Since we're freeing multiple skbs, we might as well use bulk free to save a
few cycles. Use the same conditions for bulk free as in napi_consume_skb.
Signed-off-by: Felix Fietkau
---
net/core/skbuff.c | 35 +++
1 file changed, 31 insertions(+), 4 deletions(-)
dif
On 03/23/2019 06:14 PM, Lawrence Brakmo wrote:
> There was no eBPF at that time. We like the flexibility we get by programing
> the algorithms in eBPF.
eBPF being there does not mean we can adopt whatever research work,
we want evaluations of the costs and benefits.
>
> These are not intrusi
On 03/23/2019 08:41 AM, Alexei Starovoitov wrote:
> On Sat, Mar 23, 2019 at 02:12:39AM -0700, Eric Dumazet wrote:
>> Please think about offloading the pacing/throttling in the NIC,
>> there is no way we will report back to tcp stack instant notificatiions.
>
> I don't think 'offload to google
On 03/23/2019 08:41 AM, Alexei Starovoitov wrote:
> On Sat, Mar 23, 2019 at 02:12:39AM -0700, Eric Dumazet wrote:
>>
>>
>> On 03/23/2019 01:05 AM, brakmo wrote:
>>> This patchset adds support for propagating congestion notifications (cn)
>>> to TCP from cgroup inet skb egress BPF programs.
>>>
>
On Sat, Mar 23, 2019 at 09:40:23PM -0400, David Miller wrote:
> From: David Ahern
> Date: Fri, 22 Mar 2019 06:06:09 -0700
>
> > From: David Ahern
> >
> > The number of stubs is growing and has nothing to do with addrconf.
> > Move the definition of the stubs to a separate header file and update
Signed-off-by: Vladimir Oltean
---
drivers/net/dsa/sja1105/Makefile | 1 +
drivers/net/dsa/sja1105/sja1105.h | 6 +
drivers/net/dsa/sja1105/sja1105_ethtool.c | 420 ++
drivers/net/dsa/sja1105/sja1105_main.c| 3 +
4 files changed, 430 insertions(+)
c
This provides an unified API for accessing register bit fields
regardless of memory layout. The basic unit of data for these API
functions is the u64. The process of transforming an u64 from native CPU
encoding into the peripheral's encoding is called 'pack', and
transforming it from peripheral to
In order to support this, we are creating a make-shift switch tag out of
a VLAN trunk configured on the CPU port. Termination on switch ports
only works when not under a vlan_filtering bridge. We are making use of
the generic CONFIG_NET_DSA_TAG_8021Q code and leveraging it from our own
CONFIG_NET_D
This allows the driver to perform some manipulations of its own during
setup, using generic code.
One current usage scenario is for the driver to request DSA to set up
802.1Q based switch tagging for its ports.
Signed-off-by: Vladimir Oltean
---
net/dsa/dsa2.c | 8
1 file changed, 4 ins
Signed-off-by: Vladimir Oltean
---
drivers/net/dsa/sja1105/sja1105_main.c | 108 ++---
1 file changed, 99 insertions(+), 9 deletions(-)
diff --git a/drivers/net/dsa/sja1105/sja1105_main.c
b/drivers/net/dsa/sja1105/sja1105_main.c
index 92dc58afd74e..448ab0e71827 100644
--- a/
This allows drivers to query the VLAN setting imposed by the bridge
driver directly from DSA, instead of keeping their own state based on
the .port_vlan_filtering callback.
Signed-off-by: Vladimir Oltean
---
include/net/dsa.h | 1 +
net/dsa/port.c| 12
2 files changed, 9 insert
This patch provides generic DSA code for using VLAN (802.1Q) tags for
the same purpose as a dedicated switch tag for injection/extraction.
It is based on the discussions and interest that has been so far
expressed in https://www.spinics.net/lists/netdev/msg556125.html.
Unlike all other DSA-support
VLAN filtering cannot be properly disabled in SJA1105. So in order to
emulate the "no VLAN awareness" behavior (not dropping traffic that is
tagged with a VID that isn't configured on the port), we need to hack
another switch feature: programmable TPID (which is 0x8100 for 802.1Q).
We are reprogram
Currently only the (more difficult) first generation E/T series is
supported. Here the TCAM is only 4-way associative, and to know where
the hardware will search for a FDB entry, we need to perform the same
hash algorithm in order to install the entry in the correct bin.
On P/Q/R/S, the TCAM shoul
Signed-off-by: Vladimir Oltean
---
.../devicetree/bindings/net/dsa/sja1105.txt | 123 ++
1 file changed, 123 insertions(+)
create mode 100644 Documentation/devicetree/bindings/net/dsa/sja1105.txt
diff --git a/Documentation/devicetree/bindings/net/dsa/sja1105.txt
b/Documentati
Signed-off-by: Vladimir Oltean
---
Documentation/networking/dsa/sja1105.txt | 83
1 file changed, 83 insertions(+)
create mode 100644 Documentation/networking/dsa/sja1105.txt
diff --git a/Documentation/networking/dsa/sja1105.txt
b/Documentation/networking/dsa/sja1105.t
This patchset adds a DSA driver for the SPI-managed NXP SJA1105 driver.
Due to the hardware's unfriendliness, most of its state needs to be
shadowed in kernel memory by the driver. To support this and keep a
decent amount of cleanliness in the code, a new generic API for
converting between CPU-acce
This refactors the two-phase transaction from dsa_slave_vlan_rx_add_vid
and also makes that code available for other functions from within DSA.
The newly exposed function either adds or deletes the specified VLAN
entry based on a boolean argument.
Signed-off-by: Vladimir Oltean
---
net/dsa/dsa_p
On Sun, 24 Mar 2019 00:21:03 +0100, Jiri Pirko wrote:
> From: Jiri Pirko
>
> In dumpit, unlike doit, the check for info_get op being defined
> is missing. Add it and avoid null pointer dereference in case driver
> does not define this op.
>
> Fixes: f9cf22882c60 ("devlink: add device information
On Sat, 23 Mar 2019 23:47:50 +0100, Jiri Pirko wrote:
> >type setting. I'd rather those calls happened on fully initialized
> >netdev (and port), could you perhaps just take this patch into the
>
> I don't get it, you are commenting cleanup part. On init, the type set
> happens on fully initial
From: Igor Russkikh
Date: Sat, 23 Mar 2019 15:23:29 +
> Here is a set of patches targeting for performance improvement
> on various platforms and protocols.
...
Series applied, thanks.
From: Saeed Mahameed
Date: Fri, 22 Mar 2019 12:12:41 -0700
> This series includes mlx5 updates.
> For more information please see tag log below.
>
> Please pull and let me know if there is any problem.
>
> v1->v2:
> Remove NFP specific vxlan/geneve port #defines based upon Jakub's
> feedba
From: Jeff Kirsher
Date: Fri, 22 Mar 2019 10:33:13 -0700
> This series contains updates to ice driver only.
Pulled, thanks Jeff.
From: Eric Dumazet
Date: Fri, 22 Mar 2019 08:56:37 -0700
> On hosts with many cpus we can observe a very serious contention
> on spinlocks used in mm slab layer.
>
> The following can happen quite often :
>
> 1) TX path
> sendmsg() allocates one (fclone) skb on CPU A, sends a clone.
> ACK i
From: Paolo Abeni
Date: Fri, 22 Mar 2019 16:01:54 +0100
> This patch series is aimed at improving xmit performances of lockless qdisc
> in the uncontended scenario.
>
> After the lockless refactor pfifo_fast can't leverage the BYPASS optimization.
> Due to retpolines the overhead for the avoidab
From: Soheil Hassas Yeganeh
Date: Fri, 22 Mar 2019 10:59:47 -0400
> From: Soheil Hassas Yeganeh
>
> Add documentation to the tcp_ca_state enum, since this enum is
> exposed in uapi.
>
> Signed-off-by: Neal Cardwell
> Signed-off-by: Yuchung Cheng
> Signed-off-by: Eric Dumazet
> Signed-off-by
From: Eric Dumazet
Date: Fri, 22 Mar 2019 06:26:29 -0700
> tcp_clock_ns() (aka ktime_get_ns()) is using monotonic clock,
> so the checks we had in tcp_mstamp_refresh() are no longer
> relevant.
>
> This patch removes cpu stall (when the cache line is not hot)
>
> Signed-off-by: Eric Dumazet
A
From: David Ahern
Date: Fri, 22 Mar 2019 06:06:09 -0700
> From: David Ahern
>
> The number of stubs is growing and has nothing to do with addrconf.
> Move the definition of the stubs to a separate header file and update
> users. In the move, drop the vxlan specific comment before ipv6_stub.
>
From: Heiner Kallweit
Date: Fri, 22 Mar 2019 07:39:35 +0100
> Bartek reported that after few cable unplug/replug cycles suddenly
> replug isn't detected any longer. His system uses a RTL8106, I wasn't
> able to reproduce the issue with RTL8168g. According to his bisect
> the referenced commit cau
On 3/23/2019 3:18 PM, Vladimir Oltean wrote:
> Previously the green and amber LEDs on this quad PHY were solid, to
> indicate an encoding of the link speed (10/100/1000).
>
> This keeps the LEDs always on just as before, but now they flash on
> Rx/Tx activity.
>
We should probably revamp the
On 3/23/2019 4:24 PM, Vladimir Oltean wrote:
> When phylink_of_phy_connect fails, dsa_slave_phy_setup tries to save the
> day by connecting to an alternative PHY, none other than a PHY on the
> switch's internal MDIO bus, at an address equal to the port's index.
>
> However this does not take i
On 3/23/19, 10:12 AM, "Eric Dumazet" wrote:
On 03/23/2019 01:05 AM, brakmo wrote:
> This patchset adds support for propagating congestion notifications (cn)
> to TCP from cgroup inet skb egress BPF programs.
>
> Current cgroup skb BPF programs cannot trigger TCP c
On 3/23/2019 6:16 AM, Heiner Kallweit wrote:
> The AQCS109 supports a proprietary 2-pair 1Gbps mode. The standard
> registers don't allow to tell between 1000BaseT and 1000BaseT2.
> Add reporting this proprietary mode based on a vendor register.
>
> Signed-off-by: Heiner Kallweit
Reviewed-by:
On 3/23/2019 6:15 AM, Heiner Kallweit wrote:
> Add reporting firmware details. These details are available only once
> the firmware has finished initializing the chip. This can take some
> time and we need to poll for init completion.
>
> Signed-off-by: Heiner Kallweit
> ---
[snip]
>
> +/*
On 3/23/2019 6:14 AM, Heiner Kallweit wrote:
> If both link partners are Aquantia PHY's then additional information is
> exchanged as part of the auto-negotiation. Report remote capabilities
> if link partner is Aquantia PHY.
>
> Signed-off-by: Heiner Kallweit
Reviewed-by: Florian Fainelli
On 3/22/2019 12:00 PM, Heiner Kallweit wrote:
> So far we effectively clear the BMCR register. Some PHY's can deal
> with this (e.g. because they reset BMCR to a default as part of a
> soft-reset) whilst on others this causes issues because e.g. the
> autoneg bit is cleared. Marvell is an exampl
From: Bart Van Assche
Date: Thu, 21 Mar 2019 15:19:39 -0700
> While inspecting the code of inet6_pton() I found an open-coded strlcpy()
> implementation in that function. The last patch in this series simplifies
> inet6_pton() and the first six patches address the compiler warnings reported
> whe
Test access to file_pos field of bpf_sysctl context, both read (incl.
narrow read) and write.
# ./test_sysctl
...
Test case: ctx:file_pos sysctl:read read ok .. [PASS]
Test case: ctx:file_pos sysctl:read read ok narrow .. [PASS]
Test case: ctx:file_pos sysctl:read write ok .. [PASS]
..
Currently the way to pass result from BPF helper to BPF program is to
provide memory area defined by pointer and size: func(void *, size_t).
It works great for generic use-case, but for simple types, such as int,
it's overkill and consumes two arguments when it could use just one.
Introduce new a
Add bpf_strtol and bpf_strtoul to convert a string to long and unsigned
long correspondingly. It's similar to user space strtol(3) and
strtoul(3) with a few changes to the API:
* instead of NUL-terminated C string the helpers expect buffer and
buffer length;
* resulting long or unsigned long is
Test that new value provided by user space on sysctl write can be read
by bpf_sysctl_get_new_value and overridden by bpf_sysctl_set_new_value.
# ./test_sysctl
...
Test case: sysctl_get_new_value sysctl:read EINVAL .. [PASS]
Test case: sysctl_get_new_value sysctl:write ok .. [PASS]
Test c
Sync bpf_strtoX related bpf UAPI changes to tools/.
Signed-off-by: Andrey Ignatov
---
tools/include/uapi/linux/bpf.h | 51 +-
1 file changed, 50 insertions(+), 1 deletion(-)
diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/bpf.h
index 01faa2
Add unit test for BPF_PROG_TYPE_CGROUP_SYSCTL program type.
Test that program can allow/deny access.
Test both valid and invalid accesses to ctx->write.
Example of output:
# ./test_sysctl
Test case: sysctl wrong attach_type .. [PASS]
Test case: sysctl:read allow all .. [PASS]
Test case: s
Test w/ and w/o BPF_F_SYSCTL_BASE_NAME, buffers with enough space and
too small buffers to get E2BIG and truncated result, etc.
# ./test_sysctl
...
Test case: sysctl_get_name sysctl_value:base ok .. [PASS]
Test case: sysctl_get_name sysctl_value:base E2BIG truncated .. [PASS]
Test case:
Test sysctl_get_current_value on sysctl read and write, buffers with
enough space and too small buffers to get E2BIG and truncated result,
etc.
# ./test_sysctl
...
Test case: sysctl_get_current_value sysctl:read ok, gt .. [PASS]
Test case: sysctl_get_current_value sysctl:read ok, eq .. [PA
Add unit test to verify that program and attach types are properly
identified for "cgroup/sysctl" section name.
Signed-off-by: Andrey Ignatov
---
tools/testing/selftests/bpf/test_section_names.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/tools/testing/selftests/bpf/test_section_nam
Support BPF_PROG_TYPE_CGROUP_SYSCTL program in libbpf: identifying
program and attach types by section name, probe.
Signed-off-by: Andrey Ignatov
---
tools/lib/bpf/libbpf.c| 3 +++
tools/lib/bpf/libbpf_probes.c | 1 +
2 files changed, 4 insertions(+)
diff --git a/tools/lib/bpf/libbpf.c
Currently kernel/bpf/cgroup.c contains only one program type and one
proto function cgroup_dev_func_proto(). It'd be useful to have base
proto function that can be reused for new cgroup-bpf program types
coming soon.
Introduce cgroup_base_func_proto().
Signed-off-by: Andrey Ignatov
---
kernel/b
Add bpf_sysctl_* and bpf_strtoX helpers to bpf_helpers.h.
Signed-off-by: Andrey Ignatov
---
tools/testing/selftests/bpf/bpf_helpers.h | 19 +++
1 file changed, 19 insertions(+)
diff --git a/tools/testing/selftests/bpf/bpf_helpers.h
b/tools/testing/selftests/bpf/bpf_helpers.h
in
Add helpers to work with new value being written to sysctl by user
space.
bpf_sysctl_get_new_value() copies value being written to sysctl into
provided buffer.
bpf_sysctl_set_new_value() overrides new value being written by user
space with a one from provided buffer. Buffer should contain string
Test that bpf_strtol and bpf_strtoul helpers can be used to convert
provided buffer to long or unsigned long correspondingly and return both
correct result and number of consumed bytes, or proper errno.
Example of output:
# ./test_sysctl
..
Test case: bpf_strtoul one number string .. [PASS]
Add C based test for a few bpf_sysctl_* helpers and bpf_strtoul.
Make sure that sysctl can be identified by name and that multiple
integers can be parsed from sysctl value with bpf_strtoul.
net/ipv4/tcp_mem is chosen as a testing sysctl, it contains 3 unsigned
longs, they all are parsed and compa
Sync BPF_PROG_TYPE_CGROUP_SYSCTL related bpf UAPI changes to tools/.
Signed-off-by: Andrey Ignatov
---
tools/include/uapi/linux/bpf.h | 90 +-
1 file changed, 89 insertions(+), 1 deletion(-)
diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/b
Test that verifier handles new argument types properly, including
uninitialized or partially initialized value, misaligned stack access,
etc.
Example of output:
#456/p ARG_PTR_TO_LONG uninitialized OK
#457/p ARG_PTR_TO_LONG half-uninitialized OK
#458/p ARG_PTR_TO_LONG misaligned OK
#459/p
Add bpf_sysctl_get_current_value() helper to copy current sysctl value
into provided by BPF_PROG_TYPE_CGROUP_SYSCTL program buffer.
It provides same string as user space can see by reading corresponding
file in /proc/sys/, including new line, etc.
Documentation for the new helper is provided in b
Add file_pos field to bpf_sysctl context to read and write sysctl file
position at which sysctl is being accessed (read or written).
The field can be used to e.g. override whole sysctl value on write to
sysctl even when sys_write is called by user space with file_pos > 0. Or
BPF program may reject
Containerized applications may run as root and it may create problems
for whole host. Specifically such applications may change a sysctl and
affect applications in other containers.
Furthermore in existing infrastructure it may not be possible to just
completely disable writing to sysctl, instead
Add bpf_sysctl_get_name() helper to copy sysctl name (/proc/sys/ entry)
into provided by BPF_PROG_TYPE_CGROUP_SYSCTL program buffer.
By default full name (w/o /proc/sys/) is copied, e.g. "net/ipv4/tcp_mem".
If BPF_F_SYSCTL_BASE_NAME flag is set, only base name will be copied,
e.g. "tcp_mem".
Doc
The patch set introduces new BPF hook for sysctl.
It adds new program type BPF_PROG_TYPE_CGROUP_SYSCTL and attach type
BPF_CGROUP_SYSCTL.
BPF_CGROUP_SYSCTL hook is placed before calling to sysctl's proc_handler so
that accesses (read/write) to sysctl can be controlled for specific cgroup
and eith
When phylink_of_phy_connect fails, dsa_slave_phy_setup tries to save the
day by connecting to an alternative PHY, none other than a PHY on the
switch's internal MDIO bus, at an address equal to the port's index.
However this does not take into consideration the scenario when the
switch that failed
From: Jiri Pirko
In dumpit, unlike doit, the check for info_get op being defined
is missing. Add it and avoid null pointer dereference in case driver
does not define this op.
Fixes: f9cf22882c60 ("devlink: add device information API")
Reported-by: Ido Schimmel
Signed-off-by: Jiri Pirko
---
ne
Sat, Mar 23, 2019 at 09:31:02PM CET, jakub.kicin...@netronome.com wrote:
>On Fri, 22 Mar 2019 17:56:32 +0100, Jiri Pirko wrote:
>> From: Jiri Pirko
>>
>> Similar to other driver, move the port type set after netdev registration
>> is done. Along with that, clear the type before unregistration.
>>
Previously the green and amber LEDs on this quad PHY were solid, to
indicate an encoding of the link speed (10/100/1000).
This keeps the LEDs always on just as before, but now they flash on
Rx/Tx activity.
Signed-off-by: Vladimir Oltean
---
drivers/net/phy/broadcom.c | 13 +
include
From: Boris Pismenny
Date: Fri, 22 Mar 2019 14:32:45 +
> The source for this is not known to me.
> I've saw that this issue was mentioned and fixed in libc:
> http://sourceware.org/ml/libc-alpha/2016-05/msg00322.html
...
> Maybe libc maintainers need to be notified whenever linux/socket.h is
On Sat, Mar 23, 2019 at 2:18 AM wenxu wrote:
>
> On 2019/3/23 下午3:50, Pravin Shelar wrote:
>
> On Thu, Mar 21, 2019 at 3:34 AM wrote:
>
> From: wenxu
>
> There is currently no support for the multicasti/broadcst aspects
> of VXLAN in ovs. In the datapath flow the tun_dst must specific.
> But in
On Fri, 22 Mar 2019 17:56:32 +0100, Jiri Pirko wrote:
> From: Jiri Pirko
>
> Similar to other driver, move the port type set after netdev registration
> is done. Along with that, clear the type before unregistration.
>
> Signed-off-by: Jiri Pirko
> diff --git a/drivers/net/ethernet/netronome/nf
Update man page to reflect the changes made in Linux.
Signed-off-by: Leslie Monis
---
man/man8/tc-pie.8 | 40 ++--
man/man8/tc.8 | 1 +
2 files changed, 19 insertions(+), 22 deletions(-)
diff --git a/man/man8/tc-pie.8 b/man/man8/tc-pie.8
index 278293bd..
tc_pie_xstats->prob has a maximum value of (2^64 - 1).
Signed-off-by: Leslie Monis
---
tc/q_pie.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tc/q_pie.c b/tc/q_pie.c
index f7924ef5..6566ca10 100644
--- a/tc/q_pie.c
+++ b/tc/q_pie.c
@@ -198,7 +198,7 @@ static int pie_print
The PIE qdisc implementation in Linux has recently been
aligned with RFC 8033. This patch series makes tc/q_pie.c
consistent with that implementation and also updates the man
page tc-pie(8).
Leslie Monis (2):
tc: pie: change maximum integer value of tc_pie_xstats->prob
tc: pie: update man page
This partially reverts ed8fe20205ac ("net: dsa: mv88e6xxx: prevent
interrupt storm caused by mv88e6390x_port_set_cmode"). I missed
that chip->ports[].cmode is overwritten anyway by the cmode
caching in mv88e6xxx_setup().
Signed-off-by: Heiner Kallweit
---
The old code causes no issues, therefore
This patches fixes few issues in mv88e6390x_port_set_cmode().
1. When entering the function the old cmode may be 0, in this case
mv88e6390x_serdes_get_lane() returns -ENODEV. As result we bail
out and have no chance to set a new mode. Therefore deal properly
with -ENODEV.
2. Once we have
Hi Cong,
Thanks for your responses so far.
> On 22 Mar 2019, at 23:09, Cong Wang wrote:
>
> On Fri, Mar 22, 2019 at 3:06 PM Kevin 'ldir' Darbyshire-Bryant
> wrote:
>>
>>
>>
>>> On 22 Mar 2019, at 21:31, Cong Wang wrote:
>>>
>>> On Fri, Mar 22, 2019 at 1:50 PM Kevin 'ldir' Darbyshire-Bryan
/linux/commits/Yi-Hung-Wei/netfilter-Export-nf_ct_-set-destroy-_timeout/20190323-195349
base: https://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-next.git master
config: x86_64-randconfig-b0-03232323 (attached as .config)
compiler: gcc-7 (Debian 7.3.0-1) 7.3.0
reproduce:
# save the
From: Willem de Bruijn
These three variables are set in one branch and used in another with
the same condition. But on some architectures they still generate
compiler warnings of the kind:
warning: 'inner_trans' may be used uninitialized in this function
[-Wmaybe-uninitialized]
Silence these
On 03/22, Alexei Starovoitov wrote:
> On Fri, Mar 22, 2019 at 06:19:57PM -0700, Stanislav Fomichev wrote:
> > Are we ok with breaking api in this case? I'm all in on removing this
> > extra information. We can always put it back if somebody complains (and
> > manually parse in eth_get_headlen case)
On Sat, Mar 23, 2019 at 6:53 AM Daniel Borkmann wrote:
>
> Hey Willem,
>
> On 03/22/2019 07:32 PM, Willem de Bruijn wrote:
> > From: Willem de Bruijn
> >
> > BPF allows for dynamic tunneling, choosing the tunnel destination and
> > features on-demand. Extend bpf_skb_adjust_room to allow for effic
From: Willem de Bruijn
These three variables are set in one branch and used in another with
the same condition. But on some architectures they still generate
compiler warnings of the kind:
warning: 'inner_trans' may be used uninitialized in this function
[-Wmaybe-uninitialized]
Silence these
On Sat, Mar 23, 2019 at 02:12:39AM -0700, Eric Dumazet wrote:
>
>
> On 03/23/2019 01:05 AM, brakmo wrote:
> > This patchset adds support for propagating congestion notifications (cn)
> > to TCP from cgroup inet skb egress BPF programs.
> >
> > Current cgroup skb BPF programs cannot trigger TCP c
This correlates with default internet MTU. This also allows page
flip/reuse to be activated, since each allocated RX page now serves for
two frags/packets.
Signed-off-by: Igor Russkikh
---
drivers/net/ethernet/aquantia/atlantic/aq_cfg.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
From: Nikita Danilov
Default LRO HW configuration was very conservative.
Low Number of Descriptors per LRO Sequence, small session
timeout, inefficient settings in interrupt generation logic.
Change max number of LRO descriptors from 2 to 16 to
increase performance. Increase maximum coalescing
Atlantic driver used 14 bytes preallocated skb size. That made L3 protocol
processing inefficient because pskb_pull had to fetch all the L3/L4 headers
from extra fragments.
Specially on UDP flows that caused extra packet drops because CPU was
overloaded with pskb_pull.
This patch uses eth_get_hea
Here is a set of patches targeting for performance improvement
on various platforms and protocols.
Our main target was rx performance on iommu systems, notably
NVIDIA Jetson TX2 and NVIDIA Xavier platforms.
We introduce page reuse strategy to better deal with iommu dma mapping costs.
With it we s
We introduce internal aq_rxpage wrapper over regular page
where extra field is tracked: rxpage offset inside of allocated page.
This offset allows to reuse one page for multiple packets.
When needed (for example with large frames processing), allocated
pageorder could be customized. This gives eve
The driver is now constantly tested in our lab on aarch64 hardware:
Jetson tx2, Pascal and Xavier tegra based hardware.
Many of tegra smmu related HW bugs were fixed or workarounded already.
Thus, add ARM64 into Kconfig.
Add also COMPILE_TEST dependency.
Signed-off-by: Igor Russkikh
---
driver
Before that, we've refilled ring even on single descriptor move.
Under high packet load that caused page allocation logic to be triggered
too often. That made overall ring processing slower.
Moreover, with page buffer reuse implemented, we should give a chance
higher networking levels to process r
For multigig rates 1K ring size is often not enough and causes extra
packet drops in hardware.
Signed-off-by: Igor Russkikh
---
drivers/net/ethernet/aquantia/atlantic/aq_cfg.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/aquantia/atlantic/aq_cfg.h
b/d
/linux/commits/Yi-Hung-Wei/netfilter-Export-nf_ct_-set-destroy-_timeout/20190323-195349
base: https://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-next.git master
config: x86_64-rhel (attached as .config)
compiler: gcc-7 (Debian 7.3.0-1) 7.3.0
reproduce:
# save the attached .config to linux
The ipip tunnel introduced in commit dd9ee3444014 ("vti4: Fix a ipip
packet processing bug in 'IPCOMP' virtual tunnel") largely duplicated
the existing vti_input and vti_recv functions. Refactored to
deduplicate the common code.
Signed-off-by: Jeremy Sowden
---
net/ipv4/ip_vti.c | 60 ++
/linux/commits/Yi-Hung-Wei/netfilter-Export-nf_ct_-set-destroy-_timeout/20190323-195349
base: https://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-next.git master
config: i386-randconfig-n1-201911 (attached as .config)
compiler: gcc-7 (Debian 7.3.0-1) 7.3.0
reproduce:
# save the attached
From: Dmitry Bezrukov
New device of QNAP based on aqc111u
Add this ID to blacklist of cdc_ether driver as well
Signed-off-by: Dmitry Bezrukov
---
drivers/net/usb/aqc111.c| 15 +++
drivers/net/usb/cdc_ether.c | 8
2 files changed, 23 insertions(+)
diff --git a/drivers
RFC 8033 replaces the IETF draft for PIE
Signed-off-by: Leslie Monis
---
net/sched/Kconfig | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/net/sched/Kconfig b/net/sched/Kconfig
index 1b9afdee5ba9..5c02ad97ef23 100644
--- a/net/sched/Kconfig
+++ b/net/sched/Kconfig
@@ -358,8
Simplify aqr_config_aneg().
Signed-off-by: Heiner Kallweit
---
drivers/net/phy/aquantia_main.c | 45 ++---
1 file changed, 13 insertions(+), 32 deletions(-)
diff --git a/drivers/net/phy/aquantia_main.c b/drivers/net/phy/aquantia_main.c
index 1013711e1..37032f3ba 1006
Add reporting firmware details. These details are available only once
the firmware has finished initializing the chip. This can take some
time and we need to poll for init completion.
Signed-off-by: Heiner Kallweit
---
drivers/net/phy/aquantia_main.c | 58 +
1 fil
The AQCS109 supports a proprietary 2-pair 1Gbps mode. The standard
registers don't allow to tell between 1000BaseT and 1000BaseT2.
Add reporting this proprietary mode based on a vendor register.
Signed-off-by: Heiner Kallweit
---
drivers/net/phy/aquantia_main.c | 14 +-
1 file change
If both link partners are Aquantia PHY's then additional information is
exchanged as part of the auto-negotiation. Report remote capabilities
if link partner is Aquantia PHY.
Signed-off-by: Heiner Kallweit
---
drivers/net/phy/aquantia_main.c | 49 +
1 file changed
This series detects and reports quite some Aquantia-specific settings
and features.
Heiner Kallweit (3):
net: phy: aquantia: report remote capabilities if link partner is
Aquantia PHY
net: phy: aquantia: add debug message with PHY details like firmware
version
net: phy: aquantia: inf
After upgrading iproute2 from 4.20 to 5.0 the following error occurs:
$ ip route show table default
Error: ipv4: FIB table does not exist.
Dump terminated
The command works for all tables other than 'default' one. It seems related to
this commit[1]
I also saw "Error: ipv4: MR table does not exi
From: wenxu
There is currently no support for the multicast/broadcast aspects
of VXLAN in ovs. In the datapath flow the tun_dst must specific.
But in the IP_TUNNEL_INFO_BRIDGE mode the tun_dst can not be specific.
And the packet can forward through the fdb table of vxlan devcice. In
this mode the
Hey Willem,
On 03/22/2019 07:32 PM, Willem de Bruijn wrote:
> From: Willem de Bruijn
>
> BPF allows for dynamic tunneling, choosing the tunnel destination and
> features on-demand. Extend bpf_skb_adjust_room to allow for efficient
> tunneling at the TC hooks.
Patch 9 in this series caused the f
On 2019/3/23 下午3:50, Pravin Shelar wrote:
>
>> diff --git a/net/openvswitch/flow_netlink.c b/net/openvswitch/flow_netlink.c
>> index 691da85..033df5c 100644
>> --- a/net/openvswitch/flow_netlink.c
>> +++ b/net/openvswitch/flow_netlink.c
>> @@ -403,6 +403,7 @@ size_t ovs_key_attr_size(void)
>>
1 - 100 of 113 matches
Mail list logo
