* Herbert Xu <[EMAIL PROTECTED]> 2006-08-16 21:12
> On Wed, Aug 16, 2006 at 12:58:56PM +0200, Thomas Graf wrote:
> >
> > All route and tc notifications already use the pid so applications
> > can decide whether the event was caused by them. A notification
> > is a
* Herbert Xu <[EMAIL PROTECTED]> 2006-08-16 21:57
> On Wed, Aug 16, 2006 at 01:40:03PM +0200, Thomas Graf wrote:
> >
> > It was added to help quagga identify which route modifications
> > are self caused. It's not possible to use rtm_protocol for this
> > pu
* jamal <[EMAIL PROTECTED]> 2006-08-16 08:04
> current->pid i think is coming out to be a bad idea. Thomas' patches
> revert it out. Again this has everything to do with the original idea
> what maps to pid now changing to socketid.
It probably developed from autobind using current->tid.
-
To unsu
* Herbert Xu <[EMAIL PROTECTED]> 2006-08-16 21:39
> For a broadcast notification, the nlmsg_pid field is meaningless
> because the nlmsg_seq field is also meaningless. I'm not denying
> that it wouldn't be useful to have the originator's socket address
> in there. What I'm saying is that it's the
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19.git/net/core/rtnetlink.c
===
--- net-2.6.19.git.orig/net/core/rtnetlink.c
+++ net-2.6.19.git/net/core/rtnetlink.c
@@ -188,22 +188,27 @@ void rtnl_set_sk_err(u32
Provide a simple ip6_ins_rt() for the majority of users and
an alternative for the exception via netlink. Avoids code
obfuscation.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19.git/include/net/ip6_r
authorship information for notifications.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19.git/include/net/ip6_fib.h
===
--- net-2.6.19.git.orig/include/net/ip6_fib.h
+++ net-2.6.19.git/include/net/ip6_fib.h
@@
Same approach as for IPv4, simplifies the interface
and makes it extendable without breaking the ioctl
interface.
Same memory corruption fixes in RTM_GETROUTE as for IPv4.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majord
Fixes various unvalidated netlink attributes causing memory
corruptions when left empty by userspace applications.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19.git/net/ipv6/route.c
===
--- net-2.6.19.git.or
Provide a simple ip6_del_rt() for the majority of users and
an alternative for the exception via netlink. Avoids code
obfuscation.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19.git/include/net/ip6_r
Converts existing NLA_STRING attributes to use the new
validation features, saving a couple of temporary buffers.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19.git/net/core/rtnetlink.c
===
--- net-2.6.19.gi
Validation of netlink string attributes was weak, forcing everyone
to use nla_strlcpy() to copy the attribute into a temporary buffer.
This patchset implements length validation checks for existing
NLA_STRING attributes and adds a new type NLA_NUL_STRING for
NUL terminated strings.
-
To unsubscri
ify a maximum length of the
string.
Aims at easing the pain with using nla_strlcpy() on temporary
buffers.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19.git/include/net/netlink.h
===
--- net-2.6.19.git.orig/
t required to be NUL terminated
is extended to provide means to specify a maximum length of the
string.
Aims at easing the pain with using nla_strlcpy() on temporary
buffers.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19
* H. Peter Anvin <[EMAIL PROTECTED]> 2006-08-22 17:31
> Specifically, Linux will not ProxyARP for an address unless it has a
> route for it, *and* that route either has a DNAT marking or points to a
> different interface than the input interface:
I can think of a very ugly way: Use netfilter to
* David Miller <[EMAIL PROTECTED]> 2006-08-23 15:14
> From: Thomas Graf <[EMAIL PROTECTED]>
> Date: Wed, 23 Aug 2006 21:14:25 +0200
>
> > * H. Peter Anvin <[EMAIL PROTECTED]> 2006-08-22 17:31
> > > Specifically, Linux will not ProxyARP for an address
* Johannes Berg <[EMAIL PROTECTED]> 2006-08-22 15:52
> +static struct nla_policy nl80211_policy[NL80211_ATTR_MAX+1] __read_mostly = {
> + [NL80211_ATTR_IFINDEX] = { .type = NLA_U32 },
> + [NL80211_ATTR_WIPHY] = { .type = NLA_U32 },
> + [NL80211_ATTR_FLAGS] = { .type = NLA_U32 },
> +
* Johannes Berg <[EMAIL PROTECTED]> 2006-08-24 18:07
> +static int nl80211_get_cmdlist(struct sk_buff *skb, struct genl_info *info)
> +{
> + struct nl80211_registered_driver *drv;
> + struct sk_buff *msg;
> + void *hdr;
> + int err;
> + struct nlattr *start;
> + u8 *data;
>
* Michael Buesch <[EMAIL PROTECTED]> 2006-08-24 19:09
> On Thursday 24 August 2006 18:07, Johannes Berg wrote:
> > + pkt = alloc_skb(framelen, GFP_KERNEL);
>
> if (!pkt)
> return -ENOMEM;
>
> > + pktdata = skb_put(pkt, framelen);
> > + memcpy(pktdata, frame, framelen);
>
> having two
* Johannes Berg <[EMAIL PROTECTED]> 2006-08-25 11:04
> I completely reworked that now so it will:
> * create a nested NL80211_ATTR_INTERFACE_LIST with nested {
>* 1..N attributes, with nested {
> * ATTR_IFINDEX and
> * ATTR_IFNAME
>}
> }
>
> how does that sound? Maybe I should
* Patrick McHardy <[EMAIL PROTECTED]> 2006-08-25 12:29
> This patch adds support to mask the nfmark value before the lookup
> the the fw classifier. Unfortunately it has some drawbacks, so I'd
> be interested if anyone can think of a better way.
>
> The problem is that in order to avoid walking th
* YOSHIFUJI Hideaki / ?$B5HF#1QL@ <[EMAIL PROTECTED]> 2006-08-25 17:21
> commit 10204d532f5f8bb379009ba0bee2113bafda72be
> Author: YOSHIFUJI Hideaki <[EMAIL PROTECTED]>
> Date: Mon Aug 21 19:22:01 2006 +0900
>
> [IPV6] ROUTE: Routing by FWMARK.
>
> Based on patch by Jean Lorchat <[E
* YOSHIFUJI Hideaki / ?$B5HF#1QL@ <[EMAIL PROTECTED]> 2006-08-26 00:08
> Hello.
>
> Here's some IPv6 policy rouging fixes on top of net-2.6.19 tree.
>
> [PATCH 1/3] [IPV6] ROUTE: Fix FWMARK support.
> [PATCH 2/3] [IPV6] ROUTE: Fix size of fib6_rule_policy.
>
> If we accept Patrick's IPv4 fwm
* Olaf Kirch <[EMAIL PROTECTED]> 2006-08-29 10:24
> we had bug reports from people seeing lots of spurious messages
> like the following:
>
> kernel: peth0: received packet with own address as source address.
>
> and
>
> xenbr0: duplicate address detected!
>
> This is on a Xen enabled machine,
* Andrew Morton <[EMAIL PROTECTED]> 2006-08-29 09:16
>
>
> Begin forwarded message:
>
> Date: Tue, 29 Aug 2006 10:25:07 -0400
> From: [EMAIL PROTECTED]
> To: Andrew Morton <[EMAIL PROTECTED]>
> Cc: linux-kernel@vger.kernel.org
> Subject: 2.6.18-rc4-mm3: BUG: warning at include/net/dst.h:154/dst_
]
op unknown (0x03) [DOIT]
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/include/linux/genetlink.h
===
--- net-2.6.19.orig/include/linux/genetlink.h 2006-08-31 16:01:16.0
A typo introduced by myself which leads to freeing the skb
containing the netlink message when it should free the newly
allocated skb for the reply.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/net/core/rtnet
Converts all of the IPv6 address configuration code to the new
netlink api and removes a bogus check on IFF_UP when adding
new addresses.
--
VGER BF report: U 0.467674
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/net/ipv6/addrconf.c
===
--- net-2.6.19.orig/net/ipv6/addrconf.c 2006-09-01 23:15:14.0 +0200
+++ net-2.6.19/net/ipv6/addrconf.c 2006-09-01
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/net/ipv6/addrconf.c
===
--- net-2.6.19.orig/net/ipv6/addrconf.c 2006-09-01 23:14:47.0 +0200
+++ net-2.6.19/net/ipv6/addrconf.c 2006-09-01
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/net/ipv6/addrconf.c
===
--- net-2.6.19.orig/net/ipv6/addrconf.c 2006-09-01 23:15:18.0 +0200
+++ net-2.6.19/net/ipv6/addrconf.c 2006-09-01
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/net/ipv6/addrconf.c
===
--- net-2.6.19.orig/net/ipv6/addrconf.c 2006-09-01 23:15:21.0 +0200
+++ net-2.6.19/net/ipv6/addrconf.c 2006-09-01
e the NLM_F_CREATE check is removed.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/net/ipv6/addrconf.c
===
--- net-2.6.19.orig/net/ipv6/addrconf.c 2006-09-01 23:15:25.0 +0200
+++ net-2.6.19/net/ipv6/addrconf.c
Same behaviour as IPv4, using IFF_UP is a no-no anyway.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/net/ipv6/addrconf.c
===
--- net-2.6.19.orig/net/ipv6/addrconf.c 2006-09-01 23:15:24.0 +0200
+
Replaces INET6_IFADDR_RTA_SPACE with a new function calculating
the total required message size for all address messages.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/net/ipv6/addrconf.c
===
--- net-2.6.1
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.19/net/ipv6/addrconf.c
===
--- net-2.6.19.orig/net/ipv6/addrconf.c 2006-09-01 23:15:19.0 +0200
+++ net-2.6.19/net/ipv6/addrconf.c 2006-09-01
* H. Peter Anvin <[EMAIL PROTECTED]> 2006-09-07 15:28
> Thomas Graf wrote:
> >What about adding blackhole device to be used for such routes.
> >I believe it would be good architecture to always use devices
> >to state directions packets are being received from and sen
* Johannes Berg <[EMAIL PROTECTED]> 2006-09-14 09:44
> This patch adds an NLA_CUSTOM_CHECK type for netlink attributes
> in order to be able to centrally define new attribute structures
> instead of having to check these special types in each function
> that uses such an attribute.
>
> nl80211 wil
* YOSHIFUJI Hideaki / ?$B5HF#1QL@ <[EMAIL PROTECTED]> 2006-09-19 00:08
> [NET]: Move netlink interface bits to linux/if_link.h.
>
> Moving netlink interface bits to linux/if.h is rather troublesome for
> applications including both linux/if.h (which was changed to be included
>
* David Miller <[EMAIL PROTECTED]> 2006-09-18 14:22
> From: Alexey Kuznetsov <[EMAIL PROTECTED]>
> Date: Tue, 19 Sep 2006 01:03:21 +0400
>
> > 1. It even does not disable possibility to record timestamp inside
> >driver, which Alan was afraid of. The sequence is:
> >
> > if (!skb->tstamp.
* Johannes Berg <[EMAIL PROTECTED]> 2006-09-14 11:21
> On Thu, 2006-09-14 at 10:14 +0200, Thomas Graf wrote:
>
> > Looks good, we have to watch the size of struct nla_policy though.
> > This bumps the size from 4 bytes to 16 bytes on 64bit architectures
> > which m
* Johannes Berg <[EMAIL PROTECTED]> 2006-09-14 12:53
> This is some preliminary code how I'm currently thinking (and that might
> change radically :) ) configuration might look like.
>
> It uses the patch I previously posted to make genetlink attributes
> custom-definable.
>
> --- wireless-dev.or
* Johannes Berg <[EMAIL PROTECTED]> 2006-09-20 09:03
> > Just use a nested attribute here, this new array format you introduce
> > having 1 byte ID, 1 byte len is equivalent to using a set of nested
> > attributes with nla_type=id, nla_len=len.
>
> No, it is only validated, it is then supposed to
* James Morris <[EMAIL PROTECTED]> 2006-05-27 13:21
> Actually, a possible solution here is to completely remove all internal
> knowledge of netlink messages from SELinux and have the netfilter
> framework and protocols provide methods to determine message types and
> permissions.
Right, regard
* jamal <[EMAIL PROTECTED]> 2006-05-31 08:20
> The challenge is how to inform SELinux of these permissions.
> The access limit could be done by putting a SELinux hook at the time the
> skb gets to the generic netlink code?
> Note: There's actually two things that can be classified for access
> con
* jamal <[EMAIL PROTECTED]> 2006-05-31 08:00
> We could start by just adding a check for NETLINK_GENERIC in your table
> (as is done generally for other netlink families/protocols with SELinux)
> and then do the fine-grained stuff. I think that checking for attributes
> instead of types will need t
* James Morris <[EMAIL PROTECTED]> 2006-05-31 11:42
> On Wed, 31 May 2006, jamal wrote:
>
> > To also answer your other email:
> > Look at security/selinux/nlmsgtab.c for example for NETLINK_ROUTE
> > and compare with NETLINK_GENERIC to see the hole. I was suggesting if
> > we started by just add
* jamal <[EMAIL PROTECTED]> 2006-06-19 09:41
> // the attributes you want to own
>
> enum {
> FOOBAR_ATTR_UNSPEC,
> FOOBAR_ATTR_TYPE,
> FOOBAR_ATTR_TYPEID,
> FOOBAR_ATTR_TYPENAME,
> FOOBAR_ATTR_OPER,
> /* add future attributes here */
> __FOOBA
Hello
> > > TODO:
> > > a) Add a more complete compiling kernel module with events.
> > > Have Thomas put his Mashimaro example and point to it.
> >
> > I guess we have a legal issue here ;)
> >
>
> change the name ;->
Ask Mr. Mashimaro has become my replacement for 8ball. Renaming
it would le
This patchset transforms skb->input_dev based on a device
reference to skb->iif based on an interface index moving
towards accurate iif information for routing and classification
through the following changesets:
[NET]: Use interface index to keep input dev
Updating iif to the VLAN device helps keeping routing
namespaces defined in case packets from multiple VLANs
collapse on a single device again.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.git/net/8021q/vlan
Using the interface index instead of a direct reference
allows a safe usage beyond the scope where an interface
could disappear.
The old input_dev field was incorrectly made dependant
on CONFIG_NET_CLS_ACT in skb_copy().
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.git/i
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6.git/include/linux/tc_ematch/tc_em_meta.h
===
--- net-2.6.git.orig/include/linux/tc_ematch/tc_em_meta.h
+++ net-2.6.git/include/linux/tc_ematch/tc_em_meta.h
@@ -81,6
* David Miller <[EMAIL PROTECTED]> 2006-06-26 10:46
> From: Patrick McHardy <[EMAIL PROTECTED]>
> Date: Mon, 26 Jun 2006 19:04:15 +0200
>
> > I know this was discussed before, but I can't remember the
> > exact outcome. Why don't we just unconditionally update iif
> > in netif_receive_skb()?
>
>
* Patrick McHardy <[EMAIL PROTECTED]> 2006-06-27 00:29
> Doesn't sound too silly (actually quite nifty in my opinion),
> but I'm not sure I understand correctly, binding to ifb doesn't
> work since it changes both skb->dev and skb->input_dev.
I don't understand this concern. So far the mirred acti
* Thomas Graf <[EMAIL PROTECTED]> 2006-06-26 16:54
> This patchset transforms skb->input_dev based on a device
> reference to skb->iif based on an interface index moving
> towards accurate iif information for routing and classification
> through the following changeset
* jamal <[EMAIL PROTECTED]> 2006-06-27 09:07
> I am reading the thread backwards, so i may miss some of the obvious.
> I also dont remember the exact discussion we had - but the consensus was
> to leave the field setting as is.
> Note the meta-setter (been sitting on it for too long) also set the
>
* jamal <[EMAIL PROTECTED]> 2006-06-28 08:22
> Let me see if i understood correctly:
> you have a device from both vlan1 and vlan2 both being redirected to
> ifb1? and vlan 3, 4=> ifb2?
> And i take it to make it interesting vlan1,2 on eth0 and vlan3,4 ontop
> of eth1? note the iflink would/should
These patches have been floating around during the enlarge cb[]
discussion a few months back but I never had the drive to
actually submit them officially. Inspired by dave's skb writings
I now do.
Further additional ideas to reduce sk_buff in size:
o If we can't get rid of mac_len it could at le
Reduce local_df to a bit field and ip_summed to a 2 bits
field thus saving 13 bits. Move bit fields, packet type,
and protocol into the spare area between the priority
and the destructor. Saves 4 bytes on both, 32bit and
64bit architectures.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6/include/linux/skbuff.h
===
--- net-2.6.orig/include/linux/skbuff.h
+++ net-2.6/include/linux/skbuff.h
@@ -183,7 +183,6 @@ struct skb_shared_info {
* @pr
Patches 1-3 should be straight forward. Patch 4 adds a new qdisc
implementing a black hole, nothing spectacular but quite useful.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordo
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6/include/net/pkt_sched.h
===
--- net-2.6.orig/include/net/pkt_sched.h
+++ net-2.6/include/net/pkt_sched.h
@@ -207,8 +207,6 @@ psched_tod_diff(int delta_sec, int
Adds qdisc_alloc() to share code between qdisc_create()
and qdisc_create_dflt(). Hides the qdisc alignment behind
macros and makes use of them.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6/include/net/sch_gen
operation fail so the user is able to fix
his mistake.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6/net/sched/sch_api.c
===
--- net-2.6.orig/net/sched/sch_api.c
+++ net-2.6/net/sched/sch_api.c
@@ -454,15 +
Useful in combination with classful qdiscs to drop or
temporary disable certain flows, e.g. one could block
specific ds flows with dsmark.
Unlike the noop qdisc it can be controlled by the user and
statistic accounting is done.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2
* jamal <[EMAIL PROTECTED]> 2005-07-04 09:16
> On Mon, 2005-04-07 at 13:48 +0200, Thomas Graf wrote:
> > plain text document attachment (sch_blackhole)
> > Useful in combination with classful qdiscs to drop or
> > temporary disable certain flows, e.g. one could block
* jamal <[EMAIL PROTECTED]> 2005-07-05 08:09
> On Mon, 2005-04-07 at 15:49 +0200, Thomas Graf wrote:
> > Assume you have a classful setup with per client classes and you
> > want to disable one such class for a moment (maybe the customer
> > didn't pay the bills ;
* jamal <[EMAIL PROTECTED]> 2005-07-05 09:04
> On Tue, 2005-05-07 at 14:30 +0200, Thomas Graf wrote:
>
> >
> > It is really argueable, my point is that it adds quite some flexibility
> > at no cost and I don't want to miss it anymore. ;-> Having a way to
&
* Patrick McHardy <[EMAIL PROTECTED]> 2005-07-05 16:06
> Thomas Graf wrote:
> >* jamal <[EMAIL PROTECTED]> 2005-07-05 09:04
> >
> >>BTW, enqueue returning the code by calling the qdisc::drop() - is that
> >>gonna cause some issues?
> >
> &g
* David S. Miller <[EMAIL PROTECTED]> 2005-07-05 14:19
> From: jamal <[EMAIL PROTECTED]>
> Date: Tue, 05 Jul 2005 11:13:04 -0400
>
> > On Tue, 2005-05-07 at 17:03 +0200, Patrick McHardy wrote:
> >
> > > Sounds reasonable. As you said, blackhole implies NET_XMIT_SUCCESS,
> > > so I propose to use
* David S. Miller <[EMAIL PROTECTED]> 2005-07-05 14:35
> From: Thomas Graf <[EMAIL PROTECTED]>
> Date: Tue, 5 Jul 2005 23:28:10 +0200
>
> > Can we agree on NET_XMIT_CN? The dev_xmit callers handle it the same
> > way as NET_XMIT_SUCCESS but we still have a chanc
Useful in combination with classful qdiscs to drop or
temporary disable certain flows, e.g. one could block
specific ds flows with dsmark.
Unlike the noop qdisc it can be controlled by the user and
statistic accounting is done.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2
* Patrick McHardy <[EMAIL PROTECTED]> 2005-07-17 04:09
> Kill skb->tc_classid. It leaves a hole on 64 bit machines, I didn't
> try to plug it since with all the planned changes we probably need
> some restructuring at the end anyway.
I have no objections to this, looks perfectly
fine to me.
> @@
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6/net/sched/em_text.c
===
--- net-2.6.orig/net/sched/em_text.c2005-06-24 13:17:08.0 +0200
+++ net-2.6/net/sched/em_text.c 2005-07-17 19:33:30.0
e loop,
this approach preserves the possibility to increase the number
of bands at compile time.
Signed-off-by: Thomas Graf <[EMAIL PROTECTED]>
Index: net-2.6/net/sched/sch_generic.c
===
--- net-2.6.orig/net/sched/sch_generic.c
* David S. Miller <[EMAIL PROTECTED]> 2005-07-21 17:07
> Thomas, this kills the TCF_META_ID_REALDEV stuff, so we should
> kill it in 2.6.13-rcX too so that nobody starts using it in
> userspace ok?
Sure. I was just thinking that maybe we should delay
the iproute2 release with the ematch bits until
> Hopefully we can weed out the unusable ematch bits before 2.6.13 is
> released. Therefore, once 2.6.13 goes out the iproute2 update should
> be OK.
Sounds good.
> I'm hoping that since we're doing the SKB shrinking in parallel in the
> net-2.6.14 tree with the ongoing 2.6.13 bug fixing, we sho
* Stephen Hemminger <[EMAIL PROTECTED]> 2005-07-22 13:48
> I don't see how the ematch iproute2 stuff depends on SKB shrinking.
> The CVS repository has the latest ematch stuff, just testing and
> checking before the next drop.
We cannot remove TCF_META_ID_* ids without breaking backwards
compatibi
* Patrick McHardy <[EMAIL PROTECTED]> 2005-07-23 06:39
> I'm trying to convert tc_index to a u16, which looks OK in all places
> but sched_gred. I can't make sense of this code:
>
> /* fix tc_index? --could be controvesial but needed for
>requeueing */
> skb->tc_index=(
* Jamal Hadi Salim <[EMAIL PROTECTED]> 2005-07-23 09:32
> @@ -205,7 +205,6 @@ struct sk_buff {
> struct sock *sk;
> struct timeval stamp;
> struct net_device *dev;
> - struct net_device *input_dev;
> struct net_device *real_dev;
>
* Jamal Hadi Salim <[EMAIL PROTECTED]> 2005-07-23 10:37
> ifindex has to be 32 bit because most management apps (SNMP etc) expect
> it to be 32 bit. I havent scanned the code but making int actually
> doesnt seem right - unless some code is expecting it to be a -ve number
> somewhere; it should be
* Jamal Hadi Salim <[EMAIL PROTECTED]> 2005-07-23 10:14
> Most of these pieces of h/ware have different approaches to how those
> underlying rings are used sending out to the wire. As an example, it is
> clear the wireless folks have time slots + prios as the desired qos
> scheme. This will map nic
* Patrick McHardy <[EMAIL PROTECTED]> 2005-07-23 17:58
> Jamal Hadi Salim wrote:
> >Did you check if there was no side effect with the macros that
> >manipulate tc_verdict?
>
> Of course, these macros are basically the only user. They could use
> some cleanup btw, I think I'll do that later.
Hold
* David S. Miller <[EMAIL PROTECTED]> 2005-07-24 18:08
> The key is what should happen when the ring for prio X fills
> up? netif_stop_queue() in it's current form is the wrong
> thing to do, because it prevents lower priority packets from
> being queued which is exactly what we want to do if thos
* Thomas Graf <[EMAIL PROTECTED]> 2005-07-23 18:23
> Hold on, I'm working on a patch to introduce TC_CB() so we can
> move some of the fields of tc_verd into the control buffer so
> I can do the cleanup at the same time.
I tried but none of the bits can be moved into cb as-is
> I am not against reporting back (NETDEV_TX_CONTINUE or
> NETDEV_TX_GIVE_ME_THE_SAME_Q_AGAIN for example);
> what worries me is that the stopping of the queue being complex. As i
> said, I may be worrying too much and i will be proved wrong.
I understand, this schema will probably work perfectly
* Jamal Hadi Salim <[EMAIL PROTECTED]> 2005-07-25 09:50
> Thomas:
> bit 0 of tc_verd can be moved into cb;
Any reason why we don't see OK2MUNGE in the action itself but
have it triggered by setting MUNGED first?
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a m
* David S. Miller <[EMAIL PROTECTED]> 2005-07-25 11:46
> There is, of course, still the issue of refcounting. If we go
> the ifindex route for skb->input_dev, it might make sense to
> translate device specifications in action rules into ifindex.
> But things could break if you down then up a devic
* David S. Miller <[EMAIL PROTECTED]> 2005-07-25 12:09
> Ok, so can we agree on the following patch? ing_filter() should
> never see a NULL ->input_dev, and we could add a BUG() check
> there for that invariant if you want.
This looks good to me, a input_dev==NULL means the packets
origin is loca
* David S. Miller <[EMAIL PROTECTED]> 2005-07-25 13:14
> From: Thomas Graf <[EMAIL PROTECTED]>
> Date: Mon, 25 Jul 2005 22:05:21 +0200
>
> > The meta ematch which will the primary user of input_dev
> > provides both, filtering by ifindex and by name so I'm fi
* jamal <[EMAIL PROTECTED]> 2005-07-25 18:22
> > The code block which uses input_dev is dead, TC_NCLS is never
> > set as of now. Jamal might have use for it though? Generally,
> > tcf_action_exec() may be called on both ingress and egress.
>
> I mentioned the meta action already; the dummy device
* jamal <[EMAIL PROTECTED]> 2005-07-25 18:28
> So ifindex maybe the simplest way to go. If someone unloads a module
> for the netdevice then loads it back or pops out a hotplug net card
> and then pops it in - could we safely assume they are talking about
> another device? ;->
We can also use ifin
* jamal <[EMAIL PROTECTED]> 2005-07-25 18:40
> On Tue, 2005-26-07 at 00:32 +0200, Thomas Graf wrote:
> > I was referring to the tc_verd & TC_NCLS check in tcf_action_exec()
> > which uses input_dev in a printk. I assume dave was referring to the
> > same code
* David S. Miller <[EMAIL PROTECTED]> 2005-07-25 15:56
> From: Thomas Graf <[EMAIL PROTECTED]>
> Date: Tue, 26 Jul 2005 00:35:05 +0200
>
> > We can also use ifindex and hold a reference. We won't access input_dev
> > too often so a __dev_get_by_ifin
* David S. Miller <[EMAIL PROTECTED]> 2005-07-25 16:18
> From: Thomas Graf <[EMAIL PROTECTED]>
> Date: Tue, 26 Jul 2005 01:11:13 +0200
>
> > Not sure but as I see it:
> ...
> > The __dev_get_by_index() can be expensive with
> > more than a 1k interfaces
* Patrick McHardy <[EMAIL PROTECTED]> 2005-07-26 01:46
> Netlink users don't have to care about shared or cloned skbs. I don't
> think its a big issue to use alloc_skb and then the usual netlink
> macros. Thomas added a number of macros that simplfiy use a lot.
Once I've finished the generic netli
* Patrick McHardy <[EMAIL PROTECTED]> 2005-07-26 02:16
> Thomas Graf wrote:
> >* Patrick McHardy <[EMAIL PROTECTED]> 2005-07-26 01:46
> >
> >>You still have to take care of mixed 64/32 bit environments, u64 fields
> >>for example are differently alli
* Evgeniy Polyakov <[EMAIL PROTECTED]> 2005-07-26 08:45
> On Tue, Jul 26, 2005 at 01:46:04AM +0200, Patrick McHardy ([EMAIL PROTECTED])
> wrote:
> > Usually netlink is easily extendable by using nested TLVs. By hiding
> > this you basically remove this extensibility.
>
> Current netlink is not ex
* Anand SVR <[EMAIL PROTECTED]> 2005-07-27 23:17
> As a starting point, I would like to define classification rules such
> that web access to
> *.edu OR *.net OR *.org can be put under one bandwidth chunk. Public
> mail sites such as *.yahoo.com OR gmail.com OR *hotmail.com under a
> different ch
801 - 900 of 1030 matches
Mail list logo