Commits · 051ba67447de1294aaacc59752a37f72091d15ed · Kirill Smelkov / linux

11 Sep, 2019 40 commits

tcp: force a PSH flag on TSO packets · 051ba674

Eric Dumazet authored Sep 10, 2019

When tcp sends a TSO packet, adding a PSH flag on it
reduces the sojourn time of GRO packet in GRO receivers.

This is particularly the case under pressure, since RX queues
receive packets for many concurrent flows.

A sender can give a hint to GRO engines when it is
appropriate to flush a super-packet, especially when pacing
is in the picture, since next packet is probably delayed by
one ms.

Having less packets in GRO engine reduces chance
of LRU eviction or inflated RTT, and reduces GRO cost.

We found recently that we must not set the PSH flag on
individual full-size MSS segments [1] :

 Under pressure (CWR state), we better let the packet sit
 for a small delay (depending on NAPI logic) so that the
 ACK packet is delayed, and thus next packet we send is
 also delayed a bit. Eventually the bottleneck queue can
 be drained. DCTCP flows with CWND=1 have demonstrated
 the issue.

This patch allows to slowdown the aggregate traffic without
involving high resolution timers on senders and/or
receivers.

It has been used at Google for about four years,
and has been discussed at various networking conferences.

[1] segments smaller than MSS already have PSH flag set
    by tcp_sendmsg() / tcp_mark_push(), unless MSG_MORE
    has been requested by the user.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Soheil Hassas Yeganeh <soheil@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Tariq Toukan <tariqt@mellanox.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

051ba674

ipv6: Don't use dst gateway directly in ip6_confirm_neigh() · cbfd6891

Stefano Brivio authored Sep 09, 2019

This is the equivalent of commit 2c6b55f4 ("ipv6: fix neighbour
resolution with raw socket") for ip6_confirm_neigh(): we can send a
packet with MSG_CONFIRM on a raw socket for a connected route, so the
gateway would be :: here, and we should pick the next hop using
rt6_nexthop() instead.

This was found by code review and, to the best of my knowledge, doesn't
actually fix a practical issue: the destination address from the packet
is not considered while confirming a neighbour, as ip6_confirm_neigh()
calls choose_neigh_daddr() without passing the packet, so there are no
similar issues as the one fixed by said commit.

A possible source of issues with the existing implementation might come
from the fact that, if we have a cached dst, we won't consider it,
while rt6_nexthop() takes care of that. I might just not be creative
enough to find a practical problem here: the only way to affect this
with cached routes is to have one coming from an ICMPv6 redirect, but
if the next hop is a directly connected host, there should be no
topology for which a redirect applies here, and tests with redirected
routes show no differences for MSG_CONFIRM (and MSG_PROBE) packets on
raw sockets destined to a directly connected host.

However, directly using the dst gateway here is not consistent anymore
with neighbour resolution, and, in general, as we want the next hop,
using rt6_nexthop() looks like the only sane way to fetch it.
Reported-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Acked-by: Guillaume Nault <gnault@redhat.com>
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

cbfd6891

net: stmmac: pci: Add HAPS support using GMAC5 · ebecb860

Jose Abreu authored Sep 09, 2019

Add the support for Synopsys HAPS board that uses GMAC5.
Signed-off-by: Jose Abreu <joabreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ebecb860

net: phy: dp83867: Add SGMII mode type switching · 507ddd5c

Vitaly Gaiduk authored Sep 09, 2019

This patch adds ability to switch beetween two PHY SGMII modes.
Some hardware, for example, FPGA IP designs may use 6-wire mode
which enables differential SGMII clock to MAC.
Signed-off-by: Vitaly Gaiduk <vitaly.gaiduk@cloudbear.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>

507ddd5c

net: phy: dp83867: Add documentation for SGMII mode type · a2111c46

Vitaly Gaiduk authored Sep 09, 2019

Add documentation of ti,sgmii-ref-clock-output-enable
which can be used to select SGMII mode type (4 or 6-wire).
Signed-off-by: Vitaly Gaiduk <vitaly.gaiduk@cloudbear.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>

a2111c46

dt-bindings: net: dwmac: document 'mac-mode' property · 9c15d359

Alexandru Ardelean authored Sep 06, 2019

This change documents the 'mac-mode' property that was introduced in the
'stmmac' driver to support passive mode converters that can sit in-between
the MAC & PHY.
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

9c15d359

net: stmmac: implement support for passive mode converters via dt · 0060c878

Alexandru Ardelean authored Sep 06, 2019

In-between the MAC & PHY there can be a mode converter, which converts one
mode to another (e.g. GMII-to-RGMII).

The converter, can be passive (i.e. no driver or OS/SW information
required), so the MAC & PHY need to be configured differently.

For the `stmmac` driver, this is implemented via a `mac-mode` property in
the device-tree, which configures the MAC into a certain mode, and for the
PHY a `phy_interface` field will hold the mode of the PHY. The mode of the
PHY will be passed to the PHY and from there-on it work in a different
mode. If unspecified, the default `phy-mode` will be used for both.
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

0060c878

Merge branch 'qed-Fix-series' · c1609946

David S. Miller authored Sep 11, 2019

Sudarsana Reddy Kalluru says:

====================
qed* Fix series.

The patch series addresses couple of issues in the recent commits.
Patch (1) populates the actual dump-size of config attribute instead of
providing a fixed size value.
Patch(2) updates frame format of flash config buffer as required by
management FW (MFW).

Please consider applying it to net-next.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

c1609946

qed: Fix Config attribute frame format. · 2da244a5

Sudarsana Reddy Kalluru authored Sep 11, 2019

MFW associates the entity id to a config attribute instead of assigning
one entity id for all the config attributes.
This patch incorporates driver changes to link entity id to a config id
attribute.

Fixes: 0dabbe1b ("qed: Add driver API for flashing the config attributes.")
Signed-off-by: Sudarsana Reddy Kalluru <skalluru@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

2da244a5

qed*: Fix size of config attribute dump. · 9e54ba7c

Sudarsana Reddy Kalluru authored Sep 11, 2019

Driver currently returns max-buf-size as size of the config attribute.
This patch incorporates changes to read this value from MFW (if available)
and provide it to the user. Also did a trivial clean up in this path.

Fixes: d44a3ced ("qede: Add support for reading the config id attributes.")
Signed-off-by: Sudarsana Reddy Kalluru <skalluru@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

9e54ba7c

Merge branch 'devlink-unknown' · 2f2fa16e

David S. Miller authored Sep 11, 2019

Simon Horman says:

====================
devlink: add unknown 'fw_load_policy' value

Dirk says:

Recently we added an unknown value for the 'reset_dev_on_drv_probe' devlink
parameter. Extend the 'fw_load_policy' parameter in the same way.

The only driver that uses this right now is the nfp driver.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

2f2fa16e

nfp: devlink: set unknown fw_load_policy · 44798ece

Dirk van der Merwe authored Sep 11, 2019

If the 'app_fw_from_flash' HWinfo key is invalid, set the
'fw_load_policy' devlink parameter value to unknown.
Suggested-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

44798ece

devlink: add unknown 'fw_load_policy' value · 64f658de

Dirk van der Merwe authored Sep 11, 2019

Similar to the 'reset_dev_on_drv_probe' devlink parameter, it is useful
to have an unknown value which can be used by drivers to report that the
hardware value isn't recognized or is otherwise invalid instead of
failing the operation.

This is especially useful for u8/enum parameters.
Suggested-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

64f658de

Merge tag 'mac80211-next-for-davem-2019-09-11' of... · c1b3ddf7

David S. Miller authored Sep 11, 2019

Merge tag 'mac80211-next-for-davem-2019-09-11' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next

Johannes Berg says:

====================
We have a number of changes, but things are settling down:
 * a fix in the new 6 GHz channel support
 * a fix for recent minstrel (rate control) updates
   for an infinite loop
 * handle interface type changes better wrt. management frame
   registrations (for management frames sent to userspace)
 * add in-BSS RX time to survey information
 * handle HW rfkill properly if !CONFIG_RFKILL
 * send deauth on IBSS station expiry, to avoid state mismatches
 * handle deferred crypto tailroom updates in mac80211 better
   when device restart happens
 * fix a spectre-v1 - really a continuation of a previous patch
 * advertise NL80211_CMD_UPDATE_FT_IES as supported if so
 * add some missing parsing in VHT extended NSS support
 * support HE in mac80211_hwsim
 * let mac80211 drivers determine the max MTU themselves
along with the usual cleanups etc.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

c1b3ddf7

cfg80211: Purge frame registrations on iftype change · c1d3ad84

Denis Kenzior authored Aug 28, 2019

Currently frame registrations are not purged, even when changing the
interface type. This can lead to potentially weird situations where
frames possibly not allowed on a given interface type remain registered
due to the type switching happening after registration.

The kernel currently relies on userspace apps to actually purge the
registrations themselves, this is not something that the kernel should
rely on.

Add a call to cfg80211_mlme_purge_registrations() to forcefully remove
any registrations left over prior to switching the iftype.

Cc: stable@vger.kernel.org
Signed-off-by: Denis Kenzior <denkenz@gmail.com>
Link: https://lore.kernel.org/r/20190828211110.15005-1-denkenz@gmail.comSigned-off-by: Johannes Berg <johannes.berg@intel.com>

c1d3ad84

Merge tag 'mlx5-updates-2019-09-10' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux · 747f3cf4

David S. Miller authored Sep 11, 2019

Saeed Mahameed says:

====================
mlx5-updates-2019-09-10

Misc build warnings cleanup for mlx5:

1) Reduce stack usage in FW trace
2) Fix addr's type in mlx5dr_icm_dm
3) Fix rt's type in dr_action_create_reformat_action
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

747f3cf4

Merge branch 'stmmac-next' · 6d47b4bc

David S. Miller authored Sep 11, 2019

Jose Abreu says:

====================
net: stmmac: Improvements for -next

Misc patches for -next. It includes:
 - Two fixes for features in -next only
 - New features support for GMAC cores (which includes GMAC4 and GMAC5)
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

6d47b4bc

net: stmmac: ARP Offload for GMAC4+ Cores · c9b10043