On 2021-04-09 9:43 a.m., Petr Machata wrote:
Jamal Hadi Salim writes:
Does the spectrum not support multiple actions?
e.g with a policy like:
match blah action trap action drop skip_sw
Trap drops implicitly. We need a "trap, but don't drop". Expressed in
terms of exis
On 2021-04-09 7:03 a.m., Petr Machata wrote:
Jamal Hadi Salim writes:
I am concerned about adding new opcodes which only make sense if you
offload (or make sense only if you are running in s/w).
Those opcodes are intended to be generic abstractions so the dispatcher
can decide what to do
On 2021-04-08 5:25 p.m., Jakub Kicinski wrote:
On Thu, 8 Apr 2021 10:05:07 -0400 Jamal Hadi Salim wrote:
On 2021-04-08 9:38 a.m., Petr Machata wrote:
The TC action "trap" is used to instruct the HW datapath to drop the
matched packet and transfer it for processing in the SW pi
s to "trap" and the second is to "drop".
This is no different semantically than say "mirror and drop"
offload being enunciated by "skip_sw".
Does the spectrum not support multiple actions?
e.g with a policy like:
match blah action trap action drop skip_sw
cheers,
jamal
On 2021-04-08 3:50 a.m., Vlad Buslov wrote:
On Thu 08 Apr 2021 at 02:50, Cong Wang wrote:
Origins of setting ovr based on NLM_F_REPLACE are lost since this code
goes back to Linus' Linux-2.6.12-rc2 commit. Jamal, do you know if this
is the expected behavior or just something unint
t may be feasible to fix "replace" but not sure how many scripts over
the years are now dependent on that behavior.
cheers,
jamal
On 2021-03-26 9:09 a.m., Simon Horman wrote:
From: Baowen Zheng
Add selftest cases in action police for packets per second.
These tests depend on corresponding iproute2 command support.
Signed-off-by: Baowen Zheng
Signed-off-by: Simon Horman
Gracias.
Reviewed-by: Jamal Hadi Salim
rate and burst
parameters.
...
Sorry, I missed CCing a number of interested parties when posting
this patch-set. I've added them to this email.
Ref:
https://lore.kernel.org/netdev/20210312140831.23346-1-simon.hor...@netronome.com/
For 2/3:
Reviewed-by: Jamal Hadi Salim
If you submit o
the mailing
list peo...@netdevconf.info or see us at the twitters
as @netdev01
cheers,
jamal (on behalf of the Netdev Society)
On 2021-02-04 11:50 a.m., Phil Sutter wrote:
On Thu, Feb 04, 2021 at 10:28:26AM -0500, Jamal Hadi Salim wrote:
On 2021-02-04 9:50 a.m., Phil Sutter wrote:
On Thu, Feb 04, 2021 at 09:34:01AM -0500, Jamal Hadi Salim wrote:
On 2021-02-04 9:04 a.m., Phil Sutter wrote:
Jamal,
On Thu, Feb 04
On 2021-02-04 9:50 a.m., Phil Sutter wrote:
On Thu, Feb 04, 2021 at 09:34:01AM -0500, Jamal Hadi Salim wrote:
On 2021-02-04 9:04 a.m., Phil Sutter wrote:
Jamal,
On Thu, Feb 04, 2021 at 08:19:55AM -0500, Jamal Hadi Salim wrote:
I couldnt tell by inspection if what used to work before
sorry - meant to say, tdc is in kernel tree:
tools/testing/selftests/tc-testing
cheers,
jamal
On 2021-02-04 9:04 a.m., Phil Sutter wrote:
Jamal,
On Thu, Feb 04, 2021 at 08:19:55AM -0500, Jamal Hadi Salim wrote:
I couldnt tell by inspection if what used to work before continues to.
In particular the kernel version does consider the divisor when folding.
That's correct. And so do
\
action ok
Probably a simple meaning of "working" is:
the values before and after (your changes) are consistent.
If also you will do us a kindness and add maybe a testcase in tdc?
This way next person wanting to fix it can run the tests first before
posting a patch.
cheers,
jama
On 2021-02-01 7:33 a.m., Simon Horman wrote:
On Fri, Jan 29, 2021 at 09:30:00AM -0500, Jamal Hadi Salim wrote:
Ido's comment is important: Why not make packet rate vs byte rate
mutually exclusive? If someone uses packet rate then you make sure
they dont interleave with attributes for
On 2021-01-29 10:14 p.m., Alexei Starovoitov wrote:
On Fri, Jan 29, 2021 at 6:14 AM Jamal Hadi Salim wrote:
On 2021-01-29 9:06 a.m., Jamal Hadi Salim wrote:
Which leads to:
Why not extend the general feature so one can register for optional
callbacks not just for expire but also add/del
Hi Phil,
On 2021-01-22 8:59 a.m., Phil Sutter wrote:
Jamal,
On Fri, Jan 22, 2021 at 06:25:22AM -0500, Jamal Hadi Salim wrote:
[...]
My gut feel is user space is the right/easier spot to fix this
as long as it doesnt break the working setup of 8b.
One last attempt at clarifying the situation
Hi Phil,
On 2021-01-20 10:23 a.m., Phil Sutter wrote:
Hi Jamal,
On Wed, Jan 20, 2021 at 08:55:11AM -0500, Jamal Hadi Salim wrote:
On 2021-01-18 6:29 a.m., Phil Sutter wrote:
Hi!
Playing with u32 filter's hash table I noticed it is not possible to use
'sample' option with k
e are still going to extract only 8 bits for the bucket.
Can you provide an example of what wouldnt work?
cheers,
jamal
Turns out key hashing in kernel and iproute2 differ:
* net/sched/cls_u32.c (kernel) basically does:
hash = ntohl(key & mask);
hash >>= ffs(ntohl(mask)) - 1;
hash &
is patch has been tested by syzbot and tested with tdc.py by me.
LGTM.
Initially i was worried about performance impact but i found nothing
observable. We need to add a tdc test for batch (I can share how i did
batch testing at next meet).
Tested-by: Jamal Hadi Salim
Acked-by: Jamal Hadi Salim
cheers,
jamal
agically goes away :-) .
So: How does one know that the stack-upcall dropped a packet because
of GRO issues? Debugging with kprobe or traces doesnt count as an
answer.
cheers,
jamal
While we bid goodbye to 2020, here's some good hopeful news for 2021.
cheers,
jamal
Forwarded Message
Subject: [NetDev-People] Announcing Netdev 0x15
Date: Sun, 27 Dec 2020 15:54:31 -0700
From: Tom Herbert via people
Reply-To: Tom Herbert
To: peo...@netdevconf.info
scratch
pad is written by one program in the array and read by another later on.
skb->cb allows for that. Unless you mean i can already write to some
XDP data_meta area?
cheers,
jamal
space? Something equivalent to skb->cb which is lacking in
XDP.
cheers,
jamal
On 2020-12-16 6:47 a.m., Maxim Mikityanskiy wrote:
On 2020-12-15 18:37, Jamal Hadi Salim wrote:
[..]
Same question above:
Is there a limit to the number of classes that can be created?
Yes, the commit message of the mlx5 patch lists the limitations of our
NICs. Basically, it's 256
, it may make sense to provide a knob to query
(maybe ethtool) and if such limits can be adjusted it may be worth
looking at providing interfaces via devlink.
cheers,
jamal
cheers,
jamal
act_ct gets loaded, it configs that hook.
The frag xmit hook maybe reused by other modules.
Signed-off-by: wenxu
LGMT. Cong, Jamal still fine by you guys?
Yes, I do not look much into detail, but overall it definitely looks good.
This is targeting net-next, so it is fine to fix anything we miss
802.1Q and vlan_ethtype 0x800.
am i missing something? the rule should look different now?
Thanks,
Roi
Hi,
I debugged this and it break vlan rules.
The issue is from this part of the change
We have a test for this in tdc. Maybe we should make it a rule to
run tdc tests on changes like these?
cheers,
jamal
rs. Create aliases
for existing action flags constants with "TCA_ACT_" prefix.
Signed-off-by: Vlad Buslov
Are we expecting to add both aliases for all new flags?
I don't think it makes sense to have both aliases for any new flags.
Agreed.
cheers,
jamal
slov
LGTM, thanks for the effort
.
Acked-by: Jamal Hadi Salim
cheers,
jamal
ort to fix the over mtu for
defrag in act_ct.
Overall it looks much better to me now, so:
Acked-by: Cong Wang
LGTM as well.
Acked-by: Jamal Hadi Salim
cheers,
jamal
Hi,
On 2020-11-20 2:32 a.m., Xiaoliang Yang wrote:
Hi Jamal,
On 2020-11-19 0:11, Jamal Hadi Salim wrote:
The 11/17/2020 14:30, Xiaoliang Yang wrote:
EXTERNAL EMAIL: Do not click links or open attachments unless you
[..]
We already have mirroring + ability to add/pop tags.
Would the
ol 802.1Q flower \
skip_hw dst_mac 00:01:02:03:04:06 vlan_id 1 \
actopm vlan pop
cheers,
jamal
This nagged me:
What happens if all the frags dont make it out?
Should you at least return an error code(from tcf_fragment?)
and get the action err counters incremented?
cheers,
jamal
On 2020-11-15 8:05 a.m., wenxu wrote:
在 2020/11/15 2:05, Cong Wang 写道:
On Wed, Nov 11, 2020 at 9:44 PM
FWIW, I see the pragmatic need for this and so; cant think
of a better way to do this. Patch one could probably go in its
own merit.
Wenxu, please Cc maintainers in the future - makes it easier
to get feedback.
Acked-by: Jamal Hadi Salim
cheers,
jamal
On 2020-11-12 4:43 a.m., we...@ucloud.cn
On 2020-11-05 4:01 p.m., Andrii Nakryiko wrote:
On Thu, Nov 5, 2020 at 6:05 AM Jamal Hadi Salim wrote:
On 2020-11-04 10:19 p.m., David Ahern wrote:
[..]
[..]
2cents feedback from a dabbler in ebpf on user experience:
What David described above *has held me back*.
Over time it seems
that people can cutnpaste from. Every time you add a new feature
put some sample code in the examples.
cheers,
jamal
the action). This
is different from filter terse dump where index is excluded (filter can be
identified by its own handle).
Move tcf_action_dump_terse() function to the beginning of source file in
order to call it from tcf_dump_walker().
Signed-off-by: Vlad Buslov
Suggested-by: Jamal Hadi Salim
CA_FLAG_LARGE_DUMP_ON
in the uapi header will help. Of course that would be a separate
patch which will require conversion code in both the kernel and user
space.
FWIW, the patch is good for what i tested. So even if you do send an
update with a name change please add:
Acked-by: Jamal Hadi Salim
cheers,
jamal
; !from_act) {
> + if (nla_put(skb, TCA_ACT_COOKIE, cookie->len, cookie->data)) {
> + rcu_read_unlock();
> + goto nla_put_failure;
> + }
cheers,
jamal
On 2020-10-26 2:03 p.m., Vlad Buslov wrote:
On Mon 26 Oct 2020 at 20:01, Jamal Hadi Salim wrote:
On 2020-10-26 1:46 p.m., Vlad Buslov wrote:
yeah, something like TCA_ACT_FLAGS_TERSE.
new tcf_action_dump_terse() takes one more field which says to
include or not the cookies since that is
On 2020-10-26 1:46 p.m., Vlad Buslov wrote:
On Mon 26 Oct 2020 at 19:12, Jamal Hadi Salim wrote:
On 2020-10-26 7:28 a.m., Vlad Buslov wrote:
On Sat 24 Oct 2020 at 20:40, Jamal Hadi Salim wrote:
[..]
Yes, that makes sense. I guess introducing something like 'tc action -br
ls ..&
On 2020-10-26 7:28 a.m., Vlad Buslov wrote:
On Sat 24 Oct 2020 at 20:40, Jamal Hadi Salim wrote:
[..]
Yes, that makes sense. I guess introducing something like 'tc action -br
ls ..' mode implemented by means of existing terse flag + new 'also
output action index' fla
ion kind before returning from action_until->print_aopt()
callbacks. This is necessary to support terse dump mode in following patch
in the series.
Signed-off-by: Vlad Buslov
Suggested-by: Jamal Hadi Salim
Acked-by: Jamal Hadi Salim
cheers,
jamal
On 2020-10-23 8:48 a.m., Vlad Buslov wrote:
On Thu 22 Oct 2020 at 17:05, Jamal Hadi Salim wrote:
On 2020-10-21 4:19 a.m., Vlad Buslov wrote:
On Tue 20 Oct 2020 at 15:29, Jamal Hadi Salim wrote:
On 2020-10-19 11:18 a.m., Vlad Buslov wrote:
My worry is you have a very specific use case for
On 2020-10-21 4:19 a.m., Vlad Buslov wrote:
On Tue 20 Oct 2020 at 15:29, Jamal Hadi Salim wrote:
On 2020-10-19 11:18 a.m., Vlad Buslov wrote:
My worry is you have a very specific use case for your hardware or
maybe it is ovs - where counters are uniquely tied to filters and
there is no
On 2020-10-19 11:18 a.m., Vlad Buslov wrote:
On Mon 19 Oct 2020 at 16:48, Jamal Hadi Salim wrote:
On 2020-10-18 8:16 a.m., Vlad Buslov wrote:
[..]
That could be a good thing, no? you get to see the action name with the
error. Its really not a big deal if you decide to do a->terse_pr
On 2020-10-18 8:16 a.m., Vlad Buslov wrote:
On Sat 17 Oct 2020 at 14:20, Jamal Hadi Salim wrote:
On 2020-10-16 12:42 p.m., Vlad Buslov wrote:
Either one sounds appealing - the refactoring feels simpler
as opposed to a->terse_print().
With such refactoring we action type will be prin
arameter
when actions are shared by filters (since they will have the same
index).
Am i missing something?
cheers,
jamal
e at least show up?
cheers,
jamal
directly on the 0x14 playlist here:
https://bit.ly/2SHJWow
cheers,
jamal
PS: apologies if you receive this multiple times...
On 2020-10-09 8:15 a.m., Vlad Buslov wrote:
On Fri 09 Oct 2020 at 15:03, Jamal Hadi Salim wrote:
Which test(s)?
I see for example d45e mentioning terse but i dont see corresponding
code in the iproute2 tree i just pulled.
Yes. The tests d45e and 7c65 were added as a part of kernel series
On 2020-10-08 11:34 a.m., Vlad Buslov wrote:
On Thu 08 Oct 2020 at 15:58, Jamal Hadi Salim wrote:
Hi Jamal,
The existing terse dump tdc tests will have to be changed according with
new iproute2 tc syntax for brief(terse) output.
Which test(s)?
I see for example d45e mentioning terse but
f patching and testing. Having said that:
The differentiation via TCA_DUMP_FLAGS_TERSE in the request
is in my opinion sufficient to accept the patch.
Also, assuming you have tested with outstanding tc tests for the
first patch i think it looks reasoan
cheers,
jamal
-sessions.html
cheers,
jamal
PS: apologies if you receive this multiple times...
Ethtool reporting zero. ip -s reporting non-zero.
Which is right? See attached for more details.
cheers,
jamal
some bus info
b3:00.0 Ethernet controller: Intel Corporation Ethernet Controller XXV710 for
25GbE SFP28 (rev 02)
Subsystem: Super Micro Computer Inc Ethernet Controller
On 2020-08-17 3:47 p.m., Cong Wang wrote:
On Mon, Aug 17, 2020 at 4:19 AM Jamal Hadi Salim wrote:
[..]
There is no ambiguity of intent in the fw case, there is only one field.
In the case of having multiple fields it is ambigious if you
unconditionally look.
Example: policy says to match
On 2020-08-16 2:59 p.m., Cong Wang wrote:
On Thu, Aug 13, 2020 at 5:52 AM Jamal Hadi Salim wrote:
[..]
How do you know whether to use hash or mark or both
for that specific key?
Hmm, you can just unconditionally pass skb->hash and skb->mark,
no? Something lik
On 2020-08-11 7:25 p.m., Cong Wang wrote:
On Sun, Aug 9, 2020 at 4:41 PM Jamal Hadi Salim wrote:
[..]
Not sure if I get you correctly, but with a combined implementation
you can do above too, right? Something like:
(AND case)
$TC filter add dev $DEV1 parent : protocol ip prio 3 handle
On 2020-08-09 2:15 p.m., Cong Wang wrote:
On Fri, Aug 7, 2020 at 3:28 PM Jamal Hadi Salim wrote:
From: Jamal Hadi Salim
his classifier, in the same spirit as the tc skb mark classifier,
provides a generic (fast lookup) approach to filter on the hash value
and optional mask.
like skb->m
From: Jamal Hadi Salim
This classifier - in the same spirit as the tc skb mark classifier - provides
a generic (fast lookup) approach to filter on the hash value and optional
mask.
like skb->mark, skb->hash could be set by multiple entities in the datapath
including but not limited to ha
From: Jamal Hadi Salim
his classifier, in the same spirit as the tc skb mark classifier,
provides a generic (fast lookup) approach to filter on the hash value
and optional mask.
like skb->mark, skb->hash could be set by multiple entities in the
datapath including but not limited to ha
rcu_head rcu;
u32 mask;
struct fw_filter __rcu *ht[HTSIZE];
- struct rcu_head rcu;
};
Does not fix it..
cheers,
jamal
On 2020-08-07 11:16 a.m., Jamal Hadi Salim wrote:
Made this small change:
--
diff --git a/net/sched/cls_fw.c b
e hash table got too large.
Smells like hard coded expectation?
How to fix?
cheers,
jamal
On 2020-07-10 8:04 a.m., Jamal Hadi Salim wrote:
On 2020-07-09 8:19 a.m., Jiri Pirko wrote:
Thu, Jul 09, 2020 at 01:00:26PM CEST, j...@mojatatu.com wrote:
Main point is: classifying based on hash(and for that
matter any other metadata like mark) is needed as a general
utility for the system
netdevconf.info/0x14/registration.html
cheers,
jamal
desire to have flower as a one stop shop for all things classification
but this is at the expense of other classifiers. I too need this for
offloading as well as getting the RSS proper feature i described.ets
make progress.
You go ahead - i will submit a version to add it as a separate
hash classifier.
cheers,
jamal
existing things like XDP, tc ebpf and the action
posted by Ariel.
cheers,
jamal
job there.
For humans, agreed everything in one place is convinient.
Note: your arguement could be used for "ls" to include "grep"
functionality because in my scripts I do both most of the time.
cheers,
jamal
+ Nothing stops you from creating separate classifier to match on hash
as you wanted to :)
Hi Ariel,
On 2020-07-05 1:26 p.m., Ariel Levkovich wrote:
On 7/3/20 7:22 AM, Jamal Hadi Salim wrote:
[..]
Hi Jamal,
I agree that using skbedit makes some sense and can provide the same
functionality.
However I believe that from a concept point of view, using it is wrong.
In my honest
current modularity as opposed
to dumping everything into flower.
Ive always wanted to write the skbhash classifier but
time was scarce. At one point i had some experiment
where I would copy skb hash into mark in the driver
and use fw classifier for further processing.
It was ugly.
cheers,
jamal
On
code
can also find it?
Any objections to this? It's not actually clear to me how the discussion
you quoted above landed; but this will at least make things consistent
across all the different actions/etc.
I didnt follow the original discussion - will try to read in order
to form an opinion.
cheers,
jamal
+ counters.red_frames_count -
filter->stats.drops;
Thanks for the explanation.
What is filter->stats?
The rest of those counters seem related to the gate action.
How do you account for policing actions?
cheers,
jamal
On 2020-06-23 7:52 p.m., Po Liu wrote:
Hi Jamal,
My question: Is this any different from how stats are structured?
[..]
My question: Why cant you apply the same semantics for the counters?
Does your hardware have an indexed counter/stats table? If yes then you
Yes,
That is the
x 5
infact they use the same counter.
cheers,
jamal
stats.
cheers,
jamal
On 2020-06-23 2:34 a.m., Po Liu wrote:
From: Po Liu
Hardware may own many entries for police flow. So that make one(or
multi) flow to be policed by one hardware entry. This patch add the
police action index provide to the driver side make it mapping the
driver hardware entry
On 2020-05-27 6:13 p.m., Stephen Hemminger wrote:
On Tue, 26 May 2020 18:04:09 +0200
Andrea Claudi wrote:
Jamal reported a segfault in bpf_make_custom_path() when custom pinning is
used. This is caused by commit c0325b06382cb ("bpf: replace snprintf with
asprintf when dealing with
On 2020-05-25 4:53 a.m., Andrea Claudi wrote:
Reverting c0325b06382c will for sure fix the segfault identified by
Jamal and get rid of the problems highlighted by Daniel and others.
To fix the s[n]printf truncation warning we can simply check for its
return value. From the snprintf man page
, when I discussed my initial terse
dump idea with Jamal he asked me not to ossify such behavior to allow
for implementation of offloaded shared actions in future.
Trying to recollect our discussion (please forgive me if i am
rehashing). old skule hardware model typically is ACL style with one
On 2020-05-22 9:33 p.m., Daniel Borkmann wrote:
On 5/18/20 3:00 PM, Jamal Hadi Salim wrote:
ping?
Note: these are trivial bug fixes.
Looking at c0325b06382c ("bpf: replace snprintf with asprintf when
dealing with long buffers"),
I wonder whether it's best to just revert a
On 2020-05-18 11:38 a.m., David Ahern wrote:
On 5/18/20 7:10 AM, Jamal Hadi Salim wrote:
On 2020-05-17 9:28 a.m., Roman Mashak wrote:
Have print_tm() dump firstuse value along with install, lastuse
and expires.
Signed-off-by: Roman Mashak
---
tc/tc_util.c | 5 +
1 file changed, 5
lternatively just print 0 (to be backward compatible on old
kernels it will never be zero).
cheers,
jamal
ping?
Note: these are trivial bug fixes.
cheers,
jamal
On 2020-04-28 12:15 p.m., Jamal Hadi Salim wrote:
Stephen,
What happened to this?
cheers,
jamal
On 2020-04-23 1:58 p.m., Jamal Hadi Salim wrote:
From: Jamal Hadi Salim
Changes from V2:
1) Dont initialize tmp on stack (Stephen)
2
eck for the corner case in the kernel.
1)Zero is a legit value when dumping or
getting (example an action was just hit when you dumped).
2) zero also now means "has not been used".
My suggestion is to leave this alone in the kernel.
In user space/iproute2 check if lastused and created
are equal and declare "has not been used".
cheers,
jamal
Hi,
Apologies for taking longer than promised to provide an
update on the conference.
We finally have a decision - the conference is going fully
virtual. For details please see:
https://netdevconf.info/0x14/news.html?virtual-netdev-0x14
cheers,
jamal
and parent from caller down to
tcf_fill_node")
Fixes: edf6711c9840 ("net: sched: remove classid and q fields from tcf_proto")
Cc: Jamal Hadi Salim
Cc: Jiri Pirko
Signed-off-by: Cong Wang
Acked-by: Jamal Hadi Salim
cheers,
jamal
On 2020-05-03 8:02 a.m., Jamal Hadi Salim wrote:
On 2020-05-02 10:28 p.m., Cong Wang wrote:
On Sat, May 2, 2020 at 2:19 AM Jamal Hadi Salim wrote:
On 2020-05-02 4:48 a.m., Jamal Hadi Salim wrote:
[..]
Note:
tc filter show dev dummy0 root
should not show that filter. OTOH,
tc filter show
On 2020-05-02 10:28 p.m., Cong Wang wrote:
On Sat, May 2, 2020 at 2:19 AM Jamal Hadi Salim wrote:
On 2020-05-02 4:48 a.m., Jamal Hadi Salim wrote:
[..]
Note:
tc filter show dev dummy0 root
should not show that filter. OTOH,
tc filter show dev dummy0 parent :
should.
Hmm, but we use
On 2020-05-02 4:48 a.m., Jamal Hadi Salim wrote:
On 2020-04-30 11:53 p.m., Cong Wang wrote:
[..]
Steps to reproduce this:
ip li set dev dummy0 up
tc qd add dev dummy0 ingress
tc filter add dev dummy0 parent : protocol arp basic action pass
tc filter show dev dummy0 root
Before
show that filter. OTOH,
tc filter show dev dummy0 parent :
should.
root and : are distinct/unique installation hooks.
cheers,
jamal
Stephen,
What happened to this?
cheers,
jamal
On 2020-04-23 1:58 p.m., Jamal Hadi Salim wrote:
From: Jamal Hadi Salim
Changes from V2:
1) Dont initialize tmp on stack (Stephen)
2) Dont look at the return code of snprintf (Dominique)
3) Set errno to EINVAL instead of returning -EINVAL
On 2019-10-23 9:04 a.m., Vlad Buslov wrote:
On Wed 23 Oct 2019 at 15:49, Jamal Hadi Salim wrote:
Hi Vlad,
I understand your use case being different since it is for h/w
offload. If you have time can you test with batching many actions
and seeing the before/after improvement?
Will do
r space resolution.
I expect a native xdp implementation using a hash map to be
inline with ovs kmod if not better.
Hashes are good for datapath use cases but not when you consider
a holistic access where you have to worry about control aspect.
cheers,
jamal
ion is if this uglier than introducing a new TLV for every
action.
cheers,
jamal
Given TCA_ACT_FLAGS_FAST_INIT is common to all actions would it make
sense to use Could you have used a TLV in the namespace of TCA_ACT_MAX
(outer TLV)? You will have to pass a param to ->init().
cheers,
jamal
n_ops of vlan action, which calculates
the action size, and passes size to tcf_add_notify()/tcf_del_notify().
patch 2 updates the TDC test suite with relevant vlan test cases.
For the series:
Acked-by: Jamal Hadi Salim
cheers,
jamal
example test case file to
demonstrate how the scapy block works in the test cases.
Shouldve said V3 in the subject line - but fwiw,
ACKed-by: Jamal Hadi Salim
cheers,
jamal
ave an error talking to the kernel
%
patch 1 adds callback in tc_action_ops of mirred action, which calculates
the action size, and passes size to tcf_add_notify()/tcf_del_notify().
patch 2 updates the TDC test suite with relevant test cases.
Acked-by: Jamal Hadi Salim
cheers,
jamal
e a test scenario where
this could be demonstrated it would help.
cheers,
jamal
to use skb metadata
approach which is performance unfriendly (could be argued to more
correct).
Also: Please consider submitting a test case or two for tdc.
cheers,
jamal
1 - 100 of 2311 matches
Mail list logo
