> + }
> +
> + if (!ethtype)
> + return -EINVAL;
> +
> + /* By default do not copy to cpu */
> + copy_to_cpu = false;
> +
> + for (i = 0; flow->actions && flow->actions[i].uid; i++) {
> + switch (flow->actions[i].uid) {
> + case ACTION_COPY_TO_CPU:
> + copy_to_cpu = true;
> + break;
> + default:
> + return -EINVAL;
> + }
> + }
> +
> + err = rocker_flow_tbl_term_mac(rocker_port, in_lport, in_lport_mask,
> + ethtype, eth_dst, eth_dst_mask,
> + vlan_id, vlan_id_mask,
> + copy_to_cpu, flags);
> + return err;
> +}
> +
> +static int rocker_flow_set_ucast_routing(struct net_device *dev,
> + struct net_flow_flow *flow)
> +{
> + return -EOPNOTSUPP;
> +}
> +
> +static int rocker_flow_set_mcast_routing(struct net_device *dev,
> + struct net_flow_flow *flow)
> +{
> + return -EOPNOTSUPP;
> +}
> +
> +static int rocker_flow_set_bridge(struct net_device *dev,
> + struct net_flow_flow *flow)
> +{
> + enum rocker_of_dpa_table_id goto_tbl;
> + struct rocker_port *rocker_port = netdev_priv(dev);
> + u32 in_lport, in_lport_mask, group_id, tunnel_id;
> + __be16 vlan_id, vlan_id_mask;
> + const u8 *eth_dst, *eth_dst_mask;
> + int i, err = 0, flags = 0;
> + bool copy_to_cpu;
> +
> + err = is_valid_net_flow(&bridge_table, flow);
> + if (err)
> + return err;
> +
> + goto_tbl = ROCKER_OF_DPA_TABLE_ID_ACL_POLICY;
> +
> + /* If user does not specify vid match default to any */
> + vlan_id = rocker_port->internal_vlan_id;
> + vlan_id_mask = 0;
> +
> + /* If user does not specify in_lport match default to any */
> + in_lport = rocker_port->lport;
> + in_lport_mask = 0;
> +
> + /* If user does not specify a mac address match any */
> + eth_dst = rocker_port->dev->dev_addr;
> + eth_dst_mask = NULL;
> +
> + /* Do not support for tunnel_id yet. */
> + tunnel_id = 0;
> +
> + for (i = 0; flow->matches && flow->matches[i].instance; i++) {
> + switch (flow->matches[i].instance) {
> + case HEADER_INSTANCE_IN_LPORT:
> + in_lport = flow->matches[i].value_u32;
> + in_lport_mask = flow->matches[i].mask_u32;
> + break;
> + case HEADER_INSTANCE_VLAN_OUTER:
> + if (flow->matches[i].field != HEADER_VLAN_VID)
> + break;
> +
> + vlan_id = htons(flow->matches[i].value_u16);
> + vlan_id_mask = htons(flow->matches[i].mask_u16);
> + break;
> + case HEADER_INSTANCE_ETHERNET:
> + switch (flow->matches[i].field) {
> + case HEADER_ETHERNET_DST_MAC:
> + eth_dst = (u8 *)&flow->matches[i].value_u64;
> + eth_dst_mask = (u8 *)&flow->matches[i].mask_u64;
> + break;
> + default:
> + return -EINVAL;
> + }
> + break;
> + default:
> + return -EINVAL;
> + }
> + }
> +
> + /* By default do not copy to cpu and skip group assignment */
> + copy_to_cpu = false;
> + group_id = ROCKER_GROUP_NONE;
> +
> + for (i = 0; flow->actions && flow->actions[i].uid; i++) {
> + struct net_flow_action_arg *arg = &flow->actions[i].args[0];
> +
> + switch (flow->actions[i].uid) {
> + case ACTION_SET_GOTO_TABLE:
> + goto_tbl = rocker_goto_value(arg->value_u16);
> + break;
> + case ACTION_COPY_TO_CPU:
> + copy_to_cpu = true;
> + break;
> + case ACTION_SET_GROUP_ID:
> + group_id = arg->value_u32;
> + break;
> + default:
> + return -EINVAL;
> + }
> + }
> +
> + /* Ignoring eth_dst_mask it seems to cause a EINVAL return code */
> + err = rocker_flow_tbl_bridge(rocker_port, flags,
> + eth_dst, eth_dst_mask,
> + vlan_id, tunnel_id,
> + goto_tbl, group_id, copy_to_cpu);
> + return err;
> +}
> +
> +static int rocker_flow_set_acl(struct net_device *dev,
> + struct net_flow_flow *flow)
> +{
> + struct rocker_port *rocker_port = netdev_priv(dev);
> + u32 in_lport, in_lport_mask, group_id, tunnel_id;
> + __be16 vlan_id, vlan_id_mask, ethtype = 0;
> + const u8 *eth_dst, *eth_src, *eth_dst_mask, *eth_src_mask;
> + u8 protocol, protocol_mask, dscp, dscp_mask;
> + int i, err = 0, flags = 0;
> +
> + err = is_valid_net_flow(&bridge_table, flow);
> + if (err)
> + return err;
> +
> + /* If user does not specify vid match default to any */
> + vlan_id = rocker_port->internal_vlan_id;
> + vlan_id_mask = 0;
> +
> + /* If user does not specify in_lport match default to any */
> + in_lport = rocker_port->lport;
> + in_lport_mask = 0;
> +
> + /* If user does not specify a mac address match any */
> + eth_dst = rocker_port->dev->dev_addr;
> + eth_src = zero_mac;
> + eth_dst_mask = NULL;
> + eth_src_mask = NULL;
> +
> + /* If user does not set protocol/dscp mask them out */
> + protocol = 0;
> + dscp = 0;
> + protocol_mask = 0;
> + dscp_mask = 0;
> +
> + /* Do not support for tunnel_id yet. */
> + tunnel_id = 0;
> +
> + for (i = 0; flow->matches && flow->matches[i].instance; i++) {
> + switch (flow->matches[i].instance) {
> + case HEADER_INSTANCE_IN_LPORT:
> + in_lport = flow->matches[i].value_u32;
> + in_lport_mask = flow->matches[i].mask_u32;
> + break;
> + case HEADER_INSTANCE_VLAN_OUTER:
> + if (flow->matches[i].field != HEADER_VLAN_VID)
> + break;
> +
> + vlan_id = htons(flow->matches[i].value_u16);
> + vlan_id_mask = htons(flow->matches[i].mask_u16);
> + break;
> + case HEADER_INSTANCE_ETHERNET:
> + switch (flow->matches[i].field) {
> + case HEADER_ETHERNET_SRC_MAC:
> + eth_src = (u8 *)&flow->matches[i].value_u64;
> + eth_src_mask = (u8 *)&flow->matches[i].mask_u64;
> + break;
> + case HEADER_ETHERNET_DST_MAC:
> + eth_dst = (u8 *)&flow->matches[i].value_u64;
> + eth_dst_mask = (u8 *)&flow->matches[i].mask_u64;
> + break;
> + case HEADER_ETHERNET_ETHERTYPE:
> + ethtype = htons(flow->matches[i].value_u16);
> + break;
> + default:
> + return -EINVAL;
> + }
> + break;
> + case HEADER_INSTANCE_IPV4:
> + switch (flow->matches[i].field) {
> + case HEADER_IPV4_PROTOCOL:
> + protocol = flow->matches[i].value_u8;
> + protocol_mask = flow->matches[i].mask_u8;
> + break;
> + case HEADER_IPV4_DSCP:
> + dscp = flow->matches[i].value_u8;
> + dscp_mask = flow->matches[i].mask_u8;
> + break;
> + default:
> + return -EINVAL;
> + }
> + default:
> + return -EINVAL;
> + }
> + }
> +
> + /* By default do not copy to cpu and skip group assignment */
> + group_id = ROCKER_GROUP_NONE;
> +
> + for (i = 0; flow->actions && flow->actions[i].uid; i++) {
> + switch (flow->actions[i].uid) {
> + case ACTION_SET_GROUP_ID:
> + group_id = flow->actions[i].args[0].value_u32;
> + break;
> + default:
> + return -EINVAL;
> + }
> + }
> +
> + err = rocker_flow_tbl_acl(rocker_port, flags,
> + in_lport, in_lport_mask,
> + eth_src, eth_src_mask,
> + eth_dst, eth_dst_mask, ethtype,
> + vlan_id, vlan_id_mask,
> + protocol, protocol_mask,
> + dscp, dscp_mask,
> + group_id);
> + return err;
> +}
> +
> +static int rocker_set_flows(struct net_device *dev,
> + struct net_flow_flow *flow)
> +{
> + int err = -EINVAL;
> +
> + if (!flow->matches || !flow->actions)
> + return -EINVAL;
> +
> + switch (flow->table_id) {
> + case ROCKER_FLOW_TABLE_ID_INGRESS_PORT:
> + err = rocker_flow_set_ig_port(dev, flow);
> + break;
> + case ROCKER_FLOW_TABLE_ID_VLAN:
> + err = rocker_flow_set_vlan(dev, flow);
> + break;
> + case ROCKER_FLOW_TABLE_ID_TERMINATION_MAC:
> + err = rocker_flow_set_term_mac(dev, flow);
> + break;
> + case ROCKER_FLOW_TABLE_ID_UNICAST_ROUTING:
> + err = rocker_flow_set_ucast_routing(dev, flow);
> + break;
> + case ROCKER_FLOW_TABLE_ID_MULTICAST_ROUTING:
> + err = rocker_flow_set_mcast_routing(dev, flow);
> + break;
> + case ROCKER_FLOW_TABLE_ID_BRIDGING:
> + err = rocker_flow_set_bridge(dev, flow);
> + break;
> + case ROCKER_FLOW_TABLE_ID_ACL_POLICY:
> + err = rocker_flow_set_acl(dev, flow);
> + break;
> + default:
> + break;
> + }
> +
> + return err;
> +}
> +
> +static int rocker_del_flows(struct net_device *dev,
> + struct net_flow_flow *flow)
> +{
> + return -EOPNOTSUPP;
> +}
> #endif
>
> static const struct net_device_ops rocker_port_netdev_ops = {
> @@ -3828,6 +4342,9 @@ static const struct net_device_ops rocker_port_netdev_ops = {
> .ndo_flow_get_actions = rocker_get_actions,
> .ndo_flow_get_tbl_graph = rocker_get_tgraph,
> .ndo_flow_get_hdr_graph = rocker_get_hgraph,
> +
> + .ndo_flow_set_flows = rocker_set_flows,
> + .ndo_flow_del_flows = rocker_del_flows,
> #endif
> };
Looks good overall to me
> diff --git a/drivers/net/ethernet/rocker/rocker_pipeline.h b/drivers/net/ethernet/rocker/rocker_pipeline.h
> index 9544339..701e139 100644
> --- a/drivers/net/ethernet/rocker/rocker_pipeline.h
> +++ b/drivers/net/ethernet/rocker/rocker_pipeline.h
> @@ -527,6 +527,7 @@ enum rocker_flow_table_id_space {
> ROCKER_FLOW_TABLE_ID_VLAN,
> ROCKER_FLOW_TABLE_ID_TERMINATION_MAC,
> ROCKER_FLOW_TABLE_ID_UNICAST_ROUTING,
> + ROCKER_FLOW_TABLE_ID_MULTICAST_ROUTING,
> ROCKER_FLOW_TABLE_ID_BRIDGING,
> ROCKER_FLOW_TABLE_ID_ACL_POLICY,
> ROCKER_FLOW_TABLE_NULL = 0,
> @@ -588,7 +589,7 @@ struct net_flow_table acl_table = {
>
> struct net_flow_table null_table = {
> .name = "",
> - .uid = 0,
> + .uid = ROCKER_FLOW_TABLE_NULL,
> .source = 0,
> .size = 0,
> .matches = NULL,
>
Move these changes to previous patch?
^ permalink raw reply
* Re: [PATCH] GMAC: fix simple_return.cocci warnings
From: Joe Perches @ 2015-01-06 7:26 UTC (permalink / raw)
To: Roger
Cc: David Miller, fengguang.wu, kbuild-all, peppe.cavallaro, netdev,
linux-kernel
In-Reply-To: <54AB8B23.3080303@rock-chips.com>
On Tue, 2015-01-06 at 15:13 +0800, Roger wrote:
> What should I do now?
I think it would be better to change
"int gmac_clk_enable" to "void gmac_clk_enable"
(it always returns 0)
This function should simply call gmac_clk_enable
and return 0;
> >>> --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> >>> +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> >>> @@ -422,11 +422,7 @@ static int rk_gmac_init(struct platform_
> >>> if (ret)
> >>> return ret;
> >>>
> >>> - ret = gmac_clk_enable(bsp_priv, true);
> >>> - if (ret)
> >>> - return ret;
> >>> -
> >>> - return 0;
> >>> + return gmac_clk_enable(bsp_priv, true);
> >> I think this change is not particularly better.
> >>
> >> When the pattern is multiply repeated like:
> > ...
> >> I think it's better to not change the last
> >> test in the sequence just to minimize overall
> >> line count.
> > I think it's a wash and that both ways are about the same to me.
> >
> > I won't apply this, sorry.
^ permalink raw reply
* [PATCH net-next 5/5] tipc: convert tipc reference table to use generic rhashtable
From: Ying Xue @ 2015-01-06 7:23 UTC (permalink / raw)
To: tgraf; +Cc: jon.maloy, netdev, Paul.Gortmaker, tipc-discussion, davem
In-Reply-To: <1420529003-22244-1-git-send-email-ying.xue@windriver.com>
As tipc reference table is statically allocated, its memory size
requested on stack initialization stage is quite big even if the
maximum port number is just restricted to 8191 currently, however,
the number already becomes insufficient in practice. But if the
maximum ports is allowed to its theory value - 2^32, its consumed
memory size will reach a ridiculously unacceptable value. Apart from
this, heavy tipc users spend a considerable amount of time in
tipc_sk_get() due to the read-lock on ref_table_lock.
If tipc reference table is converted with generic rhashtable, above
mentioned both disadvantages would be resolved respectively: making
use of the new resizable hash table can avoid locking on the lookup;
smaller memory size is required at initial stage, for example, 256
hash bucket slots are requested at the beginning phase instead of
allocating the entire 8191 slots in old mode. The hash table will
grow if entries exceeds 75% of table size up to a total table size
of 1M, and it will automatically shrink if usage falls below 30%,
but the minimum table size is allowed down to 256.
Also converts ref_table_lock to a separate mutex to protect hash table
mutations on write side. Lastly defers the release of the socket
reference using call_rcu() to allow using an RCU read-side protected
call to rhashtable_lookup().
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Erik Hugne <erik.hugne@ericsson.com>
Cc: Thomas Graf <tgraf@suug.ch>
---
net/tipc/Kconfig | 12 --
net/tipc/config.c | 24 +--
net/tipc/core.c | 10 +-
net/tipc/core.h | 3 -
net/tipc/socket.c | 480 +++++++++++++++++++----------------------------------
net/tipc/socket.h | 4 +-
6 files changed, 180 insertions(+), 353 deletions(-)
diff --git a/net/tipc/Kconfig b/net/tipc/Kconfig
index c890848..91c8a8e 100644
--- a/net/tipc/Kconfig
+++ b/net/tipc/Kconfig
@@ -20,18 +20,6 @@ menuconfig TIPC
If in doubt, say N.
-config TIPC_PORTS
- int "Maximum number of ports in a node"
- depends on TIPC
- range 127 65535
- default "8191"
- help
- Specifies how many ports can be supported by a node.
- Can range from 127 to 65535 ports; default is 8191.
-
- Setting this to a smaller value saves some memory,
- setting it to higher allows for more ports.
-
config TIPC_MEDIA_IB
bool "InfiniBand media type support"
depends on TIPC && INFINIBAND_IPOIB
diff --git a/net/tipc/config.c b/net/tipc/config.c
index 876f4c6..0b3a90e 100644
--- a/net/tipc/config.c
+++ b/net/tipc/config.c
@@ -183,22 +183,6 @@ static struct sk_buff *cfg_set_own_addr(void)
return tipc_cfg_reply_error_string("cannot change to network mode");
}
-static struct sk_buff *cfg_set_max_ports(void)
-{
- u32 value;
-
- if (!TLV_CHECK(req_tlv_area, req_tlv_space, TIPC_TLV_UNSIGNED))
- return tipc_cfg_reply_error_string(TIPC_CFG_TLV_ERROR);
- value = ntohl(*(__be32 *)TLV_DATA(req_tlv_area));
- if (value == tipc_max_ports)
- return tipc_cfg_reply_none();
- if (value < 127 || value > 65535)
- return tipc_cfg_reply_error_string(TIPC_CFG_INVALID_VALUE
- " (max ports must be 127-65535)");
- return tipc_cfg_reply_error_string(TIPC_CFG_NOT_SUPPORTED
- " (cannot change max ports while TIPC is active)");
-}
-
static struct sk_buff *cfg_set_netid(void)
{
u32 value;
@@ -285,15 +269,9 @@ struct sk_buff *tipc_cfg_do_cmd(u32 orig_node, u16 cmd, const void *request_area
case TIPC_CMD_SET_NODE_ADDR:
rep_tlv_buf = cfg_set_own_addr();
break;
- case TIPC_CMD_SET_MAX_PORTS:
- rep_tlv_buf = cfg_set_max_ports();
- break;
case TIPC_CMD_SET_NETID:
rep_tlv_buf = cfg_set_netid();
break;
- case TIPC_CMD_GET_MAX_PORTS:
- rep_tlv_buf = tipc_cfg_reply_unsigned(tipc_max_ports);
- break;
case TIPC_CMD_GET_NETID:
rep_tlv_buf = tipc_cfg_reply_unsigned(tipc_net_id);
break;
@@ -317,6 +295,8 @@ struct sk_buff *tipc_cfg_do_cmd(u32 orig_node, u16 cmd, const void *request_area
case TIPC_CMD_SET_REMOTE_MNG:
case TIPC_CMD_GET_REMOTE_MNG:
case TIPC_CMD_DUMP_LOG:
+ case TIPC_CMD_SET_MAX_PORTS:
+ case TIPC_CMD_GET_MAX_PORTS:
rep_tlv_buf = tipc_cfg_reply_error_string(TIPC_CFG_NOT_SUPPORTED
" (obsolete command)");
break;
diff --git a/net/tipc/core.c b/net/tipc/core.c
index a5737b8..71b2ada 100644
--- a/net/tipc/core.c
+++ b/net/tipc/core.c
@@ -34,6 +34,8 @@
* POSSIBILITY OF SUCH DAMAGE.
*/
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
#include "core.h"
#include "name_table.h"
#include "subscr.h"
@@ -47,7 +49,6 @@ int tipc_random __read_mostly;
/* configurable TIPC parameters */
u32 tipc_own_addr __read_mostly;
-int tipc_max_ports __read_mostly;
int tipc_net_id __read_mostly;
int sysctl_tipc_rmem[3] __read_mostly; /* min/default/max */
@@ -84,9 +85,9 @@ static void tipc_core_stop(void)
tipc_netlink_stop();
tipc_subscr_stop();
tipc_nametbl_stop();
- tipc_sk_ref_table_stop();
tipc_socket_stop();
tipc_unregister_sysctl();
+ tipc_sk_rht_destroy();
}
/**
@@ -98,7 +99,7 @@ static int tipc_core_start(void)
get_random_bytes(&tipc_random, sizeof(tipc_random));
- err = tipc_sk_ref_table_init(tipc_max_ports, tipc_random);
+ err = tipc_sk_rht_init();
if (err)
goto out_reftbl;
@@ -138,7 +139,7 @@ out_socket:
out_netlink:
tipc_nametbl_stop();
out_nametbl:
- tipc_sk_ref_table_stop();
+ tipc_sk_rht_destroy();
out_reftbl:
return err;
}
@@ -150,7 +151,6 @@ static int __init tipc_init(void)
pr_info("Activated (version " TIPC_MOD_VER ")\n");
tipc_own_addr = 0;
- tipc_max_ports = CONFIG_TIPC_PORTS;
tipc_net_id = 4711;
sysctl_tipc_rmem[0] = TIPC_CONN_OVERLOAD_LIMIT >> 4 <<
diff --git a/net/tipc/core.h b/net/tipc/core.h
index 8460213..56fe422 100644
--- a/net/tipc/core.h
+++ b/net/tipc/core.h
@@ -37,8 +37,6 @@
#ifndef _TIPC_CORE_H
#define _TIPC_CORE_H
-#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
-
#include <linux/tipc.h>
#include <linux/tipc_config.h>
#include <linux/tipc_netlink.h>
@@ -79,7 +77,6 @@ int tipc_snprintf(char *buf, int len, const char *fmt, ...);
* Global configuration variables
*/
extern u32 tipc_own_addr __read_mostly;
-extern int tipc_max_ports __read_mostly;
extern int tipc_net_id __read_mostly;
extern int sysctl_tipc_rmem[3] __read_mostly;
extern int sysctl_tipc_named_timeout __read_mostly;
diff --git a/net/tipc/socket.c b/net/tipc/socket.c
index 4731cad..3335ccf 100644
--- a/net/tipc/socket.c
+++ b/net/tipc/socket.c
@@ -34,22 +34,25 @@
* POSSIBILITY OF SUCH DAMAGE.
*/
+#include <linux/rhashtable.h>
+#include <linux/jhash.h>
#include "core.h"
#include "name_table.h"
#include "node.h"
#include "link.h"
-#include <linux/export.h>
#include "config.h"
#include "socket.h"
-#define SS_LISTENING -1 /* socket is listening */
-#define SS_READY -2 /* socket is connectionless */
+#define SS_LISTENING -1 /* socket is listening */
+#define SS_READY -2 /* socket is connectionless */
-#define CONN_TIMEOUT_DEFAULT 8000 /* default connect timeout = 8s */
-#define CONN_PROBING_INTERVAL 3600000 /* [ms] => 1 h */
-#define TIPC_FWD_MSG 1
-#define TIPC_CONN_OK 0
-#define TIPC_CONN_PROBING 1
+#define CONN_TIMEOUT_DEFAULT 8000 /* default connect timeout = 8s */
+#define CONN_PROBING_INTERVAL 3600000 /* [ms] => 1 h */
+#define TIPC_FWD_MSG 1
+#define TIPC_CONN_OK 0
+#define TIPC_CONN_PROBING 1
+#define TIPC_MAX_PORT 0xffffffff
+#define TIPC_MIN_PORT 1
/**
* struct tipc_sock - TIPC socket structure
@@ -59,7 +62,7 @@
* @conn_instance: TIPC instance used when connection was established
* @published: non-zero if port has one or more associated names
* @max_pkt: maximum packet size "hint" used when building messages sent by port
- * @ref: unique reference to port in TIPC object registry
+ * @portid: unique port identity in TIPC socket hash table
* @phdr: preformatted message header used when sending messages
* @port_list: adjacent ports in TIPC's global list of ports
* @publications: list of publications for port
@@ -74,6 +77,8 @@
* @link_cong: non-zero if owner must sleep because of link congestion
* @sent_unacked: # messages sent by socket, and not yet acked by peer
* @rcv_unacked: # messages read by user, but not yet acked back to peer
+ * @node: hash table node
+ * @rcu: rcu struct for tipc_sock
*/
struct tipc_sock {
struct sock sk;
@@ -82,7 +87,7 @@ struct tipc_sock {
u32 conn_instance;
int published;
u32 max_pkt;
- u32 ref;
+ u32 portid;
struct tipc_msg phdr;
struct list_head sock_list;
struct list_head publications;
@@ -95,6 +100,8 @@ struct tipc_sock {
bool link_cong;
uint sent_unacked;
uint rcv_unacked;
+ struct rhash_head node;
+ struct rcu_head rcu;
};
static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *skb);
@@ -103,16 +110,14 @@ static void tipc_write_space(struct sock *sk);
static int tipc_release(struct socket *sock);
static int tipc_accept(struct socket *sock, struct socket *new_sock, int flags);
static int tipc_wait_for_sndmsg(struct socket *sock, long *timeo_p);
-static void tipc_sk_timeout(unsigned long ref);
+static void tipc_sk_timeout(unsigned long portid);
static int tipc_sk_publish(struct tipc_sock *tsk, uint scope,
struct tipc_name_seq const *seq);
static int tipc_sk_withdraw(struct tipc_sock *tsk, uint scope,
struct tipc_name_seq const *seq);
-static u32 tipc_sk_ref_acquire(struct tipc_sock *tsk);
-static void tipc_sk_ref_discard(u32 ref);
-static struct tipc_sock *tipc_sk_get(u32 ref);
-static struct tipc_sock *tipc_sk_get_next(u32 *ref);
-static void tipc_sk_put(struct tipc_sock *tsk);
+static struct tipc_sock *tipc_sk_lookup(u32 portid);
+static int tipc_sk_insert(struct tipc_sock *tsk);
+static void tipc_sk_remove(struct tipc_sock *tsk);
static const struct proto_ops packet_ops;
static const struct proto_ops stream_ops;
@@ -174,6 +179,9 @@ static const struct nla_policy tipc_nl_sock_policy[TIPC_NLA_SOCK_MAX + 1] = {
* - port reference
*/
+/* Protects tipc socket hash table mutations */
+static struct rhashtable tipc_sk_rht;
+
static u32 tsk_peer_node(struct tipc_sock *tsk)
{
return msg_destnode(&tsk->phdr);
@@ -305,7 +313,6 @@ static int tipc_sk_create(struct net *net, struct socket *sock,
struct sock *sk;
struct tipc_sock *tsk;
struct tipc_msg *msg;
- u32 ref;
/* Validate arguments */
if (unlikely(protocol != 0))
@@ -339,24 +346,22 @@ static int tipc_sk_create(struct net *net, struct socket *sock,
return -ENOMEM;
tsk = tipc_sk(sk);
- ref = tipc_sk_ref_acquire(tsk);
- if (!ref) {
- pr_warn("Socket create failed; reference table exhausted\n");
- return -ENOMEM;
- }
tsk->max_pkt = MAX_PKT_DEFAULT;
- tsk->ref = ref;
INIT_LIST_HEAD(&tsk->publications);
msg = &tsk->phdr;
tipc_msg_init(msg, TIPC_LOW_IMPORTANCE, TIPC_NAMED_MSG,
NAMED_H_SIZE, 0);
- msg_set_origport(msg, ref);
/* Finish initializing socket data structures */
sock->ops = ops;
sock->state = state;
sock_init_data(sock, sk);
- k_init_timer(&tsk->timer, (Handler)tipc_sk_timeout, ref);
+ if (tipc_sk_insert(tsk)) {
+ pr_warn("Socket create failed; port numbrer exhausted\n");
+ return -EINVAL;
+ }
+ msg_set_origport(msg, tsk->portid);
+ k_init_timer(&tsk->timer, (Handler)tipc_sk_timeout, tsk->portid);
sk->sk_backlog_rcv = tipc_backlog_rcv;
sk->sk_rcvbuf = sysctl_tipc_rmem[1];
sk->sk_data_ready = tipc_data_ready;
@@ -442,6 +447,13 @@ int tipc_sock_accept_local(struct socket *sock, struct socket **newsock,
return ret;
}
+static void tipc_sk_callback(struct rcu_head *head)
+{
+ struct tipc_sock *tsk = container_of(head, struct tipc_sock, rcu);
+
+ sock_put(&tsk->sk);
+}
+
/**
* tipc_release - destroy a TIPC socket
* @sock: socket to destroy
@@ -491,7 +503,7 @@ static int tipc_release(struct socket *sock)
(sock->state == SS_CONNECTED)) {
sock->state = SS_DISCONNECTING;
tsk->connected = 0;
- tipc_node_remove_conn(dnode, tsk->ref);
+ tipc_node_remove_conn(dnode, tsk->portid);
}
if (tipc_msg_reverse(skb, &dnode, TIPC_ERR_NO_PORT))
tipc_link_xmit_skb(skb, dnode, 0);
@@ -499,16 +511,16 @@ static int tipc_release(struct socket *sock)
}
tipc_sk_withdraw(tsk, 0, NULL);
- tipc_sk_ref_discard(tsk->ref);
k_cancel_timer(&tsk->timer);
+ tipc_sk_remove(tsk);
if (tsk->connected) {
skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE, TIPC_CONN_MSG,
SHORT_H_SIZE, 0, dnode, tipc_own_addr,
tsk_peer_port(tsk),
- tsk->ref, TIPC_ERR_NO_PORT);
+ tsk->portid, TIPC_ERR_NO_PORT);
if (skb)
- tipc_link_xmit_skb(skb, dnode, tsk->ref);
- tipc_node_remove_conn(dnode, tsk->ref);
+ tipc_link_xmit_skb(skb, dnode, tsk->portid);
+ tipc_node_remove_conn(dnode, tsk->portid);
}
k_term_timer(&tsk->timer);
@@ -518,7 +530,8 @@ static int tipc_release(struct socket *sock)
/* Reject any messages that accumulated in backlog queue */
sock->state = SS_DISCONNECTING;
release_sock(sk);
- sock_put(sk);
+
+ call_rcu(&tsk->rcu, tipc_sk_callback);
sock->sk = NULL;
return 0;
@@ -611,7 +624,7 @@ static int tipc_getname(struct socket *sock, struct sockaddr *uaddr,
addr->addr.id.ref = tsk_peer_port(tsk);
addr->addr.id.node = tsk_peer_node(tsk);
} else {
- addr->addr.id.ref = tsk->ref;
+ addr->addr.id.ref = tsk->portid;
addr->addr.id.node = tipc_own_addr;
}
@@ -946,7 +959,7 @@ static int tipc_sendmsg(struct kiocb *iocb, struct socket *sock,
}
new_mtu:
- mtu = tipc_node_get_mtu(dnode, tsk->ref);
+ mtu = tipc_node_get_mtu(dnode, tsk->portid);
__skb_queue_head_init(&head);
rc = tipc_msg_build(mhdr, m, 0, dsz, mtu, &head);
if (rc < 0)
@@ -955,7 +968,7 @@ new_mtu:
do {
skb = skb_peek(&head);
TIPC_SKB_CB(skb)->wakeup_pending = tsk->link_cong;
- rc = tipc_link_xmit(&head, dnode, tsk->ref);
+ rc = tipc_link_xmit(&head, dnode, tsk->portid);
if (likely(rc >= 0)) {
if (sock->state != SS_READY)
sock->state = SS_CONNECTING;
@@ -1028,7 +1041,7 @@ static int tipc_send_stream(struct kiocb *iocb, struct socket *sock,
struct tipc_msg *mhdr = &tsk->phdr;
struct sk_buff_head head;
DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
- u32 ref = tsk->ref;
+ u32 portid = tsk->portid;
int rc = -EINVAL;
long timeo;
u32 dnode;
@@ -1067,7 +1080,7 @@ next:
goto exit;
do {
if (likely(!tsk_conn_cong(tsk))) {
- rc = tipc_link_xmit(&head, dnode, ref);
+ rc = tipc_link_xmit(&head, dnode, portid);
if (likely(!rc)) {
tsk->sent_unacked++;
sent += send;
@@ -1076,7 +1089,7 @@ next:
goto next;
}
if (rc == -EMSGSIZE) {
- tsk->max_pkt = tipc_node_get_mtu(dnode, ref);
+ tsk->max_pkt = tipc_node_get_mtu(dnode, portid);
goto next;
}
if (rc != -ELINKCONG)
@@ -1130,8 +1143,8 @@ static void tipc_sk_finish_conn(struct tipc_sock *tsk, u32 peer_port,
tsk->probing_state = TIPC_CONN_OK;
tsk->connected = 1;
k_start_timer(&tsk->timer, tsk->probing_interval);
- tipc_node_add_conn(peer_node, tsk->ref, peer_port);
- tsk->max_pkt = tipc_node_get_mtu(peer_node, tsk->ref);
+ tipc_node_add_conn(peer_node, tsk->portid, peer_port);
+ tsk->max_pkt = tipc_node_get_mtu(peer_node, tsk->portid);
}
/**
@@ -1238,7 +1251,7 @@ static void tipc_sk_send_ack(struct tipc_sock *tsk, uint ack)
if (!tsk->connected)
return;
skb = tipc_msg_create(CONN_MANAGER, CONN_ACK, INT_H_SIZE, 0, dnode,
- tipc_own_addr, peer_port, tsk->ref, TIPC_OK);
+ tipc_own_addr, peer_port, tsk->portid, TIPC_OK);
if (!skb)
return;
msg = buf_msg(skb);
@@ -1552,7 +1565,7 @@ static int filter_connect(struct tipc_sock *tsk, struct sk_buff **buf)
tsk->connected = 0;
/* let timer expire on it's own */
tipc_node_remove_conn(tsk_peer_node(tsk),
- tsk->ref);
+ tsk->portid);
}
retval = TIPC_OK;
}
@@ -1743,7 +1756,7 @@ int tipc_sk_rcv(struct sk_buff *skb)
u32 dnode;
/* Validate destination and message */
- tsk = tipc_sk_get(dport);
+ tsk = tipc_sk_lookup(dport);
if (unlikely(!tsk)) {
rc = tipc_msg_eval(skb, &dnode);
goto exit;
@@ -1763,7 +1776,7 @@ int tipc_sk_rcv(struct sk_buff *skb)
rc = -TIPC_ERR_OVERLOAD;
}
spin_unlock_bh(&sk->sk_lock.slock);
- tipc_sk_put(tsk);
+ sock_put(sk);
if (likely(!rc))
return 0;
exit:
@@ -2050,20 +2063,20 @@ restart:
goto restart;
}
if (tipc_msg_reverse(skb, &dnode, TIPC_CONN_SHUTDOWN))
- tipc_link_xmit_skb(skb, dnode, tsk->ref);
- tipc_node_remove_conn(dnode, tsk->ref);
+ tipc_link_xmit_skb(skb, dnode, tsk->portid);
+ tipc_node_remove_conn(dnode, tsk->portid);
} else {
dnode = tsk_peer_node(tsk);
skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE,
TIPC_CONN_MSG, SHORT_H_SIZE,
0, dnode, tipc_own_addr,
tsk_peer_port(tsk),
- tsk->ref, TIPC_CONN_SHUTDOWN);
- tipc_link_xmit_skb(skb, dnode, tsk->ref);
+ tsk->portid, TIPC_CONN_SHUTDOWN);
+ tipc_link_xmit_skb(skb, dnode, tsk->portid);
}
tsk->connected = 0;
sock->state = SS_DISCONNECTING;
- tipc_node_remove_conn(dnode, tsk->ref);
+ tipc_node_remove_conn(dnode, tsk->portid);
/* fall through */
case SS_DISCONNECTING:
@@ -2084,14 +2097,14 @@ restart:
return res;
}
-static void tipc_sk_timeout(unsigned long ref)
+static void tipc_sk_timeout(unsigned long portid)
{
struct tipc_sock *tsk;
struct sock *sk;
struct sk_buff *skb = NULL;
u32 peer_port, peer_node;
- tsk = tipc_sk_get(ref);
+ tsk = tipc_sk_lookup(portid);
if (!tsk)
return;
@@ -2108,20 +2121,20 @@ static void tipc_sk_timeout(unsigned long ref)
/* Previous probe not answered -> self abort */
skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE, TIPC_CONN_MSG,
SHORT_H_SIZE, 0, tipc_own_addr,
- peer_node, ref, peer_port,
+ peer_node, portid, peer_port,
TIPC_ERR_NO_PORT);
} else {
skb = tipc_msg_create(CONN_MANAGER, CONN_PROBE, INT_H_SIZE,
0, peer_node, tipc_own_addr,
- peer_port, ref, TIPC_OK);
+ peer_port, portid, TIPC_OK);
tsk->probing_state = TIPC_CONN_PROBING;
k_start_timer(&tsk->timer, tsk->probing_interval);
}
bh_unlock_sock(sk);
if (skb)
- tipc_link_xmit_skb(skb, peer_node, ref);
+ tipc_link_xmit_skb(skb, peer_node, portid);
exit:
- tipc_sk_put(tsk);
+ sock_put(sk);
}
static int tipc_sk_publish(struct tipc_sock *tsk, uint scope,
@@ -2132,12 +2145,12 @@ static int tipc_sk_publish(struct tipc_sock *tsk, uint scope,
if (tsk->connected)
return -EINVAL;
- key = tsk->ref + tsk->pub_count + 1;
- if (key == tsk->ref)
+ key = tsk->portid + tsk->pub_count + 1;
+ if (key == tsk->portid)
return -EADDRINUSE;
publ = tipc_nametbl_publish(seq->type, seq->lower, seq->upper,
- scope, tsk->ref, key);
+ scope, tsk->portid, key);
if (unlikely(!publ))
return -EINVAL;
@@ -2188,9 +2201,9 @@ static int tipc_sk_show(struct tipc_sock *tsk, char *buf,
ret = tipc_snprintf(buf, len, "<%u.%u.%u:%u>:",
tipc_zone(tipc_own_addr),
tipc_cluster(tipc_own_addr),
- tipc_node(tipc_own_addr), tsk->ref);
+ tipc_node(tipc_own_addr), tsk->portid);
else
- ret = tipc_snprintf(buf, len, "%-10u:", tsk->ref);
+ ret = tipc_snprintf(buf, len, "%-10u:", tsk->portid);
if (tsk->connected) {
u32 dport = tsk_peer_port(tsk);
@@ -2224,13 +2237,15 @@ static int tipc_sk_show(struct tipc_sock *tsk, char *buf,
struct sk_buff *tipc_sk_socks_show(void)
{
+ const struct bucket_table *tbl;
+ struct rhash_head *pos;
struct sk_buff *buf;
struct tlv_desc *rep_tlv;
char *pb;
int pb_len;
struct tipc_sock *tsk;
int str_len = 0;
- u32 ref = 0;
+ int i;
buf = tipc_cfg_reply_alloc(TLV_SPACE(ULTRA_STRING_MAX_LEN));
if (!buf)
@@ -2239,14 +2254,18 @@ struct sk_buff *tipc_sk_socks_show(void)
pb = TLV_DATA(rep_tlv);
pb_len = ULTRA_STRING_MAX_LEN;
- tsk = tipc_sk_get_next(&ref);
- for (; tsk; tsk = tipc_sk_get_next(&ref)) {
- lock_sock(&tsk->sk);
- str_len += tipc_sk_show(tsk, pb + str_len,
- pb_len - str_len, 0);
- release_sock(&tsk->sk);
- tipc_sk_put(tsk);
+ rcu_read_lock();
+ tbl = rht_dereference_rcu((&tipc_sk_rht)->tbl, &tipc_sk_rht);
+ for (i = 0; i < tbl->size; i++) {
+ rht_for_each_entry_rcu(tsk, pos, tbl, i, node) {
+ spin_lock_bh(&tsk->sk.sk_lock.slock);
+ str_len += tipc_sk_show(tsk, pb + str_len,
+ pb_len - str_len, 0);
+ spin_unlock_bh(&tsk->sk.sk_lock.slock);
+ }
}
+ rcu_read_unlock();
+
str_len += 1; /* for "\0" */
skb_put(buf, TLV_SPACE(str_len));
TLV_SET(rep_tlv, TIPC_TLV_ULTRA_STRING, NULL, str_len);
@@ -2259,255 +2278,91 @@ struct sk_buff *tipc_sk_socks_show(void)
*/
void tipc_sk_reinit(void)
{
+ const struct bucket_table *tbl;
+ struct rhash_head *pos;
+ struct tipc_sock *tsk;
struct tipc_msg *msg;
- u32 ref = 0;
- struct tipc_sock *tsk = tipc_sk_get_next(&ref);
+ int i;
- for (; tsk; tsk = tipc_sk_get_next(&ref)) {
- lock_sock(&tsk->sk);
- msg = &tsk->phdr;
- msg_set_prevnode(msg, tipc_own_addr);
- msg_set_orignode(msg, tipc_own_addr);
- release_sock(&tsk->sk);
- tipc_sk_put(tsk);
+ rcu_read_lock();
+ tbl = rht_dereference_rcu((&tipc_sk_rht)->tbl, &tipc_sk_rht);
+ for (i = 0; i < tbl->size; i++) {
+ rht_for_each_entry_rcu(tsk, pos, tbl, i, node) {
+ spin_lock_bh(&tsk->sk.sk_lock.slock);
+ msg = &tsk->phdr;
+ msg_set_prevnode(msg, tipc_own_addr);
+ msg_set_orignode(msg, tipc_own_addr);
+ spin_unlock_bh(&tsk->sk.sk_lock.slock);
+ }
}
+ rcu_read_unlock();
}
-/**
- * struct reference - TIPC socket reference entry
- * @tsk: pointer to socket associated with reference entry
- * @ref: reference value for socket (combines instance & array index info)
- */
-struct reference {
- struct tipc_sock *tsk;
- u32 ref;
-};
-
-/**
- * struct tipc_ref_table - table of TIPC socket reference entries
- * @entries: pointer to array of reference entries
- * @capacity: array index of first unusable entry
- * @init_point: array index of first uninitialized entry
- * @first_free: array index of first unused socket reference entry
- * @last_free: array index of last unused socket reference entry
- * @index_mask: bitmask for array index portion of reference values
- * @start_mask: initial value for instance value portion of reference values
- */
-struct ref_table {
- struct reference *entries;
- u32 capacity;
- u32 init_point;
- u32 first_free;
- u32 last_free;
- u32 index_mask;
- u32 start_mask;
-};
-
-/* Socket reference table consists of 2**N entries.
- *
- * State Socket ptr Reference
- * ----- ---------- ---------
- * In use non-NULL XXXX|own index
- * (XXXX changes each time entry is acquired)
- * Free NULL YYYY|next free index
- * (YYYY is one more than last used XXXX)
- * Uninitialized NULL 0
- *
- * Entry 0 is not used; this allows index 0 to denote the end of the free list.
- *
- * Note that a reference value of 0 does not necessarily indicate that an
- * entry is uninitialized, since the last entry in the free list could also
- * have a reference value of 0 (although this is unlikely).
- */
-
-static struct ref_table tipc_ref_table;
-
-static DEFINE_RWLOCK(ref_table_lock);
-
-/**
- * tipc_ref_table_init - create reference table for sockets
- */
-int tipc_sk_ref_table_init(u32 req_sz, u32 start)
+static struct tipc_sock *tipc_sk_lookup(u32 portid)
{
- struct reference *table;
- u32 actual_sz;
-
- /* account for unused entry, then round up size to a power of 2 */
-
- req_sz++;
- for (actual_sz = 16; actual_sz < req_sz; actual_sz <<= 1) {
- /* do nothing */
- };
-
- /* allocate table & mark all entries as uninitialized */
- table = vzalloc(actual_sz * sizeof(struct reference));
- if (table == NULL)
- return -ENOMEM;
-
- tipc_ref_table.entries = table;
- tipc_ref_table.capacity = req_sz;
- tipc_ref_table.init_point = 1;
- tipc_ref_table.first_free = 0;
- tipc_ref_table.last_free = 0;
- tipc_ref_table.index_mask = actual_sz - 1;
- tipc_ref_table.start_mask = start & ~tipc_ref_table.index_mask;
+ struct tipc_sock *tsk;
- return 0;
-}
+ rcu_read_lock();
+ tsk = rhashtable_lookup(&tipc_sk_rht, &portid);
+ if (tsk)
+ sock_hold(&tsk->sk);
+ rcu_read_unlock();
-/**
- * tipc_ref_table_stop - destroy reference table for sockets
- */
-void tipc_sk_ref_table_stop(void)
-{
- if (!tipc_ref_table.entries)
- return;
- vfree(tipc_ref_table.entries);
- tipc_ref_table.entries = NULL;
+ return tsk;
}
-/* tipc_ref_acquire - create reference to a socket
- *
- * Register an socket pointer in the reference table.
- * Returns a unique reference value that is used from then on to retrieve the
- * socket pointer, or to determine if the socket has been deregistered.
- */
-u32 tipc_sk_ref_acquire(struct tipc_sock *tsk)
+static int tipc_sk_insert(struct tipc_sock *tsk)
{
- u32 index;
- u32 index_mask;
- u32 next_plus_upper;
- u32 ref = 0;
- struct reference *entry;
-
- if (unlikely(!tsk)) {
- pr_err("Attempt to acquire ref. to non-existent obj\n");
- return 0;
- }
- if (unlikely(!tipc_ref_table.entries)) {
- pr_err("Ref. table not found in acquisition attempt\n");
- return 0;
- }
-
- /* Take a free entry, if available; otherwise initialize a new one */
- write_lock_bh(&ref_table_lock);
- index = tipc_ref_table.first_free;
- entry = &tipc_ref_table.entries[index];
+ u32 remaining = (TIPC_MAX_PORT - TIPC_MIN_PORT) + 1;
+ u32 portid = prandom_u32() % remaining + TIPC_MIN_PORT;
- if (likely(index)) {
- index = tipc_ref_table.first_free;
- entry = &tipc_ref_table.entries[index];
- index_mask = tipc_ref_table.index_mask;
- next_plus_upper = entry->ref;
- tipc_ref_table.first_free = next_plus_upper & index_mask;
- ref = (next_plus_upper & ~index_mask) + index;
- entry->tsk = tsk;
- } else if (tipc_ref_table.init_point < tipc_ref_table.capacity) {
- index = tipc_ref_table.init_point++;
- entry = &tipc_ref_table.entries[index];
- ref = tipc_ref_table.start_mask + index;
+ while (remaining--) {
+ portid++;
+ if ((portid < TIPC_MIN_PORT) || (portid > TIPC_MAX_PORT))
+ portid = TIPC_MIN_PORT;
+ tsk->portid = portid;
+ sock_hold(&tsk->sk);
+ if (rhashtable_lookup_insert(&tipc_sk_rht, &tsk->node))
+ return 0;
+ sock_put(&tsk->sk);
}
- if (ref) {
- entry->ref = ref;
- entry->tsk = tsk;
- }
- write_unlock_bh(&ref_table_lock);
- return ref;
+ return -1;
}
-/* tipc_sk_ref_discard - invalidate reference to an socket
- *
- * Disallow future references to an socket and free up the entry for re-use.
- */
-void tipc_sk_ref_discard(u32 ref)
+static void tipc_sk_remove(struct tipc_sock *tsk)
{
- struct reference *entry;
- u32 index;
- u32 index_mask;
-
- if (unlikely(!tipc_ref_table.entries)) {
- pr_err("Ref. table not found during discard attempt\n");
- return;
- }
-
- index_mask = tipc_ref_table.index_mask;
- index = ref & index_mask;
- entry = &tipc_ref_table.entries[index];
-
- write_lock_bh(&ref_table_lock);
+ struct sock *sk = &tsk->sk;
- if (unlikely(!entry->tsk)) {
- pr_err("Attempt to discard ref. to non-existent socket\n");
- goto exit;
- }
- if (unlikely(entry->ref != ref)) {
- pr_err("Attempt to discard non-existent reference\n");
- goto exit;
+ if (rhashtable_remove(&tipc_sk_rht, &tsk->node)) {
+ WARN_ON(atomic_read(&sk->sk_refcnt) == 1);
+ __sock_put(sk);
}
-
- /* Mark entry as unused; increment instance part of entry's
- * reference to invalidate any subsequent references
- */
-
- entry->tsk = NULL;
- entry->ref = (ref & ~index_mask) + (index_mask + 1);
-
- /* Append entry to free entry list */
- if (unlikely(tipc_ref_table.first_free == 0))
- tipc_ref_table.first_free = index;
- else
- tipc_ref_table.entries[tipc_ref_table.last_free].ref |= index;
- tipc_ref_table.last_free = index;
-exit:
- write_unlock_bh(&ref_table_lock);
}
-/* tipc_sk_get - find referenced socket and return pointer to it
- */
-struct tipc_sock *tipc_sk_get(u32 ref)
+int tipc_sk_rht_init(void)
{
- struct reference *entry;
- struct tipc_sock *tsk;
+ struct rhashtable_params rht_params = {
+ .nelem_hint = 256,
+ .head_offset = offsetof(struct tipc_sock, node),
+ .key_offset = offsetof(struct tipc_sock, portid),
+ .key_len = sizeof(u32), /* portid */
+ .hashfn = jhash,
+ .max_shift = 20, /* 1M */
+ .min_shift = 8, /* 256 */
+ .grow_decision = rht_grow_above_75,
+ .shrink_decision = rht_shrink_below_30,
+ };
- if (unlikely(!tipc_ref_table.entries))
- return NULL;
- read_lock_bh(&ref_table_lock);
- entry = &tipc_ref_table.entries[ref & tipc_ref_table.index_mask];
- tsk = entry->tsk;
- if (likely(tsk && (entry->ref == ref)))
- sock_hold(&tsk->sk);
- else
- tsk = NULL;
- read_unlock_bh(&ref_table_lock);
- return tsk;
+ return rhashtable_init(&tipc_sk_rht, &rht_params);
}
-/* tipc_sk_get_next - lock & return next socket after referenced one
-*/
-struct tipc_sock *tipc_sk_get_next(u32 *ref)
+void tipc_sk_rht_destroy(void)
{
- struct reference *entry;
- struct tipc_sock *tsk = NULL;
- uint index = *ref & tipc_ref_table.index_mask;
+ /* Wait for socket readers to complete */
+ synchronize_net();
- read_lock_bh(&ref_table_lock);
- while (++index < tipc_ref_table.capacity) {
- entry = &tipc_ref_table.entries[index];
- if (!entry->tsk)
- continue;
- tsk = entry->tsk;
- sock_hold(&tsk->sk);
- *ref = entry->ref;
- break;
- }
- read_unlock_bh(&ref_table_lock);
- return tsk;
-}
-
-static void tipc_sk_put(struct tipc_sock *tsk)
-{
- sock_put(&tsk->sk);
+ rhashtable_destroy(&tipc_sk_rht);
}
/**
@@ -2829,7 +2684,7 @@ static int __tipc_nl_add_sk(struct sk_buff *skb, struct netlink_callback *cb,
attrs = nla_nest_start(skb, TIPC_NLA_SOCK);
if (!attrs)
goto genlmsg_cancel;
- if (nla_put_u32(skb, TIPC_NLA_SOCK_REF, tsk->ref))
+ if (nla_put_u32(skb, TIPC_NLA_SOCK_REF, tsk->portid))
goto attr_msg_cancel;
if (nla_put_u32(skb, TIPC_NLA_SOCK_ADDR, tipc_own_addr))
goto attr_msg_cancel;
@@ -2859,22 +2714,29 @@ int tipc_nl_sk_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
int err;
struct tipc_sock *tsk;
- u32 prev_ref = cb->args[0];
- u32 ref = prev_ref;
-
- tsk = tipc_sk_get_next(&ref);
- for (; tsk; tsk = tipc_sk_get_next(&ref)) {
- lock_sock(&tsk->sk);
- err = __tipc_nl_add_sk(skb, cb, tsk);
- release_sock(&tsk->sk);
- tipc_sk_put(tsk);
- if (err)
- break;
+ const struct bucket_table *tbl;
+ struct rhash_head *pos;
+ u32 prev_portid = cb->args[0];
+ u32 portid = prev_portid;
+ int i;
- prev_ref = ref;
+ rcu_read_lock();
+ tbl = rht_dereference_rcu((&tipc_sk_rht)->tbl, &tipc_sk_rht);
+ for (i = 0; i < tbl->size; i++) {
+ rht_for_each_entry_rcu(tsk, pos, tbl, i, node) {
+ spin_lock_bh(&tsk->sk.sk_lock.slock);
+ portid = tsk->portid;
+ err = __tipc_nl_add_sk(skb, cb, tsk);
+ spin_unlock_bh(&tsk->sk.sk_lock.slock);
+ if (err)
+ break;
+
+ prev_portid = portid;
+ }
}
+ rcu_read_unlock();
- cb->args[0] = prev_ref;
+ cb->args[0] = prev_portid;
return skb->len;
}
@@ -2962,12 +2824,12 @@ static int __tipc_nl_list_sk_publ(struct sk_buff *skb,
int tipc_nl_publ_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
int err;
- u32 tsk_ref = cb->args[0];
+ u32 tsk_portid = cb->args[0];
u32 last_publ = cb->args[1];
u32 done = cb->args[2];
struct tipc_sock *tsk;
- if (!tsk_ref) {
+ if (!tsk_portid) {
struct nlattr **attrs;
struct nlattr *sock[TIPC_NLA_SOCK_MAX + 1];
@@ -2984,13 +2846,13 @@ int tipc_nl_publ_dump(struct sk_buff *skb, struct netlink_callback *cb)
if (!sock[TIPC_NLA_SOCK_REF])
return -EINVAL;
- tsk_ref = nla_get_u32(sock[TIPC_NLA_SOCK_REF]);
+ tsk_portid = nla_get_u32(sock[TIPC_NLA_SOCK_REF]);
}
if (done)
return 0;
- tsk = tipc_sk_get(tsk_ref);
+ tsk = tipc_sk_lookup(tsk_portid);
if (!tsk)
return -EINVAL;
@@ -2999,9 +2861,9 @@ int tipc_nl_publ_dump(struct sk_buff *skb, struct netlink_callback *cb)
if (!err)
done = 1;
release_sock(&tsk->sk);
- tipc_sk_put(tsk);
+ sock_put(&tsk->sk);
- cb->args[0] = tsk_ref;
+ cb->args[0] = tsk_portid;
cb->args[1] = last_publ;
cb->args[2] = done;
diff --git a/net/tipc/socket.h b/net/tipc/socket.h
index d340893..c7d46d06 100644
--- a/net/tipc/socket.h
+++ b/net/tipc/socket.h
@@ -46,8 +46,8 @@ int tipc_sk_rcv(struct sk_buff *buf);
struct sk_buff *tipc_sk_socks_show(void);
void tipc_sk_mcast_rcv(struct sk_buff *buf);
void tipc_sk_reinit(void);
-int tipc_sk_ref_table_init(u32 requested_size, u32 start);
-void tipc_sk_ref_table_stop(void);
+int tipc_sk_rht_init(void);
+void tipc_sk_rht_destroy(void);
int tipc_nl_sk_dump(struct sk_buff *skb, struct netlink_callback *cb);
int tipc_nl_publ_dump(struct sk_buff *skb, struct netlink_callback *cb);
--
1.7.9.5
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
^ permalink raw reply related
* [PATCH net-next 4/5] rhashtable: involve rhashtable_lookup_insert routine
From: Ying Xue @ 2015-01-06 7:23 UTC (permalink / raw)
To: tgraf; +Cc: jon.maloy, netdev, Paul.Gortmaker, tipc-discussion, davem
In-Reply-To: <1420529003-22244-1-git-send-email-ying.xue@windriver.com>
Involve a new function called rhashtable_lookup_insert() which makes
lookup and insertion atomic under bucket lock protection, helping us
avoid to introduce an extra lock when we search and insert an object
into hash table.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
---
include/linux/rhashtable.h | 1 +
lib/rhashtable.c | 98 ++++++++++++++++++++++++++++++++++++--------
2 files changed, 83 insertions(+), 16 deletions(-)
diff --git a/include/linux/rhashtable.h b/include/linux/rhashtable.h
index de1459c7..73c913f 100644
--- a/include/linux/rhashtable.h
+++ b/include/linux/rhashtable.h
@@ -168,6 +168,7 @@ int rhashtable_shrink(struct rhashtable *ht);
void *rhashtable_lookup(struct rhashtable *ht, const void *key);
void *rhashtable_lookup_compare(struct rhashtable *ht, const void *key,
bool (*compare)(void *, void *), void *arg);
+bool rhashtable_lookup_insert(struct rhashtable *ht, struct rhash_head *obj);
void rhashtable_destroy(struct rhashtable *ht);
diff --git a/lib/rhashtable.c b/lib/rhashtable.c
index f5288b1..c77b472 100644
--- a/lib/rhashtable.c
+++ b/lib/rhashtable.c
@@ -504,8 +504,27 @@ static void rhashtable_wakeup_worker(struct rhashtable *ht)
schedule_delayed_work(&ht->run_work, 0);
}
+static void __rhashtable_insert(struct rhashtable *ht, struct rhash_head *obj,
+ struct bucket_table *tbl, u32 hash)
+{
+ struct rhash_head *head = rht_dereference_bucket(tbl->buckets[hash],
+ tbl, hash);
+
+ if (rht_is_a_nulls(head))
+ INIT_RHT_NULLS_HEAD(obj->next, ht, hash);
+ else
+ RCU_INIT_POINTER(obj->next, head);
+
+ rcu_assign_pointer(tbl->buckets[hash], obj);
+
+ atomic_inc(&ht->nelems);
+
+ /* Only grow the table if no resizing is currently in progress. */
+ rhashtable_wakeup_worker(ht);
+}
+
/**
- * rhashtable_insert - insert object into hash hash table
+ * rhashtable_insert - insert object into hash table
* @ht: hash table
* @obj: pointer to hash head inside object
*
@@ -522,7 +541,6 @@ static void rhashtable_wakeup_worker(struct rhashtable *ht)
void rhashtable_insert(struct rhashtable *ht, struct rhash_head *obj)
{
struct bucket_table *tbl;
- struct rhash_head *head;
spinlock_t *lock;
unsigned hash;
@@ -533,20 +551,9 @@ void rhashtable_insert(struct rhashtable *ht, struct rhash_head *obj)
lock = bucket_lock(tbl, hash);
spin_lock_bh(lock);
- head = rht_dereference_bucket(tbl->buckets[hash], tbl, hash);
- if (rht_is_a_nulls(head))
- INIT_RHT_NULLS_HEAD(obj->next, ht, hash);
- else
- RCU_INIT_POINTER(obj->next, head);
-
- rcu_assign_pointer(tbl->buckets[hash], obj);
+ __rhashtable_insert(ht, obj, tbl, hash);
spin_unlock_bh(lock);
- atomic_inc(&ht->nelems);
-
- /* Only grow the table if no resizing is currently in progress. */
- rhashtable_wakeup_worker(ht);
-
rcu_read_unlock();
}
EXPORT_SYMBOL_GPL(rhashtable_insert);
@@ -560,7 +567,7 @@ EXPORT_SYMBOL_GPL(rhashtable_insert);
* walk the bucket chain upon removal. The removal operation is thus
* considerable slow if the hash table is not correctly sized.
*
- * Will automatically shrink the table via rhashtable_expand() if the the
+ * Will automatically shrink the table via rhashtable_expand() if the
* shrink_decision function specified at rhashtable_init() returns true.
*
* The caller must ensure that no concurrent table mutations occur. It is
@@ -641,7 +648,7 @@ static bool rhashtable_compare(void *ptr, void *arg)
* for a entry with an identical key. The first matching entry is returned.
*
* This lookup function may only be used for fixed key hash table (key_len
- * paramter set). It will BUG() if used inappropriately.
+ * parameter set). It will BUG() if used inappropriately.
*
* Lookups may occur in parallel with hashtable mutations and resizing.
*/
@@ -702,6 +709,65 @@ restart:
}
EXPORT_SYMBOL_GPL(rhashtable_lookup_compare);
+/**
+ * rhashtable_lookup_insert - lookup and insert object into hash table
+ * @ht: hash table
+ * @obj: pointer to hash head inside object
+ *
+ * Locks down the bucket chain in both the old and new table if a resize
+ * is in progress to ensure that writers can't remove from the old table
+ * and can't insert to the new table during the atomic operation of search
+ * and insertion. Searches for duplicates in both the old and new table if
+ * a resize is in progress.
+ *
+ * This lookup function may only be used for fixed key hash table (key_len
+ * parameter set). It will BUG() if used inappropriately.
+ *
+ * It is safe to call this function from atomic context.
+ *
+ * Will trigger an automatic deferred table resizing if the size grows
+ * beyond the watermark indicated by grow_decision() which can be passed
+ * to rhashtable_init().
+ */
+bool rhashtable_lookup_insert(struct rhashtable *ht, struct rhash_head *obj)
+{
+ struct bucket_table *new_tbl, *old_tbl;
+ spinlock_t *new_bucket_lock, *old_bucket_lock;
+ u32 new_hash, old_hash;
+ bool success = true;
+
+ BUG_ON(!ht->p.key_len);
+
+ rcu_read_lock();
+
+ old_tbl = rht_dereference_rcu(ht->tbl, ht);
+ old_hash = head_hashfn(ht, old_tbl, obj);
+ old_bucket_lock = bucket_lock(old_tbl, old_hash);
+ spin_lock_bh(old_bucket_lock);
+
+ new_tbl = rht_dereference_rcu(ht->future_tbl, ht);
+ new_hash = head_hashfn(ht, new_tbl, obj);
+ new_bucket_lock = bucket_lock(new_tbl, new_hash);
+ if (unlikely(old_tbl != new_tbl))
+ spin_lock_bh_nested(new_bucket_lock, RHT_LOCK_NESTED);
+
+ if (rhashtable_lookup(ht, rht_obj(ht, obj) + ht->p.key_offset)) {
+ success = false;
+ goto exit;
+ }
+
+ __rhashtable_insert(ht, obj, new_tbl, new_hash);
+
+exit:
+ if (unlikely(old_tbl != new_tbl))
+ spin_unlock_bh(new_bucket_lock);
+ spin_unlock_bh(old_bucket_lock);
+ rcu_read_unlock();
+
+ return success;
+}
+EXPORT_SYMBOL_GPL(rhashtable_lookup_insert);
+
static size_t rounded_hashtable_size(struct rhashtable_params *params)
{
return max(roundup_pow_of_two(params->nelem_hint * 4 / 3),
--
1.7.9.5
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
^ permalink raw reply related
* [PATCH net-next 3/5] rhashtable: use future table size to make expansion decision
From: Ying Xue @ 2015-01-06 7:23 UTC (permalink / raw)
To: tgraf; +Cc: jon.maloy, netdev, Paul.Gortmaker, tipc-discussion, davem
In-Reply-To: <1420529003-22244-1-git-send-email-ying.xue@windriver.com>
Should use future table size instead of old table size to decide
whether hash table is worth being expanded.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Thomas Graf <tgraf@suug.ch>
---
lib/rhashtable.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/lib/rhashtable.c b/lib/rhashtable.c
index 6eda22f..f5288b1 100644
--- a/lib/rhashtable.c
+++ b/lib/rhashtable.c
@@ -478,13 +478,14 @@ EXPORT_SYMBOL_GPL(rhashtable_shrink);
static void rht_deferred_worker(struct work_struct *work)
{
struct rhashtable *ht;
- struct bucket_table *tbl;
+ struct bucket_table *tbl, *new_tbl;
ht = container_of(work, struct rhashtable, run_work.work);
mutex_lock(&ht->mutex);
tbl = rht_dereference(ht->tbl, ht);
+ new_tbl = rht_dereference(ht->future_tbl, ht);
- if (ht->p.grow_decision && ht->p.grow_decision(ht, tbl->size))
+ if (ht->p.grow_decision && ht->p.grow_decision(ht, new_tbl->size))
rhashtable_expand(ht);
else if (ht->p.shrink_decision && ht->p.shrink_decision(ht, tbl->size))
rhashtable_shrink(ht);
--
1.7.9.5
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
^ permalink raw reply related
* [PATCH net-next 2/5] rhashtable: introduce rhashtable_wakeup_worker helper function
From: Ying Xue @ 2015-01-06 7:23 UTC (permalink / raw)
To: tgraf; +Cc: jon.maloy, netdev, Paul.Gortmaker, tipc-discussion, davem
In-Reply-To: <1420529003-22244-1-git-send-email-ying.xue@windriver.com>
Introduce rhashtable_wakeup_worker() helper function to reduce
duplicated code where to wake up worker.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Thomas Graf <tgraf@suug.ch>
---
lib/rhashtable.c | 20 +++++++++++++-------
1 file changed, 13 insertions(+), 7 deletions(-)
diff --git a/lib/rhashtable.c b/lib/rhashtable.c
index f2fdd7a..6eda22f 100644
--- a/lib/rhashtable.c
+++ b/lib/rhashtable.c
@@ -492,6 +492,17 @@ static void rht_deferred_worker(struct work_struct *work)
mutex_unlock(&ht->mutex);
}
+static void rhashtable_wakeup_worker(struct rhashtable *ht)
+{
+ struct bucket_table *tbl = rht_dereference_rcu(ht->tbl, ht);
+ struct bucket_table *new_tbl = rht_dereference_rcu(ht->future_tbl, ht);
+
+ if (ht->tbl != ht->future_tbl &&
+ ((ht->p.grow_decision && ht->p.grow_decision(ht, new_tbl->size)) ||
+ (ht->p.shrink_decision && ht->p.shrink_decision(ht, tbl->size))))
+ schedule_delayed_work(&ht->run_work, 0);
+}
+
/**
* rhashtable_insert - insert object into hash hash table
* @ht: hash table
@@ -533,9 +544,7 @@ void rhashtable_insert(struct rhashtable *ht, struct rhash_head *obj)
atomic_inc(&ht->nelems);
/* Only grow the table if no resizing is currently in progress. */
- if (ht->tbl != ht->future_tbl &&
- ht->p.grow_decision && ht->p.grow_decision(ht, tbl->size))
- schedule_delayed_work(&ht->run_work, 0);
+ rhashtable_wakeup_worker(ht);
rcu_read_unlock();
}
@@ -584,10 +593,7 @@ restart:
spin_unlock_bh(lock);
- if (ht->tbl != ht->future_tbl &&
- ht->p.shrink_decision &&
- ht->p.shrink_decision(ht, tbl->size))
- schedule_delayed_work(&ht->run_work, 0);
+ rhashtable_wakeup_worker(ht);
rcu_read_unlock();
--
1.7.9.5
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox