On Thu, Sep 08, 2016 at 08:30:52AM +0200, Ingo Molnar wrote:
> > Thanks for the fix. Just saw build bot complaining last night and
> > by the morning your fix is already here. Thanks!
>
> Guys, this is the first time I saw this patch (and presumably there's some
> other
> patch as well?) - could
From: Florian Westphal
Don't acquire the readlock anymore and rely on rcu alone.
In case writer on other CPU changed policy at the wrong moment (after we
obtained sk policy pointer but before we could obtain the reference)
just repeat the lookup.
Signed-off-by: Florian Westphal
Signed-off-by:
From: Florian Westphal
This is required once we allow lockless readers.
Signed-off-by: Florian Westphal
Signed-off-by: Steffen Klassert
---
net/xfrm/xfrm_policy.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
From: Florian Westphal
If we don't hold the policy lock anymore the refcnt might
already be 0, i.e. policy struct is about to be free'd.
Switch to atomic_inc_not_zero to avoid this.
On removal policies are already unlinked from the tables (lists)
before the last _put occurs so we are not suppos
From: Florian Westphal
After earlier patches conversions all spots acquire the writer lock and
we can now convert this to a normal spinlock.
Signed-off-by: Florian Westphal
Signed-off-by: Steffen Klassert
---
include/net/netns/xfrm.h | 2 +-
net/xfrm/xfrm_policy.c | 68
On Wed, Sep 07, 2016 at 05:25:18PM -0700, David Miller wrote:
> From: Subash Abhinov Kasiviswanathan
> Date: Tue, 6 Sep 2016 18:09:31 -0600
>
> > proc_dointvec limits the values to INT_MAX in u32 sysctl entries.
> > proc_douintvec allows to write upto UINT_MAX.
> >
> > Signed-off-by: Subash Abh
1) Constify the xfrm_replay structures. From Julia Lawall
2) Protect xfrm state hash tables with rcu, lookups
can be done now without acquiring xfrm_state_lock.
From Florian Westphal.
3) Protect xfrm policy hash tables with rcu, lookups
can be done now without acquiring xfrm_policy_lock.
From: Julia Lawall
The xfrm_replay structures are never modified, so declare them as const.
Done with the help of Coccinelle.
Signed-off-by: Julia Lawall
Signed-off-by: Steffen Klassert
---
include/net/xfrm.h | 2 +-
net/xfrm/xfrm_replay.c | 6 +++---
2 files changed, 4 insertions(+), 4
From: Florian Westphal
push the lock down, after earlier patches we can rely on rcu to
make sure state struct won't go away.
Signed-off-by: Florian Westphal
Signed-off-by: Steffen Klassert
---
include/net/netns/xfrm.h | 6 +++---
net/xfrm/xfrm_state.c| 6 --
2 files changed, 7 inserti
On Thu, Sep 8, 2016 at 1:36 AM, David Miller wrote:
> From: Or Gerlitz
> Date: Wed, 7 Sep 2016 23:51:56 +0300
>
>> On Wed, Sep 7, 2016 at 11:27 PM, David Miller wrote:
>>> From: Saeed Mahameed
>>> Date: Wed, 7 Sep 2016 23:16:13 +0300
>>>
Hi Dave, Sorry to bother, but i would like to drop t
An earlier patch accidentally replaced a write_lock_bh
with a spin_unlock_bh. Fix this by using spin_lock_bh
instead.
Fixes: 9d0380df6217 ("xfrm: policy: convert policy_lock to spinlock")
Signed-off-by: Steffen Klassert
---
net/xfrm/xfrm_policy.c | 2 +-
1 file changed, 1 insertion(+), 1 deletio
From: Florian Westphal
Since commit 56f047305dd4b6b617
("xfrm: add rcu grace period in xfrm_policy_destroy()") xfrm policy
objects are already free'd via rcu.
In order to make more places lockless (i.e. use rcu_read_lock instead of
grabbing read-side of policy rwlock) we only need to:
- use rcu
From: Florian Westphal
side effect: no longer disables BH (should be fine).
Signed-off-by: Florian Westphal
Signed-off-by: Steffen Klassert
---
include/net/netns/xfrm.h | 2 +-
net/xfrm/xfrm_policy.c | 4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/include/net/netns/
From: Florian Westphal
This is required once we allow lockless access of bydst/bysrc hash tables.
Signed-off-by: Florian Westphal
Signed-off-by: Steffen Klassert
---
net/xfrm/xfrm_state.c | 38 +++---
1 file changed, 19 insertions(+), 19 deletions(-)
diff --gi
From: Florian Westphal
After commit 5b8ef3415a21f173
("xfrm: Remove ancient sleeping when the SA is in acquire state")
gc does not need any per-netns data anymore.
As far as gc is concerned all state structs are the same, so we
can use a global work struct for it.
Signed-off-by: Florian Westpha
From: Florian Westphal
It doesn't seem that important.
We now get inconsistent view of the counters, but those are stale anyway
right after we drop the lock.
Signed-off-by: Florian Westphal
Signed-off-by: Steffen Klassert
---
net/xfrm/xfrm_policy.c | 2 --
1 file changed, 2 deletions(-)
dif
From: Florian Westphal
Once xfrm_state_lookup_byaddr no longer acquires the state lock another
cpu might be freeing the state entry at the same time.
To detect this we use atomic_inc_not_zero, we then signal -EAGAIN to
caller in case our result was stale.
Signed-off-by: Florian Westphal
Signed
From: Florian Westphal
The hash table backend memory and the state structs are free'd via
kfree/vfree.
Once we only rely on rcu during lookups we have to make sure no other cpu
is currently accessing this before doing the free.
Free operations already happen from worker so we can use synchroniz
From: Florian Westphal
Once xfrm_policy_lookup_bytype doesn't grab xfrm_policy_lock anymore its
possible for a hash resize to occur in parallel.
Use sequence counter to block lookup in case a resize is in
progress and to also re-lookup in case hash table was altered
in the mean time (might cause
From: Florian Westphal
Once xfrm_state_find is lockless we have to cope with a concurrent
resize opertion.
We use a sequence counter to block in case a resize is in progress
and to detect if we might have missed a state that got moved to
a new hash table.
Signed-off-by: Florian Westphal
Signed
From: Florian Westphal
Before xfrm_state_find() can use rcu_read_lock instead of xfrm_state_lock
we need to switch users of the hash table to assign/obtain the pointers
with the appropriate rcu helpers.
Signed-off-by: Florian Westphal
Signed-off-by: Steffen Klassert
---
net/xfrm/xfrm_state.c
On Thu, Sep 08, 2016 at 08:16:01AM +0200, Jiri Pirko wrote:
> From: Jiri Pirko
>
> When neigh_init fails, we have to do proper cleanup including
> router_fini call.
>
> Fixes: 6cf3c971dc84cb ("mlxsw: spectrum_router: Add private neigh table")
> Signed-off-by: Jiri Pirko
Acked-by: Ido Schimmel
On Thu, Sep 8, 2016 at 3:32 AM, David Miller wrote:
> From: Hadar Hen Zion
> Date: Wed, 7 Sep 2016 11:08:02 +0300
>
>> This patchset introduces ip tunnel manipulation support using the TC
>> subsystem.
>
> Please address the feedback given by Eric Dumazet for patch #4,
> thank you.
Sure, I'll
From: Nogah Frankel
Add a nested attribute of offload stats to if_stats_msg
named IFLA_STATS_LINK_OFFLOAD_XSTATS.
Under it, add SW stats, meaning stats only per packets that went via
slowpath to the cpu, named IFLA_OFFLOAD_XSTATS_CPU_HIT.
Signed-off-by: Nogah Frankel
Signed-off-by: Jiri Pirko
From: Jiri Pirko
The problem we try to handle is about offloaded forwarded packets
which are not seen by kernel. Let me try to draw it:
port1 port2 (HW stats are counted here)
\ /
\/
\
From: Nogah Frankel
Change the default statistics ndo to return HW statistics
(like the one returned by ethtool_ops).
The HW stats are collected to a cache by delayed work every 1 sec.
Implement the offload stat ndo.
Add a function to get SW statistics, to be called from this function.
Signed-of
From: Nogah Frankel
Add a new ndo to return statistics for offloaded operation.
Since there can be many different offloaded operation with many
stats types, the ndo gets an attribute id by which it knows which
stats are wanted. The ndo also gets a void pointer to be cast according
to the attribut
* Alexei Starovoitov wrote:
> On Tue, Sep 06, 2016 at 03:10:22PM +0200, Arnd Bergmann wrote:
> > The newly added bpf_overflow_handler function is only built of both
> > CONFIG_EVENT_TRACING and CONFIG_BPF_SYSCALL are enabled, but the caller
> > only checks the latter:
> >
> > kernel/events/core
From: Alexey Kodanev
Running LTP 'icmp-uni-basic.sh -6 -p ipcomp -m tunnel' test over
openvswitch + veth can trigger kernel panic:
BUG: unable to handle kernel NULL pointer dereference
at 00e0 IP: [] xfrm_input+0x82/0x750
...
[] xfrm6_rcv_spi+0x1e/0x20
[] xfrm6_tunnel_rcv+0
From: Vegard Nossum
During fuzzing I regularly run into this WARN(). According to Herbert Xu,
this "certainly shouldn't be a WARN, it probably shouldn't print anything
either".
Cc: Stephen Hemminger
Cc: Steffen Klassert
Cc: Herbert Xu
Signed-off-by: Vegard Nossum
Signed-off-by: Steffen Klass
From: Vegard Nossum
AFAICT this message is just printed whenever input validation fails.
This is a normal failure and we shouldn't be dumping the stack over it.
Looks like it was originally a printk that was maybe incorrectly
upgraded to a WARN:
commit 62db5cfd70b1ef53aa21f144a806fe3b78c84fab
A
From: David Ahern
Subash reported that commit 42a7b32b73d6 ("xfrm: Add oif to dst lookups")
broke a wifi use case that uses fib rules and xfrms. The intent of
42a7b32b73d6 was driven by VRFs with IPsec. As a compromise relax the
use of oif in xfrm lookups to L3 master devices only (ie., oif is ei
From: Vegard Nossum
If we hit any of the error conditions inside xfrm_dump_sa(), then
xfrm_state_walk_init() never gets called. However, we still call
xfrm_state_walk_done() from xfrm_dump_sa_done(), which will crash
because the state walk was never initialized properly.
We can fix this by setti
1) Fix a crash when xfrm_dump_sa returns an error.
From Vegard Nossum.
2) Remove some incorrect WARN() on normal error handling.
From Vegard Nossum.
3) Ignore socket policies when rebuilding hash tables,
socket policies are not inserted into the hash tables.
From Tobias Brunner.
4) I
From: Tobias Brunner
Whenever thresholds are changed the hash tables are rebuilt. This is
done by enumerating all policies and hashing and inserting them into
the right table according to the thresholds and direction.
Because socket policies are also contained in net->xfrm.policy_all but
no has
From: Steffen Klassert
Date: Thu, 8 Sep 2016 07:51:17 +0200
> On Wed, Sep 07, 2016 at 05:25:18PM -0700, David Miller wrote:
>> From: Subash Abhinov Kasiviswanathan
>> Date: Tue, 6 Sep 2016 18:09:31 -0600
>>
>> > proc_dointvec limits the values to INT_MAX in u32 sysctl entries.
>> > proc_douint
From: Sriharsha Basavapatna
Date: Wed, 7 Sep 2016 19:57:48 +0530
> The following patch set contains an error recovery feature and a few
> bug fixes. Please consider applying this to the net-next tree. Thanks.
>
> Patch-1 Supports HW error recovery in Skyhawk/BEx adapters
> Patch-2 Fixes driver
On 16-09-06 07:52 AM, Eric Dumazet wrote:
> On Thu, 2016-09-01 at 22:57 -0700, Cong Wang wrote:
>
>
>
> Missing changelog ?
>
Yeah this stuff is a bit tricky more notes to walk us through
this would be helpful. (btw you can disregard my comment from
earlier this morning I've tracked most of th
From: Jiri Pirko
Couple of fixes from Ido and myself.
Ido Schimmel (1):
mlxsw: spectrum: Set port type before setting its address
Jiri Pirko (1):
mlxsw: spectrum_router: Fix error path in mlxsw_sp_router_init
drivers/net/ethernet/mellanox/mlxsw/spectrum.c| 18 +-
d
From: Jiri Pirko
When neigh_init fails, we have to do proper cleanup including
router_fini call.
Fixes: 6cf3c971dc84cb ("mlxsw: spectrum_router: Add private neigh table")
Signed-off-by: Jiri Pirko
---
drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 9 -
1 file changed, 8 insert
From: Ido Schimmel
During port init, we currently set the port's type to Ethernet after
setting its MAC address. However, the hardware documentation states this
should be the other way around.
Align the driver with the hardware documentation and set the port's MAC
address after setting its type.
On Wed, 7 Sep 2016 20:21:24 -0700 Tom Herbert wrote:
> On Wed, Sep 7, 2016 at 7:58 PM, John Fastabend
> wrote:
> > On 16-09-07 11:22 AM, Jesper Dangaard Brouer wrote:
> >>
> >> On Wed, 7 Sep 2016 19:57:19 +0300 Saeed Mahameed
> >> wrote:
> >>> On Wed, Sep 7, 2016 at 6:32 PM, Eric Dumazet
From: Eric Dumazet
In commit f02db315b8d8 ("ipv4: IP_TOS and IP_TTL can be specified as
ancillary data") Francesco added IP_TOS values specified as integer.
However, kernel sends to userspace (at recvmsg() time) an IP_TOS value
in a single byte, when IP_RECVTOS is set on the socket.
It can be v
Hi,
On Tue, Sep 6, 2016 at 9:18 PM, Richard Cochran
wrote:
>
>> +#define GEM_TISUBN 0x01bc /* 1588 Timer Increment Sub-ns */
>
> This regsiter does not exist. Looking at
>
>Zynq-7000 AP SoC Technical Reference Manual
>UG585 (v1.10) February 23, 2015
>
> starting on page 1273 w
net/bluetooth/mgmt.c:905:6-13: WARNING: kzalloc should be used for rp, instead
of kmalloc/memset
Use kzalloc rather than kmalloc followed by memset with 0
This considers some simple cases that are common and easy to validate
Note in particular that there are no ...s in the rule, so all of th
From: Linus Walleij
Date: Wed, 7 Sep 2016 15:53:42 +0200
> On some systems (such as the Qualcomm APQ8060 Dragonboard) the
> RESET signal of the SMSC911x is not pulled up by a resistor (or
> the internal pull-up that will pull it up if the pin is not
> even connected) but instead connected to a G
From: Linus Walleij
Date: Wed, 7 Sep 2016 15:53:31 +0200
> This adds device tree bindings for:
>
> - An optional GPIO line for releasing the RESET signal to the
> SMSC911x devices
>
> - An optional PME (power management event) interrupt line that
> can be utilized to wake up the system on
On Wed, Sep 7, 2016 at 7:58 PM, John Fastabend wrote:
> On 16-09-07 11:22 AM, Jesper Dangaard Brouer wrote:
>>
>> On Wed, 7 Sep 2016 19:57:19 +0300 Saeed Mahameed
>> wrote:
>>> On Wed, Sep 7, 2016 at 6:32 PM, Eric Dumazet wrote:
On Wed, 2016-09-07 at 18:08 +0300, Saeed Mahameed wrote:
David Miller [mailto:da...@davemloft.net]
> Sent: Thursday, September 08, 2016 8:38 AM
[...]
> By forcing a certain mode via a Kconfig value, you are basically making it
> impossible for distributions to do something reasonable here.
The request is always from some manufacturers, not end users.
Th
On 16-09-07 11:22 AM, Jesper Dangaard Brouer wrote:
>
> On Wed, 7 Sep 2016 19:57:19 +0300 Saeed Mahameed
> wrote:
>> On Wed, Sep 7, 2016 at 6:32 PM, Eric Dumazet wrote:
>>> On Wed, 2016-09-07 at 18:08 +0300, Saeed Mahameed wrote:
On Wed, Sep 7, 2016 at 5:41 PM, Eric Dumazet
wrote:
Bjørn Mork [mailto:bj...@mork.no]
> Sent: Wednesday, September 07, 2016 9:51 PM
[...]
> So this adds a lot of code to work around the issues you introduced by
> unnecessarily blacklisting the CDC ECM configuration earlier, and still
> makes the r8152 driver handle the device even in ECM mode.
I su
ftell() may return -1 in error case, which is not handled and therefore pass a
negative offset to fseek(). The return code of fseek() is also not checked.
Reported-by: Phil Sutter
Signed-off-by: Hangbin Liu
---
ip/iproute.c | 10 +-
1 file changed, 9 insertions(+), 1 deletion(-)
diff -
On 2016/09/08 3:22, Michal Soltys wrote:
...
> 4.7.2
> git describe on that commit suggests it's been available since 4.6.x
>
> What I did in details:
>
> ip li add name port1b type veth peer name port1e
> ip li add br0 type bridge
> ip li set dev br0 type bridge vlan_default_pvid 0
> ip li set d
During namespace cleanup, if ‘dst’ subsystem is holding
a reference to the loopback interface in the namespace, it does not get
released. This is because in the case where the net_device held by 'dst'
is same as the namespace's loopback net_device, current code first
does a ’dev_hold’ on the same d
> This is definitely going to require a few passes... (Working my way
> through every driver with an ndo_change_mtu wired up right now to
> see just how crazy this might get).
It might be something Coccinelle can help you with. Try describing the
transformation you want to do, to their mailing lis
On Wed, Sep 7, 2016 at 9:56 AM, Eric Garver wrote:
> Add support for 802.1ad including the ability to push and pop double
> tagged vlans. Add support for 802.1ad to netlink parsing and flow
> conversion. Uses double nested encap attributes to represent double
> tagged vlan. Inner TPID encoded alon
On Wed, Sep 7, 2016 at 9:56 AM, Eric Garver wrote:
> From: Thomas F Herbert
>
> openvswitch: Add support for 8021.AD
>
> Change the description of the VLAN tpid field.
>
> Signed-off-by: Thomas F Herbert
Acked-by: Pravin B Shelar
On Wed, Sep 7, 2016 at 9:56 AM, Eric Garver wrote:
> This is to simplify using double tagged vlans. This function allows all
> valid vlan ethertypes to be checked in a single function call.
> Also replace some instances that check for both ETH_P_8021Q and
> ETH_P_8021AD.
>
> Patch based on one ori
From: Tomer Tayar
Date: Wed, 7 Sep 2016 16:36:23 +0300
> This patch series adds the support of debug data collection in the qed driver,
> and the means to extract it in the qede driver via the get_regs operation.
...
> Please consider applying this to 'net-next'.
Series applied, thank you.
From: Oliver Neukum
Date: Wed, 7 Sep 2016 15:27:09 +0200
> SOme statements in the driver only served to inform
> which functions were entered. Ftrace can do that just as good without
> needing memory. Remove the statements.
>
> Signed-off-by: Oliver Neukum
Applied to net-next, thanks Oliver.
From: Yuval Mintz
Date: Wed, 7 Sep 2016 11:55:34 +
>> We get a few warnings when building kernel with W=1:
>> drivers/net/ethernet/qlogic/qed/qed_l2.c:112:5: warning: no previous
>> prototype for 'qed_sp_vport_start' [-Wmissing-prototypes]
>>
>>
>> In fact, these functions are only used
From: Baoyou Xie
Date: Wed, 7 Sep 2016 19:07:00 +0800
> We get 4 warnings when building kernel with W=1:
> drivers/net/ethernet/qlogic/qed/qed_selftest.c:6:5: warning: no previous
> prototype for 'qed_selftest_memory' [-Wmissing-prototypes]
> drivers/net/ethernet/qlogic/qed/qed_selftest.c:19:5:
From: Hayes Wang
Date: Wed, 7 Sep 2016 16:12:19 +0800
> Some people prefer to use ECM mode rather than vendor mode. Therefore, I add
> CONFIG_RTL8152_CONFIG_VALUE in Kconfig. Then, the users could choose the USB
> configuration value which they want. The default is to support vendor mode
> only.
From: Hadar Hen Zion
Date: Wed, 7 Sep 2016 11:08:02 +0300
> This patchset introduces ip tunnel manipulation support using the TC
> subsystem.
Please address the feedback given by Eric Dumazet for patch #4,
thank you.
From: Andrei Vagin
Date: Tue, 6 Sep 2016 21:31:17 -0700
> This bug was detected by kmemleak:
> unreferenced object 0x8804269cc3c0 (size 64):
> comm "criu", pid 1042, jiffies 4294907360 (age 13.713s)
> hex dump (first 32 bytes):
> a0 32 cc 2c 04 88 ff ff 00 00 00 00 00 00 00 00 .2.,.
From: Florian Fainelli
Date: Tue, 6 Sep 2016 20:22:21 -0700
> Signed-off-by: Florian Fainelli
Since MAINTAINERS entry accuracy is very important for bug reporting
and patch CC:'ing, I tend to apply these kind of patches always to
'net'.
And I have done so in this case too.
Thanks.
From: "Francis Y. Yan"
Date: Tue, 6 Sep 2016 18:32:40 -0700
> diff --git a/include/linux/tcp.h b/include/linux/tcp.h
> index 7be9b12..f5b588e 100644
> --- a/include/linux/tcp.h
> +++ b/include/linux/tcp.h
> @@ -176,6 +176,7 @@ struct tcp_sock {
>* were acked.
>
From: Subash Abhinov Kasiviswanathan
Date: Tue, 6 Sep 2016 18:09:31 -0600
> proc_dointvec limits the values to INT_MAX in u32 sysctl entries.
> proc_douintvec allows to write upto UINT_MAX.
>
> Signed-off-by: Subash Abhinov Kasiviswanathan
I am assuming Steffen will pick this up.
From: David Howells
Date: Wed, 07 Sep 2016 16:25:50 +0100
>
> Here's a patch that adds a tracepoint that allows local aborts to be
> debugged. This needs to be applied on top of the just-posted call refcount
> overhaul patch.
>
> The patches can be found here also:
>
>
> http://git.ker
ged thusly:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git
> rxrpc-rewrite-20160907-1
Pulled.
On Thu, 2016-09-08 at 01:45 +0200, Michal Soltys wrote:
> > Keep in mind that TCP stack can send packets, responding to incoming
> > ACK.
> >
> > So you might check that incoming ACK are handled by the 'right' cpu.
> >
> > Without RFS, there is no such guarantee.
> Did some more testing today, i
On 2016-09-07 02:19, Eric Dumazet wrote:
> On Tue, 2016-09-06 at 23:00 +0200, Michal Soltys wrote:
>> On 2016-09-06 22:21, Alexander Duyck wrote:
>> > On Tue, Sep 6, 2016 at 11:46 AM, Michal Soltys wrote:
>> >> Hi,
>> >>
>> >> I've been testing different configurations and I didn't manage to get X
On Wed, Sep 07, 2016 at 01:35:35PM -0700, Stephen Hemminger wrote:
> On Wed, 7 Sep 2016 15:53:56 -0400
> Jarod Wilson wrote:
>
> > --- a/net/core/dev.c
> > +++ b/net/core/dev.c
> > @@ -6466,9 +6466,17 @@ int dev_set_mtu(struct net_device *dev, int new_mtu)
> > if (new_mtu == dev->mtu)
> >
On Wed, Sep 07, 2016 at 10:31:12PM +0200, Andrew Lunn wrote:
> Hi Jarod
>
> > - /* MTU must be positive.*/
> > - if (new_mtu < 0)
> > + if (new_mtu < dev->min_mtu) {
> > + netdev_err(dev, "Invalid MTU %d requested, hw min %d\n",
> > + new_mtu, dev->min
Hi,
On 09/07/2016 08:53 AM, Linus Walleij wrote:
On some systems (such as the Qualcomm APQ8060 Dragonboard) the
RESET signal of the SMSC911x is not pulled up by a resistor (or
the internal pull-up that will pull it up if the pin is not
even connected) but instead connected to a GPIO line, so tha
From: Zubair Lutfullah Kakakhel
Date: Mon, 5 Sep 2016 13:07:54 +0100
> The MIPS based xilfpga platform uses this driver.
> Enable it for MIPS
>
> Signed-off-by: Zubair Lutfullah Kakakhel
>
> ---
> V1 -> V6 are from a series that has gotten too big.
> So I have split this patch and am sending i
LLVM can generate code that tests for direct packet access via
skb->data/data_end in a way that currently gets rejected by the
verifier, example:
[...]
7: (61) r3 = *(u32 *)(r6 +80)
8: (61) r9 = *(u32 *)(r6 +76)
9: (bf) r2 = r9
10: (07) r2 += 54
11: (3d) if r3 >= r2 goto pc+12
R1
From: Or Gerlitz
Date: Wed, 7 Sep 2016 23:51:56 +0300
> On Wed, Sep 7, 2016 at 11:27 PM, David Miller wrote:
>> From: Saeed Mahameed
>> Date: Wed, 7 Sep 2016 23:16:13 +0300
>>
>>> Hi Dave, Sorry to bother, but i would like to drop this series for now
>>> Huy is working to define and come up wit
On Wed, 2016-09-07 at 15:26 -0700, Stephen Hemminger wrote:
> How much does this grow the size of tcp socket structure?
This actually shrinks it by 8 bytes, or more on debug kernels where
sizeof(spinlock_t) > 4
Before :
struct sk_buff_head out_of_order_queue; // At least 24 bytes on 64bit
A
On Wed, 07 Sep 2016 14:49:28 -0700
Eric Dumazet wrote:
> From: Yaogong Wang
>
> Over the years, TCP BDP has increased by several orders of magnitude,
> and some people are considering to reach the 2 Gbytes limit.
>
> Even with current window scale limit of 14, ~1 Gbytes maps to ~740,000
> MSS.
On Wed, Sep 7, 2016 at 11:55 PM, Or Gerlitz via iovisor-dev
wrote:
> On Wed, Sep 7, 2016 at 3:42 PM, Saeed Mahameed wrote:
>> From: Rana Shahout
>>
>> Add support for the BPF_PROG_TYPE_PHYS_DEV hook in mlx5e driver.
>>
>> When XDP is on we make sure to change channels RQs type to
>> MLX5_WQ_TYPE
From: Yaogong Wang
Over the years, TCP BDP has increased by several orders of magnitude,
and some people are considering to reach the 2 Gbytes limit.
Even with current window scale limit of 14, ~1 Gbytes maps to ~740,000
MSS.
In presence of packet losses (or reorders), TCP stores incoming p
On 09/06, Martin Blumenstingl wrote:
> This exposes the MPLL2 clock as this is one of the input clocks of the
> ethernet controller's internal mux.
>
> Signed-off-by: Martin Blumenstingl
> ---
Acked-by: Stephen Boyd
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux
On Wed, Sep 07, 2016 at 01:55:19PM -0700, Daniel Walker wrote:
> So the code only allows second granularity negative updates,
No.
> or the seconds
> component is the only part which needs to actually be negative ?
Yes.
See the function, clockadj_step, in linuxptp/clockadj.c for a working
user s
The netdevice type structure for macsec was being defined but never used.
To set the network device type the macro SET_NETDEV_DEVTYPE must be called.
Compile tested only, I don't use macsec.
Signed-off-by: Stephen Hemminger
--- a/drivers/net/macsec.c 2016-08-25 17:20:20.671031972 -0700
+++
On 09/07/2016 01:48 PM, Richard Cochran wrote:
On Wed, Sep 07, 2016 at 01:40:59PM -0700, Daniel Walker wrote:
There is a test (below) , which prevents negative nanosecond updates. The
code below would force a negative update to always return more than
NSEC_PER_SEC. It should be using abs() inste
This structure is defined but never used. Flagged with W=1
Signed-off-by: Stephen Hemminger
--- a/net/core/rtnetlink.c 2016-09-07 13:55:17.030107744 -0700
+++ b/net/core/rtnetlink.c 2016-09-07 13:55:45.190236855 -0700
@@ -3669,10 +3669,6 @@ nla_put_failure:
return -EMSGSIZE;
}
On Wed, Sep 7, 2016 at 3:42 PM, Saeed Mahameed wrote:
> From: Rana Shahout
>
> Add support for the BPF_PROG_TYPE_PHYS_DEV hook in mlx5e driver.
>
> When XDP is on we make sure to change channels RQs type to
> MLX5_WQ_TYPE_LINKED_LIST rather than "striding RQ" type to
> ensure "page per packet".
>
On Wed, Sep 7, 2016 at 11:27 PM, David Miller wrote:
> From: Saeed Mahameed
> Date: Wed, 7 Sep 2016 23:16:13 +0300
>
>> Hi Dave, Sorry to bother, but i would like to drop this series for now
>> Huy is working to define and come up with a better mechanism to
>> enable/disable the new DCBX hybrid m
On Wed, Sep 07, 2016 at 01:40:59PM -0700, Daniel Walker wrote:
> There is a test (below) , which prevents negative nanosecond updates. The
> code below would force a negative update to always return more than
> NSEC_PER_SEC. It should be using abs() instead which would return the value
> desired.
On Wed, 7 Sep 2016 15:53:56 -0400
Jarod Wilson wrote:
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -6466,9 +6466,17 @@ int dev_set_mtu(struct net_device *dev, int new_mtu)
> if (new_mtu == dev->mtu)
> return 0;
>
> - /* MTU must be positive.*/
> - if (
From: Saeed Mahameed
Date: Wed, 7 Sep 2016 23:16:13 +0300
> Hi Dave, Sorry to bother, but i would like to drop this series for now
> Huy is working to define and come up with a better mechanism to
> enable/disable the new DCBX hybrid mode he is adding.
I can't "just drop" any change that's in my
Hi Jarod
> - /* MTU must be positive.*/
> - if (new_mtu < 0)
> + if (new_mtu < dev->min_mtu) {
> + netdev_err(dev, "Invalid MTU %d requested, hw min %d\n",
> +new_mtu, dev->min_mtu);
> return -EINVAL;
> + }
> +
> + if (
On Tue, Sep 6, 2016 at 7:04 PM, Saeed Mahameed wrote:
> Hi Dave,
>
Hi Dave, Sorry to bother, but i would like to drop this series for now
Huy is working to define and come up with a better mechanism to
enable/disable the new DCBX hybrid mode he is adding.
Thanks for understanding,
Saeed.
> This
On Wed, Sep 7, 2016 at 9:19 PM, Eric Dumazet wrote:
> On Wed, 2016-09-07 at 19:57 +0300, Saeed Mahameed wrote:
>
>> Jesper has a similar Idea to make the qdisc think it is under
>> pressure, when the device
>> TX ring is idle most of the time, i think his idea can come in handy here.
>> I am not f
On 09/07/2016 10:34 AM, Eric Dumazet wrote:
From: Eric Dumazet
When DATA and/or FIN are carried in a SYN/ACK message or SYN message,
we append an skb in socket receive queue, but we forget to call
sk_forced_mem_schedule().
Effect is that the socket has a negative sk->sk_forward_alloc as long a
On Tue, Sep 06, 2016 at 04:55:29PM -0700, David Miller wrote:
> From: Jarod Wilson
> Date: Fri, 2 Sep 2016 13:07:42 -0400
>
> > In any case, the number of "mtu < 68" and "#define FOO_MIN_MTU 68", or
> > variations thereof, under drivers/net/ is kind of crazy.
>
> Agreed, we can have a default an
On Wed, 7 Sep 2016 15:42:25 +0300 Saeed Mahameed wrote:
> For non-striding RQ configuration before this patch we had a ring
> with pre-allocated SKBs and mapped the SKB->data buffers for
> device.
>
> For robustness and better RX data buffers management, we allocate a
> page per packet and bui
On Wed, 7 Sep 2016 15:42:22 +0300 Saeed Mahameed wrote:
> From: Tariq Toukan
>
> To improve the memory consumption scheme, we omit the flow that
> demands and splits high-order pages in Striding RQ, and stay
> with a single Striding RQ flow that uses order-0 pages.
Thanks you for doing this!
Due to a mistake in the hardware manual, the FWSLC and POST1-4 registers
were not documented and left out of the driver for RZ/A making the CAM
feature non-operational.
Additionally, when the offset values for POST1-4 are left blank, the driver
attempts to set them using an offset of 0x which c
1 - 100 of 231 matches
Mail list logo
