* Re: [PATCH v3] xmit_compl_seq: information to reclaim vmsplice buffers
From: Michael S. Tsirkin @ 2010-09-27 17:23 UTC (permalink / raw)
To: Tom Herbert; +Cc: netdev, davem, sridharr
In-Reply-To: <alpine.DEB.1.00.1009231426450.11579@pokey.mtv.corp.google.com>
On Thu, Sep 23, 2010 at 02:35:16PM -0700, Tom Herbert wrote:
> In this patch we propose to adds some socket API to retrieve the
> "transmit completion sequence number", essentially a byte counter
> for the number of bytes that have been transmitted and will not be
> retransmitted. In the case of TCP, this should correspond to snd_una.
>
> The purpose of this API is to provide information to userspace about
> which buffers can be reclaimed when sending with vmsplice() on a
> socket.
>
> There are two methods for retrieving the completed sequence number:
> through a simple getsockopt (implemented here for TCP), as well as
> returning the value in the ancilary data of a recvmsg.
>
> The expected flow would be something like:
> - Connect is created
> - Initial completion seq # is retrieved through the sockopt, and is
> stored in userspace "compl_seq" variable for the connection.
> - Whenever a send is done, compl_seq += # bytes sent.
> - When doing a vmsplice the completion sequence number is saved
> for each user space buffer, buffer_compl_seq = compl_seq.
> - When recvmsg returns with a completion sequence number in
> ancillary data, any buffers cover by that sequence number
> (where buffer_compl_seq < recvmsg_compl_seq) are reclaimed
> and can be written to again.
> - If no data is receieved on a connection (recvmsg does not
> return), a timeout can be used to call the getsockopt and
> reclaim buffers as a fallback.
>
> Using recvmsg data in this manner is sort of a cheap way to get a
> "callback" for when a vmspliced buffer is consumed. It will work
> well for a client where the response causes recvmsg to return.
> On the server side it works well if there are a sufficient
> number of requests coming on the connection (resorting to the
> timeout if necessary as described above).
>
> Signed-off-by: Tom Herbert <therbert@google.com>
> Signed-off-by: Sridhar Raman <sridharr@google.com>
Can not packets referencing this memory
still be outstanding at the NIC device, if retransmit happens
before the ack but after the packet was passed to a device?
It's true that the reftransmit will likely get discarded
by the remote end, but this might be a security issue
if an application puts sensitive data in the buffer
and that gets inadvertently sent on the wire, can it not?
--
MST
^ permalink raw reply
* [PATCH linux-2.6 v2] IPv6: Temp addresses are immediately deleted.
From: Glenn Wurster @ 2010-09-27 17:10 UTC (permalink / raw)
To: David S. Miller, Alexey Kuznetsov, Pekka Savola (ipv6),
James Morris, Hideaki YOSHIFUJI <yosh
Cc: netdev, linux-kernel
There is a bug in the interaction between ipv6_create_tempaddr and
addrconf_verify. Because ipv6_create_tempaddr uses the cstamp and tstamp
from the public address in creating a private address, if we have not
received a router advertisement in a while, tstamp + temp_valid_lft might be
< now. If this happens, the new address is created inside
ipv6_create_tempaddr, then the loop within addrconf_verify starts again and
the address is immediately deleted. We are left with no temporary addresses
on the interface, and no more will be created until the public IP address is
updated. To avoid this, set the expiry time to be the minimum of the time
left on the public address or the config option PLUS the current age of the
public interface.
Version 2, now with 100% fewer line wraps. Thanks to David Miller for
pointing out the line wrapping issue.
Signed-off-by: Glenn Wurster <gwurster@scs.carleton.ca>
---
net/ipv6/addrconf.c | 7 ++++---
1 files changed, 4 insertions(+), 3 deletions(-)
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index cfee6ae..9c74454 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -836,7 +836,7 @@ static int ipv6_create_tempaddr(struct inet6_ifaddr *ifp, struct inet6_ifaddr *i
{
struct inet6_dev *idev = ifp->idev;
struct in6_addr addr, *tmpaddr;
- unsigned long tmp_prefered_lft, tmp_valid_lft, tmp_cstamp, tmp_tstamp;
+ unsigned long tmp_prefered_lft, tmp_valid_lft, tmp_cstamp, tmp_tstamp, age;
unsigned long regen_advance;
int tmp_plen;
int ret = 0;
@@ -886,12 +886,13 @@ retry:
goto out;
}
memcpy(&addr.s6_addr[8], idev->rndid, 8);
+ age = (jiffies - ifp->tstamp) / HZ;
tmp_valid_lft = min_t(__u32,
ifp->valid_lft,
- idev->cnf.temp_valid_lft);
+ idev->cnf.temp_valid_lft + age);
tmp_prefered_lft = min_t(__u32,
ifp->prefered_lft,
- idev->cnf.temp_prefered_lft -
+ idev->cnf.temp_prefered_lft + age -
idev->cnf.max_desync_factor);
tmp_plen = ifp->prefix_len;
max_addresses = idev->cnf.max_addresses;
--
1.5.6.5
^ permalink raw reply related
* Re: [PATCH v6 0/8] ptp: IEEE 1588 hardware clock support
From: Alan Cox @ 2010-09-27 17:05 UTC (permalink / raw)
To: Christoph Lameter
Cc: John Stultz, Rodolfo Giometti, Arnd Bergmann, Peter Zijlstra,
linux-api, devicetree-discuss, linux-kernel, David Miller, netdev,
Thomas Gleixner, linuxppc-dev, Richard Cochran, linux-arm-kernel,
Krzysztof Halasa
In-Reply-To: <alpine.DEB.2.00.1009271054320.9258@router.home>
On Mon, 27 Sep 2010 10:56:09 -0500 (CDT)
Christoph Lameter <cl@linux.com> wrote:
>
> On Fri, 24 Sep 2010, Alan Cox wrote:
>
> > Whether you add new syscalls or do the fd passing using flags and hide
> > the ugly bits in glibc is another question.
>
> Use device specific ioctls instead of syscalls?
Some of the ioctls are probably not device specific, the job of the OS in
part is to present a unified interface. We already have a mess of HPET
and RTC driver ioctls.
Some of it undoubtedly is device specific.
Alan
^ permalink raw reply
* [PATCH linux-2.6 v2] IPv6: Create temporary address if none exists.
From: Glenn Wurster @ 2010-09-27 17:04 UTC (permalink / raw)
To: David S. Miller, Alexey Kuznetsov, Pekka Savola (ipv6),
James Morris, Hideaki YOSHIFUJI <yosh
Cc: netdev, linux-kernel
If privacy extentions are enabled, but no current temporary address exists,
then create one when we get a router advertisement.
Version 2, now with 100% fewer line wraps. Thanks to David Miller for
pointing out the line wrapping issue.
Signed-off-by: Glenn Wurster <gwurster@scs.carleton.ca>
---
net/ipv6/addrconf.c | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index ab70a3f..cfee6ae 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -2022,10 +2022,11 @@ ok:
ipv6_ifa_notify(0, ift);
}
- if (create && in6_dev->cnf.use_tempaddr > 0) {
+ if ((create || list_empty(&in6_dev->tempaddr_list)) && in6_dev->cnf.use_tempaddr > 0) {
/*
* When a new public address is created as described in [ADDRCONF],
- * also create a new temporary address.
+ * also create a new temporary address. Also create a temporary
+ * address if it's enabled but no temporary address currently exists.
*/
read_unlock_bh(&in6_dev->lock);
ipv6_create_tempaddr(ifp, NULL);
--
1.5.6.5
^ permalink raw reply related
* Re: net-next rcu_dereference_check() warning
From: Eric Dumazet @ 2010-09-27 16:45 UTC (permalink / raw)
To: Dimitris Michailidis; +Cc: netdev
In-Reply-To: <4CA0C691.1090001@chelsio.com>
Le lundi 27 septembre 2010 à 09:30 -0700, Dimitris Michailidis a écrit :
> With today's net-next I am getting the below. It appears to be recent, it
> wasn't happening with the net-next I was using last week.
>
> ===================================================
> [ INFO: suspicious rcu_dereference_check() usage. ]
> ---------------------------------------------------
> /build/net-next-2.6/include/linux/netpoll.h:67 invoked
> rcu_dereference_check() without protection!
>
> other info that might help us debug this:
>
>
> rcu_scheduler_active = 1, debug_locks = 0
> 1 lock held by ypbind/5046:
> #0: (sk_lock-AF_INET){+.+.+.}, at: [<ffffffff81503139>]
> udp_sendmsg+0x3fd/0x5dc
>
> stack backtrace:
> Pid: 5046, comm: ypbind Not tainted 2.6.36-rc3+ #16
> Call Trace:
> [<ffffffff810668e4>] lockdep_rcu_dereference+0xaa/0xb3
> [<ffffffff814b8d09>] netif_rx+0x53/0x1cf
> [<ffffffff810313b7>] ? get_parent_ip+0x11/0x41
> [<ffffffff814b9068>] netif_rx_ni+0x23/0x5d
> [<ffffffff814e57a9>] ip_dev_loopback_xmit+0xc5/0xcb
> [<ffffffff814e593c>] ip_mc_output+0x18d/0x1a5
> [<ffffffff814e30ec>] ip_local_out+0x62/0x66
> [<ffffffff814e33cb>] ip_push_pending_frames+0x2db/0x337
> [<ffffffff81501b55>] udp_push_pending_frames+0x2de/0x34c
> [<ffffffff81503139>] ? udp_sendmsg+0x3fd/0x5dc
> [<ffffffff8150322a>] udp_sendmsg+0x4ee/0x5dc
> [<ffffffff8150aa74>] inet_sendmsg+0x13a/0x145
> [<ffffffff8150a93a>] ? inet_sendmsg+0x0/0x145
> [<ffffffff814a6d24>] sock_sendmsg+0xc0/0xd9
> [<ffffffff8106aad4>] ? lock_release_non_nested+0xe5/0x296
> [<ffffffff810c8fb0>] ? might_fault+0x68/0xb8
> [<ffffffff810c8fb0>] ? might_fault+0x68/0xb8
> [<ffffffff810c8ff9>] ? might_fault+0xb1/0xb8
> [<ffffffff810c8fb0>] ? might_fault+0x68/0xb8
> [<ffffffff814a6672>] ? move_addr_to_kernel+0x3f/0x56
> [<ffffffff814a7458>] sys_sendto+0xe0/0x108
> [<ffffffff810f9a76>] ? d_kill+0x6c/0x74
> [<ffffffff810e99df>] ? fput+0x1ce/0x1dd
> [<ffffffff81002c9b>] system_call_fastpath+0x16/0x1b
Hi Dimitris
Fix is known and under its way.
http://kerneltrap.org/mailarchive/linux-kernel/2010/9/23/4623037
Thanks !
^ permalink raw reply
* net-next rcu_dereference_check() warning
From: Dimitris Michailidis @ 2010-09-27 16:30 UTC (permalink / raw)
To: netdev
With today's net-next I am getting the below. It appears to be recent, it
wasn't happening with the net-next I was using last week.
===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
/build/net-next-2.6/include/linux/netpoll.h:67 invoked
rcu_dereference_check() without protection!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 0
1 lock held by ypbind/5046:
#0: (sk_lock-AF_INET){+.+.+.}, at: [<ffffffff81503139>]
udp_sendmsg+0x3fd/0x5dc
stack backtrace:
Pid: 5046, comm: ypbind Not tainted 2.6.36-rc3+ #16
Call Trace:
[<ffffffff810668e4>] lockdep_rcu_dereference+0xaa/0xb3
[<ffffffff814b8d09>] netif_rx+0x53/0x1cf
[<ffffffff810313b7>] ? get_parent_ip+0x11/0x41
[<ffffffff814b9068>] netif_rx_ni+0x23/0x5d
[<ffffffff814e57a9>] ip_dev_loopback_xmit+0xc5/0xcb
[<ffffffff814e593c>] ip_mc_output+0x18d/0x1a5
[<ffffffff814e30ec>] ip_local_out+0x62/0x66
[<ffffffff814e33cb>] ip_push_pending_frames+0x2db/0x337
[<ffffffff81501b55>] udp_push_pending_frames+0x2de/0x34c
[<ffffffff81503139>] ? udp_sendmsg+0x3fd/0x5dc
[<ffffffff8150322a>] udp_sendmsg+0x4ee/0x5dc
[<ffffffff8150aa74>] inet_sendmsg+0x13a/0x145
[<ffffffff8150a93a>] ? inet_sendmsg+0x0/0x145
[<ffffffff814a6d24>] sock_sendmsg+0xc0/0xd9
[<ffffffff8106aad4>] ? lock_release_non_nested+0xe5/0x296
[<ffffffff810c8fb0>] ? might_fault+0x68/0xb8
[<ffffffff810c8fb0>] ? might_fault+0x68/0xb8
[<ffffffff810c8ff9>] ? might_fault+0xb1/0xb8
[<ffffffff810c8fb0>] ? might_fault+0x68/0xb8
[<ffffffff814a6672>] ? move_addr_to_kernel+0x3f/0x56
[<ffffffff814a7458>] sys_sendto+0xe0/0x108
[<ffffffff810f9a76>] ? d_kill+0x6c/0x74
[<ffffffff810e99df>] ? fput+0x1ce/0x1dd
[<ffffffff81002c9b>] system_call_fastpath+0x16/0x1b
^ permalink raw reply
* Re: [PATCH RFC] netns: keep vlan slaves on master netns move
From: Eric W. Biederman @ 2010-09-27 16:23 UTC (permalink / raw)
To: Patrick McHardy; +Cc: David Miller, netdev, Daniel Lezcano, David Lamparter
In-Reply-To: <4C9363CB.7050302@trash.net>
Patrick McHardy <kaber@trash.net> writes:
> Am 15.09.2010 04:10, schrieb Eric W. Biederman:
>>
>> Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com>
>>
>> My inclination is that the best way to handle this is to would be to
>> push the device deletion for vlan and macvlan devices into the device
>> core, where we could get the advantage of batched deletions.
>
> We've added batched deletion to both about a year ago, what exactly
> is the problem?
My problem is that while the both have dellink implemented we
don't really batch their deletes as best we can.
For macvlan when the underlying device is unregistered we delete
each macvlan device one by one.
For vlans we do batch the deletes, but not in the same batch as
the underlying device.
>> Right
>> now vlan and macvlan devices are the only devices I know of that cause
>> other devices to be removed during unregister, so removing that
>> specialness seems reasonable.
>
> Actually all devices can cause this when used as a lower device by
> vlan or macvlan. Both vlan and macvlan are useless without a lower
> device, so I don't see why we shouldn't remove them when the lower
> device is unregistered.
I definitely think we should remove the devices when the lower device
is destroyed/removed. However unregistered is a slightly different
concept, and arguably macvlan and vlan devices are much more intimately
connected to their underlying device than that.
What I meant is we should implement the deletion differently. By doing
something like having a list of all derived devices for a device, that
we could walk and call dellink on from rollback_registered_many. In
theory that could destroy a whole pile of macvlan and vlan devices with
various different stackings one on the other all in the same batch.
Eric
^ permalink raw reply
* Re: l2tpv3 support not yet in iproute2?
From: Eric Dumazet @ 2010-09-27 16:23 UTC (permalink / raw)
To: jpo234; +Cc: netdev, James Chapman
In-Reply-To: <loom.20100927T170521-559@post.gmane.org>
Le lundi 27 septembre 2010 à 15:22 +0000, jpo234 a écrit :
> Hello all,
> some time ago James Chapman sent patches that add L2TPv3 support to iproute2:
> http://kerneltrap.org/mailarchive/linux-netdev/2010/3/18/6272266/thread
> While the kernel support arrived in 2.6.35 the user space configuration support
> still seems to be missing:
> iproute2-2.6.35$ find . -type f -iname "*.c" | xargs fgrep -i l2tp
> doesn't find anything.
>
> Can somebody clarify the current situation?
Hmm, this reminds me l2tp locking is buggy and not yet fixed.
http://kerneltrap.org/mailarchive/linux-netdev/2010/4/20/6275106
James, any chance you could fix the locking ?
^ permalink raw reply
* Re: [PATCH v6 0/8] ptp: IEEE 1588 hardware clock support
From: M. Warner Losh @ 2010-09-27 16:14 UTC (permalink / raw)
To: cl
Cc: johnstul, giometti, peterz, linux-api, devicetree-discuss,
linux-kernel, davem, netdev, tglx, linuxppc-dev, richardcochran,
linux-arm-kernel, khc
In-Reply-To: <alpine.DEB.2.00.1009271038150.9258@router.home>
In message: <alpine.DEB.2.00.1009271038150.9258@router.home>
Christoph Lameter <cl@linux.com> writes:
: On Thu, 23 Sep 2010, john stultz wrote:
: > The design actually avoids most userland induced latency.
: >
: > 1) On the PTP hardware syncing point, the reference packet gets
: > timestamped with the PTP hardware time on arrival. This allows the
: > offset calculation to be done in userland without introducing latency.
:
: The timestamps allows the calculation of the network transmission time I
: guess and therefore its more accurate to calculate that effect out. Ok but
: then the overhead of getting to code in user space (that does the proper
: clock adjustments) is resulting in the addition of a relatively long time
: that is subject to OS scheduling latencies and noises.
The timestamps at the hardware level allow you to factor out variation
caused by OS Scheduling, OS network stack delay and internal buffering
on the NIC. Variation in measurements is what kills accuracy.
When steering a clock by making an error measurement of the phase and
frequency of it, the latency induced by OS scheduling tends to be
unimportant. It is far more important to know when you steered the
clock (called adjtime or friends) than to steer it at any fixed
latency to when the data for the measurements was made. Measuring the
time of steer can tolerate errors in the range of OS scheduling
latencies easily, since that tends to produce a very small effect. It
introduces an error in your expected phase for the next measurement on
the order of the product of the time of steer error times the change
in fractional frequency (abs( 1 - (nu_new / nu_old))). Even if the
estimate is really bad at 100ms, most steers are on the order about
one part per million. This leads to a sub-nanosecond phase error
estimate in the next measurement cycle (a non-accumulating error). A
1ms error leads to maybe tens of picoseconds of estimate error.
This is a common error that I've seen repeated in this thread. The
only reason that it has historically been important is because when
you are doing timestamping in software based on an interrupt, that
stuff does matter.
Warner
^ permalink raw reply
* [PATCH net-next-2.6 v2] net: sk_{detach|attach}_filter() rcu fixes
From: Eric Dumazet @ 2010-09-27 16:07 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Paul E. McKenney
In-Reply-To: <1285603064.3017.42.camel@edumazet-laptop>
Le lundi 27 septembre 2010 à 17:57 +0200, Eric Dumazet a écrit :
> sk_attach_filter() is run with socket locked.
>
> Use the appropriate rcu_dereference_protected() instead of blocking BH,
> and rcu_dereference_bh().
> There is no point adding BH prevention and memory barrier.
Hmm, same thing can be done in sk_detach_filter, here is a v2.
Thanks
[PATCH net-next-2.6 v2] net: sk_{detach|attach}_filter() rcu fixes
sk_attach_filter() and sk_detach_filter() are run with socket locked.
Use the appropriate rcu_dereference_protected() instead of blocking BH,
and rcu_dereference_bh().
There is no point adding BH prevention and memory barrier.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
net/core/filter.c | 10 ++++------
1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/net/core/filter.c b/net/core/filter.c
index 52b051f..7adf503 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -638,10 +638,9 @@ int sk_attach_filter(struct sock_fprog *fprog, struct sock *sk)
return err;
}
- rcu_read_lock_bh();
- old_fp = rcu_dereference_bh(sk->sk_filter);
+ old_fp = rcu_dereference_protected(sk->sk_filter,
+ sock_owned_by_user(sk));
rcu_assign_pointer(sk->sk_filter, fp);
- rcu_read_unlock_bh();
if (old_fp)
sk_filter_delayed_uncharge(sk, old_fp);
@@ -654,14 +653,13 @@ int sk_detach_filter(struct sock *sk)
int ret = -ENOENT;
struct sk_filter *filter;
- rcu_read_lock_bh();
- filter = rcu_dereference_bh(sk->sk_filter);
+ filter = rcu_dereference_protected(sk->sk_filter,
+ sock_owned_by_user(sk));
if (filter) {
rcu_assign_pointer(sk->sk_filter, NULL);
sk_filter_delayed_uncharge(sk, filter);
ret = 0;
}
- rcu_read_unlock_bh();
return ret;
}
EXPORT_SYMBOL_GPL(sk_detach_filter);
^ permalink raw reply related
* [PATCH net-next-2.6] net: sk_attach_filter() rcu fix
From: Eric Dumazet @ 2010-09-27 15:57 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Paul E. McKenney
sk_attach_filter() is run with socket locked.
Use the appropriate rcu_dereference_protected() instead of blocking BH,
and rcu_dereference_bh().
There is no point adding BH prevention and memory barrier.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
net/core/filter.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/net/core/filter.c b/net/core/filter.c
index 52b051f..93ac5d2 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -638,10 +638,9 @@ int sk_attach_filter(struct sock_fprog *fprog, struct sock *sk)
return err;
}
- rcu_read_lock_bh();
- old_fp = rcu_dereference_bh(sk->sk_filter);
+ old_fp = rcu_dereference_protected(sk->sk_filter,
+ sock_owned_by_user(sk));
rcu_assign_pointer(sk->sk_filter, fp);
- rcu_read_unlock_bh();
if (old_fp)
sk_filter_delayed_uncharge(sk, old_fp);
^ permalink raw reply related
* Re: [PATCH v6 0/8] ptp: IEEE 1588 hardware clock support
From: Christoph Lameter @ 2010-09-27 15:56 UTC (permalink / raw)
To: Alan Cox
Cc: Richard Cochran, linux-kernel, devicetree-discuss, linux-api,
linux-arm-kernel, linuxppc-dev, netdev, Arnd Bergmann,
David Miller, John Stultz, Krzysztof Halasa, Peter Zijlstra,
Rodolfo Giometti, Thomas Gleixner
In-Reply-To: <20100924150246.0e6064b6@lxorguk.ukuu.org.uk>
On Fri, 24 Sep 2010, Alan Cox wrote:
> Whether you add new syscalls or do the fd passing using flags and hide
> the ugly bits in glibc is another question.
Use device specific ioctls instead of syscalls?
^ permalink raw reply
* Re: linux-next: build failure after merge of the final tree (net tree related)
From: Ohad Ben-Cohen @ 2010-09-27 15:55 UTC (permalink / raw)
To: Stephen Rothwell
Cc: David Miller, netdev, linux-next, linux-kernel, John W. Linville
In-Reply-To: <20100927154408.55a28127.sfr@canb.auug.org.au>
On Mon, Sep 27, 2010 at 7:44 AM, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> Assembler messages:
> Fatal error: can't create drivers/net/wireless/wl12xx/.tmp_wl12xx_platform_data.o: No such file or directory
> drivers/net/wireless/wl12xx/wl12xx_platform_data.c: In function 'wl12xx_get_platform_data':
> drivers/net/wireless/wl12xx/wl12xx_platform_data.c:28: error: cannot open drivers/net/wireless/wl12xx/.tmp_wl12xx_platform_data.gcno
> drivers/net/wireless/wl12xx/wl12xx_platform_data.c:28: confused by earlier errors, bailing out
>
> Presumably caused by commit 61ee7007a5d61aa066076da578e8e8084e122d7d
> ("wl12xx: add platform data passing support").
>
> I do my builds with a separate object directory (which may be a reason
> you don't see this).
Thank you, Stephen.
I have just tried building with a separate object directory, and I was
then able to reproduce it and fix it with:
From fb0d1dc5ce6c258b3ecb0a8997791a77be3c5307 Mon Sep 17 00:00:00 2001
From: Ohad Ben-Cohen <ohad@wizery.com>
Date: Mon, 27 Sep 2010 17:33:57 +0200
Subject: [PATCH] wl12xx: fix separate-object-folder builds
Make this go away (happens when building with a separate object directory):
Assembler messages:
Fatal error: can't create
drivers/net/wireless/wl12xx/.tmp_wl12xx_platform_data.o: No such file
or directory
drivers/net/wireless/wl12xx/wl12xx_platform_data.c: In function
'wl12xx_get_platform_data':
drivers/net/wireless/wl12xx/wl12xx_platform_data.c:28: error: cannot
open drivers/net/wireless/wl12xx/.tmp_wl12xx_platform_data.gcno
drivers/net/wireless/wl12xx/wl12xx_platform_data.c:28: confused by
earlier errors, bailing out
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
---
drivers/net/wireless/Makefile | 2 +-
.../wireless/{wl12xx => }/wl12xx_platform_data.c | 0
2 files changed, 1 insertions(+), 1 deletions(-)
rename drivers/net/wireless/{wl12xx => }/wl12xx_platform_data.c (100%)
diff --git a/drivers/net/wireless/Makefile b/drivers/net/wireless/Makefile
index 85af697..0a64bd5 100644
--- a/drivers/net/wireless/Makefile
+++ b/drivers/net/wireless/Makefile
@@ -51,6 +51,6 @@ obj-$(CONFIG_MAC80211_HWSIM) += mac80211_hwsim.o
obj-$(CONFIG_WL12XX) += wl12xx/
# small builtin driver bit
-obj-$(CONFIG_WL12XX_PLATFORM_DATA) += wl12xx/wl12xx_platform_data.o
+obj-$(CONFIG_WL12XX_PLATFORM_DATA) += wl12xx_platform_data.o
obj-$(CONFIG_IWM) += iwmc3200wifi/
diff --git a/drivers/net/wireless/wl12xx/wl12xx_platform_data.c
b/drivers/net/wireless/wl12xx_platform_data.c
similarity index 100%
rename from drivers/net/wireless/wl12xx/wl12xx_platform_data.c
rename to drivers/net/wireless/wl12xx_platform_data.c
--
1.7.0.4
>
> I have reverted that commit for today (and commits
> 09cecc340b3b4d9960b039c0f576548bbf857f5a ("wl1271: take irq info from
> private board data") and 15cea99306ae14ce5f7c3d3989bcc17202e2b0be
> ("wl1271: make ref_clock configurable by board") which follow it).
> --
> Cheers,
> Stephen Rothwell sfr@canb.auug.org.au
> http://www.canb.auug.org.au/~sfr/
>
^ permalink raw reply related
* Re: [PATCH v6 0/8] ptp: IEEE 1588 hardware clock support
From: Christoph Lameter @ 2010-09-27 15:52 UTC (permalink / raw)
To: john stultz
Cc: Rodolfo Giometti, Peter Zijlstra,
linux-api-u79uwXL29TY76Z2rM5mHXA,
devicetree-discuss-uLR06cmDAlY/bJ5BZ2RsiQ,
linux-kernel-u79uwXL29TY76Z2rM5mHXA, David Miller,
netdev-u79uwXL29TY76Z2rM5mHXA, Thomas Gleixner,
linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ, Richard Cochran,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
Krzysztof Halasa
In-Reply-To: <1285278136.2587.154.camel-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org>
On Thu, 23 Sep 2010, john stultz wrote:
> > > 3) Further, the PTP hardware counter can be simply set to a new offset
> > > to put it in line with the network time. This could cause trouble with
> > > timekeeping much like unsynced TSCs do.
> >
> > You can do the same for system time.
>
> Settimeofday does allow CLOCK_REALTIME to jump, but the CLOCK_MONOTONIC
> time cannot jump around. Having a clocksource that is non-monotonic
> would break this.
Currently time runs at the same speed. CLOCK_MONOTONIC runs at a offset
to CLOCK_REALTIME. We are creating APIs here that allow time to run at
different speeds.
> The design actually avoids most userland induced latency.
>
> 1) On the PTP hardware syncing point, the reference packet gets
> timestamped with the PTP hardware time on arrival. This allows the
> offset calculation to be done in userland without introducing latency.
The timestamps allows the calculation of the network transmission time I
guess and therefore its more accurate to calculate that effect out. Ok but
then the overhead of getting to code in user space (that does the proper
clock adjustments) is resulting in the addition of a relatively long time
that is subject to OS scheduling latencies and noises.
> 2) On the system syncing side, the proposal for the PPS interrupt allows
> the PTP hardware to trigger an interrupt on the second boundary that
> would take a timestamp of the system time. Then the pps interface allows
> for the timestamp to be read from userland allowing the offset to be
> calculated without introducing additional latency.
Sorry dont really get the whole picture here it seems. Sounds like one is
going through additional unnecessary layers. Why would the PTP hardware
triggger an interrupt? I thought the PTP messages came in via
timestamping and are then processed by software. Then the software is
issuing a hardware interrupt that then triggers the PPS subsystem. And
that is supposed to be better than directly interfacing with the PTP?
> Additionally, even just in userland, it would be easy to bracket two
> reads of the system time around one read of the PTP clock to bound any
> userland latency fairly well. It may not be as good as the PPS interface
> (although that depends on the interrupt latency), but if the accesses
> are all local, it probably could get fairly close.
That sounds hacky.
> > Ok maybe we need some sort of control interface to manage the clock like
> > the others have.
>
> That's what the clock_adjtime call provides.
Ummm... You are managing a hardware device with hardware (driver) specific
settings. That is currently being done via ioctls. Why generalize it?
> > The posix clocks today assumes one notion of real "time" in the kernel.
> > All clocks increase in lockstep (aside from offset updates).
>
> Not true. The cputime clockids do not increment at the same rate (as the
> apps don't always run). Further CLOCK_MONOTONIC_RAW provides a non-freq
> corrected view of CLOCK_MONOTONIC, so it increments at a slightly
> different rate.
cputime clockids are not tracking time but cpu resource use.
> Re-using the fairly nice (Alan of course disagrees :) posix interface
> seems at least a little better for application developers who actually
> have to use the hardware.
Well it may also be confusing for others. The application developers also
will have a hard time using a generic clock interface to control PTP
device specific things like frequencies, rates etc etc. So you always need
to ioctl/device specific control interface regardless.
^ permalink raw reply
* l2tpv3 support not yet in iproute2?
From: jpo234 @ 2010-09-27 15:22 UTC (permalink / raw)
To: netdev
Hello all,
some time ago James Chapman sent patches that add L2TPv3 support to iproute2:
http://kerneltrap.org/mailarchive/linux-netdev/2010/3/18/6272266/thread
While the kernel support arrived in 2.6.35 the user space configuration support
still seems to be missing:
iproute2-2.6.35$ find . -type f -iname "*.c" | xargs fgrep -i l2tp
doesn't find anything.
Can somebody clarify the current situation?
Thanks in advance
Joerg
^ permalink raw reply
* Re: [PATCH v6 0/8] ptp: IEEE 1588 hardware clock support
From: Christoph Lameter @ 2010-09-27 15:37 UTC (permalink / raw)
To: Christian Riesch
Cc: Alan Cox, john stultz, Richard Cochran,
linux-kernel-u79uwXL29TY76Z2rM5mHXA,
devicetree-discuss-uLR06cmDAlY/bJ5BZ2RsiQ,
linux-api-u79uwXL29TY76Z2rM5mHXA,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ,
netdev-u79uwXL29TY76Z2rM5mHXA, Arnd Bergmann, David Miller,
Krzysztof Halasa, Peter Zijlstra, Rodolfo Giometti,
Thomas Gleixner
In-Reply-To: <4C9BC7CE.8020400-U4+xqT1Vg0VeoWH0uzbU5w@public.gmane.org>
On Thu, 23 Sep 2010, Christian Riesch wrote:
> > > It implies clock tuning in userspace for a potential sub microsecond
> > > accurate clock. The clock accuracy will be limited by user space
> > > latencies and noise. You wont be able to discipline the system clock
> > > accurately.
> >
> > Noise matters, latency doesn't.
>
> Well put! That's why we need hardware support for PTP timestamping to reduce
> the noise, but get along well with the clock servo that is steering the PHC in
> user space.
Even if I buy into the catch phrase above: User space is subject to noise
that the in kernel code is not. If you do the tuning over long intervals
then it hopefully averages out but it still causes jitter effects that
affects the degree of accuracy (or sync) that you can reach. And the noise
varies with the load on the system.
^ permalink raw reply
* [PATCH net-next-2.6] fib: use atomic_inc_not_zero() in fib_rules_lookup
From: Eric Dumazet @ 2010-09-27 14:18 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Paul E. McKenney
It seems we dont use appropriate refcount increment in an
rcu_read_lock() protected section.
fib_rule_get() might increment a null refcount and bad things could
happen.
While fib_nl_delrule() respects an rcu grace period before calling
fib_rule_put(), fib_rules_cleanup_ops() calls fib_rule_put() without a
grace period.
Note : after this patch, we might avoid the synchronize_rcu() call done
in fib_nl_delrule()
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
net/core/fib_rules.c | 8 +++++---
1 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index 42e84e0..910eac3 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -225,9 +225,11 @@ jumped:
err = ops->action(rule, fl, flags, arg);
if (err != -EAGAIN) {
- fib_rule_get(rule);
- arg->rule = rule;
- goto out;
+ if (likely(atomic_inc_not_zero(&rule->refcnt))) {
+ arg->rule = rule;
+ goto out;
+ }
+ break;
}
}
^ permalink raw reply related
* [patch] ipvs: Keep track of backlog connections
From: Simon Horman @ 2010-09-27 14:05 UTC (permalink / raw)
To: lvs-devel, netfilter-devel, netdev
Cc: Sven Wegener, Wensong Zhang, Julian Anastasov,
Venkata Mohan Reddy Koppula, Patrick McHardy
From: Sven Wegener <sven.wegener@stealer.net>
A backlog connection is a connection that is on its way from inactive to
active. Speaking in TCP language, a connection from which we've seen the
initial SYN packet, but the three-way handshake hasn't finished yet.
These connections are expected to move to active soon. When a
destination is overloaded or isn't able to successfully establish
connections for various reasons, this count increases quickly and is an
indication for a problem.
Signed-off-by: Sven Wegener <sven.wegener@stealer.net>
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Acked-by: Simon Horman <horms@verge.net.au>
---
Patrick, please consider this for nf-next
diff -urp net-next-2.6-e548833-nfct/linux/include/linux/ip_vs.h linux/include/linux/ip_vs.h
--- net-next-2.6-e548833-nfct/linux/include/linux/ip_vs.h 2010-09-15 11:28:02.000000000 +0300
+++ linux/include/linux/ip_vs.h 2010-09-26 15:21:53.507865229 +0300
@@ -91,6 +91,7 @@
/* Flags that are not sent to backup server start from bit 16 */
#define IP_VS_CONN_F_NFCT (1 << 16) /* use netfilter conntrack */
+#define IP_VS_CONN_F_BACKLOG (1 << 17) /* backlog connection */
/* Connection flags from destination that can be changed by user space */
#define IP_VS_CONN_F_DEST_MASK (IP_VS_CONN_F_FWD_MASK | \
@@ -360,6 +361,7 @@ enum {
IPVS_DEST_ATTR_PERSIST_CONNS, /* persistent connections */
IPVS_DEST_ATTR_STATS, /* nested attribute for dest stats */
+ IPVS_DEST_ATTR_BACKLOG_CONNS, /* backlog connections */
__IPVS_DEST_ATTR_MAX,
};
diff -urp net-next-2.6-e548833-nfct/linux/include/net/ip_vs.h linux/include/net/ip_vs.h
--- net-next-2.6-e548833-nfct/linux/include/net/ip_vs.h 2010-09-16 08:53:04.000000000 +0300
+++ linux/include/net/ip_vs.h 2010-09-26 15:31:27.369865994 +0300
@@ -501,6 +501,7 @@ struct ip_vs_dest {
/* connection counters and thresholds */
atomic_t activeconns; /* active connections */
atomic_t inactconns; /* inactive connections */
+ atomic_t backlogconns; /* backlog connections */
atomic_t persistconns; /* persistent connections */
__u32 u_threshold; /* upper threshold */
__u32 l_threshold; /* lower threshold */
diff -urp net-next-2.6-e548833-nfct/linux/net/netfilter/ipvs/ip_vs_conn.c linux/net/netfilter/ipvs/ip_vs_conn.c
--- net-next-2.6-e548833-nfct/linux/net/netfilter/ipvs/ip_vs_conn.c 2010-09-15 11:14:13.000000000 +0300
+++ linux/net/netfilter/ipvs/ip_vs_conn.c 2010-09-26 15:24:48.292865793 +0300
@@ -611,6 +611,8 @@ static inline void ip_vs_unbind_dest(str
} else {
atomic_dec(&dest->activeconns);
}
+ if (cp->flags & IP_VS_CONN_F_BACKLOG)
+ atomic_dec(&dest->backlogconns);
} else {
/* It is a persistent connection/template, so decrease
the peristent connection counter */
diff -urp net-next-2.6-e548833-nfct/linux/net/netfilter/ipvs/ip_vs_ctl.c linux/net/netfilter/ipvs/ip_vs_ctl.c
--- net-next-2.6-e548833-nfct/linux/net/netfilter/ipvs/ip_vs_ctl.c 2010-09-16 08:56:34.000000000 +0300
+++ linux/net/netfilter/ipvs/ip_vs_ctl.c 2010-09-26 15:26:45.407867200 +0300
@@ -2593,6 +2593,7 @@ static const struct nla_policy ip_vs_des
[IPVS_DEST_ATTR_INACT_CONNS] = { .type = NLA_U32 },
[IPVS_DEST_ATTR_PERSIST_CONNS] = { .type = NLA_U32 },
[IPVS_DEST_ATTR_STATS] = { .type = NLA_NESTED },
+ [IPVS_DEST_ATTR_BACKLOG_CONNS] = { .type = NLA_U32 },
};
static int ip_vs_genl_fill_stats(struct sk_buff *skb, int container_type,
@@ -2840,6 +2841,8 @@ static int ip_vs_genl_fill_dest(struct s
atomic_read(&dest->activeconns));
NLA_PUT_U32(skb, IPVS_DEST_ATTR_INACT_CONNS,
atomic_read(&dest->inactconns));
+ NLA_PUT_U32(skb, IPVS_DEST_ATTR_BACKLOG_CONNS,
+ atomic_read(&dest->backlogconns));
NLA_PUT_U32(skb, IPVS_DEST_ATTR_PERSIST_CONNS,
atomic_read(&dest->persistconns));
diff -urp net-next-2.6-e548833-nfct/linux/net/netfilter/ipvs/ip_vs_proto_tcp.c linux/net/netfilter/ipvs/ip_vs_proto_tcp.c
--- net-next-2.6-e548833-nfct/linux/net/netfilter/ipvs/ip_vs_proto_tcp.c 2010-09-10 08:27:33.000000000 +0300
+++ linux/net/netfilter/ipvs/ip_vs_proto_tcp.c 2010-09-26 15:29:18.425865407 +0300
@@ -510,6 +510,15 @@ set_tcp_state(struct ip_vs_protocol *pp,
atomic_dec(&dest->inactconns);
cp->flags &= ~IP_VS_CONN_F_INACTIVE;
}
+ if (new_state == IP_VS_TCP_S_SYN_RECV &&
+ !(cp->flags & IP_VS_CONN_F_BACKLOG)) {
+ atomic_inc(&dest->backlogconns);
+ cp->flags |= IP_VS_CONN_F_BACKLOG;
+ } else if (new_state == IP_VS_TCP_S_ESTABLISHED &&
+ cp->flags & IP_VS_CONN_F_BACKLOG) {
+ atomic_dec(&dest->backlogconns);
+ cp->flags &= ~IP_VS_CONN_F_BACKLOG;
+ }
}
}
--
To unsubscribe from this list: send the line "unsubscribe lvs-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [patch v2] ipvs: IPv6 tunnel mode
From: Simon Horman @ 2010-09-27 13:59 UTC (permalink / raw)
To: lvs-devel, netfilter-devel, netdev
Cc: Hans Schillstrom, Julian Anastasov, lvs-devel, Julius Volz,
Wensong Zhang, Patrick McHardy
From: Julian Anastasov <ja@ssi.bg>
Tunnel mode for IPv6 doesn't work.
IPv6 encapsulation uses a bad source address for the tunnel.
i.e. VIP will be used as local-addr and encap. dst addr.
Decapsulation will not accept this.
Example
LVS (eth1 2003::2:0:1/96, VIP 2003::2:0:100)
(eth0 2003::1:0:1/96)
RS (ethX 2003::1:0:5/96)
tcpdump
2003::2:0:100 > 2003::1:0:5:
IP6 (hlim 63, next-header TCP (6) payload length: 40)
2003::3:0:10.50991 > 2003::2:0:100.http: Flags [S], cksum 0x7312
(correct), seq 3006460279, win 5760, options [mss 1440,sackOK,TS val
1904932 ecr 0,nop,wscale 3], length 0
In Linux IPv6 impl. you can't have a tunnel with an any cast address
receiving packets (I have not tried to interpret RFC 2473)
To have receive capabilities the tunnel must have:
- Local address set as multicast addr or an unicast addr
- Remote address set as an unicast addr.
- Loop back addres or Link local address are not allowed.
This causes us to setup a tunnel in the Real Server with the
LVS as the remote address, here you can't use the VIP address since it's
used inside the tunnel.
Solution
Use outgoing interface IPv6 address (match against the destination).
i.e. use ip6_route_output() to look up the route cache and
then use ipv6_dev_get_saddr(...) to set the source address of the
encapsulated packet.
Additionally, cache the results in new destination
fields: dst_cookie and dst_saddr and properly check the
returned dst from ip6_route_output. We now add xfrm_lookup
call only for the tunneling method where the source address
is a local one.
Original patch by Hans Schillstrom.
Check dst state and cache results for IPv6 by Julian Anastasov.
Tested-by: Hans Schillstrom <hans.schillstrom@ericsson.com>
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
* v1
This is Julian's patch with a slightly edited version of the description
from Hans's original patch.
* v2
Updated changelog as per commends from Julian
Patrick, please consider this for nf-next.
diff -urp net-next-2.6-e548833-nfct_snat_reroute/linux/include/net/ip_vs.h linux/include/net/ip_vs.h
--- net-next-2.6-e548833-nfct_snat_reroute/linux/include/net/ip_vs.h 2010-09-16 09:03:48.000000000 +0300
+++ linux/include/net/ip_vs.h 2010-09-22 10:50:18.548963467 +0300
@@ -509,6 +509,10 @@ struct ip_vs_dest {
spinlock_t dst_lock; /* lock of dst_cache */
struct dst_entry *dst_cache; /* destination cache entry */
u32 dst_rtos; /* RT_TOS(tos) for dst */
+ u32 dst_cookie;
+#ifdef CONFIG_IP_VS_IPV6
+ struct in6_addr dst_saddr;
+#endif
/* for virtual service */
struct ip_vs_service *svc; /* service it belongs to */
diff -urp net-next-2.6-e548833-nfct_snat_reroute/linux/net/netfilter/ipvs/ip_vs_xmit.c linux/net/netfilter/ipvs/ip_vs_xmit.c
--- net-next-2.6-e548833-nfct_snat_reroute/linux/net/netfilter/ipvs/ip_vs_xmit.c 2010-09-16 09:02:25.000000000 +0300
+++ linux/net/netfilter/ipvs/ip_vs_xmit.c 2010-09-22 16:29:43.271964521 +0300
@@ -26,6 +26,7 @@
#include <net/route.h> /* for ip_route_output */
#include <net/ipv6.h>
#include <net/ip6_route.h>
+#include <net/addrconf.h>
#include <linux/icmpv6.h>
#include <linux/netfilter.h>
#include <linux/netfilter_ipv4.h>
@@ -37,26 +38,27 @@
* Destination cache to speed up outgoing route lookup
*/
static inline void
-__ip_vs_dst_set(struct ip_vs_dest *dest, u32 rtos, struct dst_entry *dst)
+__ip_vs_dst_set(struct ip_vs_dest *dest, u32 rtos, struct dst_entry *dst,
+ u32 dst_cookie)
{
struct dst_entry *old_dst;
old_dst = dest->dst_cache;
dest->dst_cache = dst;
dest->dst_rtos = rtos;
+ dest->dst_cookie = dst_cookie;
dst_release(old_dst);
}
static inline struct dst_entry *
-__ip_vs_dst_check(struct ip_vs_dest *dest, u32 rtos, u32 cookie)
+__ip_vs_dst_check(struct ip_vs_dest *dest, u32 rtos)
{
struct dst_entry *dst = dest->dst_cache;
if (!dst)
return NULL;
- if ((dst->obsolete
- || (dest->af == AF_INET && rtos != dest->dst_rtos)) &&
- dst->ops->check(dst, cookie) == NULL) {
+ if ((dst->obsolete || rtos != dest->dst_rtos) &&
+ dst->ops->check(dst, dest->dst_cookie) == NULL) {
dest->dst_cache = NULL;
dst_release(dst);
return NULL;
@@ -66,15 +68,16 @@ __ip_vs_dst_check(struct ip_vs_dest *des
}
static struct rtable *
-__ip_vs_get_out_rt(struct ip_vs_conn *cp, u32 rtos)
+__ip_vs_get_out_rt(struct sk_buff *skb, struct ip_vs_conn *cp, u32 rtos)
{
+ struct net *net = dev_net(skb->dev);
struct rtable *rt; /* Route to the other host */
struct ip_vs_dest *dest = cp->dest;
if (dest) {
spin_lock(&dest->dst_lock);
if (!(rt = (struct rtable *)
- __ip_vs_dst_check(dest, rtos, 0))) {
+ __ip_vs_dst_check(dest, rtos))) {
struct flowi fl = {
.oif = 0,
.nl_u = {
@@ -84,13 +87,13 @@ __ip_vs_get_out_rt(struct ip_vs_conn *cp
.tos = rtos, } },
};
- if (ip_route_output_key(&init_net, &rt, &fl)) {
+ if (ip_route_output_key(net, &rt, &fl)) {
spin_unlock(&dest->dst_lock);
IP_VS_DBG_RL("ip_route_output error, dest: %pI4\n",
&dest->addr.ip);
return NULL;
}
- __ip_vs_dst_set(dest, rtos, dst_clone(&rt->dst));
+ __ip_vs_dst_set(dest, rtos, dst_clone(&rt->dst), 0);
IP_VS_DBG(10, "new dst %pI4, refcnt=%d, rtos=%X\n",
&dest->addr.ip,
atomic_read(&rt->dst.__refcnt), rtos);
@@ -106,7 +109,7 @@ __ip_vs_get_out_rt(struct ip_vs_conn *cp
.tos = rtos, } },
};
- if (ip_route_output_key(&init_net, &rt, &fl)) {
+ if (ip_route_output_key(net, &rt, &fl)) {
IP_VS_DBG_RL("ip_route_output error, dest: %pI4\n",
&cp->daddr.ip);
return NULL;
@@ -117,62 +120,79 @@ __ip_vs_get_out_rt(struct ip_vs_conn *cp
}
#ifdef CONFIG_IP_VS_IPV6
+
+static struct dst_entry *
+__ip_vs_route_output_v6(struct net *net, struct in6_addr *daddr,
+ struct in6_addr *ret_saddr, int do_xfrm)
+{
+ struct dst_entry *dst;
+ struct flowi fl = {
+ .oif = 0,
+ .nl_u = {
+ .ip6_u = {
+ .daddr = *daddr,
+ },
+ },
+ };
+
+ dst = ip6_route_output(net, NULL, &fl);
+ if (dst->error)
+ goto out_err;
+ if (!ret_saddr)
+ return dst;
+ if (ipv6_addr_any(&fl.fl6_src) &&
+ ipv6_dev_get_saddr(net, ip6_dst_idev(dst)->dev,
+ &fl.fl6_dst, 0, &fl.fl6_src) < 0)
+ goto out_err;
+ if (do_xfrm && xfrm_lookup(net, &dst, &fl, NULL, 0) < 0)
+ goto out_err;
+ ipv6_addr_copy(ret_saddr, &fl.fl6_src);
+ return dst;
+
+out_err:
+ dst_release(dst);
+ IP_VS_DBG_RL("ip6_route_output error, dest: %pI6\n", daddr);
+ return NULL;
+}
+
static struct rt6_info *
-__ip_vs_get_out_rt_v6(struct ip_vs_conn *cp)
+__ip_vs_get_out_rt_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
+ struct in6_addr *ret_saddr, int do_xfrm)
{
+ struct net *net = dev_net(skb->dev);
struct rt6_info *rt; /* Route to the other host */
struct ip_vs_dest *dest = cp->dest;
+ struct dst_entry *dst;
if (dest) {
spin_lock(&dest->dst_lock);
- rt = (struct rt6_info *)__ip_vs_dst_check(dest, 0, 0);
+ rt = (struct rt6_info *)__ip_vs_dst_check(dest, 0);
if (!rt) {
- struct flowi fl = {
- .oif = 0,
- .nl_u = {
- .ip6_u = {
- .daddr = dest->addr.in6,
- .saddr = {
- .s6_addr32 =
- { 0, 0, 0, 0 },
- },
- },
- },
- };
+ u32 cookie;
- rt = (struct rt6_info *)ip6_route_output(&init_net,
- NULL, &fl);
- if (!rt) {
+ dst = __ip_vs_route_output_v6(net, &dest->addr.in6,
+ &dest->dst_saddr,
+ do_xfrm);
+ if (!dst) {
spin_unlock(&dest->dst_lock);
- IP_VS_DBG_RL("ip6_route_output error, dest: %pI6\n",
- &dest->addr.in6);
return NULL;
}
- __ip_vs_dst_set(dest, 0, dst_clone(&rt->dst));
- IP_VS_DBG(10, "new dst %pI6, refcnt=%d\n",
- &dest->addr.in6,
+ rt = (struct rt6_info *) dst;
+ cookie = rt->rt6i_node ? rt->rt6i_node->fn_sernum : 0;
+ __ip_vs_dst_set(dest, 0, dst_clone(&rt->dst), cookie);
+ IP_VS_DBG(10, "new dst %pI6, src %pI6, refcnt=%d\n",
+ &dest->addr.in6, &dest->dst_saddr,
atomic_read(&rt->dst.__refcnt));
}
+ if (ret_saddr)
+ ipv6_addr_copy(ret_saddr, &dest->dst_saddr);
spin_unlock(&dest->dst_lock);
} else {
- struct flowi fl = {
- .oif = 0,
- .nl_u = {
- .ip6_u = {
- .daddr = cp->daddr.in6,
- .saddr = {
- .s6_addr32 = { 0, 0, 0, 0 },
- },
- },
- },
- };
-
- rt = (struct rt6_info *)ip6_route_output(&init_net, NULL, &fl);
- if (!rt) {
- IP_VS_DBG_RL("ip6_route_output error, dest: %pI6\n",
- &cp->daddr.in6);
+ dst = __ip_vs_route_output_v6(net, &cp->daddr.in6, ret_saddr,
+ do_xfrm);
+ if (!dst)
return NULL;
- }
+ rt = (struct rt6_info *) dst;
}
return rt;
@@ -248,6 +268,7 @@ int
ip_vs_bypass_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
struct ip_vs_protocol *pp)
{
+ struct net *net = dev_net(skb->dev);
struct rtable *rt; /* Route to the other host */
struct iphdr *iph = ip_hdr(skb);
u8 tos = iph->tos;
@@ -263,7 +284,7 @@ ip_vs_bypass_xmit(struct sk_buff *skb, s
EnterFunction(10);
- if (ip_route_output_key(&init_net, &rt, &fl)) {
+ if (ip_route_output_key(net, &rt, &fl)) {
IP_VS_DBG_RL("%s(): ip_route_output error, dest: %pI4\n",
__func__, &iph->daddr);
goto tx_error_icmp;
@@ -313,25 +334,18 @@ int
ip_vs_bypass_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
struct ip_vs_protocol *pp)
{
+ struct net *net = dev_net(skb->dev);
+ struct dst_entry *dst;
struct rt6_info *rt; /* Route to the other host */
struct ipv6hdr *iph = ipv6_hdr(skb);
int mtu;
- struct flowi fl = {
- .oif = 0,
- .nl_u = {
- .ip6_u = {
- .daddr = iph->daddr,
- .saddr = { .s6_addr32 = {0, 0, 0, 0} }, } },
- };
EnterFunction(10);
- rt = (struct rt6_info *)ip6_route_output(&init_net, NULL, &fl);
- if (!rt) {
- IP_VS_DBG_RL("%s(): ip6_route_output error, dest: %pI6\n",
- __func__, &iph->daddr);
+ dst = __ip_vs_route_output_v6(net, &iph->daddr, NULL, 0);
+ if (!dst)
goto tx_error_icmp;
- }
+ rt = (struct rt6_info *) dst;
/* MTU checking */
mtu = dst_mtu(&rt->dst);
@@ -397,7 +411,7 @@ ip_vs_nat_xmit(struct sk_buff *skb, stru
IP_VS_DBG(10, "filled cport=%d\n", ntohs(*p));
}
- if (!(rt = __ip_vs_get_out_rt(cp, RT_TOS(iph->tos))))
+ if (!(rt = __ip_vs_get_out_rt(skb, cp, RT_TOS(iph->tos))))
goto tx_error_icmp;
/* MTU checking */
@@ -472,7 +486,7 @@ ip_vs_nat_xmit_v6(struct sk_buff *skb, s
IP_VS_DBG(10, "filled cport=%d\n", ntohs(*p));
}
- rt = __ip_vs_get_out_rt_v6(cp);
+ rt = __ip_vs_get_out_rt_v6(skb, cp, NULL, 0);
if (!rt)
goto tx_error_icmp;
@@ -557,7 +571,6 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, s
struct iphdr *old_iph = ip_hdr(skb);
u8 tos = old_iph->tos;
__be16 df = old_iph->frag_off;
- sk_buff_data_t old_transport_header = skb->transport_header;
struct iphdr *iph; /* Our new IP header */
unsigned int max_headroom; /* The extra header space needed */
int mtu;
@@ -572,7 +585,7 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, s
goto tx_error;
}
- if (!(rt = __ip_vs_get_out_rt(cp, RT_TOS(tos))))
+ if (!(rt = __ip_vs_get_out_rt(skb, cp, RT_TOS(tos))))
goto tx_error_icmp;
tdev = rt->dst.dev;
@@ -616,7 +629,7 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, s
old_iph = ip_hdr(skb);
}
- skb->transport_header = old_transport_header;
+ skb->transport_header = skb->network_header;
/* fix old IP header checksum */
ip_send_check(old_iph);
@@ -670,9 +683,9 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb
struct ip_vs_protocol *pp)
{
struct rt6_info *rt; /* Route to the other host */
+ struct in6_addr saddr; /* Source for tunnel */
struct net_device *tdev; /* Device to other host */
struct ipv6hdr *old_iph = ipv6_hdr(skb);
- sk_buff_data_t old_transport_header = skb->transport_header;
struct ipv6hdr *iph; /* Our new IP header */
unsigned int max_headroom; /* The extra header space needed */
int mtu;
@@ -687,17 +700,17 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb
goto tx_error;
}
- rt = __ip_vs_get_out_rt_v6(cp);
+ rt = __ip_vs_get_out_rt_v6(skb, cp, &saddr, 1);
if (!rt)
goto tx_error_icmp;
tdev = rt->dst.dev;
mtu = dst_mtu(&rt->dst) - sizeof(struct ipv6hdr);
- /* TODO IPv6: do we need this check in IPv6? */
- if (mtu < 1280) {
+ if (mtu < IPV6_MIN_MTU) {
dst_release(&rt->dst);
- IP_VS_DBG_RL("%s(): mtu less than 1280\n", __func__);
+ IP_VS_DBG_RL("%s(): mtu less than %d\n", __func__,
+ IPV6_MIN_MTU);
goto tx_error;
}
if (skb_dst(skb))
@@ -730,7 +743,7 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb
old_iph = ipv6_hdr(skb);
}
- skb->transport_header = old_transport_header;
+ skb->transport_header = skb->network_header;
skb_push(skb, sizeof(struct ipv6hdr));
skb_reset_network_header(skb);
@@ -750,8 +763,8 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb
be16_add_cpu(&iph->payload_len, sizeof(*old_iph));
iph->priority = old_iph->priority;
memset(&iph->flow_lbl, 0, sizeof(iph->flow_lbl));
- iph->daddr = rt->rt6i_dst.addr;
- iph->saddr = cp->vaddr.in6; /* rt->rt6i_src.addr; */
+ ipv6_addr_copy(&iph->daddr, &rt->rt6i_dst.addr);
+ ipv6_addr_copy(&iph->saddr, &saddr);
iph->hop_limit = old_iph->hop_limit;
/* Another hack: avoid icmp_send in ip_fragment */
@@ -791,7 +804,7 @@ ip_vs_dr_xmit(struct sk_buff *skb, struc
EnterFunction(10);
- if (!(rt = __ip_vs_get_out_rt(cp, RT_TOS(iph->tos))))
+ if (!(rt = __ip_vs_get_out_rt(skb, cp, RT_TOS(iph->tos))))
goto tx_error_icmp;
/* MTU checking */
@@ -843,7 +856,7 @@ ip_vs_dr_xmit_v6(struct sk_buff *skb, st
EnterFunction(10);
- rt = __ip_vs_get_out_rt_v6(cp);
+ rt = __ip_vs_get_out_rt_v6(skb, cp, NULL, 0);
if (!rt)
goto tx_error_icmp;
@@ -919,7 +932,7 @@ ip_vs_icmp_xmit(struct sk_buff *skb, str
* mangle and send the packet here (only for VS/NAT)
*/
- if (!(rt = __ip_vs_get_out_rt(cp, RT_TOS(ip_hdr(skb)->tos))))
+ if (!(rt = __ip_vs_get_out_rt(skb, cp, RT_TOS(ip_hdr(skb)->tos))))
goto tx_error_icmp;
/* MTU checking */
@@ -993,7 +1006,7 @@ ip_vs_icmp_xmit_v6(struct sk_buff *skb,
* mangle and send the packet here (only for VS/NAT)
*/
- rt = __ip_vs_get_out_rt_v6(cp);
+ rt = __ip_vs_get_out_rt_v6(skb, cp, NULL, 0);
if (!rt)
goto tx_error_icmp;
^ permalink raw reply
* Re: [PATCH -next 2/4] ip_gre: percpu stats accounting
From: Eric Dumazet @ 2010-09-27 13:57 UTC (permalink / raw)
To: Ben Hutchings; +Cc: David Miller, netdev
In-Reply-To: <1285594140.2263.0.camel@achroite.uk.solarflarecom.com>
Le lundi 27 septembre 2010 à 14:29 +0100, Ben Hutchings a écrit :
> > diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
> > index 5d6ddcb..de39b22 100644
> > --- a/net/ipv4/ip_gre.c
> > +++ b/net/ipv4/ip_gre.c
> [...]
> > @@ -377,7 +405,7 @@ static struct ip_tunnel *ipgre_tunnel_locate(struct net *net,
> > if (parms->name[0])
> > strlcpy(name, parms->name, IFNAMSIZ);
> > else
> > - sprintf(name, "gre%%d");
> > + strcpy(name, "gre%d");
> >
> > dev = alloc_netdev(sizeof(*t), name, ipgre_tunnel_setup);
> > if (!dev)
> [...]
>
> This is a valid fix, but doesn't belong in this patch!
>
Sorry ? It was not a fix, but at most a cleanup ;)
Anyway I forgot the gretap case...
[PATCH 2/4 v2] ip_gre: percpu stats accounting
Maintain per_cpu tx_bytes, tx_packets, rx_bytes, rx_packets.
Other seldom used fields are kept in netdev->stats structure, possibly
unsafe.
This is a preliminary work to support lockless transmit path, and
correct RX stats, that are already unsafe.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
v2: gretap should be handled too
net/ipv4/ip_gre.c | 143 ++++++++++++++++++++++++++++++++------------
1 files changed, 104 insertions(+), 39 deletions(-)
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index 5d6ddcb..a1b5d5e 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -165,6 +165,34 @@ struct ipgre_net {
#define for_each_ip_tunnel_rcu(start) \
for (t = rcu_dereference(start); t; t = rcu_dereference(t->next))
+/* often modified stats are per cpu, other are shared (netdev->stats) */
+struct pcpu_tstats {
+ unsigned long rx_packets;
+ unsigned long rx_bytes;
+ unsigned long tx_packets;
+ unsigned long tx_bytes;
+};
+
+static struct net_device_stats *ipgre_get_stats(struct net_device *dev)
+{
+ struct pcpu_tstats sum = { 0 };
+ int i;
+
+ for_each_possible_cpu(i) {
+ const struct pcpu_tstats *tstats = per_cpu_ptr(dev->tstats, i);
+
+ sum.rx_packets += tstats->rx_packets;
+ sum.rx_bytes += tstats->rx_bytes;
+ sum.tx_packets += tstats->tx_packets;
+ sum.tx_bytes += tstats->tx_bytes;
+ }
+ dev->stats.rx_packets = sum.rx_packets;
+ dev->stats.rx_bytes = sum.rx_bytes;
+ dev->stats.tx_packets = sum.tx_packets;
+ dev->stats.tx_bytes = sum.tx_bytes;
+ return &dev->stats;
+}
+
/* Given src, dst and key, find appropriate for input tunnel. */
static struct ip_tunnel * ipgre_tunnel_lookup(struct net_device *dev,
@@ -584,7 +612,7 @@ static int ipgre_rcv(struct sk_buff *skb)
if ((tunnel = ipgre_tunnel_lookup(skb->dev,
iph->saddr, iph->daddr, key,
gre_proto))) {
- struct net_device_stats *stats = &tunnel->dev->stats;
+ struct pcpu_tstats *tstats;
secpath_reset(skb);
@@ -608,22 +636,22 @@ static int ipgre_rcv(struct sk_buff *skb)
/* Looped back packet, drop it! */
if (skb_rtable(skb)->fl.iif == 0)
goto drop;
- stats->multicast++;
+ tunnel->dev->stats.multicast++;
skb->pkt_type = PACKET_BROADCAST;
}
#endif
if (((flags&GRE_CSUM) && csum) ||
(!(flags&GRE_CSUM) && tunnel->parms.i_flags&GRE_CSUM)) {
- stats->rx_crc_errors++;
- stats->rx_errors++;
+ tunnel->dev->stats.rx_crc_errors++;
+ tunnel->dev->stats.rx_errors++;
goto drop;
}
if (tunnel->parms.i_flags&GRE_SEQ) {
if (!(flags&GRE_SEQ) ||
(tunnel->i_seqno && (s32)(seqno - tunnel->i_seqno) < 0)) {
- stats->rx_fifo_errors++;
- stats->rx_errors++;
+ tunnel->dev->stats.rx_fifo_errors++;
+ tunnel->dev->stats.rx_errors++;
goto drop;
}
tunnel->i_seqno = seqno + 1;
@@ -632,8 +660,8 @@ static int ipgre_rcv(struct sk_buff *skb)
/* Warning: All skb pointers will be invalidated! */
if (tunnel->dev->type == ARPHRD_ETHER) {
if (!pskb_may_pull(skb, ETH_HLEN)) {
- stats->rx_length_errors++;
- stats->rx_errors++;
+ tunnel->dev->stats.rx_length_errors++;
+ tunnel->dev->stats.rx_errors++;
goto drop;
}
@@ -642,13 +670,17 @@ static int ipgre_rcv(struct sk_buff *skb)
skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);
}
- skb_tunnel_rx(skb, tunnel->dev);
+ tstats = this_cpu_ptr(tunnel->dev->tstats);
+ tstats->rx_packets++;
+ tstats->rx_bytes += skb->len;
+
+ __skb_tunnel_rx(skb, tunnel->dev);
skb_reset_network_header(skb);
ipgre_ecn_decapsulate(iph, skb);
if (netif_rx(skb) == NET_RX_DROP)
- stats->rx_dropped++;
+ tunnel->dev->stats.rx_dropped++;
rcu_read_unlock();
return 0;
@@ -665,8 +697,7 @@ drop_nolock:
static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct ip_tunnel *tunnel = netdev_priv(dev);
- struct net_device_stats *stats = &dev->stats;
- struct netdev_queue *txq = netdev_get_tx_queue(dev, 0);
+ struct pcpu_tstats *tstats;
struct iphdr *old_iph = ip_hdr(skb);
struct iphdr *tiph;
u8 tos;
@@ -694,7 +725,7 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
/* NBMA tunnel */
if (skb_dst(skb) == NULL) {
- stats->tx_fifo_errors++;
+ dev->stats.tx_fifo_errors++;
goto tx_error;
}
@@ -740,14 +771,20 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
}
{
- struct flowi fl = { .oif = tunnel->parms.link,
- .nl_u = { .ip4_u =
- { .daddr = dst,
- .saddr = tiph->saddr,
- .tos = RT_TOS(tos) } },
- .proto = IPPROTO_GRE };
+ struct flowi fl = {
+ .oif = tunnel->parms.link,
+ .nl_u = {
+ .ip4_u = {
+ .daddr = dst,
+ .saddr = tiph->saddr,
+ .tos = RT_TOS(tos)
+ }
+ },
+ .proto = IPPROTO_GRE
+ }
+;
if (ip_route_output_key(dev_net(dev), &rt, &fl)) {
- stats->tx_carrier_errors++;
+ dev->stats.tx_carrier_errors++;
goto tx_error;
}
}
@@ -755,7 +792,7 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
if (tdev == dev) {
ip_rt_put(rt);
- stats->collisions++;
+ dev->stats.collisions++;
goto tx_error;
}
@@ -818,7 +855,7 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
dev->needed_headroom = max_headroom;
if (!new_skb) {
ip_rt_put(rt);
- txq->tx_dropped++;
+ dev->stats.tx_dropped++;
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
@@ -885,15 +922,15 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
}
nf_reset(skb);
-
- IPTUNNEL_XMIT();
+ tstats = this_cpu_ptr(dev->tstats);
+ __IPTUNNEL_XMIT(tstats, &dev->stats);
return NETDEV_TX_OK;
tx_error_icmp:
dst_link_failure(skb);
tx_error:
- stats->tx_errors++;
+ dev->stats.tx_errors++;
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
@@ -913,13 +950,19 @@ static int ipgre_tunnel_bind_dev(struct net_device *dev)
/* Guess output device to choose reasonable mtu and needed_headroom */
if (iph->daddr) {
- struct flowi fl = { .oif = tunnel->parms.link,
- .nl_u = { .ip4_u =
- { .daddr = iph->daddr,
- .saddr = iph->saddr,
- .tos = RT_TOS(iph->tos) } },
- .proto = IPPROTO_GRE };
+ struct flowi fl = {
+ .oif = tunnel->parms.link,
+ .nl_u = {
+ .ip4_u = {
+ .daddr = iph->daddr,
+ .saddr = iph->saddr,
+ .tos = RT_TOS(iph->tos)
+ }
+ },
+ .proto = IPPROTO_GRE
+ };
struct rtable *rt;
+
if (!ip_route_output_key(dev_net(dev), &rt, &fl)) {
tdev = rt->dst.dev;
ip_rt_put(rt);
@@ -1171,13 +1214,19 @@ static int ipgre_open(struct net_device *dev)
struct ip_tunnel *t = netdev_priv(dev);
if (ipv4_is_multicast(t->parms.iph.daddr)) {
- struct flowi fl = { .oif = t->parms.link,
- .nl_u = { .ip4_u =
- { .daddr = t->parms.iph.daddr,
- .saddr = t->parms.iph.saddr,
- .tos = RT_TOS(t->parms.iph.tos) } },
- .proto = IPPROTO_GRE };
+ struct flowi fl = {
+ .oif = t->parms.link,
+ .nl_u = {
+ .ip4_u = {
+ .daddr = t->parms.iph.daddr,
+ .saddr = t->parms.iph.saddr,
+ .tos = RT_TOS(t->parms.iph.tos)
+ }
+ },
+ .proto = IPPROTO_GRE
+ };
struct rtable *rt;
+
if (ip_route_output_key(dev_net(dev), &rt, &fl))
return -EADDRNOTAVAIL;
dev = rt->dst.dev;
@@ -1217,12 +1266,19 @@ static const struct net_device_ops ipgre_netdev_ops = {
.ndo_start_xmit = ipgre_tunnel_xmit,
.ndo_do_ioctl = ipgre_tunnel_ioctl,
.ndo_change_mtu = ipgre_tunnel_change_mtu,
+ .ndo_get_stats = ipgre_get_stats,
};
+static void ipgre_dev_free(struct net_device *dev)
+{
+ free_percpu(dev->tstats);
+ free_netdev(dev);
+}
+
static void ipgre_tunnel_setup(struct net_device *dev)
{
dev->netdev_ops = &ipgre_netdev_ops;
- dev->destructor = free_netdev;
+ dev->destructor = ipgre_dev_free;
dev->type = ARPHRD_IPGRE;
dev->needed_headroom = LL_MAX_HEADER + sizeof(struct iphdr) + 4;
@@ -1260,6 +1316,10 @@ static int ipgre_tunnel_init(struct net_device *dev)
} else
dev->header_ops = &ipgre_header_ops;
+ dev->tstats = alloc_percpu(struct pcpu_tstats);
+ if (!dev->tstats)
+ return -ENOMEM;
+
return 0;
}
@@ -1446,6 +1506,10 @@ static int ipgre_tap_init(struct net_device *dev)
ipgre_tunnel_bind_dev(dev);
+ dev->tstats = alloc_percpu(struct pcpu_tstats);
+ if (!dev->tstats)
+ return -ENOMEM;
+
return 0;
}
@@ -1456,6 +1520,7 @@ static const struct net_device_ops ipgre_tap_netdev_ops = {
.ndo_set_mac_address = eth_mac_addr,
.ndo_validate_addr = eth_validate_addr,
.ndo_change_mtu = ipgre_tunnel_change_mtu,
+ .ndo_get_stats = ipgre_get_stats,
};
static void ipgre_tap_setup(struct net_device *dev)
@@ -1464,7 +1529,7 @@ static void ipgre_tap_setup(struct net_device *dev)
ether_setup(dev);
dev->netdev_ops = &ipgre_tap_netdev_ops;
- dev->destructor = free_netdev;
+ dev->destructor = ipgre_dev_free;
dev->iflink = 0;
dev->features |= NETIF_F_NETNS_LOCAL;
^ permalink raw reply related
* Re: [PATCH v11 12/17] Add a kconfig entry and make entry for mp device.
From: Ben Hutchings @ 2010-09-27 13:56 UTC (permalink / raw)
To: xiaohui.xin; +Cc: netdev, kvm, linux-kernel, mingo, davem, herbert, jdike
In-Reply-To: <0eed02838e581f2ef2247be8e69bc34cdecce6b2.1285385607.git.xiaohui.xin@intel.com>
This patch is in the wrong position in the sequence. It needs to be
applied after mpassthru.c is created, not before.
Ben.
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [PATCH v11 11/17] Add header file for mp device.
From: Ben Hutchings @ 2010-09-27 13:55 UTC (permalink / raw)
To: xiaohui.xin; +Cc: netdev, kvm, linux-kernel, mingo, davem, herbert, jdike
In-Reply-To: <924a3a20630b9b0b7f3c66633c792360bbcc790d.1285385607.git.xiaohui.xin@intel.com>
On Sat, 2010-09-25 at 12:27 +0800, xiaohui.xin@intel.com wrote:
> From: Xin Xiaohui <xiaohui.xin@intel.com>
>
> Signed-off-by: Xin Xiaohui <xiaohui.xin@intel.com>
> Signed-off-by: Zhao Yu <yzhao81new@gmail.com>
> Reviewed-by: Jeff Dike <jdike@linux.intel.com>
> ---
> include/linux/mpassthru.h | 25 +++++++++++++++++++++++++
> 1 files changed, 25 insertions(+), 0 deletions(-)
> create mode 100644 include/linux/mpassthru.h
>
> diff --git a/include/linux/mpassthru.h b/include/linux/mpassthru.h
> new file mode 100644
> index 0000000..ba8f320
> --- /dev/null
> +++ b/include/linux/mpassthru.h
> @@ -0,0 +1,25 @@
> +#ifndef __MPASSTHRU_H
> +#define __MPASSTHRU_H
> +
> +#include <linux/types.h>
> +#include <linux/if_ether.h>
> +
> +/* ioctl defines */
> +#define MPASSTHRU_BINDDEV _IOW('M', 213, int)
> +#define MPASSTHRU_UNBINDDEV _IO('M', 214)
[...]
You need to include <linux/ioctl.h> first!
Ben.
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [PATCH v11 04/17]Add a function make external buffer owner to query capability.
From: Ben Hutchings @ 2010-09-27 13:45 UTC (permalink / raw)
To: xiaohui.xin; +Cc: netdev, kvm, linux-kernel, mingo, davem, herbert, jdike
In-Reply-To: <6df1b1c2d2ffe70da911984c7fe6fab16f23043b.1285385607.git.xiaohui.xin@intel.com>
On Sat, 2010-09-25 at 12:27 +0800, xiaohui.xin@intel.com wrote:
[...]
> diff --git a/net/core/dev.c b/net/core/dev.c
> index 264137f..636f11b 100644
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -2468,6 +2468,55 @@ void netif_nit_deliver(struct sk_buff *skb)
> rcu_read_unlock();
> }
>
> +/* To support meidate passthru(zero-copy) with NIC driver,
> + * we'd better query NIC driver for the capability it can
> + * provide, especially for packet split mode, now we only
> + * query for the header size, and the payload a descriptor
> + * may carry. If a driver does not use the API to export,
> + * then we may try to use a default value, currently,
> + * we use the default value from an IGB driver. Now,
> + * it's only called by mpassthru device.
> + */
> +#if defined(CONFIG_MEDIATE_PASSTHRU) || defined(CONFIG_MEDIATE_PASSTHRU_MODULE)
> +int netdev_mp_port_prep(struct net_device *dev,
> + struct mpassthru_port *port)
> +{
> + int rc;
> + int npages, data_len;
> + const struct net_device_ops *ops = dev->netdev_ops;
> +
> + if (ops->ndo_mp_port_prep) {
> + rc = ops->ndo_mp_port_prep(dev, port);
> + if (rc)
> + return rc;
> + } else {
> + /* If the NIC driver did not report this,
> + * then we try to use default value.
> + */
> + port->hdr_len = 128;
> + port->data_len = 2048;
> + port->npages = 1;
> + }
[...]
Is it really necessary to have a default?
Also have you considered an API for changing the header/data split?
Ben.
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [PATCH v11 03/17] Add a ndo_mp_port_prep pointer to net_device_ops.
From: Ben Hutchings @ 2010-09-27 13:42 UTC (permalink / raw)
To: xiaohui.xin; +Cc: netdev, kvm, linux-kernel, mingo, davem, herbert, jdike
In-Reply-To: <79267b300503d98d43ccfcf11d1374f59fec8578.1285385607.git.xiaohui.xin@intel.com>
On Sat, 2010-09-25 at 12:27 +0800, xiaohui.xin@intel.com wrote:
> From: Xin Xiaohui <xiaohui.xin@intel.com>
>
> If the driver want to allocate external buffers,
> then it can export it's capability, as the skb
> buffer header length, the page length can be DMA, etc.
> The external buffers owner may utilize this.
[...]
This information needs to be included in the comment above struct
net_device_ops, not just in the commit message.
Ben.
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [PATCH v11 02/17] Add a new struct for device to manipulate external buffer.
From: Ben Hutchings @ 2010-09-27 13:41 UTC (permalink / raw)
To: xiaohui.xin; +Cc: netdev, kvm, linux-kernel, mingo, davem, herbert, jdike
In-Reply-To: <c6751a0126ef4d3d1ce75e596ccf9cebd9757052.1285385607.git.xiaohui.xin@intel.com>
On Sat, 2010-09-25 at 12:27 +0800, xiaohui.xin@intel.com wrote:
> From: Xin Xiaohui <xiaohui.xin@intel.com>
>
> Signed-off-by: Xin Xiaohui <xiaohui.xin@intel.com>
> Signed-off-by: Zhao Yu <yzhao81new@gmail.com>
> Reviewed-by: Jeff Dike <jdike@linux.intel.com>
> ---
> include/linux/netdevice.h | 22 +++++++++++++++++++++-
> 1 files changed, 21 insertions(+), 1 deletions(-)
>
> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
> index fa8b476..ba582e1 100644
> --- a/include/linux/netdevice.h
> +++ b/include/linux/netdevice.h
> @@ -530,6 +530,25 @@ struct netdev_queue {
> unsigned long tx_dropped;
> } ____cacheline_aligned_in_smp;
>
> +/* Add a structure in structure net_device, the new field is
> + * named as mp_port. It's for mediate passthru (zero-copy).
That belongs in the commit message.
> + * It contains the capability for the net device driver,
> + * a socket, and an external buffer creator, external means
> + * skb buffer belongs to the device may not be allocated from
> + * kernel space.
Who sets which fields in this structure? Can you make this a kernel-doc
comment specifying the use of each field?
Ben.
> + */
> +struct mpassthru_port {
> + int hdr_len;
> + int data_len;
> + int npages;
> + unsigned flags;
> + struct socket *sock;
> + int vnet_hlen;
> + struct skb_ext_page *(*ctor)(struct mpassthru_port *,
> + struct sk_buff *, int);
> + struct skb_ext_page *(*hash)(struct net_device *,
> + struct page *);
> +};
>
> /*
> * This structure defines the management hooks for network devices.
> @@ -952,7 +971,8 @@ struct net_device {
> struct macvlan_port *macvlan_port;
> /* GARP */
> struct garp_port *garp_port;
> -
> + /* mpassthru */
> + struct mpassthru_port *mp_port;
> /* class/net/name entry */
> struct device dev;
> /* space for optional device, statistics, and wireless sysfs groups */
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox