On 6/30/18, 5:26 PM, "netdev-ow...@vger.kernel.org on behalf of Neal Cardwell"
wrote:
On Fri, Jun 29, 2018 at 9:48 PM Lawrence Brakmo wrote:
>
> When have observed high tail latencies when using DCTCP for RPCs as
> compared to using Cubic. For example, in one setup there are 2 h
On Fri, Jun 29, 2018 at 3:52 PM Ilpo Järvinen wrote:
> > +.005 < . 1:1(0) ack 2001 win 257
>
> Why did the receiver send a cumulative ACK only for 2001?
Sorry, you are right Ilpo. Upon further reflection, the packetdrill
scenario I posted is not a realistic one, and I agree we should not
worry ab
On 6/30/18, 11:23 AM, "Neal Cardwell" wrote:
On Fri, Jun 29, 2018 at 9:48 PM Lawrence Brakmo wrote:
>
> We observed high 99 and 99.9% latencies when doing RPCs with DCTCP. The
> problem is triggered when the last packet of a request arrives CE
> marked. The reply will carry t
From: Daniel Borkmann
Date: Sun, 1 Jul 2018 01:56:37 +0200
> The following pull-request contains BPF updates for your *net* tree.
>
> The main changes are:
...
> Please consider pulling these changes from:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf.git
Pulled, thanks Daniel.
On Fri, Jun 29, 2018 at 9:48 PM Lawrence Brakmo wrote:
>
> When have observed high tail latencies when using DCTCP for RPCs as
> compared to using Cubic. For example, in one setup there are 2 hosts
> sending to a 3rd one, with each sender having 3 flows (1 stream,
> 1 1MB back-to-back RPCs and 1 1
On 06/30/2018 03:17 PM, John Fastabend wrote:
> This addresses two syzbot issues that lead to identifying (by Eric and
> Wei) a class of bugs where we don't correctly check for IPv4/v6
> sockets and their associated state. The second issue was a locking
> omission in sockhash.
>
> The first patch
Hi David,
The following pull-request contains BPF updates for your *net* tree.
The main changes are:
1) A bpf_fib_lookup() helper fix to change the API before freeze to
return an encoding of the FIB lookup result and return the nexthop
device index in the params struct (instead of device i
On 06/29/2018 02:48 PM, Mauricio Vasquez B wrote:
> Decrement the number of elements in the map in case the allocation
> of a new node fails.
>
> Signed-off-by: Mauricio Vasquez B
Thanks for the fix, Mauricio!
Could you reply with a Fixes: tag in order to track the commit originally
introducing
On 06/28/2018 11:41 PM, Jakub Kicinski wrote:
> Hi!
>
> Set of random updates to bpftool and libbpf. I'm preparing for
> extending bpftool prog load, but there is a good number of
> improvements that can be made before bpf -> bpf-next merge
> helping to keep the later patch set to a manageable si
This hack (affecting the non-PCIe models only) was introduced in 2004
to deal with link negotiation failures in 1GBit mode. Based on a
comment in the r8169 vendor driver I assume the issue affects RTL8169sb
in combination with particular 1GBit switch models.
Resetting the PHY every 10s and hoping
On 6/29/18 4:10 PM, Ben Greear wrote:
> Hello,
>
> We're trying to create lots of strongswan VPN tunnels on network devices
> bound to different VRFs. We are using Fedora-24 on the client side,
> with a 4.16.15+ kernel
> and updated 'ip' package, etc.
>
> So far, no luck getting it to work.
>
>
On 06/30/18 19:06, Boris Pismenny wrote:
On 06/30/18 01:19, Willem de Bruijn wrote:
On Fri, Jun 29, 2018 at 2:24 AM Saeed Mahameed
wrote:
From: Boris Pismenny
This patch enables UDP GSO support. We enable this by using two WQEs
the first is a UDP LSO WQE for all segments with equal lengt
On Fri, Jun 29, 2018 at 9:48 PM Lawrence Brakmo wrote:
>
> We observed high 99 and 99.9% latencies when doing RPCs with DCTCP. The
> problem is triggered when the last packet of a request arrives CE
> marked. The reply will carry the ECE mark causing TCP to shrink its cwnd
> to 1 (because there ar
Instead of just specify that a PHY is connected at some
speed, also specify which one. This is helpful with several
PHYs on the system.
Signed-off-by: Linus Walleij
---
drivers/net/ethernet/cortina/gemini.c | 13 -
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/drivers
The initialization sequence for the ethernet, setting up
interrupt routing and such things, need to be done after
both the ports are clocked and reset. Before this the
config will not "take". Move the initialization to the
port probe function and keep track of init status in
the state.
Signed-off-
The code was not tested with two ports actually in use at
the same time. (I blame this on lack of actual hardware using
that feature.) Now after locating a system using both ports,
add necessary fix to make both ports come up.
Signed-off-by: Linus Walleij
---
drivers/net/ethernet/cortina/gemini.
The hardware supposedly handles frames up to 10236 bytes and
implements .ndo_change_mtu() so accept 10236 minus the ethernet
header for a VLAN tagged frame on the netdevices.
Signed-off-by: Linus Walleij
---
drivers/net/ethernet/cortina/gemini.c | 5 +
1 file changed, 5 insertions(+)
diff -
The code to calculate the hardware register enumerator
for the maximum L3 length isn't entirely simple to read.
Use the existing defines and rewrite the function into a
table look-up.
Signed-off-by: Linus Walleij
---
drivers/net/ethernet/cortina/gemini.c | 61 ---
1 file
On 06/30/18 01:19, Willem de Bruijn wrote:
On Fri, Jun 29, 2018 at 2:24 AM Saeed Mahameed wrote:
From: Boris Pismenny
This patch enables UDP GSO support. We enable this by using two WQEs
the first is a UDP LSO WQE for all segments with equal length, and the
second is for the last segment i
Since the addition of GRO for ESP, gro_receive can consume the skb and
return -EINPROGRESS. In that case, the lower layer GRO handler cannot
touch the skb anymore.
Commit 5f114163f2f5 ("net: Add a skb_gro_flush_final helper.") converted
some of the gro_receive handlers that can lead to ESP's gro_r
The current code, in the error path of sock_hash_ctx_update_elem,
checks if the sock has a psock in the user data and if so decrements
the reference count of the psock. However, if the error happens early
in the error path we may have never incremented the psock reference
count and if the psock exi
This applies on top of "BPF fixes for sockhash" I just didn't
want to confuse that series yet again by re-ordering/adding
these patches in it
I missed fixing the error path in the sockhash code to align with
supporting socks in multiple maps. Simply checking if the psock is
present does not
This removes locking from readers of RCU hash table. Its not
necessary.
Fixes: 81110384441a ("bpf: sockmap, add hash map support")
Signed-off-by: John Fastabend
---
0 files changed
diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c
index 63fb047..12ac10a 100644
--- a/kernel/bpf/sockmap.c
If a hashmap is free'd with open socks it removes the reference to
the hash entry from the psock. If that is the last reference to the
psock then it will also be free'd by the reference counting logic.
However the current logic that removes the hash reference from the
list of references is broken.
First the sk_callback_lock() was being used to protect both the
sock callback hooks and the psock->maps list. This got overly
convoluted after the addition of sockhash (in sockmap it made
some sense because masp and callbacks were tightly coupled) so
lets split out a specific lock for maps and only
Add map_release_uref pointer to hashmap ops. This was dropped when
original sockhash code was ported into bpf-next before initial
commit.
Fixes: 81110384441a ("bpf: sockmap, add hash map support")
Acked-by: Martin KaFai Lau
Signed-off-by: John Fastabend
---
kernel/bpf/sockmap.c |1 +
1 file
This addresses two syzbot issues that lead to identifying (by Eric and
Wei) a class of bugs where we don't correctly check for IPv4/v6
sockets and their associated state. The second issue was a locking
omission in sockhash.
The first patch addresses IPv6 socks and fixing an error where
sockhash wo
This fixes a crash where we assign tcp_prot to IPv6 sockets instead
of tcpv6_prot.
Previously we overwrote the sk->prot field with tcp_prot even in the
AF_INET6 case. This patch ensures the correct tcp_prot and tcpv6_prot
are used.
Tested with 'netserver -6' and 'netperf -H [IPv6]' as well as
'ne
From: Roopa Prabhu
Date: Fri, 29 Jun 2018 14:32:15 -0700
> From: Roopa Prabhu
>
> After commit f9d4b0c1e969 ("fib_rules: move common handling of newrule
> delrule msgs into fib_nl2rule"), rule_exists got replaced by rule_find
> for existing rule lookup in both the add and del paths. While this
From: Petr Machata
Date: Sat, 30 Jun 2018 02:44:23 +0200
> There are a number of tests that check features of the Linux networking
> stack. By running them on suitable interfaces, one can exercise the
> mlxsw offloading code. However none of these tests attempts to push
> mlxsw to the limits supp
Hi Jianbo,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on net-next/master]
url:
https://github.com/0day-ci/linux/commits/Jianbo-Liu/net-flow_dissector-Save-vlan-ethertype-from-headers/20180630-180158
reproduce:
# apt-get install sparse
From: Jakub Kicinski
Date: Fri, 29 Jun 2018 17:04:33 -0700
> This set contains assorted updates to driver base and flower.
> First patch is a follow up to a fix to calculating counters which
> went into net. For ethtool counters we should also make sure
> they are visible even after ring reconfi
From: Cong Wang
Date: Fri, 29 Jun 2018 13:42:48 -0700
> As noticed by Eric, we need to switch to the helper
> dev_change_tx_queue_len() for SIOCSIFTXQLEN call path too,
> otheriwse still miss dev_qdisc_change_tx_queue_len().
>
> Fixes: 6a643ddb5624 ("net: introduce helper dev_change_tx_queue_len
From: Vakul Garg
Date: Sat, 30 Jun 2018 00:45:55 +0530
> Calling skb_unclone() is expensive as it triggers a memcpy operation.
> Instead of calling skb_unclone() unconditionally, call it only when skb
> has a shared frag_list. This improves tls rx throughout significantly.
>
> Signed-off-by: Vak
From: Julian Wiedmann
Date: Fri, 29 Jun 2018 19:45:49 +0200
> please apply a few qeth fixes for -net and your 4.17 stable queue.
>
> Patches 1-3 fix several issues wrt to MAC address management that were
> introduced during the 4.17 cycle.
> Patch 4 tackles a long-standing issue with busy multi-
From: Sabrina Dubroca
Date: Fri, 29 Jun 2018 17:51:26 +0200
> The __alx_open function can be called from ndo_open, which is called
> under RTNL, or from alx_resume, which isn't. Since commit d768319cd427,
> we're calling the netif_set_real_num_{tx,rx}_queues functions, which
> need to be called u
From: Bert Kenward
Date: Fri, 29 Jun 2018 16:29:28 +0100
> Fixes: fc7a6c287ff3 ("sfc: use a semaphore to lock farch filters too")
> Suggested-by: Joseph Korty
> Signed-off-by: Bert Kenward
Applied and queued up for -stable.
From: Keara Leibovitz
Date: Fri, 29 Jun 2018 10:47:31 -0400
> Create unittests for the tc tunnel_key action.
>
> v2:
> For the tests expecting failures, added non-zero exit codes in the
> teardowns. This prevents those tests from failing if the act_tunnel_key
> module is unloaded.
>
> Signed-of
From: Heiner Kallweit
Date: Fri, 29 Jun 2018 08:07:04 +0200
> The very first version of RTL8169 from 2002 (and only this one) has
> support for a TBI 1000BaseX fiber interface. The TBI support in the
> driver makes switching to phylib tricky, so best would be to get
> rid of it. I found no report
From: dsah...@kernel.org
Date: Thu, 28 Jun 2018 13:36:55 -0700
> From: David Ahern
>
> Sowmini reported that a recent commit broke prefix routes for linklocal
> addresses. The newly added modify_prefix_route is attempting to add a
> new prefix route when the ifp priority does not match the route
From: Heiner Kallweit
Date: Thu, 28 Jun 2018 20:46:45 +0200
> In preparation of adding phylib support to the r8169 driver we need
> PHY drivers for all chip-internal PHY types. Fortunately almost all
> of them are either supported by the Realtek PHY driver already or work
> with the genphy driver
From: Ursula Braun
Date: Thu, 28 Jun 2018 19:05:03 +0200
> SMC requires a configured pnet table to map Ethernet interfaces to
> RoCE adapter ports. For s390 there exists hardware support to group
> such devices. The first three patches cover the s390 pnetid support,
> enabling SMC-R usage on s390
From: Heiner Kallweit
Date: Thu, 28 Jun 2018 20:36:15 +0200
> I see no need to define a private debug output symbol, let's use the
> standard debug output functions instead. In this context also remove
> the deprecated PFX define.
>
> The one assertion is wrong IMO anyway, this code path is used
From: Petr Machata
Date: Thu, 28 Jun 2018 18:56:15 +0200
> The forwarding selftests infrastructure makes it possible to run the
> individual tests on a purely software netdevices. Names of interfaces to
> run the test with can be passed as command line arguments to a test.
> lib.sh then creates v
Sat, Jun 30, 2018 at 12:18:02AM CEST, xiyou.wangc...@gmail.com wrote:
>On Fri, Jun 29, 2018 at 10:06 AM Samudrala, Sridhar
> wrote:
>>
>> So instead of introducing 'chaintemplate' object in the kernel, can't we add
>> 'chain'
>> object in the kernel that takes the 'template' as an attribute?
>
>Th
To support matching on both outer and inner vlan headers,
we add new cvlan_id/cvlan_prio/cvlan_ethtype for inner vlan header.
Example:
# tc filter add dev eth0 protocol 802.1ad parent : \
flower vlan_id 1000 vlan_ethtype 802.1q \
cvlan_id 100 cvlan_ethtype ipv4 \
action vlan po
Currently the encapsulated ethertype is not dumped as it's the same as
TCA_FLOWER_KEY_ETH_TYPE keyvalue. But the dumping result is inconsistent
with input, we add dumping it with TCA_FLOWER_KEY_VLAN_ETH_TYPE.
Signed-off-by: Jianbo Liu
Acked-by: Jiri Pirko
---
net/sched/cls_flower.c | 4
1
As flow dissector stores vlan ethertype, tc flower now can match on that.
It is to make preparation for supporting QinQ.
Signed-off-by: Jianbo Liu
Acked-by: Jiri Pirko
---
net/sched/cls_flower.c | 7 +--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/net/sched/cls_flower.c b/
As support dissecting of QinQ inner and outer vlan headers, user can
add rules to match on QinQ vlan headers.
Signed-off-by: Jianbo Liu
Acked-by: Jiri Pirko
---
include/uapi/linux/pkt_cls.h | 4 +++
net/sched/cls_flower.c | 65 ++--
2 files changed
Dissect the QinQ packets to get both outer and inner vlan information,
then store to the extended flow keys.
Signed-off-by: Jianbo Liu
Acked-by: Jiri Pirko
---
include/net/flow_dissector.h | 2 ++
net/core/flow_dissector.c| 32 +---
2 files changed, 19 insertion
Currently TC flower supports only one vlan tag, it doesn't match on both outer
and inner vlan headers for QinQ. To do this, we add support to get both outer
and inner vlan headers for flow dissector, and then TC flower do matching on
those information.
We also plan to extend TC command to support
Change vlan dissector key to save vlan tpid to support both 802.1Q
and 802.1AD ethertype.
Signed-off-by: Jianbo Liu
Acked-by: Jiri Pirko
---
include/net/flow_dissector.h | 2 +-
net/core/flow_dissector.c| 2 ++
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/include/net/flow_d
From: Sudarsana Reddy Kalluru
Date: Thu, 28 Jun 2018 04:52:15 -0700
> Driver performs the internal reload when it receives tx-timeout event from
> the OS. Internal reload might fail in some scenarios e.g., fatal HW issues.
> In such cases OS still see the link, which would result in undesirable
>
From: Dan Carpenter
Date: Thu, 28 Jun 2018 12:31:25 +0300
> Static checkers complain that id_tbl->table points to longs and 4 bytes
> is smaller than sizeof(long). But the since other side is dividing by
> 32 instead of sizeof(long), that means the current code works fine.
>
> Anyway, it's more
From: Dan Carpenter
Date: Thu, 28 Jun 2018 12:24:42 +0300
> The code assumes that there is 4 bytes in a pointer and it doesn't
> allocate enough memory.
>
> Signed-off-by: Dan Carpenter
Applied.
From: Yuchung Cheng
Date: Wed, 27 Jun 2018 16:04:48 -0700
> Fast Open key could be stored in different endian based on the CPU.
> Previously hosts in different endianness in a server farm using
> the same key config (sysctl value) would produce different cookies.
> This patch fixes it by always s
From: Jose Abreu
Date: Wed, 27 Jun 2018 13:43:20 +0100
> This adds support for CBS reconfiguration using the TC application.
>
> A new callback was added to TC ops struct and another one to DMA ops to
> reconfigure the channel mode.
>
> Tested in GMAC5.10.
>
> Signed-off-by: Jose Abreu
Appli
Signed-off-by: Amritha Nambiar
---
Documentation/ABI/testing/sysfs-class-net-queues | 11
Documentation/networking/scaling.txt | 61 ++
2 files changed, 61 insertions(+), 11 deletions(-)
diff --git a/Documentation/ABI/testing/sysfs-class-net-queues
b/Doc
This patch adds support to pick Tx queue based on the Rx queue(s) map
configuration set by the admin through the sysfs attribute
for each Tx queue. If the user configuration for receive queue(s) map
does not apply, then the Tx queue selection falls back to CPU(s) map
based selection and finally to
Extend transmit queue sysfs attribute to configure Rx queue(s) map
per Tx queue. By default no receive queues are configured for the
Tx queue.
- /sys/class/net/eth0/queues/tx-*/xps_rxqs
Signed-off-by: Amritha Nambiar
---
net/core/net-sysfs.c | 83 ++
This patch series implements support for Tx queue selection based on
Rx queue(s) map. This is done by configuring Rx queue(s) map per Tx-queue
using sysfs attribute. If the user configuration for Rx queues does
not apply, then the Tx queue selection falls back to XPS using CPUs and
finally to hashi
Refactor XPS code to support Tx queue selection based on
CPU(s) map or Rx queue(s) map.
Signed-off-by: Amritha Nambiar
---
include/linux/cpumask.h | 11 ++
include/linux/netdevice.h | 98 -
net/core/dev.c| 211 ++---
Change 'skc_tx_queue_mapping' field in sock_common structure from
'int' to 'unsigned short' type with ~0 indicating unset and
other positive queue values being set. This will accommodate adding
a new 'unsigned short' field in sock_common in the next patch for
rx_queue_mapping.
Signed-off-by: Amrit
This patch adds a new field to sock_common 'skc_rx_queue_mapping'
which holds the receive queue number for the connection. The Rx queue
is marked in tcp_finish_connect() to allow a client app to do
SO_INCOMING_NAPI_ID after a connect() call to get the right queue
association for a socket. Rx queue
Use static_key for XPS maps to reduce the cost of extra map checks,
similar to how it is used for RPS and RFS. This includes static_key
'xps_needed' for XPS and another for 'xps_rxqs_needed' for XPS using
Rx queues map.
Signed-off-by: Amritha Nambiar
---
net/core/dev.c | 31 +++
From: Ka-Cheong Poon
Date: Wed, 27 Jun 2018 03:23:27 -0700
> This patch changes the internal representation of an IP address to use
> struct in6_addr. IPv4 address is stored as an IPv4 mapped address.
> All the functions which take an IP address as argument are also
> changed to use struct in6_a
On 06/29/2018 07:01 PM, Jakub Kicinski wrote:
> On Fri, 29 Jun 2018 09:04:15 +0200, Daniel Borkmann wrote:
>> On 06/28/2018 06:54 PM, Jakub Kicinski wrote:
>>> On Thu, 28 Jun 2018 09:42:06 +0200, Jiri Benc wrote:
On Wed, 27 Jun 2018 11:49:49 +0200, Daniel Borkmann wrote:
> Looks good t
Hi,
2018-06-28 15:40 GMT+09:00 Xin Long :
> On Tue, Jun 26, 2018 at 8:02 PM, 吉藤英明
> wrote:
>> 2018-06-26 13:33 GMT+09:00 Xin Long :
>>> On Tue, Jun 26, 2018 at 12:31 AM, Marcelo Ricardo Leitner
>>> wrote:
Hi,
On Tue, Jun 26, 2018 at 01:12:00AM +0900, 吉藤英明 wrote:
> Hi,
>
>>
On 6/29/2018 6:06 AM, David Miller wrote:
> From: Amritha Nambiar
> Date: Wed, 27 Jun 2018 15:31:34 -0700
>
>> @@ -1702,6 +1709,13 @@ static inline int sk_tx_queue_get(const struct sock
>> *sk)
>> return -1;
>> }
>>
>> +static inline void sk_rx_queue_set(struct sock *sk, const struct sk_
On 6/29/2018 6:05 AM, David Miller wrote:
> From: Amritha Nambiar
> Date: Wed, 27 Jun 2018 15:31:28 -0700
>
>> @@ -1681,17 +1681,25 @@ static inline int sk_receive_skb(struct sock *sk,
>> struct sk_buff *skb,
>>
>> static inline void sk_tx_queue_set(struct sock *sk, int tx_queue)
>> {
>> +
On 6/29/2018 5:59 AM, David Miller wrote:
> From: Amritha Nambiar
> Date: Wed, 27 Jun 2018 15:31:18 -0700
>
>> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
>> index c6b377a..3790ac9 100644
>> --- a/include/linux/netdevice.h
>> +++ b/include/linux/netdevice.h
> ...
>> +stati
71 matches
Mail list logo
