Eric Dumazet writes:
> On 8/25/19 7:52 PM, Cong Wang wrote:
>> On Wed, Aug 21, 2019 at 11:00 PM Akshat Kakkar wrote:
>>>
>>> On Thu, Aug 22, 2019 at 3:37 AM Cong Wang wrote:
> I am using ipset + iptables to classify and not filters. Besides, if
> tc is allowing me to define qdisc -> cl
The xfrm_if(s) on each netns can be deleted when its xfrmi dev is
deleted. xfrmi dev's removal can happen when:
1. Its phydev is being deleted and NETDEV_UNREGISTER event is
processed in xfrmi_event() from my last patch.
2. netns is being removed and all xfrmi devs will be deleted.
3. r
When creating a xfrmi dev, it always holds the phydev. The xfrmi dev should
be deleted when the phydev is being unregistered, so that the phydev can be
put on time. Otherwise the phydev's deleting will get stuck:
# ip link add dummy10 type dummy
# ip link add xfrmi10 type xfrm dev dummy10
#
sysctl net.sctp.ecn_enable is added in this patch. It will allow
users to change the default sctp ecn flag, net.sctp.ecn_enable.
This feature was also required on this thread:
http://lkml.iu.edu/hypermail/linux/kernel/0812.1/01858.html
Signed-off-by: Xin Long
---
net/sctp/sysctl.c | 7 ++
This patchset is to make ecn flag per netns and endpoint and then
add SCTP_ECN_SUPPORTED sockopt, as does for other feature flags.
Xin Long (3):
sctp: make ecn flag per netns and endpoint
sctp: allow users to set netns ecn flag with sysctl
sctp: allow users to set ep ecn flag by sockopt
in
SCTP_ECN_SUPPORTED sockopt will be added to allow users to change
ep ecn flag, and it's similar with other feature flags.
Signed-off-by: Xin Long
---
include/uapi/linux/sctp.h | 1 +
net/sctp/socket.c | 73 +++
2 files changed, 74 insertions(+
This patch is to add ecn flag for both netns_sctp and sctp_endpoint,
net->sctp.ecn_enable is set 1 by default, and ep->ecn_enable will
be initialized with net->sctp.ecn_enable.
asoc->peer.ecn_capable will be set during negotiation only when
ep->ecn_enable is set on both sides.
Signed-off-by: Xin
Tested on Bananapi-R2 (mt7623+mt7530) and Bananapi-R64 v0.1 (mt7622+rtl8367)
Tested-by: Frank Wunderlich
regards Frank
The rds6_inc_info_copy() function has a couple struct members which
are leaking stack information. The ->tos field should hold actual
information and the ->flags field needs to be zeroed out.
Fixes: 3eb450367d08 ("rds: add type of service(tos) infrastructure")
Fixes: b7ff8b1036f0 ("rds: Extend RD
Currently, at xdp_adjust_tail_kern.c, MAX_PCKT_SIZE is limited
to 600. To make this size flexible, a new map 'pcktsz' is added.
By updating new packet size to this map from the userland,
xdp_adjust_tail_kern.o will use this value as a new max_pckt_size.
If no '-P ' option is used, the size of max
On Sun 25 Aug 2019 at 21:32, Cong Wang wrote:
> On Wed, Aug 21, 2019 at 11:27 AM Vlad Buslov wrote:
>> At first I was confused why you bring up rtnl lock in commit message
>> (flower classifier has 'unlocked' flag set and can't rely on it anymore)
>> but looking at the code I see that we lost r
On Mon, Aug 26, 2019 at 04:30:01PM +0800, Xin Long wrote:
> This patchset is to make ecn flag per netns and endpoint and then
> add SCTP_ECN_SUPPORTED sockopt, as does for other feature flags.
>
> Xin Long (3):
> sctp: make ecn flag per netns and endpoint
> sctp: allow users to set netns ecn f
By adding an additional serdes_get_lane implementation (for Topaz), we
can merge the implementations of other SERDES functions (powering and
IRQs). We can skip checking port numbers, since the serdes_get_lane()
methods inform if there is no lane on a port or if the lane cannot be
used for given cmo
On Sun, 25 Aug 2019 12:12:14 -0400
Vivien Didelot wrote:
> In fact you're also relying on -ENODEV, which is what you return here (and in
> other places) instead of 0. So I'm afraid you have to address my comment
> now...
Vivien, you are right. I returned -ENODEV for Peridot when no lane was
to
Create a serdes_get_lane() method in the mv88e6xxx operations structure.
Use it instead of calling the different implementations.
Also change the methods so that their return value is used only for
error. The lane number is put into a place referred to by a pointer
given as argument. If the port do
Currently we support SERDES on the Topaz family in a limited way: no
IRQs and the cmode is not writable, thus the mode is determined by
strapping pins.
Marvell's examples though show how to make cmode writable on port 5 and
support SGMII autonegotiation. It is done by writing hidden registers,
for
This patch moves the functions operating on the hidden debug registers
into it's own file, port_hidden.c. The functions prefix is renamed from
mv88e6390_hidden_ to mv88e6xxx_port_hidden_, to be consistent with the
rest of this driver. The macros are prefixed with MV88E6XXX_ prefix, and
are changed
The mv88e6390_serdes_irq_link_sgmii IRQ handler reads the SERDES PHY
status register to determine speed, among other things. If cmode of the
port is set to 2500base-x, though, the PHY still reports 1000 Mbps (the
PHY register itself does not differentiate between 1000 Mbps and 2500
Mbps - it thinks
Hello,
this is the fourth version of changes for the Topaz/Peridot family of
switches. The patches apply on net-next.
Changes since v3:
- there was a mistake in the serdes_get_lane implementations for
6390 (patch 3/6). These methods returned -ENODEV if no lane was
to be on port, but they sh
This is a cosmetic update. We are removing the last underscore from
macros MV88E6XXX_PORT_STS_CMODE_100BASE_X and
MV88E6XXX_PORT_STS_CMODE_1000BASE_X. The 2500base-x version does not
have that underscore. Also PHY_INTERFACE_MODE_ macros do not have it
there.
Signed-off-by: Marek Behún
Reviewed-by
> -Original Message-
> From: Andrew Lunn
> Sent: Friday, August 23, 2019 7:31 PM
> To: Ioana Ciocoi Radulescu
> Cc: netdev@vger.kernel.org; da...@davemloft.net; Ioana Ciornei
>
> Subject: Re: [PATCH net-next] dpaa2-eth: Add pause frame support
>
> > --- a/drivers/net/ethernet/freescale/
In order to remove dependency on rtnl lock, modify tc_setup_flow_action()
to copy tunnel info, instead of just saving pointer to tunnel_key action
tunnel info. This is necessary to prevent concurrent action overwrite from
releasing tunnel info while it is being used by rtnl-unlocked driver.
Implem
Without rtnl lock protection filters can no longer safely manage block
offloads counter themselves. Refactor cls API to protect block offloadcnt
with tcf_block->cb_lock that is already used to protect driver callback
list and nooffloaddevcnt counter. The counter can be modified by concurrent
tasks
Currently, all cls API hardware offloads driver callbacks require caller
to hold rtnl lock when calling them. This patch set introduces new API
that allows drivers to register callbacks that are not dependent on rtnl
lock and unlocked classifiers to offload filters without obtaining rtnl
lock first
Extend struct flow_block_offload with "unlocked_driver_cb" flag to allow
registering and unregistering block hardware offload callbacks that do not
require caller to hold rtnl lock. Extend tcf_block with additional
lockeddevcnt counter that is incremented for each non-unlocked driver
callback attac
In order to allow using new flow_action infrastructure from unlocked
classifiers, modify tc_setup_flow_action() to accept new 'rtnl_held'
argument. Take rtnl lock before accessing tc_action data. This is necessary
to protect from concurrent action replace.
Signed-off-by: Vlad Buslov
Acked-by: Jir
In order to remove dependency on rtnl lock, extend tcf_block with 'cb_lock'
rwsem and use it to protect flow_block->cb_list and related counters from
concurrent modification. The lock is taken in read mode for read-only
traversal of cb_list in tc_setup_cb_call() and write mode in all other
cases. T
To remove dependency on rtnl lock, extend classifier ops with new
ops->hw_add() and ops->hw_del() callbacks. Call them from cls API while
holding cb_lock every time filter if successfully added to or deleted from
hardware.
Implement the new API in flower classifier. Use it to manage hw_filters
lis
Don't manually take rtnl lock in flower classifier before calling cls
hardware offloads API. Instead, pass rtnl lock status via 'rtnl_held'
parameter.
Signed-off-by: Vlad Buslov
Acked-by: Jiri Pirko
---
net/sched/cls_flower.c | 53 +-
1 file changed, 16 i
In order to remove dependency on rtnl lock from offloads code of
classifiers, take rtnl lock conditionally before executing driver
callbacks. Only obtain rtnl lock if block is bound to devices that require
it.
Block bind/unbind code is rtnl-locked and obtains block->cb_lock while
holding rtnl lock
In order to remove dependency on rtnl lock when calling hardware offload
API, take reference to action mirred dev when initializing flow_action
structure in tc_setup_flow_action(). Implement function
tc_cleanup_flow_action(), use it to release the device after hardware
offload API is done using it.
As a preparation for running proto ops functions without rtnl lock, change
offload counter type to atomic. This is necessary to allow updating the
counter by multiple concurrent users when offloading filters to hardware
from unlocked classifiers.
Signed-off-by: Vlad Buslov
Acked-by: Jiri Pirko
-
Hi,
In our production systems, running v4.19.y longterm kernels, we hit a
BUG_ON in 'skb_segment()'. It occurs rarely and although tried, couldn't
synthetically reproduce.
In v4.19.41 it crashes at net/core/skbuff.c:3711
while (pos < offset + len) {
if (i
Hi Marek,
On Mon, 26 Aug 2019 14:21:03 +0200, Marek Behún wrote:
> Hello,
>
> this is the fourth version of changes for the Topaz/Peridot family of
> switches. The patches apply on net-next.
> Changes since v3:
> - there was a mistake in the serdes_get_lane implementations for
>6390 (patch
On Mon, Aug 26, 2019 at 02:21:04PM +0200, Marek Behún wrote:
> The mv88e6390_serdes_irq_link_sgmii IRQ handler reads the SERDES PHY
> status register to determine speed, among other things. If cmode of the
> port is set to 2500base-x, though, the PHY still reports 1000 Mbps (the
> PHY register itse
On Mon, Aug 26, 2019 at 02:21:05PM +0200, Marek Behún wrote:
> This patch moves the functions operating on the hidden debug registers
> into it's own file, port_hidden.c. The functions prefix is renamed from
> mv88e6390_hidden_ to mv88e6xxx_port_hidden_, to be consistent with the
> rest of this dri
Hi Vladimir,
On Sun, 25 Aug 2019 21:44:54 +0300, Vladimir Oltean wrote:
> - if (enabled)
> - err = dsa_port_vid_add(upstream_dp, tx_vid, 0);
> - else
> - err = dsa_port_vid_del(upstream_dp, tx_vid);
> + err = dsa_8021q_vid_apply(ds, upstream, tx_vid, 0, enabled
On 26.08.2019 9:10, Björn Töpel wrote:
> From: Björn Töpel
>
> The state variable was read, and written outside the control mutex
> (struct xdp_sock, mutex), without proper barriers and {READ,
> WRITE}_ONCE correctness.
>
> In this commit this issue is addressed, and the state member is now
> us
> -int mv88e6390x_serdes_get_lane(struct mv88e6xxx_chip *chip, int port)
> +int mv88e6390x_serdes_get_lane(struct mv88e6xxx_chip *chip, int port, s8
> *lane)
> {
> u8 cmode_port9, cmode_port10, cmode_port;
>
> @@ -323,76 +320,80 @@ int mv88e6390x_serdes_get_lane(struct mv88e6xxx_chip
> *
On Mon, Aug 26, 2019 at 02:21:07PM +0200, Marek Behún wrote:
> By adding an additional serdes_get_lane implementation (for Topaz), we
> can merge the implementations of other SERDES functions (powering and
> IRQs). We can skip checking port numbers, since the serdes_get_lane()
> methods inform if t
> +static int mv88e6xxx_port_set_cmode(struct mv88e6xxx_chip *chip, int port,
> + phy_interface_t mode, bool allow_over_2500,
> + bool make_cmode_writable)
I don't like these two parameters. The caller of this function can do
the chec
Mon, Aug 26, 2019 at 03:44:57PM CEST, vla...@mellanox.com wrote:
>In order to remove dependency on rtnl lock, extend tcf_block with 'cb_lock'
>rwsem and use it to protect flow_block->cb_list and related counters from
>concurrent modification. The lock is taken in read mode for read-only
>traversal
Mon, Aug 26, 2019 at 03:44:59PM CEST, vla...@mellanox.com wrote:
>Without rtnl lock protection filters can no longer safely manage block
>offloads counter themselves. Refactor cls API to protect block offloadcnt
>with tcf_block->cb_lock that is already used to protect driver callback
>list and noof
Mon, Aug 26, 2019 at 03:45:00PM CEST, vla...@mellanox.com wrote:
>To remove dependency on rtnl lock, extend classifier ops with new
>ops->hw_add() and ops->hw_del() callbacks. Call them from cls API while
>holding cb_lock every time filter if successfully added to or deleted from
>hardware.
>
>Impl
On Mon, 26 Aug 2019 18:57:22 +0900
"Daniel T. Lee" wrote:
> Currently, at xdp_adjust_tail_kern.c, MAX_PCKT_SIZE is limited
> to 600. To make this size flexible, a new map 'pcktsz' is added.
>
> By updating new packet size to this map from the userland,
> xdp_adjust_tail_kern.o will use this valu
On Sun, Aug 25, 2019 at 10:37 PM Song Liu wrote:
> On Fri, Aug 23, 2019 at 7:04 PM Jakub Kicinski wrote:
> > From: Jiong Wang
> >
> > NFP is using Local Memory to model stack. LM_addr could be used as base of
> > a 16 32-bit word region of Local Memory. Then, if the stack offset is
> > beyond the
Hi Marek,
On Sun, 25 Aug 2019 18:36:09 +0200, Marek Behun wrote:
> > Aren't you relying on -ENODEV as well?
>
> Vivien, I am not relying o -ENODEV. I changed the serdes_get_lane
> semantics:
> - previously:
>- if port has a lane for current cmode, return given lane number
>- otherwise r
Tue, Aug 13, 2019 at 08:56:17AM CEST, j...@resnulli.us wrote:
>Mon, Aug 12, 2019 at 06:01:59PM CEST, dsah...@gmail.com wrote:
>>On 8/12/19 2:31 AM, Jiri Pirko wrote:
>>> Mon, Aug 12, 2019 at 03:37:26AM CEST, dsah...@gmail.com wrote:
On 8/11/19 7:34 PM, David Ahern wrote:
> On 8/10/19 12:30
Hi Vivien,
On Mon, 26 Aug 2019 at 18:20, Vivien Didelot wrote:
>
> Hi Vladimir,
>
> On Sun, 25 Aug 2019 21:44:54 +0300, Vladimir Oltean wrote:
> > - if (enabled)
> > - err = dsa_port_vid_add(upstream_dp, tx_vid, 0);
> > - else
> > - err = dsa_port_vid_del(upstream
On Mon, Aug 26, 2019 at 8:57 AM Jakub Kicinski
wrote:
>
> On Sun, Aug 25, 2019 at 10:37 PM Song Liu wrote:
> > On Fri, Aug 23, 2019 at 7:04 PM Jakub Kicinski wrote:
> > > From: Jiong Wang
> > >
> > > NFP is using Local Memory to model stack. LM_addr could be used as base of
> > > a 16 32-bit wor
Vladimir Rutsky reported stuck TCP sessions after memory pressure
events. Edge Trigger epoll() user would never receive an EPOLLOUT
notification allowing them to retry a sendmsg().
Jason tested the case of sk_stream_alloc_skb() returning NULL,
but there are other paths that could lead both sendmsg
On 8/26/19 6:18 PM, Alexei Starovoitov wrote:
On Mon, Aug 26, 2019 at 8:57 AM Jakub Kicinski
wrote:
On Sun, Aug 25, 2019 at 10:37 PM Song Liu wrote:
On Fri, Aug 23, 2019 at 7:04 PM Jakub Kicinski wrote:
From: Jiong Wang
NFP is using Local Memory to model stack. LM_addr could be used as bas
Currently, at xdp_adjust_tail_kern.c, MAX_PCKT_SIZE is limited
to 600. To make this size flexible, a new map 'pcktsz' is added.
By updating new packet size to this map from the userland,
xdp_adjust_tail_kern.o will use this value as a new max_pckt_size.
If no '-P ' option is used, the size of max
On Tue, Aug 27, 2019 at 12:54 AM Maciej Fijalkowski
wrote:
>
> On Mon, 26 Aug 2019 18:57:22 +0900
> "Daniel T. Lee" wrote:
>
> > Currently, at xdp_adjust_tail_kern.c, MAX_PCKT_SIZE is limited
> > to 600. To make this size flexible, a new map 'pcktsz' is added.
> >
> > By updating new packet size
On 2019-08-26 17:24, Ilya Maximets wrote:
This changes the error code a bit.
Previously:
umem exists + xs unbound--> EINVAL
no umem + xs unbound--> EBADF
xs bound to different dev/q --> EINVAL
With this change:
umem exists + xs unbound--> EBADF
no umem + x
On Mon, 26 Aug 2019 18:25:10 +0200, Daniel Borkmann wrote:
> On 8/26/19 6:18 PM, Alexei Starovoitov wrote:
> > On Mon, Aug 26, 2019 at 8:57 AM Jakub Kicinski
> > wrote:
> >> On Sun, Aug 25, 2019 at 10:37 PM Song Liu wrote:
> >>> On Fri, Aug 23, 2019 at 7:04 PM Jakub Kicinski wrote:
> F
On Sun, 18 Aug 2019 00:34:33 +0530
Akshat Kakkar wrote:
> My goal is not just to make as many classes as possible, but also to
> use them to do rate limiting per ip per server. Say, I have a list of
> 1 IPs and more than 100 servers. So simply if I want few IPs to
> get speed of says 1Mbps pe
On Mon, Aug 26, 2019 at 12:19 PM Eric Dumazet wrote:
>
> Vladimir Rutsky reported stuck TCP sessions after memory pressure
> events. Edge Trigger epoll() user would never receive an EPOLLOUT
> notification allowing them to retry a sendmsg().
>
> Jason tested the case of sk_stream_alloc_skb() retur
On Mon, 26 Aug 2019 18:09:16 +0200, Jiri Pirko wrote:
> DaveA, Roopa. Do you insist on doing add/remove of altnames in the
> existing setlist command using embedded message op attrs? I'm asking
> because after some time thinking about it, it still feels wrong to me :/
>
> If this would be a generi
On Mon, Aug 26, 2019 at 12:19 PM Eric Dumazet wrote:
>
> Vladimir Rutsky reported stuck TCP sessions after memory pressure
> events. Edge Trigger epoll() user would never receive an EPOLLOUT
> notification allowing them to retry a sendmsg().
>
> Jason tested the case of sk_stream_alloc_skb() retur
Simplify the code which fetches the input clock by using
devm_clk_get_optional(). This comes with a small functional change: previously
all errors were ignored when platform data is present. Now all errors are
treated as errors. If no input clock is present devm_clk_get_optional() will
return NULL
There is no need to check for regulator presence in the ->suspend()
since a wrapper does it for us. Due to this we may unconditionally set
AFTER_SUSPEND_POWER flag.
Signed-off-by: Andy Shevchenko
---
drivers/net/can/spi/mcp251x.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff
Make use of device property API in this driver so that both OF based
system and ACPI based system can use this driver.
Signed-off-by: Andy Shevchenko
---
drivers/net/can/spi/mcp251x.c | 12 +---
1 file changed, 5 insertions(+), 7 deletions(-)
diff --git a/drivers/net/can/spi/mcp251x.c b
On Mon, 26 Aug 2019 17:38:30 +0200
Andrew Lunn wrote:
> > +static int mv88e6xxx_port_set_cmode(struct mv88e6xxx_chip *chip, int port,
> > + phy_interface_t mode, bool allow_over_2500,
> > + bool make_cmode_writable)
>
> I don't like t
On Mon, 26 Aug 2019 17:38:30 +0200
Andrew Lunn wrote:
> > +static int mv88e6xxx_port_set_cmode(struct mv88e6xxx_chip *chip, int port,
> > + phy_interface_t mode, bool allow_over_2500,
> > + bool make_cmode_writable)
>
> I don't like t
On Mon, 26 Aug 2019 19:27:17 +0200, Marek Behun wrote:
> On Mon, 26 Aug 2019 17:38:30 +0200
> Andrew Lunn wrote:
>
> > > +static int mv88e6xxx_port_set_cmode(struct mv88e6xxx_chip *chip, int
> > > port,
> > > + phy_interface_t mode, bool allow_over_2500,
> > > +
On Mon, 26 Aug 2019 16:44:56 +0300, Vlad Buslov wrote:
> Currently, all cls API hardware offloads driver callbacks require caller
> to hold rtnl lock when calling them. This patch set introduces new API
> that allows drivers to register callbacks that are not dependent on rtnl
> lock and unlocked c
On 8/26/19 4:07 PM, Shmulik Ladkani wrote:
> Hi,
>
> In our production systems, running v4.19.y longterm kernels, we hit a
> BUG_ON in 'skb_segment()'. It occurs rarely and although tried, couldn't
> synthetically reproduce.
>
> In v4.19.41 it crashes at net/core/skbuff.c:3711
>
>
On Mon, 26 Aug 2019 19:31:25 +0200, Marek Behun wrote:
> On Mon, 26 Aug 2019 17:38:30 +0200
> Andrew Lunn wrote:
>
> > > +static int mv88e6xxx_port_set_cmode(struct mv88e6xxx_chip *chip, int
> > > port,
> > > + phy_interface_t mode, bool allow_over_2500,
> > > +
On Mon, 26 Aug 2019 13:44:18 -0400
Vivien Didelot wrote:
> > It can be done once at probe. At first I thought about doing this in
> > setup_errata, but this is not an erratum. So shall I create a new
> > method for this in chip operations structure? Something like
> > port_additional_setup() ?
On 25 Aug 2019, at 23:10, Björn Töpel wrote:
From: Björn Töpel
The state variable was read, and written outside the control mutex
(struct xdp_sock, mutex), without proper barriers and {READ,
WRITE}_ONCE correctness.
In this commit this issue is addressed, and the state member is now
used a
On 26 Aug 2019, at 9:34, Björn Töpel wrote:
> On 2019-08-26 17:24, Ilya Maximets wrote:
>> This changes the error code a bit.
>> Previously:
>> umem exists + xs unbound--> EINVAL
>> no umem + xs unbound--> EBADF
>> xs bound to different dev/q --> EINVAL
>>
>> With this c
Currently we support SERDES on the Topaz family in a limited way: no
IRQs and the cmode is not writable, thus the mode is determined by
strapping pins.
Marvell's examples though show how to make cmode writable on port 5 and
support SGMII autonegotiation. It is done by writing hidden registers,
for
What about this?
It adds a new chip operation (I know Vivien said not to, but I was
doing it already) port_setup_extra, and implements it for Topaz.
Also it changes the mv88e6xxx_port_set_cmode so that it does not use
those 2 additional parameters.
Should I rewrite it so that only the second cha
Mon, Aug 26, 2019 at 03:45:02PM CEST, vla...@mellanox.com wrote:
>In order to remove dependency on rtnl lock from offloads code of
>classifiers, take rtnl lock conditionally before executing driver
>callbacks. Only obtain rtnl lock if block is bound to devices that require
>it.
>
>Block bind/unbind
On Mon, 26 Aug 2019 20:03:15 +0200, Marek Behun wrote:
> What about this?
>
> It adds a new chip operation (I know Vivien said not to, but I was
> doing it already) port_setup_extra, and implements it for Topaz.
So what feedback do you expect exactly? That is *exactly* what I told
you I did not
On Mon, 26 Aug 2019 14:28:09 -0400
Vivien Didelot wrote:
> Ask yourself what is the single task achieved by this function, and name this
> operation accordingly. It seems to change the CMODE to be writable, only
> supported by certain switch models right? So in addition to port_get_cmode
> and po
> On Aug 25, 2019, at 1:40 PM, Pravin Shelar wrote:
>
> Actually I am not sure about this change. caller of this function
> (ovs_ct_execute()) does skb-pull and push of L2 header, calling
> ovs_flow_key_update() is not safe here, it expect skb data to point to
> L2 header.
Thanks for the feedb
On Mon, 26 Aug 2019 20:36:14 +0200, Marek Behun wrote:
> > Ask yourself what is the single task achieved by this function, and name
> > this
> > operation accordingly. It seems to change the CMODE to be writable, only
> > supported by certain switch models right? So in addition to port_get_cmode
On 8/26/19 12:19 PM, Eric Dumazet wrote:
> Vladimir Rutsky reported stuck TCP sessions after memory pressure
> events. Edge Trigger epoll() user would never receive an EPOLLOUT
> notification allowing them to retry a sendmsg().
>
> Jason tested the case of sk_stream_alloc_skb() returning NULL,
On 8/26/19 9:56 PM, Jason Baron wrote:
>
>
> On 8/26/19 12:19 PM, Eric Dumazet wrote:
>> Vladimir Rutsky reported stuck TCP sessions after memory pressure
>> events. Edge Trigger epoll() user would never receive an EPOLLOUT
>> notification allowing them to retry a sendmsg().
>>
>> Jason tested
On Fri, 2019-08-23 at 11:16 -0700, Jakub Kicinski wrote:
> On Fri, 23 Aug 2019 06:00:45 +, Saeed Mahameed wrote:
> > On Thu, 2019-08-22 at 18:33 -0700, Jakub Kicinski wrote:
> > > On Thu, 22 Aug 2019 23:35:52 +, Saeed Mahameed wrote:
> > > > From: Moshe Shemesh
> > > >
> > > > Added the
On Mon, 26 Aug 2019 20:14:47 +, Saeed Mahameed wrote:
> > I see thanks for the explanation and sorry for the delayed response.
> > Would it perhaps make sense to indicate the hairpin in the name?
>
> We had some internal discussion and we couldn't come up with the
> perfect name :)
>
> hair
From: Justin Pettit
Only the first fragment in a datagram contains the L4 headers. When the
Open vSwitch module parses a packet, it always sets the IP protocol
field in the key, but can only set the L4 fields on the first fragment.
The original behavior would not clear the L4 portion of the key,
When IP fragments are reassembled before being sent to conntrack, the
key from the last fragment is used. Unless there are reordering
issues, the last fragment received will not contain the L4 ports, so the
key for the reassembled datagram won't contain them. This patch updates
the key once we ha
Move the call to dma_sync_single_for_cpu after calling napi_alloc_skb.
This avoids calling dma_sync_single_for_cpu w/o handing control back
to device if the memory allocation should fail.
Signed-off-by: Heiner Kallweit
---
drivers/net/ethernet/realtek/r8169_main.c | 7 +++
1 file changed, 3
From: Andrew Lunn
Date: Mon, 26 Aug 2019 17:38:30 +0200
>> +static int mv88e6xxx_port_set_cmode(struct mv88e6xxx_chip *chip, int port,
>> +phy_interface_t mode, bool allow_over_2500,
>> +bool make_cmode_writable)
>
> I don't like th
From: Vlad Buslov
Date: Mon, 26 Aug 2019 16:44:56 +0300
> Implement following cls API changes:
>
> - Introduce new "unlocked_driver_cb" flag to struct flow_block_offload
> to allow registering and unregistering block hardware offload
> callbacks that do not require caller to hold rtnl lock.
The mv88e6390_serdes_irq_link_sgmii IRQ handler reads the SERDES PHY
status register to determine speed, among other things. If cmode of the
port is set to 2500base-x, though, the PHY still reports 1000 Mbps (the
PHY register itself does not differentiate between 1000 Mbps and 2500
Mbps - it thinks
Currently we support SERDES on the Topaz family in a limited way: no
IRQs and the cmode is not writable, thus the mode is determined by
strapping pins.
Marvell's examples though show how to make cmode writable on port 5 and
support SGMII autonegotiation. It is done by writing hidden registers,
for
This is a cosmetic update. We are removing the last underscore from
macros MV88E6XXX_PORT_STS_CMODE_100BASE_X and
MV88E6XXX_PORT_STS_CMODE_1000BASE_X. The 2500base-x version does not
have that underscore. Also PHY_INTERFACE_MODE_ macros do not have it
there.
Signed-off-by: Marek Behún
Reviewed-by
Create a serdes_get_lane() method in the mv88e6xxx operations structure.
Use it instead of calling the different implementations.
Also change the methods so that their return value is used only for
error. The lane number is put into a place referred to by a pointer
given as argument. If the port do
By adding an additional serdes_get_lane implementation (for Topaz), we
can merge the implementations of other SERDES functions (powering and
IRQs). We can skip checking port numbers, since the serdes_get_lane()
methods inform if there is no lane on a port or if the lane cannot be
used for given cmo
This patch moves the functions operating on the hidden debug registers
into it's own file, port_hidden.c. The functions prefix is renamed from
mv88e6390_hidden_ to mv88e6xxx_port_hidden_, to be consistent with the
rest of this driver. The macros are prefixed with MV88E6XXX_ prefix, and
are changed
Hello,
this is the fifth version of changes for the Topaz/Peridot family of
switches. The patches apply on net-next.
Changes since v4:
- added Reviewed-by and Tested-by tags on first 2 patches, the others
are changed are affected by changes in patch 3/6, so I did not add
the tags, except fo
Thank you for the report.
On Sun, 25 Aug 2019 22:21:50 +0530, Mallesham Jatharakonda wrote:
> Hi All,
>
> Am facing one tls double while using the Nitrox(cavium) card and n5pf
> driver over the TLS module.
>
> Please see the below details:
>
> TLS module is crashing While running SSL record enc
This is a patch series that adds the ionic driver, supporting the Pensando
ethernet device.
In this initial patchset we implement basic transmit and receive. Later
patchsets will add more advanced features.
Our thanks to Saeed Mahameed, David Miller, Andrew Lunn, Michal Kubecek,
Jacub Kicinski,
The ionic interrupt model is based on interrupt control blocks
accessed through the PCI BAR. Doorbell registers are used by
the driver to signal to the NIC that requests are waiting on
the message queues. Interrupts are used by the NIC to signal
to the driver that answers are waiting on the compl
From: Heiner Kallweit
Date: Mon, 26 Aug 2019 22:52:36 +0200
> Move the call to dma_sync_single_for_cpu after calling napi_alloc_skb.
> This avoids calling dma_sync_single_for_cpu w/o handing control back
> to device if the memory allocation should fail.
>
> Signed-off-by: Heiner Kallweit
Appli
Set up the infrastructure for managing Rx filters. We can't ask the
hardware for what filters it has, so we keep a local list of filters
that we've pushed into the HW.
Signed-off-by: Shannon Nelson
---
drivers/net/ethernet/pensando/ionic/Makefile | 2 +-
.../net/ethernet/pensando/ionic/ionic
1 - 100 of 153 matches
Mail list logo