* Re: [patch net-next 0/3] team: couple of patches
From: Jiri Pirko @ 2012-07-12 15:53 UTC (permalink / raw)
To: David Miller; +Cc: netdev
In-Reply-To: <20120712.081443.1373285885474644661.davem@davemloft.net>
Thu, Jul 12, 2012 at 05:14:43PM CEST, davem@davemloft.net wrote:
>From: David Miller <davem@davemloft.net>
>Date: Thu, 12 Jul 2012 08:11:24 -0700 (PDT)
>
>> From: Jiri Pirko <jpirko@redhat.com>
>> Date: Wed, 11 Jul 2012 17:34:01 +0200
>>
>>> Jiri Pirko (3):
>>> team: use function team_port_txable() for determing enabled and up
>>> port
>>> team: add broadcast mode
>>> team: make team_port_enabled() and team_port_txable() static inline
>>
>> All applied, thanks.
>
>Jiri, btw, any chance I can convince you to remove the EXPERIMENTAL
>Kconfig dependency?
>
>Code I've written and checked in myself over the past few days is
>several orders of magnitude more experimental than the team driver
>is :-)
Hehe :) Nevertheless, I would like to keep this flag for some more time.
I will remove that once I have all planned basic functionality in.
Jirka
^ permalink raw reply
* Re: [net-next:master 90/102] net/ipv4/route.c:1283:9: warning: unused variable 'saddr'
From: Dan Carpenter @ 2012-07-12 15:47 UTC (permalink / raw)
To: David Miller; +Cc: fengguang.wu, kernel-janitors, netdev
In-Reply-To: <20120712.074058.753681400854318989.davem@davemloft.net>
On Thu, Jul 12, 2012 at 07:40:58AM -0700, David Miller wrote:
>
> There's not need to report these to kernel-janitors if it's a
> net-next specific issue and I'm going to fix it up 5 minutes
> after you report it.
The kernel-janitors list is CC'd to prevent people from sending
duplicate messages. This has happened in the past and it's
annoying for everyone.
regards,
dan carpenter
^ permalink raw reply
* Re: [PATCH 1/2] ipvs: ip_vs_ftp depends on nf_conntrack_ftp helper
From: Pablo Neira Ayuso @ 2012-07-12 15:39 UTC (permalink / raw)
To: Simon Horman
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Hans Schillstrom, Jesper Dangaard Brouer
In-Reply-To: <1341966327-16606-2-git-send-email-horms@verge.net.au>
On Wed, Jul 11, 2012 at 09:25:26AM +0900, Simon Horman wrote:
> From: Julian Anastasov <ja@ssi.bg>
>
> The FTP application indirectly depends on the
> nf_conntrack_ftp helper for proper NAT support. If the
> module is not loaded, IPVS can resize the packets for the
> command connection, eg. PASV response but the SEQ adjustment
> logic in ipv4_confirm is not called without helper.
>
> Signed-off-by: Julian Anastasov <ja@ssi.bg>
> Signed-off-by: Simon Horman <horms@verge.net.au>
> ---
> net/netfilter/ipvs/Kconfig | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/net/netfilter/ipvs/Kconfig b/net/netfilter/ipvs/Kconfig
> index f987138..8b2cffd 100644
> --- a/net/netfilter/ipvs/Kconfig
> +++ b/net/netfilter/ipvs/Kconfig
> @@ -250,7 +250,8 @@ comment 'IPVS application helper'
>
> config IP_VS_FTP
> tristate "FTP protocol helper"
> - depends on IP_VS_PROTO_TCP && NF_CONNTRACK && NF_NAT
> + depends on IP_VS_PROTO_TCP && NF_CONNTRACK && NF_NAT && \
> + NF_CONNTRACK_FTP
If you require FTP NAT support, then this depends on NF_NAT_FTP
instead of NF_CONNTRACK_FTP.
^ permalink raw reply
* Re: [RFC PATCH 1/2] net: Add new network device function to allow for MMIO batching
From: Alexander Duyck @ 2012-07-12 15:39 UTC (permalink / raw)
To: Eric Dumazet
Cc: netdev, davem, jeffrey.t.kirsher, edumazet, bhutchings, therbert,
alexander.duyck
In-Reply-To: <1342077259.3265.8232.camel@edumazet-glaptop>
On 07/12/2012 12:14 AM, Eric Dumazet wrote:
> On Wed, 2012-07-11 at 17:26 -0700, Alexander Duyck wrote:
>> This change adds capabilities to the driver for batching the MMIO write
>> involved with transmits. Most of the logic is based off of the code for
>> the qdisc scheduling.
>>
>> What I did is break the transmit path into two parts. We already had the
>> ndo_start_xmit function which has been there all along. The part I added
>> was ndo_complete_xmit which is meant to handle notifying the hardware that
>> frames are ready for delivery.
>>
>> To control all of this I added a net sysfs value for the Tx queues called
>> dispatch_limit. When 0 it indicates that all frames will notify hardware
>> immediately. When 1 or more the netdev_complete_xmit call will queue up to
>> that number of packets, and when the value is exceeded it will notify the
>> hardware and reset the pending frame dispatch count.
>>
>> Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
>> ---
> The idea is good, but do we really need so complex schem ?
>
> Most of the transmits are done from __qdisc_run()
>
> We could add logic in __qdisc_run()/qdisc_restart()
>
> qdisc_run_end() would then have to call ndo_complete_xmit() to make
> sure the MMIO is done.
The problem is in both of the cases where I have seen the issue the
qdisc is actually empty.
In the case of pktgen it does not use the qdisc layer at all. It just
directly calls ndo_start_xmit.
In the standard networking case we never fill the qdisc because the MMIO
write stalls the entire CPU so the application never gets a chance to
get ahead of the hardware. From what I can tell the only case in which
the qdisc_run solution would work is if the ndo_start_xmit was called on
a different CPU from the application that is doing the transmitting.
Thanks,
Alex
^ permalink raw reply
* Re: [PATCH] sch_sfb: Fix missing NULL check
From: David Miller @ 2012-07-12 15:33 UTC (permalink / raw)
To: eric.dumazet; +Cc: alan, netdev
In-Reply-To: <1342101021.3265.8261.camel@edumazet-glaptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Thu, 12 Jul 2012 15:50:21 +0200
> On Thu, 2012-07-12 at 06:25 -0700, David Miller wrote:
>> From: Alan Cox <alan@lxorguk.ukuu.org.uk>
>> Date: Thu, 12 Jul 2012 14:39:11 +0100
>>
>> > Signed-off-by: Alan Cox <alna@linux.intel.com>
>> ^^^^
>>
>> I'm truly astonished that you type in signoffs by hand Alan.
>
> Weel, I do the same ;)
You guys are weird :-)
> Feel free to add my
>
> Acked-by: Eric Dumazet <edumazet@google.com>
Applied, with signoff typo fixed too :-)
^ permalink raw reply
* Re: [PATCH 2/2] net: Update alloc frag to reduce get/put page usage and recycle pages
From: Alexander Duyck @ 2012-07-12 15:33 UTC (permalink / raw)
To: Eric Dumazet
Cc: Alexander Duyck, netdev, davem, jeffrey.t.kirsher, Eric Dumazet
In-Reply-To: <1342069601.3265.8218.camel@edumazet-glaptop>
On 07/11/2012 10:06 PM, Eric Dumazet wrote:
> On Wed, 2012-07-11 at 19:02 -0700, Alexander Duyck wrote:
>
>> The gain will be minimal if any with the 1500 byte allocations, however
>> there shouldn't be a performance degradation.
>>
>> I was thinking more of the ixgbe case where we are working with only 256
>> byte allocations and can recycle pages in the case of GRO or TCP. For
>> ixgbe the advantages are significant since we drop a number of the
>> get_page calls and get the advantage of the page recycling. So for
>> example with GRO enabled we should only have to allocate 1 page for
>> headers every 16 buffers, and the 6 slots we use in that page have a
>> good likelihood of being warm in the cache since we just keep looping on
>> the same page.
>>
> Its not possible to get 16 buffers per 4096 bytes page.
Actually I was talking about buffers from the device, not buffers from
the page. However, it is possible to get 16 head_frag buffers from the
same 4K page if we consider recycling. In the case of GRO we will end
up with the first buffer keeping the head_frag, and all of the remaining
head_frags will be freed before we call netdev_alloc_frag again. So
what will end up happening is that each GRO assembled frame from ixgbe
would start with a recycled page used for the previously freed
head_frags, the page will be dropped from netdev_alloc_frag after we run
out of space, a new page will be allocated for use as head_frags, and
finally those head_frags will be freed and recycled until we hit the end
of the GRO frame and start over. So if you count them all then we end
up using the page up to 16 times, maybe even more depending on how the
page offset reset aligns with the start of the GRO frame.
> sizeof(struct skb_shared_info)=0x140 320
>
> Add 192 bytes (NET_SKB_PAD + 128)
>
> Thats a minimum of 512 bytes (but ixgbe uses more) per skb.
>
> In practice for ixgbe, its :
>
> #define IXGBE_RXBUFFER_512 512 /* Used for packet split */
> #define IXGBE_RX_HDR_SIZE IXGBE_RXBUFFER_512
>
> skb = netdev_alloc_skb_ip_align(rx_ring->netdev, IXGBE_RX_HDR_SIZE)
>
> So 4 buffers per PAGE
>
> Maybe you plan to use IXGBE_RXBUFFER_256 or IXGBE_RXBUFFER_128 ?
I have a patch that is in testing in Jeff Kirsher's tree that uses
IXGBE_RXBUFFER_256. With your recent changes it didn't make sense to
use 512 when we would only copy 256 bytes into the head. With the size
set to 256 we will get 6 buffers per page without any recycling.
Thanks,
Alex
^ permalink raw reply
* Re: [PATCH 02/16] ipv4: Deliver ICMP redirects to sockets too.
From: Hiroaki SHIMODA @ 2012-07-12 15:21 UTC (permalink / raw)
To: David Miller; +Cc: netdev
In-Reply-To: <20120712.080653.1463195798230664640.davem@davemloft.net>
On Thu, 12 Jul 2012 08:06:53 -0700 (PDT)
David Miller <davem@davemloft.net> wrote:
> From: Hiroaki SHIMODA <shimoda.hiroaki@gmail.com>
> Date: Thu, 12 Jul 2012 23:58:37 +0900
>
> > So, I think avobe deleted checks about skb->len need to move to
> > ping_err() in case of packets are malformed.
>
> You would be wrong, the check belongs in icmp_socket_deliver().
Ah, OK. Thanks ;)
^ permalink raw reply
* Re: pull-request: can-next 2012-07-11
From: David Miller @ 2012-07-12 15:19 UTC (permalink / raw)
To: mkl; +Cc: netdev, linux-can
In-Reply-To: <4FFD8825.2060109@pengutronix.de>
From: Marc Kleine-Budde <mkl@pengutronix.de>
Date: Wed, 11 Jul 2012 16:05:25 +0200
> the fourth pull request for upcoming v3.6 net-next consist of a series
> of can_gw netlink cleanups done by Thomas Graf.
Pulled, thanks.
^ permalink raw reply
* Re: [PATCH net-next 0/7] be2net updates
From: David Miller @ 2012-07-12 15:16 UTC (permalink / raw)
To: padmanabh.ratnakar; +Cc: netdev
In-Reply-To: <a8f2c513-d398-4be6-a059-242dfc2c052a@exht1.ad.emulex.com>
From: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Date: Thu, 12 Jul 2012 19:25:01 +0530
> Padmanabh Ratnakar (7):
> be2net: Fix error while toggling autoneg of pause parameters
> be2net : Fix die temperature stat for Lancer
> be2net: Fix initialization sequence for Lancer
> be2net: Activate new FW after FW download for Lancer
> be2net: Fix cleanup path when EQ creation fails
> be2net: Fix port name in message during driver load
> be2net: Enable RSS UDP hashing for Lancer and Skyhawk
All applied, but like others have said you should document in
the driver what exactly the chip uses in it's RSS calculations
and in what circumstances.
^ permalink raw reply
* Re: [patch net-next 0/3] team: couple of patches
From: David Miller @ 2012-07-12 15:14 UTC (permalink / raw)
To: jpirko; +Cc: netdev
In-Reply-To: <20120712.081124.1207448876900334978.davem@davemloft.net>
From: David Miller <davem@davemloft.net>
Date: Thu, 12 Jul 2012 08:11:24 -0700 (PDT)
> From: Jiri Pirko <jpirko@redhat.com>
> Date: Wed, 11 Jul 2012 17:34:01 +0200
>
>> Jiri Pirko (3):
>> team: use function team_port_txable() for determing enabled and up
>> port
>> team: add broadcast mode
>> team: make team_port_enabled() and team_port_txable() static inline
>
> All applied, thanks.
Jiri, btw, any chance I can convince you to remove the EXPERIMENTAL
Kconfig dependency?
Code I've written and checked in myself over the past few days is
several orders of magnitude more experimental than the team driver
is :-)
^ permalink raw reply
* Re: [PATCH net-next 01/11] sfc: Implement 128-bit writes for efx_writeo_page
From: David Miller @ 2012-07-12 15:13 UTC (permalink / raw)
To: bhutchings; +Cc: netdev, linux-net-drivers
In-Reply-To: <1342048518.2613.60.camel@bwh-desktop.uk.solarflarecom.com>
From: Ben Hutchings <bhutchings@solarflare.com>
Date: Thu, 12 Jul 2012 00:15:18 +0100
> Add support for writing a TX descriptor to the NIC in one PCIe
> transaction on x86_64 machines.
>
> Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
This absolutely does not belong in a driver.
^ permalink raw reply
* Re: [patch net-next 0/3] team: couple of patches
From: David Miller @ 2012-07-12 15:11 UTC (permalink / raw)
To: jpirko; +Cc: netdev
In-Reply-To: <1342020844-3547-1-git-send-email-jpirko@redhat.com>
From: Jiri Pirko <jpirko@redhat.com>
Date: Wed, 11 Jul 2012 17:34:01 +0200
> Jiri Pirko (3):
> team: use function team_port_txable() for determing enabled and up
> port
> team: add broadcast mode
> team: make team_port_enabled() and team_port_txable() static inline
All applied, thanks.
^ permalink raw reply
* Re: [PATCH 02/16] ipv4: Deliver ICMP redirects to sockets too.
From: David Miller @ 2012-07-12 15:06 UTC (permalink / raw)
To: shimoda.hiroaki; +Cc: netdev
In-Reply-To: <20120712235837.4d611326830a16f9a035dd75@gmail.com>
From: Hiroaki SHIMODA <shimoda.hiroaki@gmail.com>
Date: Thu, 12 Jul 2012 23:58:37 +0900
> So, I think avobe deleted checks about skb->len need to move to
> ping_err() in case of packets are malformed.
You would be wrong, the check belongs in icmp_socket_deliver().
====================
>From f0a70e902f483295a8b6d74ef4393bc577b703d7 Mon Sep 17 00:00:00 2001
From: "David S. Miller" <davem@davemloft.net>
Date: Thu, 12 Jul 2012 08:06:04 -0700
Subject: [PATCH] ipv4: Put proper checks into icmp_socket_deliver().
All handler->err() routines expect that we've done a pskb_may_pull()
test to make sure that IP header length + 8 bytes can be safely
pulled.
Reported-by: Hiroaki SHIMODA <shimoda.hiroaki@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/ipv4/icmp.c | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index d01aeb4..ea3a996 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -640,6 +640,12 @@ static void icmp_socket_deliver(struct sk_buff *skb, u32 info)
const struct net_protocol *ipprot;
int protocol = iph->protocol;
+ /* Checkin full IP header plus 8 bytes of protocol to
+ * avoid additional coding at protocol handlers.
+ */
+ if (!pskb_may_pull(skb, iph->ihl * 4 + 8))
+ return;
+
raw_icmp_error(skb, protocol, info);
rcu_read_lock();
@@ -733,12 +739,6 @@ static void icmp_unreach(struct sk_buff *skb)
goto out;
}
- /* Checkin full IP header plus 8 bytes of protocol to
- * avoid additional coding at protocol handlers.
- */
- if (!pskb_may_pull(skb, iph->ihl * 4 + 8))
- goto out;
-
icmp_socket_deliver(skb, info);
out:
--
1.7.10.4
^ permalink raw reply related
* Re: [net-next 0/5][pull request] Intel Wired LAN Driver Updates
From: David Miller @ 2012-07-12 15:01 UTC (permalink / raw)
To: jeffrey.t.kirsher; +Cc: netdev, gospo, sassmann
In-Reply-To: <1341997769-22034-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Wed, 11 Jul 2012 02:09:24 -0700
> This series contains updates to ixgbe.
>
> The following are changes since commit 4715213d9cf40285492fff4092bb1fa8e982f632:
> bridge: fix endian
> and are available in the git repository at:
> git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
>
> Alexander Duyck (5):
> ixgbe: count q_vectors instead of MSI-X vectors
> ixgbe: Add upper limit to ring features
> ixgbe: Add feature offset value to ring features
> ixgbe: Clean up a useless switch statement and dead code in
> configure_srrctl
> ixgbe: Merge RSS and flow director ring register caching and
> configuration
Pulled, thanks Jeff.
^ permalink raw reply
* Re: [PATCH 1/1] net: sched: add ipset ematch
From: David Miller @ 2012-07-12 15:00 UTC (permalink / raw)
To: fw; +Cc: netdev, kadlec
In-Reply-To: <1342040217-5637-1-git-send-email-fw@strlen.de>
From: Florian Westphal <fw@strlen.de>
Date: Wed, 11 Jul 2012 22:56:57 +0200
> Can be used to match packets against netfilter ip sets created via ipset(8).
> skb->sk_iif is used as 'incoming interface', skb->dev is 'outgoing interface'.
>
> Since ipset is usually called from netfilter, the ematch
> initializes a fake xt_action_param, pulls the ip header into the
> linear area and also sets skb->data to the IP header (otherwise
> matching Layer 4 set types doesn't work).
>
> Tested-by: Mr Dash Four <mr.dash.four@googlemail.com>
> Signed-off-by: Florian Westphal <fw@strlen.de>
Applied, thanks.
^ permalink raw reply
* Re: [PATCH 02/16] ipv4: Deliver ICMP redirects to sockets too.
From: Hiroaki SHIMODA @ 2012-07-12 14:58 UTC (permalink / raw)
To: David Miller; +Cc: netdev
In-Reply-To: <20120712.011049.831106026936792516.davem@davemloft.net>
On Thu, 12 Jul 2012 01:10:49 -0700 (PDT)
David Miller <davem@davemloft.net> wrote:
>
> And thus, we can remove the ping_err() hack.
>
> Signed-off-by: David S. Miller <davem@davemloft.net>
> ---
> net/ipv4/icmp.c | 8 +-------
> 1 file changed, 1 insertion(+), 7 deletions(-)
>
> diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
> index 18e39d1..5885146 100644
> --- a/net/ipv4/icmp.c
> +++ b/net/ipv4/icmp.c
> @@ -782,13 +782,7 @@ static void icmp_redirect(struct sk_buff *skb)
> break;
> }
>
> - /* Ping wants to see redirects.
> - * Let's pretend they are errors of sorts... */
> - if (iph->protocol == IPPROTO_ICMP &&
> - iph->ihl >= 5 &&
> - pskb_may_pull(skb, (iph->ihl<<2)+8)) {
> - ping_err(skb, icmp_hdr(skb)->un.gateway);
> - }
> + icmp_socket_deliver(skb, icmp_hdr(skb)->un.gateway);
icmp_redirect() just checks skb->len is larger than
sizeof(struct iphdr) and then ping_err() is called.
In ping_err(), *icmph is derived from following code without
sanity check of skb->len. So, I think avobe deleted checks about
skb->len need to move to ping_err() in case of packets are malformed.
struct icmphdr *icmph = (struct icmphdr *)(skb->data+(iph->ihl<<2))
^ permalink raw reply
* Re: [PATCH net-next] netxen: fix link notification order
From: David Miller @ 2012-07-12 14:57 UTC (permalink / raw)
To: fbl; +Cc: netdev, sony.chacko, rajesh.borundia
In-Reply-To: <1342033015-31442-1-git-send-email-fbl@redhat.com>
From: Flavio Leitner <fbl@redhat.com>
Date: Wed, 11 Jul 2012 15:56:55 -0300
> First update the adapter variables with the current speed and
> mode before fire the notification. Otherwise, the get_settings()
> may provide old values.
>
> Signed-off-by: Flavio Leitner <fbl@redhat.com>
Applied.
^ permalink raw reply
* Re: [PATCH net-next v2 0/7] ieee802.15.4 general fixes
From: David Miller @ 2012-07-12 14:56 UTC (permalink / raw)
To: alex.bluesman.smirnov; +Cc: eric.dumazet, netdev
In-Reply-To: <1341991368-11800-1-git-send-email-alex.bluesman.smirnov@gmail.com>
From: Alexander Smirnov <alex.bluesman.smirnov@gmail.com>
Date: Wed, 11 Jul 2012 11:22:41 +0400
> Dear David, Eric,
>
> this patch-set is mostly intended to fix sparse and LOCKDEP warnings.
> It mostly contains some my previous patches reworked and extended according
> to the hints from Eric Dumazet and Fengguang Wu. Many thanks to they!
>
> Changes since v1:
> 1. A new patch from Tony Cheneau was added. The fragmentation stops working
> after some amount of packets sent. This patch fixes this issue.
> 2. 6lowpan fragment deleting routine: I removed spinlocks from timer_expired
> handler and use spin_lock_bh to disable concurrency races with timer interrupt.
> 3. at86rf230 irq handler was a little bit modified
Series applied, but you don't need to grab a spinlock to only
load one interger from some datastructure. I mean:
lock();
ret = p->foo;
unlock();
is completely pointless.
^ permalink raw reply
* Re: [patch -next] smsc95xx: signedness bug in get_regs()
From: David Miller @ 2012-07-12 14:55 UTC (permalink / raw)
To: dan.carpenter-QHcLZuEGTsvQT0dZR+AlfA
Cc: steve.glendinning-sdUf+H5yV5I,
emeric.vigier-4ysUXcep3aM1wj+D4I0NRVaTQe2KTcn/,
netdev-u79uwXL29TY76Z2rM5mHXA, linux-usb-u79uwXL29TY76Z2rM5mHXA,
kernel-janitors-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20120711063251.GB11812-mgFCXtclrQlZLf2FXnZxJA@public.gmane.org>
From: Dan Carpenter <dan.carpenter-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
Date: Wed, 11 Jul 2012 09:32:51 +0300
> "retval" has to be a signed integer for the error handling to work.
>
> Signed-off-by: Dan Carpenter <dan.carpenter-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
Applied.
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [RFC PATCH v2] tcp: TCP Small Queues
From: Tom Herbert @ 2012-07-12 14:55 UTC (permalink / raw)
To: Eric Dumazet
Cc: David Miller, rick.jones2, ycheng, dave.taht, netdev, codel,
mattmathis, nanditad, ncardwell, andrewmcgr
In-Reply-To: <1342079487.3265.8245.camel@edumazet-glaptop>
On Thu, Jul 12, 2012 at 12:51 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> On Thu, 2012-07-12 at 00:37 -0700, David Miller wrote:
>> From: Eric Dumazet <eric.dumazet@gmail.com>
>> Date: Thu, 12 Jul 2012 09:34:19 +0200
>>
>> > On Thu, 2012-07-12 at 01:49 +0200, Eric Dumazet wrote:
>> >
>> >> The 10Gb receiver is a net-next kernel, but the 1Gb receiver is a 2.6.38
>> >> ubuntu kernel. They probably have very different TCP behavior.
>> >
>> >
>> > I tested TSQ on bnx2x and 10Gb links.
>> >
>> > I get full rate even using 65536 bytes for
>> > the /proc/sys/net/ipv4/tcp_limit_output_bytes tunable
>>
>> Great work Eric.
>
> Thanks !
>
This is indeed great work! A couple of comments...
Do you know if there are are any qdiscs that function less efficiently
when we are restricting the number of packets? For instance, will HTB
work as expected in various configurations?
One extension to this work be to make the limit dynamic and mostly
eliminate the tunable. I'm thinking we might be able to correlate the
limit to the BQL limit of the egress queue for the flow it there is
one.
Assuming all work conserving qdiscs the minimal amount of outstanding
host data for a queue could be associated with the BQL limit of the
egress NIC queue. We want to minimize the outstanding data so that:
sum(data_of_tcp_flows_share_same_queue) > bql_limit_for _queue
So this could imply a per flow limit of:
tcp_limit = max(bql_limit - bql_inflight, one_packet)
For a single active connection on a queue, the tcp_limit is equal to
the BQL limit. Once the BQL limit is hit in the NIC, we only need one
packet outstanding per flow to maintain flow control. For fairness,
we might need "one_packet" to actually be max GSO data. Also, this
disregards any latency of scheduling and running the tasklet, that
might need to be taken into account also.
Tom
^ permalink raw reply
* Re: [PATCH v2 0/2] net: support for NS8390 based ethernet on ColdFire CPU boards
From: David Miller @ 2012-07-12 14:55 UTC (permalink / raw)
To: gerg; +Cc: netdev, linux-m68k
In-Reply-To: <1341445800-396-1-git-send-email-gerg@snapgear.com>
From: <gerg@snapgear.com>
Date: Thu, 5 Jul 2012 09:49:58 +1000
>
> This is version 2 of patches that add platform support for using the NS8390
> based ethernet ports used on some ColdFire CPU boards. This version
> incorporates only minor changes from the first.
>
> Patches to use these NS8390 devices on ColdFire boards have existed
> out-of-tree for years. Some of the base IO definitions (those in
> arch/m68k/include/asm/mcfne.h) have been in mainline, but unused for most
> of that time.
>
> The first patch just neatens up mcfne.h (moving it to mcf8390.h). The
> second patch is the platform driver. The first patch would normaly just
> go through the m68knommu git tree, but I figured keeping these together made
> more sense.
Applied, thanks.
Can you explain why we've had this completely unused header
file mcfne.h in the tree? Was it used by some external driver
sources that were never merged?
^ permalink raw reply
* Re: [net-next:master 90/102] net/ipv4/route.c:1283:9: warning: unused variable 'saddr'
From: Fengguang Wu @ 2012-07-12 14:54 UTC (permalink / raw)
To: David Miller; +Cc: kernel-janitors, netdev
In-Reply-To: <20120712.074058.753681400854318989.davem@davemloft.net>
OK!
On Thu, Jul 12, 2012 at 07:40:58AM -0700, David Miller wrote:
>
> There's not need to report these to kernel-janitors if it's a
> net-next specific issue and I'm going to fix it up 5 minutes
> after you report it.
>
> ====================
> [PATCH] ipv4: Fix warnings in ip_do_redirect() for some configurations.
>
> Reported-by: Fengguang Wu <fengguang.wu@intel.com>
> Signed-off-by: David S. Miller <davem@davemloft.net>
> ---
> net/ipv4/route.c | 10 ++++++----
> 1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/net/ipv4/route.c b/net/ipv4/route.c
> index 23bbe29..9319bf1 100644
> --- a/net/ipv4/route.c
> +++ b/net/ipv4/route.c
> @@ -1275,12 +1275,9 @@ static void rt_del(unsigned int hash, struct rtable *rt)
>
> static void ip_do_redirect(struct dst_entry *dst, struct sk_buff *skb)
> {
> - const struct iphdr *iph = (const struct iphdr *) skb->data;
> __be32 new_gw = icmp_hdr(skb)->un.gateway;
> __be32 old_gw = ip_hdr(skb)->saddr;
> struct net_device *dev = skb->dev;
> - __be32 daddr = iph->daddr;
> - __be32 saddr = iph->saddr;
> struct in_device *in_dev;
> struct neighbour *n;
> struct rtable *rt;
> @@ -1336,11 +1333,16 @@ static void ip_do_redirect(struct dst_entry *dst, struct sk_buff *skb)
>
> reject_redirect:
> #ifdef CONFIG_IP_ROUTE_VERBOSE
> - if (IN_DEV_LOG_MARTIANS(in_dev))
> + if (IN_DEV_LOG_MARTIANS(in_dev)) {
> + const struct iphdr *iph = (const struct iphdr *) skb->data;
> + __be32 daddr = iph->daddr;
> + __be32 saddr = iph->saddr;
> +
> net_info_ratelimited("Redirect from %pI4 on %s about %pI4 ignored\n"
> " Advised path = %pI4 -> %pI4\n",
> &old_gw, dev->name, &new_gw,
> &saddr, &daddr);
> + }
> #endif
> ;
> }
> --
> 1.7.10.4
^ permalink raw reply
* Re: [patch -next] tcp: update a call to tcp_metric_set()
From: David Miller @ 2012-07-12 14:50 UTC (permalink / raw)
To: dan.carpenter; +Cc: kuznet, jmorris, yoshfuji, kaber, netdev, kernel-janitors
In-Reply-To: <20120712144637.GA24202@elgon.mountain>
From: Dan Carpenter <dan.carpenter@oracle.com>
Date: Thu, 12 Jul 2012 17:46:38 +0300
> We changed how the metrics were stored so RTAX_CWND needs to be changed
> to TCP_METRIC_CWND or it leads to write past the end of the
> ->tcpm_vals[] array.
>
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Update your tree, this is already fixed.
^ permalink raw reply
* [patch -next] tcp: update a call to tcp_metric_set()
From: Dan Carpenter @ 2012-07-12 14:46 UTC (permalink / raw)
To: David S. Miller
Cc: Alexey Kuznetsov, James Morris, Hideaki YOSHIFUJI,
Patrick McHardy, netdev, kernel-janitors
We changed how the metrics were stored so RTAX_CWND needs to be changed
to TCP_METRIC_CWND or it leads to write past the end of the
->tcpm_vals[] array.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
diff --git a/net/ipv4/tcp_metrics.c b/net/ipv4/tcp_metrics.c
index 1fd83d3..c6b0f20 100644
--- a/net/ipv4/tcp_metrics.c
+++ b/net/ipv4/tcp_metrics.c
@@ -412,7 +412,8 @@ void tcp_update_metrics(struct sock *sk)
max(tp->snd_cwnd >> 1, tp->snd_ssthresh));
if (!tcp_metric_locked(tm, TCP_METRIC_CWND)) {
val = tcp_metric_get(tm, TCP_METRIC_CWND);
- tcp_metric_set(tm, RTAX_CWND, (val + tp->snd_cwnd) >> 1);
+ tcp_metric_set(tm, TCP_METRIC_CWND,
+ (val + tp->snd_cwnd) >> 1);
}
} else {
/* Else slow start did not finish, cwnd is non-sense,
^ permalink raw reply related
* [patch] qlge: fix an "&&" vs "||" bug
From: Dan Carpenter @ 2012-07-12 14:47 UTC (permalink / raw)
To: Anirban Chakraborty
Cc: Jitendra Kalsaria, Ron Mercer, linux-driver, netdev,
kernel-janitors
The condition is always true so WOL will never work.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
diff --git a/drivers/net/ethernet/qlogic/qlge/qlge_ethtool.c b/drivers/net/ethernet/qlogic/qlge/qlge_ethtool.c
index 3d4462b..6f316ab 100644
--- a/drivers/net/ethernet/qlogic/qlge/qlge_ethtool.c
+++ b/drivers/net/ethernet/qlogic/qlge/qlge_ethtool.c
@@ -440,7 +440,7 @@ static int ql_set_wol(struct net_device *ndev, struct ethtool_wolinfo *wol)
unsigned short ssys_dev = qdev->pdev->subsystem_device;
/* WOL is only supported for mezz card. */
- if (ssys_dev != QLGE_MEZZ_SSYS_ID_068 ||
+ if (ssys_dev != QLGE_MEZZ_SSYS_ID_068 &&
ssys_dev != QLGE_MEZZ_SSYS_ID_180) {
netif_info(qdev, drv, qdev->ndev,
"WOL is only supported for mezz card\n");
^ permalink raw reply related
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