The use of smp_rmb() and smp_wmb() creates a Linux header dependency
on barrier.h that is uneccessary in most parts. This patch implements
the two small defines that are needed from barrier.h. As a bonus, the
new implementations are faster than the default ones as they default
to sfence and lfence
This patch removes the use of likely and unlikely in xsk.h since they
create a dependency on Linux headers as reported by several
users. There have also been reports that the use of these decreases
performance as the compiler puts the code on two different cache lines
instead of on a single one. Al
This patch set removes two dependencies on Linux kernel headers
from the XDP socket code in libbpf. A number of people have pointed
out that these two dependencies make it hard to build the XDP socket
part of libbpf without any kernel header dependencies. The two removed
dependecies are:
* Remove
I find that I do not restore every UDP paket from recvmsg when I
programming with UDP_GRO in kernel version 5.0
I can get gso_size, but I do not know the last UDP packet size in
aggregate datagram.
For example, recvmsg returns a packet size 50 of and gso_size of
24, so the UDP packets is restor
On Mon, Apr 08, 2019 at 09:04:30AM -0700, Stanislav Fomichev wrote:
> Add new set of arguments to bpf_attr for BPF_PROG_TEST_RUN:
> * ctx_in/ctx_size_in - input context
> * ctx_out/ctx_size_out - output context
>
> The intended use case is to pass some meta data to the test runs that
> operate on
Am 09.04.19 um 05:41 schrieb Cong Wang:
> On Mon, Apr 8, 2019 at 7:22 PM Reindl Harald wrote:
>> after two weeks and 27 Mio. accepted connections 5.0.4 crashed too
>>
>> "vmcore-dmesg" piped through "sort | uniq" is reduced to 399 lines
>> containing just rate-limited "-j LOG" iptables events a
Audit events could happen in a network namespace outside of a task
context due to packets received from the net that trigger an auditing
rule prior to being associated with a running task. The network
namespace could be in use by multiple containers by association to the
tasks in that network name
Some SOC like i.MX6SX clock have some limits:
- ahb clock should be disabled before ipg.
- ahb and ipg clocks are required for MAC MII bus.
So, move the ahb clock to runtime management together with
ipg clock.
Signed-off-by: Fugang Duan
---
drivers/net/ethernet/freescale/fec_main.c | 30
On Mon, Apr 8, 2019 at 7:22 PM Reindl Harald wrote:
>
>
>
> Am 25.03.19 um 23:10 schrieb Reindl Harald:
> >>> fact is that around 4.19.x the kernel had a ton of issues starting with
> >>> conncount broken over months (again: with a simple method get the
> >>> stacktrace it would have been easily d
Am 25.03.19 um 23:10 schrieb Reindl Harald:
>>> fact is that around 4.19.x the kernel had a ton of issues starting with
>>> conncount broken over months (again: with a simple method get the
>>> stacktrace it would have been easily discovered), the scheduler issue in
>>> 4.19.x eating peoples dat
Unlike '&&' operator, the '&' does not have short-circuit
evaluation semantics. IOW both sides of the operator always
get evaluated. Fix the wrong operator in
tls_is_sk_tx_device_offloaded(), which would lead to
out-of-bounds access for for non-full sockets.
Fixes: 4799ac81e52a ("tls: Add rx inl
[ Upstream commit 997dd96471641e147cb2c33ad54284000d0f5e35 ]
Currently, IPv6 defragmentation code drops non-last fragments that
are smaller than 1280 bytes: see
commit 0ed4229b08c1 ("ipv6: defrag: drop non-last frags smaller than min mtu")
This behavior is not specified in IPv6 RFCs and appears t
[ Upstream commit c23f35d19db3b36ffb9e04b08f1d91565d15f84f ]
This is a refactoring patch: without changing runtime behavior,
it moves rbtree-related code from IPv4-specific files/functions
into .h/.c defrag files shared with IPv6 defragmentation code.
Signed-off-by: Peter Oskolkov
Cc: Eric Dumaz
[ Upstream commit d4289fcc9b16b89619ee1c54f829e05e56de8b9a ]
Currently, IPv6 defragmentation code drops non-last fragments that
are smaller than 1280 bytes: see
commit 0ed4229b08c1 ("ipv6: defrag: drop non-last frags smaller than min mtu")
This behavior is not specified in IPv6 RFCs and appears t
Currently, 4.19 and earlier stable kernels contain a security fix
that is not fully IPv6 standard compliant.
This patchset backports IPv6 defrag fixes from 5.1rc that restore
standard-compliance.
Original 5.1 patchet: https://patchwork.ozlabs.org/cover/1029418/
John Masinter (captwiggum), could
On Mon, Apr 8, 2019 at 4:15 PM Sasha Levin wrote:
>
> On Mon, Apr 08, 2019 at 10:13:57AM -0700, Peter Oskolkov wrote:
> >On Mon, Apr 8, 2019 at 9:29 AM Captain Wiggum wrote:
> >>
> >> Thank you Peter!
> >>
> >> I tried the patch on 4.9.167 & 4.19.32. It's out of sync with upstream.
> >> Looks lik
From: Michael Chan
Date: Mon, 8 Apr 2019 17:39:53 -0400
> The first patch prevents possible driver crash if we get a bad RX index
> from the hardware. The second patch resets the device when the hardware
> reports buffer error to recover from the error.
>
> Please queue these for -stable also.
From: Michael Zhivich
Date: Mon, 8 Apr 2019 15:00:46 -0400
> Use constants provided by ethtool.h for speed and duplex values
> instead of raw integer constants to increase code readability.
>
> thunder_bgx already uses SPEED_UNKNOWN and DUPLEX_UNKNOWN constants,
> also provided by ethtool.h.
>
From: Heiner Kallweit
Date: Mon, 8 Apr 2019 19:28:28 +0200
> The original patch didn't set phydev->autoneg_complete in one exit path.
> Fix this.
>
> Fixes: 4950c2ba49cc ("net: phy: fix autoneg mismatch case in
> genphy_read_status")
> Reported-by: Simon Horman
> Tested-by: Simon Horman
> Sig
From: Michael Zhivich
Date: Mon, 8 Apr 2019 10:48:44 -0400
> This patch series addresses 2 related issues:
>
> 1. ethtool_validate_speed() triggers a "signed-unsigned comparison"
> warning due to type difference of SPEED_UNKNOWN constant (int)
> and argument to ethtool_validate_speed (__u32).
>
From: Ido Schimmel
Date: Mon, 8 Apr 2019 06:59:33 +
> Shalom says:
>
> Expose the following ASIC information via devlink info command:
> - Driver name
> - Hardware revision
> - Firmware PSID
> - Running firmware version
Series applied.
From: Kristian Evensen
Date: Sun, 7 Apr 2019 15:39:09 +0200
> Most, if not all, Quectel devices use dynamic interface numbers, and
> users are able to change the USB configuration at will. Matching on for
> example interface number is therefore not possible.
>
> Instead, the QMI device can be i
From: Heiner Kallweit
Date: Sun, 7 Apr 2019 13:55:01 +0200
> Recently a number of generic functions for Clause 45 PHY's has been
> added. So let's replace the old very limited genphy_10g_driver with a
> genphy_c45_driver. This driver isn't limited to 10G, however it's worth
> to be noted that Cla
From: Christopher Monsanto
Date: Sat, 6 Apr 2019 12:33:25 -0400
> It is currently impossible for the reader of an AF_UNIX stream socket to
> fully reconstruct the data sent in the presence of SCM_RIGHTS, without
> reading byte-for-byte. This prevents efficiently proxying or providing a
> high-le
From: Lorenzo Bianconi
Date: Sat, 6 Apr 2019 17:16:51 +0200
> Similar to what I did in commit bb9bd814ebf0, fix possible use-after-free
> in erspan_rcv and ip6erspan_rcv extracting tunnel metadata since the
> packet can be 'uncloned' running __iptunnel_pull_header
Series applied and queued up f
On Mon, Apr 08, 2019 at 10:13:57AM -0700, Peter Oskolkov wrote:
On Mon, Apr 8, 2019 at 9:29 AM Captain Wiggum wrote:
Thank you Peter!
I tried the patch on 4.9.167 & 4.19.32. It's out of sync with upstream.
Looks like a little different work needed for each LTS kernel.
Is someone is familiar w
From: Lucas Bates
Date: Mon, 8 Apr 2019 18:24:08 -0400
> On Sun, Apr 7, 2019 at 10:32 PM David Miller wrote:
>> > I'm not sure to understand the goal of this series:
>> > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/log/?h=45b3a3762721
>> >
>> > But after it, tc selftests a
On Fri, 5 Apr 2019 20:56:26 +0300, Vlad Buslov wrote:
> John reports:
>
> Recent refactoring of fl_change aims to use the classifier spinlock to
> avoid the need for rtnl lock. In doing so, the fl_hw_replace_filer()
> function was moved to before the lock is taken. This can create problems
> for
On Sun, Apr 7, 2019 at 10:32 PM David Miller wrote:
> > I'm not sure to understand the goal of this series:
> > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/log/?h=45b3a3762721
> >
> > But after it, tc selftests are not intuitive anymore. When naively running
> > them
> > (.
From: David Ahern
Date: Fri, 5 Apr 2019 16:30:23 -0700
> Last set of three with the end goal of enabling IPv6 gateways with IPv4
> routes.
Series applied, thanks David.
From: Heiner Kallweit
Date: Fri, 5 Apr 2019 20:46:46 +0200
> There's a significant number of reports that re-enabling ASPM causes
> different issues, ranging from decreased performance to system not
> booting at all. This affects only a minority of users, but the number
> of affected users is big
From: Heiner Kallweit
Date: Fri, 5 Apr 2019 21:23:13 +0200
> genphy_read_status() so far checks phydev->supported, not the actual
> PHY capabilities. This can make a difference if the supported speeds
> have been limited by of_set_phy_supported() or phy_set_max_speed().
>
> It seems that this is
If the RX completion indicates RX buffers errors, the RX ring will be
disabled by firmware and no packets will be received on that ring from
that point on. Recover by resetting the device.
Fixes: c0c050c58d84 ("bnxt_en: New Broadcom ethernet driver.")
Signed-off-by: Michael Chan
---
drivers/net
There is logic to check that the RX/TPA consumer index is the expected
index to work around a hardware problem. However, the potentially bad
consumer index is first used to index into an array to reference an entry.
This can potentially crash if the bad consumer index is beyond legal
range. Impro
The first patch prevents possible driver crash if we get a bad RX index
from the hardware. The second patch resets the device when the hardware
reports buffer error to recover from the error.
Please queue these for -stable also. Thanks.
Michael Chan (2):
bnxt_en: Improve RX consumer index val
From: Saeed Mahameed
Date: Fri, 5 Apr 2019 14:17:04 -0700
> This series adds misc updates to mlx5 driver.
>
> For more information please see tag log below.
>
> Please pull and let me know if there is any problem.
>
> v1->v2:
> remove inline keywords from static functions in c files.
Pulle
On 04/08/2019 09:35 AM, Paolo Abeni wrote:
> This revert commit 46b1c18f9deb ("net: sched: put back q.qlen
> into a single location").
> After the previous patch nobody accesses directly qlen for a child
> qdisc when such qdisc does per CPU stats accounting.
> In the control path nobody uses dir
The thunderx driver forbids to load an eBPF program if the MTU is too high,
but this can be circumvented by loading the eBPF, then raising the MTU.
Fix this by limiting the MTU if an eBPF program is already loaded.
Fixes: 05c773f52b96e ("net: thunderx: Add basic XDP support")
Signed-off-by: Matte
The thunderx driver splits frames bigger than 1530 bytes to multiple
pages, making impossible to run an eBPF program on it.
This leads to a maximum MTU of 1508 if QinQ is in use.
The thunderx driver forbids to load an eBPF program if the MTU is higher
than 1500 bytes. Raise the limit to 1508 so it
The thunderx driver can't use XDP with all MTU values.
This patches sets the right MTU values, and add a check to avoid setting
a wrong value which will not function.
Matteo Croce (2):
net: thunderx: raise XDP MTU to 1508
net: thunderx: don't allow jumbo frames with XDP
.../net/ethernet/cavi
On 04/08/2019 07:05 PM, Toke Høiland-Jørgensen wrote:
> From: Björn Töpel
>
> Generic XDP and driver XDP cannot be enabled at the same time. However,
> they don't share any state; let's fix that. Here, dev->xdp_prog, is
> used for both driver and generic mode. This removes the need for the
> XDP_
On Sat, 2019-04-06 at 19:20 +0200, Daniele Palmas wrote:
>
> the qmi_wwan sysfs qmap feature, being very easy to use, is serving
> well for me and customers of the company I work for (mainly directly
> with libqmi, not ModemManager),
Yeah, I don't doubt this. In fact, we could arguably provide t
On Thu, 2019-04-04 at 22:45 -0600, Subash Abhinov Kasiviswanathan wrote:
> On 2019-04-04 14:38, Johannes Berg wrote:
> > Hi,
> >
> > > The normal mode of operation of rmnet is using the rmnet netdevices
> > > in an embedded device.
> >
> > Sure. Can you say what driver this would typically live o
On Mon, Apr 8, 2019 at 1:04 PM David Miller wrote:
>
> From: Paul Thomas
> Date: Mon, 8 Apr 2019 09:20:47 -0400
>
> > + if (unlikely(skb_shinfo(skb)->tx_flags &
> > SKBTX_HW_TSTAMP) &&
> > + gem_ptp_do_txstamp(queue, skb, desc)
>
Make sure SKBTX_HW_TSTAMP (i.e. SOF_TIMESTAMPING_TX_HARDWARE) has been
enabled for this skb. It does fix the issue where normal socks that
aren't expecting a timestamp will not wake up on select, but when a
user does want a SOF_TIMESTAMPING_TX_HARDWARE it does work.
Signed-off-by: Paul Thomas
---
On Mon, Apr 08, 2019 at 03:00:46PM -0400, Michael Zhivich wrote:
> Use constants provided by ethtool.h for speed and duplex values
> instead of raw integer constants to increase code readability.
>
> thunder_bgx already uses SPEED_UNKNOWN and DUPLEX_UNKNOWN constants,
> also provided by ethtool.h.
On Mon, Apr 8, 2019 at 12:59 PM Alan Maguire wrote:
>
> Update test_tc_tunnel to verify adding inner L2 header
> encapsulation (an MPLS label or ethernet header) works.
>
> Signed-off-by: Alan Maguire
> static const int cfg_udp_src = 2;
> static const int cfg_udp_dst = ;
> +/* MPLSover
On Mon, Apr 8, 2019 at 12:59 PM Alan Maguire wrote:
>
> commit 868d523535c2 ("bpf: add bpf_skb_adjust_room encap flags")
> introduced support to bpf_skb_adjust_room for GSO-friendly GRE
> and UDP encapsulation.
>
> For GSO to work for skbs, the inner headers (mac and network) need to
> be marked.
On Mon, Apr 8, 2019 at 1:01 PM Alan Maguire wrote:
>
> commit 868d523535c2 ("bpf: add bpf_skb_adjust_room encap flags")
> introduced support to bpf_skb_adjust_room for GSO-friendly GRE
> and UDP encapsulation and later introduced associated test_tc_tunnel
> tests. Here those tests are extended to
On Mon, Apr 8, 2019 at 12:59 PM Alan Maguire wrote:
>
> Extend bpf_skb_adjust_room growth to mark inner MAC header so that
> L2 encapsulation can be used for tc tunnels.
>
> Patch #1 extends the existing test_tc_tunnel to support UDP
> encapsulation; later we want to be able to test MPLS over UDP
Use constants provided by ethtool.h for speed and duplex values
instead of raw integer constants to increase code readability.
thunder_bgx already uses SPEED_UNKNOWN and DUPLEX_UNKNOWN constants,
also provided by ethtool.h.
Signed-off-by: Michael Zhivich
---
drivers/net/ethernet/cavium/thunder/
On 08.04.2019 20:12, Andrew Lunn wrote:
>> Splitting the series and waiting for a Tested-by, as proposed by Richard,
>> may be problematic because most PHY drivers don't have a dedicated
>> maintainer, and we lack the hardware to test.
>
> Well, we could split out those we know are likely to get t
On 4/8/19, 1:55 PM, "Andrew Lunn" wrote:
>On Mon, Apr 08, 2019 at 10:48:44AM -0400, Michael Zhivich wrote:
>> This patch series addresses 2 related issues:
>>
>> 1. ethtool_validate_speed() triggers a "signed-unsigned comparison"
>> warning due to type difference of SPEED_UNKNOWN constant (int)
On Mon, 8 Apr 2019 06:59:36 +, Ido Schimmel wrote:
> From: Shalom Toledo
>
> Add firmware parameter id (fw.psid).
>
> Signed-off-by: Shalom Toledo
> Acked-by: Jiri Pirko
> Signed-off-by: Ido Schimmel
Acked-by: Jakub Kicinski
Thanks!
Hi,
On Wed, Apr 03, 2019 at 08:34:16AM +0300, Dan Carpenter wrote:
> NEXTHDR_MAX is 255. What happens here is that we take a u8 value
> "hdr->nexthdr" from the network and then look it up in
> lowpan_nexthdr_nhcs[]. The problem is that if hdr->nexthdr is 0xff then
> we read one element beyond th
> Splitting the series and waiting for a Tested-by, as proposed by Richard,
> may be problematic because most PHY drivers don't have a dedicated
> maintainer, and we lack the hardware to test.
Well, we could split out those we know are likely to get tested. I
have a number of boards with Marvell
On Mon, Apr 08, 2019 at 10:48:44AM -0400, Michael Zhivich wrote:
> This patch series addresses 2 related issues:
>
> 1. ethtool_validate_speed() triggers a "signed-unsigned comparison"
> warning due to type difference of SPEED_UNKNOWN constant (int)
> and argument to ethtool_validate_speed (__u32)
On Mon, Apr 08, 2019 at 10:48:47AM -0400, Michael Zhivich wrote:
> qlcnic driver uses u16 to store SPEED_UKNOWN ethtool constant,
> which is defined as -1, resulting in value truncation and
> thus incorrect test results against SPEED_UNKNOWN.
>
> For example, the following test will print "False":
On Mon, Apr 08, 2019 at 10:48:46AM -0400, Michael Zhivich wrote:
> tg3 driver uses u16 to store SPEED_UKNOWN ethtool constant,
> which is defined as -1, resulting in value truncation and
> thus incorrect test results against SPEED_UNKNOWN.
>
> For example, the following test will print "False":
>
On 08.04.2019 00:13, Andrew Lunn wrote:
> On Sun, Apr 07, 2019 at 11:57:13AM +0200, Heiner Kallweit wrote:
>> Recently genphy_read_abilities() has been added that dynamically detects
>> clause 22 PHY abilities. I *think* this detection should work with all
>> supported PHY's, at least for the ones
On Mon, Apr 8, 2019 at 10:07 AM Eric Dumazet wrote:
>
>
>
> On 04/08/2019 09:16 AM, Neal Cardwell wrote:
> > On Wed, Apr 3, 2019 at 8:13 PM brakmo wrote:
> >>
> >> When a packet is dropped when calling queue_xmit in __tcp_transmit_skb
> >> and packets_out is 0, it is beneficial to set a small pr
The original patch didn't set phydev->autoneg_complete in one exit path.
Fix this.
Fixes: 4950c2ba49cc ("net: phy: fix autoneg mismatch case in
genphy_read_status")
Reported-by: Simon Horman
Tested-by: Simon Horman
Signed-off-by: Heiner Kallweit
---
drivers/net/phy/phy_device.c | 10 -
On Mon, Apr 8, 2019 at 9:29 AM Captain Wiggum wrote:
>
> Thank you Peter!
>
> I tried the patch on 4.9.167 & 4.19.32. It's out of sync with upstream.
> Looks like a little different work needed for each LTS kernel.
> Is someone is familiar with it, and is available to patch it?
I'll try to backpo
The Credit Based Shaper heavily depends on link speed to calculate
the scheduling credits, we can't properly calculate the credits if the
device has failed to report the link speed.
In that case we can't dequeue packets assuming a wrong port rate that will
result into an inconsistent credit distri
The Time Aware Priority Scheduler is heavily dependent to link speed,
it relies on it to calculate transmission bytes per cycle, we can't
properly calculate the so called budget if the device has failed
to report the link speed.
In that case we can't dequeue packets assuming a wrong budget.
This p
This set fixes miscalculations based on invalid link speed values.
Changes in v6:
+ Avoid locking a spinlock while calling __ethtool_get_link_ksettings()
(suggested by: Cong Wang);
Changes in v5:
+ Don't iterate over all the net_device maintained list (suggested by: Florian
Fainelli);
Chan
From: Stefan Schmidt
Date: Mon, 8 Apr 2019 18:08:04 +0200
> When moving the documentation for the ieee802154 subsystem from
> plain text to rst the file pattern in the MAINTAINERS file got wrong.
> Updating it here to fix scripts using this file.
>
> Reported-by: Joe Perches
> Signed-off-by: S
On 04/08/2019 09:16 AM, Neal Cardwell wrote:
> On Wed, Apr 3, 2019 at 8:13 PM brakmo wrote:
>>
>> When a packet is dropped when calling queue_xmit in __tcp_transmit_skb
>> and packets_out is 0, it is beneficial to set a small probe timer.
>> Otherwise, the throughput for the flow can suffer be
From: Björn Töpel
Generic XDP and driver XDP cannot be enabled at the same time. However,
they don't share any state; let's fix that. Here, dev->xdp_prog, is
used for both driver and generic mode. This removes the need for the
XDP_QUERY_PROG command to ndo_bpf.
Signed-off-by: Björn Töpel
Signed
An XDP program can redirect packets between interfaces using either the
xdp_redirect() helper or the xdp_redirect_map() helper. Apart from the
flexibility of updating maps from userspace, the redirect_map() helper also
uses the map structure to batch packets, which results in a significant
(around
A common pattern when using xdp_redirect_map() is to create a device map
where the lookup key is simply ifindex. Because device maps are arrays,
this leaves holes in the map, and the map has to be sized to fit the
largest ifindex, regardless of how many devices actually are actually
needed in the m
This adds a new selftest checking the allocation and de-allocation of
default maps in different network namespaces. It loads the two different
kinds of programs that need allocation (programs using tail calls, and
programs using redirect), and moves interfaces around between namespaces to
make sure
The subsequent commits introducing default maps and a hash-based ifindex
devmap require a bit of refactoring of the devmap code. Perform this first
so the subsequent commits become easier to read.
Also split out the final freeing and flushing of devmaps into a workqueue,
to make it easier to queue
This series changes the xdp_redirect helper to use a hidden default map. The
redirect_map() helper also uses the map structure to batch packets, which
results in a significant (around 50%) performance boost for the _map variant.
However, the xdp_redirect() API is simpler if one just wants to redire
From: Thomas Bogendoerfer
Date: Mon, 8 Apr 2019 16:20:54 +0200
> RTC platform device will be setup by new IOC3 MFD driver, therefore
> remove it from IP27 platform code.
>
> Signed-off-by: Thomas Bogendoerfer
You cannot break the driver like this.
Your patch series must be fully bisectable,
From: Björn Töpel
With the previous patch, we no longer need the XDP_QUERY_PROG operation, so
remove the handling of that from all drivers.
Signed-off-by: Björn Töpel
Signed-off-by: Toke Høiland-Jørgensen
---
drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c |4 ---
drivers/net/ethernet/
From: Paul Thomas
Date: Mon, 8 Apr 2019 09:20:47 -0400
> + if (unlikely(skb_shinfo(skb)->tx_flags &
> SKBTX_HW_TSTAMP) &&
> + gem_ptp_do_txstamp(queue, skb, desc) ==
> 0) {
The second line should start precisely at the first colu
From: Vishal Kulkarni
Date: Mon, 8 Apr 2019 18:03:49 +0530
> During hash filter programming, driver needs to return ENOSPC error
> intead of EAGAIN when TCAM is full.
>
> Signed-off-by: Vishal Kulkarni
Applied.
Sync include/uapi/linux/bpf.h with tools/ equivalent to add
BPF_F_ADJ_ROOM_ENCAP_L2(len) macro.
Signed-off-by: Alan Maguire
---
tools/include/uapi/linux/bpf.h | 10 ++
1 file changed, 10 insertions(+)
diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/bpf.h
index 837
From: Alexandru Ardelean
Date: Mon, 8 Apr 2019 12:01:36 +0300
> This set adds a minimal set of ethtool hooks to the driver, which provide a
> decent amount of link information via ethtool.
> With this change, running `ethtool ethX` in user-space provides all the
> neatly-formatted information abo
From: Alexandru Ardelean
Date: Mon, 8 Apr 2019 12:01:57 +0300
> This hook only implements a minimal set of ioctl hooks to be able to access
> MII regs by using phytool.
> When using this simple MAC controller, it's pretty difficult to do
> debugging of the PHY chip without checking MII regs.
>
>
Update test_tc_tunnel to verify adding inner L2 header
encapsulation (an MPLS label or ethernet header) works.
Signed-off-by: Alan Maguire
---
tools/testing/selftests/bpf/progs/test_tc_tunnel.c | 232 ++---
tools/testing/selftests/bpf/test_tc_tunnel.sh | 121 +++
2 f
commit 868d523535c2 ("bpf: add bpf_skb_adjust_room encap flags")
introduced support to bpf_skb_adjust_room for GSO-friendly GRE
and UDP encapsulation.
For GSO to work for skbs, the inner headers (mac and network) need to
be marked. For L3 encapsulation using bpf_skb_adjust_room, the mac
and netwo
Extend bpf_skb_adjust_room growth to mark inner MAC header so that
L2 encapsulation can be used for tc tunnels.
Patch #1 extends the existing test_tc_tunnel to support UDP
encapsulation; later we want to be able to test MPLS over UDP and
MPLS over GRE encapsulation.
Patch #2 adds the BPF_F_ADJ_RO
commit 868d523535c2 ("bpf: add bpf_skb_adjust_room encap flags")
introduced support to bpf_skb_adjust_room for GSO-friendly GRE
and UDP encapsulation and later introduced associated test_tc_tunnel
tests. Here those tests are extended to cover UDP encapsulation also.
Signed-off-by: Alan Maguire
-
From: Paolo Abeni
Date: Mon, 8 Apr 2019 10:15:59 +0200
> After commit a297569fe00a ("net/udp: do not touch skb->peeked unless
> really needed") the 'peeked' argument of __skb_try_recv_datagram()
> and friends is always equal to !!'flags & MSG_PEEK'.
>
> Since such argument is really a boolean i
Since stats updating is always consistent with TCQ_F_CPUSTATS flag,
we can disable it at qdisc creation time flipping such bit.
In my experiments, if the NOLOCK flag is cleared, per CPU stats
accounting does not give any measurable performance gain, but it
waste some memory.
Let's clear TCQ_F_CPU
This revert commit 46b1c18f9deb ("net: sched: put back q.qlen
into a single location").
After the previous patch nobody accesses directly qlen for a child
qdisc when such qdisc does per CPU stats accounting.
In the control path nobody uses directly qlen since commit
677f1bc207c ("net: sched: introd
When checking for root qdisc queue length, do not access directly q.qlen.
In the following patches we will move back qlen accounting to per CPU
values for NOLOCK qdiscs.
Instead, prefer the qdisc_is_empty() helper usage.
Signed-off-by: Paolo Abeni
---
include/net/sch_generic.h | 2 +-
1 file ch
The core sched implementation checks independently for NOLOCK flag
to acquire/release the root spin lock and for qdisc_is_percpu_stats()
to account per CPU values in many places.
This change update the last few places checking the TCQ_F_NOLOCK to
do per CPU stats accounting according to qdisc_is_p
The commit 46b1c18f9deb ("net: sched: put back q.qlen into a single location")
introduced some measurable regression in the contended scenarios for
lock qdisc.
As Eric suggested we could replace q.qlen access with calls to qdisc_is_empty()
in the datapath and revert the above commit. The TC subsys
Such helper does not cope correctly with NOLOCK qdiscs.
In the following patches we will move back qlen to per CPU
values for such qdiscs, so qdisc_qlen_sum() is not an option,
too.
Instead, use qlen only for lock qdiscs, and always set
flow off for NOLOCK qdiscs with a not empty tx queue.
Signed-
Thank you Peter!
I tried the patch on 4.9.167 & 4.19.32. It's out of sync with upstream.
Looks like a little different work needed for each LTS kernel.
Is someone is familiar with it, and is available to patch it?
If not, I'd be happy to do this and propose a patch for at least 4.9,
4.14, 4.19.
I
On Wed, Apr 3, 2019 at 8:13 PM brakmo wrote:
>
> When a packet is dropped when calling queue_xmit in __tcp_transmit_skb
> and packets_out is 0, it is beneficial to set a small probe timer.
> Otherwise, the throughput for the flow can suffer because it may need to
> depend on the probe timer to st
When moving the documentation for the ieee802154 subsystem from
plain text to rst the file pattern in the MAINTAINERS file got wrong.
Updating it here to fix scripts using this file.
Reported-by: Joe Perches
Signed-off-by: Stefan Schmidt
---
MAINTAINERS | 2 +-
1 file changed, 1 insertion(+), 1
Simple test that sets cb to {1,2,3,4,5} and priority to 6, runs bpf
program that fails if cb is not what we expect and increments cb[i] and
priority. When the test finishes, we check that cb is now {2,3,4,5,6}
and priority is 7.
We also test the sanity checks:
* ctx_in is provided, but ctx_size_in
Support recently introduced input/output context for test runs.
We extend only bpf_prog_test_run_xattr. bpf_prog_test_run is
unextendable and left as is.
Cc: Martin Lau
Signed-off-by: Stanislav Fomichev
---
tools/include/uapi/linux/bpf.h | 7 +++
tools/lib/bpf/bpf.c| 5 +
to
Add new set of arguments to bpf_attr for BPF_PROG_TEST_RUN:
* ctx_in/ctx_size_in - input context
* ctx_out/ctx_size_out - output context
The intended use case is to pass some meta data to the test runs that
operate on skb (this has being brought up on recent LPC).
For programs that use bpf_prog_t
On Mon, Apr 8, 2019 at 8:51 AM Sasha Levin wrote:
>
> On Mon, Apr 08, 2019 at 08:49:52AM -0600, Captain Wiggum wrote:
> >Hi Sasha,
> >
> >This patch cannot be applied to upstream, the code is significantly
> >different.
> >Therefore, this un-do patch would not be seen in the upstream git log.
> >
On Mon, Apr 08, 2019 at 08:49:52AM -0600, Captain Wiggum wrote:
Hi Sasha,
This patch cannot be applied to upstream, the code is significantly different.
Therefore, this un-do patch would not be seen in the upstream git log.
It was solved there by coding a better solution, not by the un-do patch.
1 - 100 of 142 matches
Mail list logo
