Hello Stefan,
Am 06.12.2015 um 21:56 schrieb Stefan Priebe:
> i think i found the issue in 4.1 with netlink. Somebody made a mistake
> while backporting or cherry-picking your patch "netlink: Fix autobind
> race condition that leads to zero port ID" to 4.1.
>
> It misses a goto in 4.1.
>
> This
On 2015/12/6 6:19, Arnd Bergmann wrote:
> On Saturday 05 December 2015 14:10:56 yankejian wrote:
>> diff --git a/Documentation/devicetree/bindings/net/hisilicon-hns-dsaf.txt
>> b/Documentation/devicetree/bindings/net/hisilicon-hns-dsaf.txt
>> index 80411b2..ecacfa4 100644
>> --- a/Documentation/
Hi Herbert,
Am 07.12.2015 um 02:20 schrieb Herbert Xu:
> On Sun, Dec 06, 2015 at 09:56:34PM +0100, Stefan Priebe wrote:
>> Hi Herbert,
>>
>> i think i found the issue in 4.1 with netlink. Somebody made a
>> mistake while backporting or cherry-picking your patch "netlink: Fix
>> autobind race condi
On Sun, 6 Dec 2015, David Miller wrote:
> From: Julia Lawall
> Date: Sun, 6 Dec 2015 17:24:57 +0100
>
> > The hnae_buf_ops and hnae_ae_ops structures are never modified, so declare
> > them as const.
> >
> > Done with the help of Coccinelle.
> >
> > Signed-off-by: Julia Lawall
>
> This do
I can confirm that MPLS packets are sent correctly if the conditional
at mpls_iptunnel.c:57-65 is removed and just replaced with lines
58-59:
ttl = ip_hdr(skb)->ttl;
rt = (struct rtable *)dst;
Obviously this isn't sufficient for IPv6, but is a workaround that
enables MPLS-encapsulated packets to
From: Sunil Goutham
Since we have moved on to using allocated pages to carve receive
buffers instead of netdev_alloc_skb() there is no need to store
any pointers for later retrieval. Earlier we had to store
skb and skb->data pointers which later are used to handover
received packet to network sta
From: Yury Norov
The same switch-case repeates for nivc_*_intr functions.
In this patch it is moved to a helper nicvf_int_type_to_mask().
By the way:
- Unneeded write to NICVF register dropped if int_type is unknown.
- netdev_dbg() is used instead of netdev_err().
Signed-off-by: Yury Norov
S
From: Sunil Goutham
This patch series contains contains couple of cleanup patches.
Sunil Goutham (1):
net, thunderx: Remove unnecessary rcv buffer start address management
Yury Norov (1):
net: thunderx: nicvf_queues: nivc_*_intr: remove duplication
drivers/net/ethernet/cavium/thunder/nicv
From: Rainer Weikusat
Date: Sun, 06 Dec 2015 21:11:38 +
> The current unix_dgram_recvsmg code acquires the u->readlock mutex in
> order to protect access to the peek offset prior to calling
> __skb_recv_datagram for actually receiving data. This implies that a
> blocking reader will go to sle
From: Rainer Weikusat
Date: Sun, 06 Dec 2015 21:11:34 +
> The __skb_recv_datagram routine in core/ datagram.c provides a general
> skb reception factility supposed to be utilized by protocol modules
> providing datagram sockets. It encompasses both the actual recvmsg code
> and a surrounding
From: Andrew Lunn
Date: Mon, 7 Dec 2015 04:38:58 +0100
> Device tree properties for a phy device are expected to be in the phy
> node. The current code for the DP83867 also tries to look in the
> parent node. The devices binding documentation does not mention this,
> no current device tree file
From: Phil Reid
Date: Mon, 7 Dec 2015 09:38:40 +0800
> @@ -116,6 +118,9 @@ static int socfpga_dwmac_parse_data(struct socfpga_dwmac
> *dwmac, struct device *
> return -EINVAL;
> }
>
> + dwmac->f2h_ptp_ref_clk = of_property_read_bool(np, "f2h_ptp_ref_clk");
> +
> +
Too
From: Bjørn Mork
Date: Sun, 6 Dec 2015 22:47:15 +0100
> Adding a writable sysfs attribute for the "NDP to end"
> quirk flag.
>
> This makes it easier for end users to test new devices for
> this firmware bug. We've been lucky so far, but we should
> not depend on reporters capable of rebuildin
Le 06/12/2015 19:38, Andrew Lunn a écrit :
> Device tree properties for a phy device are expected to be in the phy
> node. The current code for the DP83867 also tries to look in the
> parent node. The devices binding documentation does not mention this,
> no current device tree file makes use of th
From: Julia Lawall
Date: Sun, 6 Dec 2015 17:24:57 +0100
> The hnae_buf_ops and hnae_ae_ops structures are never modified, so declare
> them as const.
>
> Done with the help of Coccinelle.
>
> Signed-off-by: Julia Lawall
This doesn't apply to net-next.
> I have the impression that this drive
From: Or Gerlitz
Date: Sun, 6 Dec 2015 18:07:37 +0200
> This series is built upon the code added in commit ce388ff "Merge branch
> 'mlx4-next'" which added HA and LAG support to mlx4 RoCE and SRIOV services.
>
> We add HA and Link Aggregation support to single ported mlx4 Ethernet VFs.
Series
From: Julia Lawall
Date: Sun, 6 Dec 2015 06:56:23 +0100 (CET)
> Remove unneeded variable used to store return value.
>
> Generated by: scripts/coccinelle/misc/returnvar.cocci
>
> CC: Asias He
> Signed-off-by: Fengguang Wu
> Signed-off-by: Julia Lawall
Applied to net-next, thanks.
--
To unsu
Device tree properties for a phy device are expected to be in the phy
node. The current code for the DP83867 also tries to look in the
parent node. The devices binding documentation does not mention this,
no current device tree file makes use of this, and it is not behaviour
we want. So remove look
From: Konstantin Shemyak
Date: Sat, 5 Dec 2015 22:55:13 +0200
> The DF bit on IPv4 GRE tunnels with fixed TTL is enforced in order
> to prevent networking loops. This was the original behavior of GRE
> tunnels, lost in the refactoring.
>
> Signed-off-by: Konstantin Shemyak
Like Pravin said, ip
From: Felix Fietkau
Date: Sat, 5 Dec 2015 13:58:11 +0100
> This semicolon causes a build error if the function call is wrapped in
> parentheses.
>
> Fixes: aabc92bbe3cf ("net: add __netdev_alloc_pcpu_stats() to indicate gfp
> flags")
> Reported-by: Imre Kaloz
> Signed-off-by: Felix Fietkau
On Sun, 2015-12-06 at 22:29 -0500, David Miller wrote:
> From: yankejian
> Date: Sat, 5 Dec 2015 15:32:29 +0800
>
> > +#if (PAGE_SIZE < 8192)
> > + if (hnae_buf_size(ring) == HNS_BUFFER_SIZE_2048) {
> > + truesize = hnae_buf_size(ring);
> > + } else {
> > + truesiz
From: Xin Long
Date: Sat, 5 Dec 2015 15:35:36 +0800
> when A sends a data to B, then A close() and enter into SHUTDOWN_PENDING
> state, if B neither claim his rwnd is 0 nor send SACK for this data, A
> will keep retransmitting this data until t5 timeout, Max.Retrans times
> can't work anymore, w
From: Bjørn Mork
Date: Sun, 6 Dec 2015 21:25:50 +0100
> The notifier calls were thrown in as a last-minute fix for an
> imagined "this device could be part of a bridge" problem. That
> revealed a certain lack of locking. Not to mention testing...
>
> Avoid this splat:
...
> Fixes: 32f7adf633b
From: yankejian
Date: Sat, 5 Dec 2015 15:32:29 +0800
> +#if (PAGE_SIZE < 8192)
> + if (hnae_buf_size(ring) == HNS_BUFFER_SIZE_2048) {
> + truesize = hnae_buf_size(ring);
> + } else {
> + truesize = ALIGN(size, L1_CACHE_BYTES);
> + last_offset = hnae_pag
On 12/6/15 6:20 PM, Alexander Duyck wrote:
That works for Linux to Linux, but what about the cases where you have
a non-Linux endpoint on the other end such as something like a Cisco
switch?
Why does is matter what kind of switch the NIC is connected to?
--
To unsubscribe from this list: send t
On 12/06/2015 06:07 PM, Peter Wu wrote:
Free skb for received frames with a wrong checksum. This can happen
pretty rapidly, exhausting all memory.
This fixes a memleak (detected with kmemleak). Originally found while
using monitor mode, but it also appears during managed mode (once the
link is u
On Sun, 2015-12-06 at 13:03 -0800, Eric Dumazet wrote:
> But then when later we promote a skb->dst to a refctounted one
> (skb_dst_force(), we might make sure we abort the operation if __refcnt
> == 0 ( and DST_NOCACHE is in dst->flags)
>
Minimum patch would be :
diff --git a/include/net/dst.h
On Sun, Dec 06, 2015 at 06:56:23AM +0100, Julia Lawall wrote:
> Remove unneeded variable used to store return value.
>
> Generated by: scripts/coccinelle/misc/returnvar.cocci
>
> CC: Asias He
> Signed-off-by: Fengguang Wu
> Signed-off-by: Julia Lawall
>
> ---
>
> vsock.c |3 +--
> 1 fil
The DSA driver needs to be passed a reference to an mdio bus. Typically
the mac is configured to use a fixed link but the mdio bus still needs
to be registered so that it con configure the switch.
This patch follows the same process as the altera tse ethernet driver for
creation of the mdio bus.
A
stmmac_config_sub_second_increment set the sub second increment to 20ns.
Driver is configured to use the fine adjustment method where the sub second
register is incremented when the acculumator incremented by the addend
register wraps overflows. This accumulator is update on every ptp clk
cycle. If
Signed-off-by: Phil Reid
---
Documentation/devicetree/bindings/net/socfpga-dwmac.txt | 2 ++
drivers/net/ethernet/stmicro/stmmac/dwmac-socfpga.c | 10 ++
2 files changed, 12 insertions(+)
diff --git a/Documentation/devicetree/bindings/net/socfpga-dwmac.txt
b/Documentation/devicetre
devm_get_clk looks in clock-name property for matching clock.
the ptp_ref_clk property is ignored.
Acked-by: Rob Herring
Signed-off-by: Phil Reid
---
Documentation/devicetree/bindings/net/stmmac.txt | 17 -
1 file changed, 8 insertions(+), 9 deletions(-)
diff --git a/Documentat
This adds a debugfs entry to view the current status of the ptp
registers.
Signed-off-by: Phil Reid
---
drivers/net/ethernet/stmicro/stmmac/stmmac.h | 1 +
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 61 +++
drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h | 5
Signed-off-by: Phil Reid
---
Documentation/devicetree/bindings/net/socfpga-dwmac.txt | 2 ++
drivers/net/ethernet/stmicro/stmmac/dwmac-socfpga.c | 9 +
2 files changed, 11 insertions(+)
diff --git a/Documentation/devicetree/bindings/net/socfpga-dwmac.txt
b/Documentation/devicetree/b
Provide ability to specify a fixed phy in the device tree and
retain the mdio bus if no phy is found. This is needed where
a dsa is connected via a fixed phy and uses the mdio bus for config.
Fixed ptp ref clock calculatins for the stmmac when ptp ref clock
is running at <= 50Mhz. Also add device
On 2015/12/6 6:15, Arnd Bergmann wrote:
> On Saturday 05 December 2015 15:56:57 yankejian wrote:
>> #size-cells = <0>;
>> compatible = "hisilicon,hns-mdio";
>> - reg = <0x0 0x803c 0x0 0x1
>> - 0x0 0x8000 0x0 0x1>;
On 2015/12/6 6:13, Arnd Bergmann wrote:
> On Saturday 05 December 2015 15:56:58 yankejian wrote:
>> +Required properties:
>> +- compatible : "hisilicon,peri-c-subctrl", "syscon";
>> +- reg : Register address and size
>> +
>> +The HiP05 PERISUB system controller is shared by peripheral controllers
On Sun, Dec 06, 2015 at 09:56:34PM +0100, Stefan Priebe wrote:
> Hi Herbert,
>
> i think i found the issue in 4.1 with netlink. Somebody made a
> mistake while backporting or cherry-picking your patch "netlink: Fix
> autobind race condition that leads to zero port ID" to 4.1.
>
> It misses a goto
On Sun, Dec 6, 2015 at 1:30 PM, Tom Herbert wrote:
> On Sun, Dec 6, 2015 at 10:44 AM, Alexander Duyck
> wrote:
>> On Sun, Dec 6, 2015 at 8:31 AM, Tom Herbert wrote:
The only spot I think you and I disagreed on was the approach. I
don't know if the hard push back does anything but puni
Free skb for received frames with a wrong checksum. This can happen
pretty rapidly, exhausting all memory.
This fixes a memleak (detected with kmemleak). Originally found while
using monitor mode, but it also appears during managed mode (once the
link is up).
Cc: sta...@vger.kernel.org
Signed-off
On Tue, Nov 17, 2015 at 2:40 PM, David Miller wrote:
> From: Sabrina Dubroca
> Date: Mon, 16 Nov 2015 22:54:20 +0100
>
>> Reset pskb in macvlan_handle_frame in case skb_share_check returned a
>> clone.
>>
>> Fixes: 8a4eb5734e8d ("net: introduce rx_handler results and logic around
>> that")
>> Si
On Sun, 2015-12-06 at 22:56 +0100, Julia Lawall wrote:
> Remove unneeded semicolon.
>
> Generated by: scripts/coccinelle/misc/semicolon.cocci
There are a lot more of these
arch/arm/mach-ixp4xx/goramo_mlr.c:78:static u32 hw_bits = 0xFFFD;/*
assume all hardware present */;
crypto/tgr192.c
On 12/06/2015 03:39 PM, Peter Wu wrote:
On Sun, Dec 06, 2015 at 02:18:36PM -0600, Larry Finger wrote:
On 12/06/2015 11:57 AM, Peter Wu wrote:
Free skb for received frames with a wrong checksum.
While using the rtl8192cu driver in monitor mode, somehow 5G of memory
was permanently lost (observa
Remove unneeded semicolon.
Generated by: scripts/coccinelle/misc/semicolon.cocci
Signed-off-by: Fengguang Wu
Signed-off-by: Julia Lawall
---
init.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/drivers/net/wireless/marvell/mwifiex/init.c
+++ b/drivers/net/wireless/marvell/mw
Adding a writable sysfs attribute for the "NDP to end"
quirk flag.
This makes it easier for end users to test new devices for
this firmware bug. We've been lucky so far, but we should
not depend on reporters capable of rebuilding the driver.
Cc: Enrico Mioso
Signed-off-by: Bjørn Mork
---
Docu
On Sun, Dec 06, 2015 at 02:18:36PM -0600, Larry Finger wrote:
> On 12/06/2015 11:57 AM, Peter Wu wrote:
> >Free skb for received frames with a wrong checksum.
> >
> >While using the rtl8192cu driver in monitor mode, somehow 5G of memory
> >was permanently lost (observable via the Available column i
Hi Anjali,
[auto build test ERROR on net/master]
[also build test ERROR on v4.4-rc3 next-20151203]
url:
https://github.com/0day-ci/linux/commits/Anjali-Singhai-Jain/net-Generalize-udp-based-tunnel-offload/20151124-090500
config: x86_64-randconfig-s0-12070507 (attached as .config)
reproduce:
On Sun, Dec 6, 2015 at 10:44 AM, Alexander Duyck
wrote:
> On Sun, Dec 6, 2015 at 8:31 AM, Tom Herbert wrote:
>>> The only spot I think you and I disagreed on was the approach. I
>>> don't know if the hard push back does anything but punish the users by
>>> delaying the time needed to find a reas
The current unix_dgram_recvsmg code acquires the u->readlock mutex in
order to protect access to the peek offset prior to calling
__skb_recv_datagram for actually receiving data. This implies that a
blocking reader will go to sleep with this mutex held if there's
presently no data to return to user
The __skb_recv_datagram routine in core/ datagram.c provides a general
skb reception factility supposed to be utilized by protocol modules
providing datagram sockets. It encompasses both the actual recvmsg code
and a surrounding 'sleep until data is available' loop. This is
inconvenient if a protoc
On Sun, 2015-12-06 at 13:26 -0500, dwil...@us.ibm.com wrote:
> Hi-
>
> I am seeing a crash on a distro V4.2.3 kernel caused by a double
> release of a dst_entry. In ipv4_dst_destroy() the call to
> list_empty() finds a poisoned next pointer, indicating the dst_entry
> has already been remov
Hi Herbert,
i think i found the issue in 4.1 with netlink. Somebody made a mistake
while backporting or cherry-picking your patch "netlink: Fix autobind
race condition that leads to zero port ID" to 4.1.
It misses a goto in 4.1.
This goto is missing in 4.1:
diff --git a/net/netlink/af_netli
The notifier calls were thrown in as a last-minute fix for an
imagined "this device could be part of a bridge" problem. That
revealed a certain lack of locking. Not to mention testing...
Avoid this splat:
RTNL: assertion failed at net/core/dev.c (1639)
CPU: 0 PID: 4293 Comm: bash Not tainted 4.4
On 12/06/2015 11:57 AM, Peter Wu wrote:
Free skb for received frames with a wrong checksum.
While using the rtl8192cu driver in monitor mode, somehow 5G of memory
was permanently lost (observable via the Available column in `free -m`).
Test scenario:
ip link set down wlan1
iw wlan1 s
On zo, 2015-12-06 at 16:29 +0100, Tilman Schmidt wrote:
> So the solution might be as simple as moving the kfree() call from
> gigaset_freecshw() to gigaset_device_release(). Something like this:
>
> --- a/drivers/isdn/gigaset/ser-gigaset.c
> +++ b/drivers/isdn/gigaset/ser-gigaset.c
> @@ -370,19 +
Hello.
On 12/05/2015 01:01 PM, Yoshihiro Kaneko wrote:
Thanks for your review.
From now on, it'll be my duty. :-)
2015-12-04 6:09 GMT+09:00 Sergei Shtylyov :
Hello.
On 12/01/2015 08:04 PM, Yoshihiro Kaneko wrote:
From: Kazuya Mizuguchi
This patch makes PTP support active in CONFIG
Hello.
On 12/06/2015 02:42 PM, Yoshihiro Kaneko wrote:
From: Kazuya Mizuguchi
AVB-DMAC Reception Warning interrupt is not enabled, so it is not
necessary to clear the interrupt.
Signed-off-by: Kazuya Mizuguchi
Signed-off-by: Yoshihiro Kaneko
In principle I agree but perhaps we should
On Sun, Dec 6, 2015 at 8:31 AM, Tom Herbert wrote:
>> The only spot I think you and I disagreed on was the approach. I
>> don't know if the hard push back does anything but punish the users by
>> delaying the time needed to find a reasonable solution. I really
>> think if we are going to get the
Hi-
I am seeing a crash on a distro V4.2.3 kernel caused by a double
release of a dst_entry. In ipv4_dst_destroy() the call to
list_empty() finds a poisoned next pointer, indicating the dst_entry
has already been removed from the list and freed. The crash occurs
18 to 24 hours into a r
From: Xin Long
Date: Sat, 5 Dec 2015 15:15:17 +0800
> When a msg is sent, sctp will hold the chunks of this msg and then try
> to enqueue them. But if the chunks are not enqueued in sctp_outq_tail()
> because of the invalid state, sctp_cmd_interpreter() may still return
> success to sctp_sendmsg
From: Xin Long
Date: Sat, 5 Dec 2015 15:19:27 +0800
> If the chunks are enqueued successfully but sctp_cmd_interpreter()
> return err to sctp_sendmsg() (mainly because of no mem), the chunks will
> get re-queued, but we are dropping the reference and freeing them.
>
> The fix is to just drop th
Free skb for received frames with a wrong checksum.
While using the rtl8192cu driver in monitor mode, somehow 5G of memory
was permanently lost (observable via the Available column in `free -m`).
Test scenario:
ip link set down wlan1
iw wlan1 set type monitor
ip link set up wlan1
From: Stefan Wahren
Date: Fri, 4 Dec 2015 16:29:10 +0100
> In case of a tx queue timeout every transmit is blocked until the
> QCA7000 resets himself and triggers a sync which makes the driver
> flushs the tx ring. So avoid this blocking situation by triggering
> the sync immediately after the t
The hnae_buf_ops and hnae_ae_ops structures are never modified, so declare
them as const.
Done with the help of Coccinelle.
Signed-off-by: Julia Lawall
---
I have the impression that this driver is not recorded properly in the
MAINTAINERS file.
drivers/net/ethernet/hisilicon/hns/hnae.c
> The only spot I think you and I disagreed on was the approach. I
> don't know if the hard push back does anything but punish the users by
> delaying the time needed to find a reasonable solution. I really
> think if we are going to get the hardware vendors to change their
> behavior we have to
From: Jeff Kirsher
Date: Sat, 5 Dec 2015 23:13:31 -0800
> On Sat, Dec 5, 2015 at 2:50 PM, David Miller wrote:
>> From: Anjali Singhai Jain
>> Date: Thu, 3 Dec 2015 23:49:31 -0800
>>
>>> This reverts commit 8fe269991aece394a7ed274f525d96c73f94109a.
>>> The case where VXLAN is a module and i40e
From: Jeff Kirsher
Date: Sun, 6 Dec 2015 00:01:19 -0800
> This series contains updates to fm10k only.
Pulled, thanks Jeff.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/ma
In HA mode, the link state for VFs for which the policy is "auto"
(i.e. follow the physical link state) should be ORed from both ports.
Signed-off-by: Or Gerlitz
Reviewed-by: Jack Morgenstein
---
drivers/net/ethernet/mellanox/mlx4/fw.c | 12 +++-
drivers/net/ethernet/mellanox/mlx4/fw.h
Under HA mode, it's possible that the VF registered its GID
(and expects to get mads through the PV scheme) on a port which is
different from the one this mad arrived on, due to HA fail over.
Therefore, if the gid is not matched on the port that the packet arrived
on, check for a match on the othe
From: Moni Shoua
Due to HW limitations, indexes to MAC and VLAN tables are always taken
from the table of the actual port. So, if a resource holds an index to
a table, it may refer to different values during the lifetime of the
resource, unless the tables are mirrored. Also, even when
driver is
From: Moni Shoua
When the mlx4 driver runs in HA mode, and all VFs are single ported
ones, we make their single port Highly-Available.
This is done by taking advantage of the HA mode properties (following
bonding changes with programming the port V2P map, etc) and adding
the missing parts which
From: Moni Shoua
Under HA mode, steering rules set by VFs should be mirrored on both
ports of the device so packets will be accepted no matter on which
port they arrived.
Since getting into HA mode is done dynamically when the user bonds mlx4
Ethernet netdevs, we keep hold of the VF DMFS rule mb
From: Moni Shoua
Under HA mode, the link down event should be sent to VFs only if both
ports are down.
Signed-off-by: Moni Shoua
Reviewed-by: Jack Morgenstein
Signed-off-by: Or Gerlitz
---
drivers/net/ethernet/mellanox/mlx4/eq.c | 24
1 file changed, 20 insertions(+)
Hi Dave,
This series is built upon the code added in commit ce388ff "Merge branch
'mlx4-next'" which added HA and LAG support to mlx4 RoCE and SRIOV services.
We add HA and Link Aggregation support to single ported mlx4 Ethernet VFs.
In this case, the PF Ethernet interfaces are bonded, the VFs
Am 06.12.2015 um 14:31 schrieb Paul Bolle:
> On wo, 2015-12-02 at 18:48 -0500, Peter Hurley wrote:
>> On 11/30/2015 01:01 PM, Paul Bolle wrote:
>>> --- a/drivers/isdn/gigaset/ser-gigaset.c
>>> +++ b/drivers/isdn/gigaset/ser-gigaset.c
>>> @@ -42,8 +42,9 @@ MODULE_PARM_DESC(cidmode, "stay in CID mod
On wo, 2015-12-02 at 18:48 -0500, Peter Hurley wrote:
> On 11/30/2015 01:01 PM, Paul Bolle wrote:
> > Should (something like) this go into stable too?
>
> Definitely for stable since it has a userspace triggerable component.
Thanks, will do.
> > --- a/drivers/isdn/gigaset/ser-gigaset.c
> > +++ b
From: Kazuya Mizuguchi
AVB-DMAC Reception Warning interrupt is not enabled, so it is not
necessary to clear the interrupt.
Signed-off-by: Kazuya Mizuguchi
Signed-off-by: Yoshihiro Kaneko
---
This patch is based on the master branch of David Miller's networking
tree.
drivers/net/ethernet/ren
tl;dr mpls_output expects skb->protocol to be set to correct
ethertype, but it isn't
https://github.com/torvalds/linux/blob/ede2059dbaf9c6557a49d466c8c7778343b208ff/net/mpls/mpls_iptunnel.c#L64
Problem:
I set up two interfaces pointed at each other, and added a static arp
entry to minimise compl
Hi,
I'm running a few Strongswan 5.3.* CentOS (Kernel 3.16.7, 4.2.6, 4.1.*)
instances on AWS to terminate VPNs between each other and/or to other
devices across the Internet.
While investigating some application issues, I've noticed that on every VPNs
I have random packet losses (from 1% to 4% ove
From: Jacob Keller
Originally this statistic was renamed because the method of dropping was
called "drop_oversized_messages", but this logic has changed much, and
this counter does actually represent messages which we failed to
transmit for a number of reasons. Rename the counter back to tx_dropp
From: Jacob Keller
Signed-off-by: Jacob Keller
Tested-by: Krishneil Singh
Signed-off-by: Jeff Kirsher
---
drivers/net/ethernet/intel/fm10k/fm10k_main.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
b/drivers/net/ethern
From: Jacob Keller
Signed-off-by: Jacob Keller
Tested-by: Krishneil Singh
Signed-off-by: Jeff Kirsher
---
drivers/net/ethernet/intel/fm10k/fm10k_pf.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pf.c
b/drivers/net/ethernet/i
From: Jacob Keller
A previous bug was uncovered by addition of a debug stat to indicate the
actual number of DWORDS we pulled from the mbmem. It turned out this was
not the same as the tx_dwords counter. While the previous bug fix should
have corrected this in all cases, add some debug stats that
From: Jacob Keller
The existing adaptive ITR algorithm is overly restrictive. It throttles
incorrectly for various traffic rates, and does not produce good
performance. The algorithm now allows for more interrupts per second,
and does some calculation to help improve for smaller packet loads. In
From: Jacob Keller
The Intel Ethernet Switch FM1 Host Interface interrupt throttle
timers are based on the PCIe link speed. Because of this, the value
being programmed into the ITR registers must be scaled accordingly.
For the PF, this is as simple as reading the PCIe link speed and storing
From: Jacob Keller
The init_hw function may fail, and in the case of VFs, it might change
the number of maximum queues available. Thus, for every flow which
checks init_hw, we need to ensure that we clear the queue scheme before,
and initialize it after. The fm10k_io_slot_reset path will end up
t
From: Jacob Keller
Define a macro for identifying when the itr value is dynamic or
adaptive. The concept was taken from i40e. This helps make clear what
the check is, and reduces the line length to something more reasonable
in a few places.
Signed-off-by: Jacob Keller
Reviewed-by: Bruce Allan
From: Jacob Keller
Since the resultant data type of the mac_update.mac_upper field is u16,
it does not make sense to typecast u8 variables to u32 first. Since
we're modifying fm10k_pf.c, also update the copyright year.
Reported-by: Matthew Vick
Signed-off-by: Jacob Keller
Reviewed-by: Bruce Al
From: Jacob Keller
Signed-off-by: Jacob Keller
Tested-by: Krishneil Singh
Signed-off-by: Jeff Kirsher
---
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
b/drivers/net/ethernet/i
From: Jacob Keller
Signed-off-by: Jacob Keller
Reviewed-by: Bruce Allan
Tested-by: Krishneil Singh
Signed-off-by: Jeff Kirsher
---
drivers/net/ethernet/intel/fm10k/fm10k.h | 2 ++
drivers/net/ethernet/intel/fm10k/fm10k_ethtool.c | 4 ++--
drivers/net/ethernet/intel/fm10k/fm10k_pci.c
From: Jacob Keller
The current default ITR for Tx is overly restrictive. Using a simple
netperf TCP_STREAM test, we top out at about 10Gb/s for a single thread
when running using 1500 byte frames. By reducing the ITR value to 25usec
(up to 40K interrupts a second from 10K), we are able to achieve
This series contains updates to fm10k only.
Jacob provides the remaining fm10k patches in the series. First change
ensures that all the logic regarding the setting of netdev features is
consolidated in one place of the driver. Fixed an issue where an assumption
was being made on how many queues
From: Jacob Keller
A recent change modified init_hw in some flows the function may fail on
VF devices. For example, if a VF doesn't yet own its own queues.
However, many callers of init_hw didn't bother to check the error code.
Other callers checked but only displayed diagnostic messages without
From: Jacob Keller
Don't change netdev hw_features later in fm10k_probe, instead set all
values inside fm10k_alloc_netdev. To do so, we need to know the MAC type
(whether it is PF or VF) in order to determine what to do. This helps
ensure that all logic regarding features is co-located.
Signed-o
From: Jacob Keller
VF drivers must detect how many queues are available. Previously, the
driver assumed that each VF has at minimum 1 queue. This assumption is
incorrect, since it is possible that the PF has not yet assigned the
queues to the VF by the time the VF checks. To resolve this, we adde
95 matches
Mail list logo
