ë(sªWby4ÊÉï¢ùÔSR°.`"Ûç:H·ÊN!ÝÎT¸÷>Ä;Ñ"Æm¥
êÍMæ2¾ YÓÈçqöq£såjö×]Ã
ö×îèÛÁÁIÙ·ÇÏ}¸ÜWïôé;hé0%\z§Q¼K7¬Þ·Â9®óYn[ððM.Y·þÕêÊ
a&É£uZ˳¢¥Âj]Öý躰¼Î¹8i0NêêÔO³ÔÄHÆK®UDßláô¿yCÐìÔp®BgÓØ0D_t/¢½ì°9þb¤¬XÉJ%Ù:jHÊø¿xѪµI¤ÌrÙX#dóµ´ç&Ú0Ôx»03øPå°ðãwvlìJÜÜ$Ù]бRJì~3än£LeÕÜrB߬þ9þ£
> -#define NXM_HEADER(CLASS, FIELD, HASMASK, LENGTH) \
> -(((CLASS) << 16) | ((FIELD) << 9) | ((HASMASK) << 8) | (LENGTH))
> +#define NXM_HEADER(VENDOR, CLASS, HASMASK, FIELD, LENGTH) \
why did you swap FIELD and HASMASK?
it's easier for me to remember if it's in th
hi Alin,
I also wanted to discuss the issues I had found while testing the vport-add
patches.
thanks,
-- Nithin
On Oct 6, 2014, at 2:08 PM, Nithin Raju wrote:
> hi Alin/Sorin,
> These were the items to discuss from out side:
>
> 1. vport add/del review comments
> 2. netlink integration status
We need to pass down the output buffer so that the kernel can return
transaction status - error or otherwise.
Signed-off-by: Nithin Raju
---
lib/netlink-socket.c |3 +--
1 files changed, 1 insertions(+), 2 deletions(-)
diff --git a/lib/netlink-socket.c b/lib/netlink-socket.c
index 1aa76ae..
On Oct 6, 2014, at 5:52 PM, Ankur Sharma
wrote:
> Hi Nithin,
>
> I have not tested these changes yet.
> I am figuring out the user-space changes needed for simulating a
> PACKET_CMD_EXECUTE.
>
> If you want to hold the review till testing is done then we can do that.
No problem. Thanks for
I applied the patches for the new vport add workflow that is out for
review, and found that some of the existing code in OvsGetExtInfoIoctl()
needs to be updated. In this patch, we add a CPP called
USE_NEW_VPORT_ADD_WORKFLOW and add the fixes under
USE_NEW_VPORT_ADD_WORKFLOW == 1. The current value
On Mon, Oct 6, 2014 at 3:51 PM, Ben Pfaff wrote:
> On Mon, Oct 06, 2014 at 10:40:50AM -0700, Gurucharan Shetty wrote:
>> Using the compiler intrinsic shows approximately around 25% speed
>> up with some classifier specific unit tests.
>>
>> Signed-off-by: Gurucharan Shetty
>
> Nice!
>
> I don't t
Hi Nithin,
I have not tested these changes yet.
I am figuring out the user-space changes needed for simulating a
PACKET_CMD_EXECUTE.
If you want to hold the review till testing is done then we can do that.
Thanks.
Regards,
Ankur
From: Nithin Raju
Sent:
On Oct 6, 2014, at 4:23 PM, Ankur Sharma
wrote:
> In this patch we have implemented the handler for
> OVS_PACKET_CMD_EXECUTE command.
>
> Signed-off-by: Ankur Sharma
Ankur,
Can you add a note on how you validated these changes?
-- Nithin
___
dev ma
In this patch we have made following changes:
OvsPacketExecute =>
Changed the data structure to have packet and
actions as pointer (instead of zero length array). It is done because
we will not do memcpy of packet now, pointer will just point
to corresponding offset in input buffer.
OvsExecuteDpI
In this patch we cover the basic registeration of OVS_PACKET_CMD_EXECUTE
command handler.
Signed-off-by: Ankur Sharma
---
datapath-windows/ovsext/Datapath.c | 18 +-
datapath-windows/ovsext/User.c | 18 ++
datapath-windows/ovsext/User.h | 5 +
3 files
In this patch we have implemented the handler for
OVS_PACKET_CMD_EXECUTE command.
Signed-off-by: Ankur Sharma
---
datapath-windows/ovsext/User.c | 74 +-
1 file changed, 73 insertions(+), 1 deletion(-)
diff --git a/datapath-windows/ovsext/User.c b/datapat
Ben,
This is much cleaner, thanks!
Acked-by: Jarno Rajahalme
On Oct 6, 2014, at 2:17 PM, Ben Pfaff wrote:
> On Wed, Oct 01, 2014 at 04:02:36PM -0700, Jarno Rajahalme wrote:
>> Depending on the CPU, count_1bits() may be more or less expensive, so
>> it is better to avoid unnecessary calls to
On Fri, Oct 03, 2014 at 01:22:07PM -0700, Jarno Rajahalme wrote:
> Replace bitwise loops with a single operation, inline all bitmap
> functions. Inlining allows the compiler to remove unnecessary code
> due to some parameters being compile-time constants.
>
> Before:
>
> $ tests/ovstest test-bit
Backported to brances 2.2 and 2.3, issue did not exist in branch 2.1.
Jarno
On Oct 6, 2014, at 3:39 PM, Jarno Rajahalme wrote:
> Thanks for the review!
>
> Pushed to master, working on the back port now.
>
> Jarno
>
> On Oct 6, 2014, at 1:27 PM, Ben Pfaff wrote:
>
>> On Wed, Oct 01, 201
On Mon, Oct 06, 2014 at 10:40:50AM -0700, Gurucharan Shetty wrote:
> Using the compiler intrinsic shows approximately around 25% speed
> up with some classifier specific unit tests.
>
> Signed-off-by: Gurucharan Shetty
Nice!
I don't think that the "& 0x" operations are
necessary
On Sat, Oct 04, 2014 at 11:28:43AM +0800, Wang Sheng-Hui wrote:
> Headroom and tailroom both are removed.
>
> Signed-off-by: Wang Sheng-Hui
Applied, thanks!
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Sat, Oct 04, 2014 at 09:22:23AM +0800, Wang Sheng-Hui wrote:
>
> The return type of ofpbuf_tail and ofpbuf_end is pointer, not byte.
>
> Signed-off-by: Wang Sheng-Hui
Applied, thanks!
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/
On Sat, Oct 04, 2014 at 10:11:25AM +0800, Wang Sheng-Hui wrote:
> Correct the old comment like ofpbuf_reserve to descript the expected
> behavior of ofpbuf_reserve_with_tailroom
>
> Signed-off-by: Wang Sheng-Hui
Applied, thanks!
___
dev mailing list
de
On Mon, Oct 06, 2014 at 01:07:23PM -0700, Nithin Raju wrote:
> validation:
> - Vport dump works now without printing any errors. I didn't go so far
> as to test if vport add works.
>
> Signed-off-by: Nithin Raju
> Acked-by: Ankur Sharma
> Acked-by: Alin Gabriel Serdean
> Tested-by: Alin Gabriel
Thanks for the review!
Pushed to master, working on the back port now.
Jarno
On Oct 6, 2014, at 1:27 PM, Ben Pfaff wrote:
> On Wed, Oct 01, 2014 at 04:02:32PM -0700, Jarno Rajahalme wrote:
>> Previously we masked labels not present in the incoming packet.
>>
>> Signed-off-by: Jarno Rajahalm
On Oct 6, 2014, at 1:39 PM, Ben Pfaff wrote:
> On Wed, Oct 01, 2014 at 04:02:35PM -0700, Jarno Rajahalme wrote:
>> We already update the count field as the last step of these functions,
>> so returning the current count is very cheap. Callers that care about
>> the count become a bit more effic
Pushed to master,
Jarno
On Oct 6, 2014, at 3:10 PM, Ben Pfaff wrote:
> On Mon, Oct 06, 2014 at 01:53:50PM -0700, Jarno Rajahalme wrote:
>>
>> On Oct 6, 2014, at 1:13 PM, Ben Pfaff wrote:
>>
>>> On Wed, Oct 01, 2014 at 04:02:31PM -0700, Jarno Rajahalme wrote:
Add function flow_wildcard
Thanks for the reviews!
Series pushed, comments below,
Jarno
On Oct 6, 2014, at 11:52 AM, Ben Pfaff wrote:
> On Wed, Sep 24, 2014 at 11:31:47AM -0700, Jarno Rajahalme wrote:
>> Batching the cmap find improves the memory behavior with large cmaps
>> and can make searches twice as fast:
>>
>>
On Mon, Oct 06, 2014 at 03:19:23PM +, Sorin Vinturis wrote:
> Acquiring a spin lock raises the IRQL to DISPATCH_LEVEL. But
> in many places of the code, while holding a spin lock, there
> are useless checks for the current IRQL against DISPATCH_LEVEL.
> Also, the dispatch flag is not correctly
On Fri, Oct 03, 2014 at 03:53:45PM -0700, Ankur Sharma wrote:
> In this patch we have implemented the flow get.
>
> Signed-off-by: Ankur Sharma
> Acked-by: Nithin Raju
All applied, thanks!
___
dev mailing list
dev@openvswitch.org
http://openvswitch.or
On Mon, Oct 06, 2014 at 12:45:15PM -0700, Gurucharan Shetty wrote:
> Upstream Linux has OVS kernel module that includes most (not all) of
> the features that comes with the kernel module from openvswitch.org.
> So, it is okay to relax the requirement for OVS userspace package
> to depend on openvsw
On Thu, Oct 02, 2014 at 09:20:32AM -0700, Jarno Rajahalme wrote:
>
> On Sep 26, 2014, at 11:33 AM, Ben Pfaff wrote:
>
> > On Wed, Sep 24, 2014 at 11:24:01AM -0700, Jarno Rajahalme wrote:
> >> Earlier I have seen the frags checking suspiciously high in perf
> >> reports, but did not understand wh
On Mon, Oct 06, 2014 at 01:53:50PM -0700, Jarno Rajahalme wrote:
>
> On Oct 6, 2014, at 1:13 PM, Ben Pfaff wrote:
>
> > On Wed, Oct 01, 2014 at 04:02:31PM -0700, Jarno Rajahalme wrote:
> >> Add function flow_wildcards_init_for_packet() that can be used to set
> >> sensible wildcards when megaflo
On Oct 6, 2014, at 11:00 AM, Ben Pfaff wrote:
> On Wed, Sep 24, 2014 at 11:31:46AM -0700, Jarno Rajahalme wrote:
>> This makes the following patch easier and cleans up the code.
>>
>> Explicit "inline" keywords seem necessary to prevent performance
>> regression on cmap_find() with GCC 4.7 -O2.
On Wed, Oct 01, 2014 at 04:02:34PM -0700, Jarno Rajahalme wrote:
> miniflow_extract_from_flow() is faster than miniflow_init(), since it
> does not scan the fields not significant for the packet type. More
> importantly, it produces the same layout as miniflow_extract, which
> does not try to compr
On Wed, Oct 01, 2014 at 04:02:33PM -0700, Jarno Rajahalme wrote:
> Working with netdev_flow_keys is simpler when the key has the fields
> present even when they are all-zeroes.
>
> Signed-off-by: Jarno Rajahalme
I don't understand the commit message. I think I need more context.
For example, wh
On Mon, Oct 6, 2014 at 11:41 AM, Alex Wang wrote:
> Thx for the fix!
>
> Since the "ERR" log is the last one, could we have this fold in?
>
> diff --git a/tests/ovs-vswitchd.at b/tests/ovs-vswitchd.at
> index 374be49..33772d8 100644
> --- a/tests/ovs-vswitchd.at
> +++ b/tests/ovs-vswitchd.at
> @@
On Wed, Oct 01, 2014 at 04:02:37PM -0700, Jarno Rajahalme wrote:
> flow inserts and removals are simplified:
>
> - No need for classifier internal mutex, as dpif-netdev already has a
> 'flow_mutex'
> - Number of memory allocations/frees can be halved
>
> Lookup code path is a bit more effcient
On Oct 6, 2014, at 10:47 AM, Ben Pfaff wrote:
> On Wed, Sep 24, 2014 at 11:31:44AM -0700, Jarno Rajahalme wrote:
>>These makes cmap_find 10% faster on GCC 4.7 (-O2 -g).
>>
>> Signed-off-by: Jarno Rajahalme
>
> I'm surprised it makes so much difference!
>
> What if we additionally change:
On Wed, Oct 01, 2014 at 04:02:36PM -0700, Jarno Rajahalme wrote:
> Depending on the CPU, count_1bits() may be more or less expensive, so
> it is better to avoid unnecessary calls to it. When the map has
> consecutive 1-bits, those can be cleared without counting.
>
> Signed-off-by: Jarno Rajahalm
hi Alin/Sorin,
These were the items to discuss from out side:
1. vport add/del review comments
2. netlink integration status: vport, packet commands
Thanks,
-- Nithin
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Oct 6, 2014, at 1:13 PM, Ben Pfaff wrote:
> On Wed, Oct 01, 2014 at 04:02:31PM -0700, Jarno Rajahalme wrote:
>> Add function flow_wildcards_init_for_packet() that can be used to set
>> sensible wildcards when megaflows are disabled. Before this, we set
>> all the mask bits to ones, which cau
Looks good, as far as it goes,
Jarno
On Sep 30, 2014, at 5:47 PM, Ben Pfaff wrote:
> This field allows a flow table to match on the output port currently in the
> action set.
>
> As-is, this patch has a number of deficiencies. I will fix them; I am
> posting for now because I feel like I ha
Upstream Linux has OVS kernel module that includes most (not all) of
the features that comes with the kernel module from openvswitch.org.
So, it is okay to relax the requirement for OVS userspace package
to depend on openvswitch-kmod package.
Signed-off-by: Gurucharan Shetty
---
rhel/openvswitch
On Wed, Oct 01, 2014 at 04:02:35PM -0700, Jarno Rajahalme wrote:
> We already update the count field as the last step of these functions,
> so returning the current count is very cheap. Callers that care about
> the count become a bit more efficient, as they avoid extra
> non-inlineable function c
On Wed, Oct 01, 2014 at 04:02:32PM -0700, Jarno Rajahalme wrote:
> Previously we masked labels not present in the incoming packet.
>
> Signed-off-by: Jarno Rajahalme
Nice improvement.
Here, I think I meant to write "second or later", which also makes
sense if we increase FLOW_MAX_MPLS_LABELS:
>
LGTM,
Acked-by: Jarno Rajahalme
On Sep 30, 2014, at 5:47 PM, Ben Pfaff wrote:
> Some of the Nicira extension actions include fixed-size 32-bit members that
> designate NXM fields. These actions can't accommodate 64-bit experimenter
> OXMs, so we need to figure out some kind of solution. This
On Wed, Oct 01, 2014 at 04:02:31PM -0700, Jarno Rajahalme wrote:
> Add function flow_wildcards_init_for_packet() that can be used to set
> sensible wildcards when megaflows are disabled. Before this, we set
> all the mask bits to ones, which caused priting tunnel, mpls, and/or
s/priting/printing/
Acked-by: Jarno Rajahalme
On Sep 30, 2014, at 5:47 PM, Ben Pfaff wrote:
> OpenFlow 1.2+ defines a means for vendors to define vendor-specific OXM
> fields, called "experimenter OXM". These OXM fields are expressed with a
> 64-bit OXM header instead of the 32-bit header used for standard OXM (a
In this patch, we make some fixes in the vport get code as well as
elevating some utility functions from static to non-static.
Signed-off-by: Nithin Raju
Acked-by: Ankur Sharma
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
---
datapath-windows/ovsext/Datapath.c | 17 +++
In this patch, we add support for family ID lookup of
OVS_WIN_NETDEV_FAMILY.
Signed-off-by: Nithin Raju
Acked-by: Ankur Sharma
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
---
lib/netlink-socket.c |5 +
1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/l
validation:
- Vport dump works now without printing any errors. I didn't go so far
as to test if vport add works.
Signed-off-by: Nithin Raju
Acked-by: Ankur Sharma
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
---
datapath-windows/ovsext/Vport.c | 242 +++
In this patch, we add a lib/netdev-windows.c which mostly contains stub
code and in subsequent patches, would use the netlink interface to query
netdev information for a vport.
The code implements netdev functionality for "internal" and "system"
types of vports.
Signed-off-by: Nithin Raju
Acked-
In this patch, we define netlink family, attributes and commands
for query the 'network device' information of VPORTs, such as
MTU, Link status, etc.
I considered adding the netdev command to the OVS_WIN_CONTROL_FAMILY
itself, but the netdev attributes are not compatible with the existing
attribut
In this patch, we add stub handlers for the netdev commands.
Signed-off-by: Nithin Raju
Acked-by: Ankur Sharma
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
---
datapath-windows/ovsext/Datapath.c | 27 +--
datapath-windows/ovsext/Vport.c| 14 ++
Primary goals of netdev-windows.c are:
1) To query the 'network device' information of a vport such as MTU, etc.
2) Monitor changes to the 'network device' information such as link
status.
In this change, we implement only #1. #2 can also be implemented, but it
does not seem to be required for the
One small matter of opinion below,
Jarno
Acked-by: Jarno Rajahalme
On Sep 30, 2014, at 5:47 PM, Ben Pfaff wrote:
> This will make it easier to support 64-bit OXM experimenter fields.
>
> Signed-off-by: Ben Pfaff
> Acked-by: YAMAMOTO Takashi
> ---
> v1->v2: Use ->s6_addr for IPv6 raw byte
On Wed, Sep 24, 2014 at 11:31:47AM -0700, Jarno Rajahalme wrote:
> Batching the cmap find improves the memory behavior with large cmaps
> and can make searches twice as fast:
>
> $ tests/ovstest test-cmap benchmark 200 8 0.1 16
> Benchmarking with n=200, 8 threads, 0.10% mutations, batch s
Acked-by: Alex Wang
On Mon, Oct 6, 2014 at 10:22 AM, Gurucharan Shetty
wrote:
> The logging options were shown as part of the "Public Key Infrastructure
> Options". Correct it.
>
> Signed-off-by: Gurucharan Shetty
> ---
> vswitchd/ovs-vswitchd.8.in |1 +
> 1 file changed, 1 insertion(+)
>
Thx for the fix!
Acked-by: Alex Wang
On Mon, Oct 6, 2014 at 10:22 AM, Gurucharan Shetty
wrote:
> The option is documented for ovsdb-server but not for ovs-vswitchd.
> This commit adds the documentation for ovs-vswitchd too.
>
> Signed-off-by: Gurucharan Shetty
> ---
> manpages.mk
Thx for the fix!
Since the "ERR" log is the last one, could we have this fold in?
diff --git a/tests/ovs-vswitchd.at b/tests/ovs-vswitchd.at
index 374be49..33772d8 100644
--- a/tests/ovs-vswitchd.at
+++ b/tests/ovs-vswitchd.at
@@ -84,7 +84,7 @@ ovs-appctl -t `pwd`/unixctl exit
# the existing ovs
Using the compiler intrinsic shows approximately around 25% speed
up with some classifier specific unit tests.
Signed-off-by: Gurucharan Shetty
---
lib/util.c |2 +-
lib/util.h | 36
2 files changed, 37 insertions(+), 1 deletion(-)
diff --git a/lib/uti
On Sep 30, 2014, at 5:47 PM, Ben Pfaff wrote:
> OpenFlow 1.5 (draft) extends the OFPAT_SET_FIELD action originally
> introduced in OpenFlow 1.2 so that it can set not just entire fields but
> any subset of bits within a field as well. This commit adds support for
> that feature when OpenFlow 1.
The logging options were shown as part of the "Public Key Infrastructure
Options". Correct it.
Signed-off-by: Gurucharan Shetty
---
vswitchd/ovs-vswitchd.8.in |1 +
1 file changed, 1 insertion(+)
diff --git a/vswitchd/ovs-vswitchd.8.in b/vswitchd/ovs-vswitchd.8.in
index 55c5fd4..35915cd 100
The option is documented for ovsdb-server but not for ovs-vswitchd.
This commit adds the documentation for ovs-vswitchd too.
Signed-off-by: Gurucharan Shetty
---
manpages.mk|2 ++
vswitchd/ovs-vswitchd.8.in |2 ++
2 files changed, 4 insertions(+)
diff --git a/manpages.mk
It is a little tricky to implement "$!" with unit tests on Windows.
This commit changes the test so that it works both on Windows
and Linux.
Signed-off-by: Gurucharan Shetty
---
tests/ovs-vswitchd.at | 14 +++---
1 file changed, 3 insertions(+), 11 deletions(-)
diff --git a/tests/ovs-
On Wed, Sep 24, 2014 at 11:31:46AM -0700, Jarno Rajahalme wrote:
> This makes the following patch easier and cleans up the code.
>
> Explicit "inline" keywords seem necessary to prevent performance
> regression on cmap_find() with GCC 4.7 -O2.
>
> Signed-off-by: Jarno Rajahalme
Do you think it'
On Wed, Sep 24, 2014 at 11:31:45AM -0700, Jarno Rajahalme wrote:
> We use the 'counter' as a "lock" providing acquire-release
> semantics. Therefore we can use normal, non-atomic access to the
> memory accesses between the atomic accesses to 'counter'. The
> cmap_node.next needs to be RCU, so tha
On Sep 30, 2014, at 5:47 PM, Ben Pfaff wrote:
> This is a first step toward improving the abstraction of OXM and NXM in the
> tree. As an immediate improvement, this commit removes all of the
> definitions of the OXM and NXM constants from the top-level header files,
> because they are no longe
On Wed, Sep 24, 2014 at 11:31:44AM -0700, Jarno Rajahalme wrote:
> These makes cmap_find 10% faster on GCC 4.7 (-O2 -g).
>
> Signed-off-by: Jarno Rajahalme
I'm surprised it makes so much difference!
What if we additionally change:
if (OVS_UNLIKELY(counter_changed(b1, c1))) {
got
>
> >
> > Not very familiar with the documentation format,
> >
> > Should we put bond_fake_iface inside 'status' column? Or we want to
> allow
> > user modify it?
>
> Do you mean bond_active_slave? Yes, put it in status colume makes more
> sense. I will move it there.
> >
>
Yes, I mean that. Thx
On Wed, Sep 24, 2014 at 11:31:43AM -0700, Jarno Rajahalme wrote:
> The test cases have been carefully crafted so that we do the same
> amount of "overhead" operations in each case. Earlier, with no
> mutations, the number of random number generations was different for
> hmap and cmap test cases.
Alex, thanks for the review. I will push it in a minute, and back port
to branch 2.3.
On Mon, Oct 6, 2014 at 9:36 AM, Alex Wang wrote:
> Thx for fixing this,
>
>
>>
>> +#include "lib/vswitch-idl.h"
>>
>
> This could be removed,
Done.
>
>
>>
>> VLOG_DEFINE_THIS_MODULE(bond);
>>
>> @@ -131,6 +132,
On Mon, Oct 6, 2014 at 2:47 AM, Tahhan, Maryam wrote:
> Hi All
> any review/update on this patch?
>
> Thanks
> Maryam
>
I am reviewing it, I should be able to finish it this week.
Thanks,
Pravin.
___
dev mailing list
dev@openvswitch.org
http://openvswi
Thx for fixing this,
> +#include "lib/vswitch-idl.h"
>
>
This could be removed,
> VLOG_DEFINE_THIS_MODULE(bond);
>
> @@ -131,6 +132,15 @@ struct bond {
> uint32_t recirc_id; /* Non zero if recirculation can be
> used.*/
> struct hmap pr_rule_ops; /* Helps to maintain p
This one fails on Windows as it is a little tricky to get it to work
with "$!".
The following works for Windows (and Linux too, atleast for one run).
If the changes that I am proposing end up not testing the original
issue the way you intended, then I am fine with adding a
AT_SKIP_IF([test "$IS_WI
Acquiring a spin lock raises the IRQL to DISPATCH_LEVEL. But
in many places of the code, while holding a spin lock, there
are useless checks for the current IRQL against DISPATCH_LEVEL.
Also, the dispatch flag is not correctly set when calling
NdisAcquireRWLockXXX() functions, which causes an extra
Can you provide an appropriate summary line for the subject? The
current one is not appropriate.
On Mon, Oct 06, 2014 at 03:05:29PM +, Sorin Vinturis wrote:
> Acquiring a spin lock raises the IRQL to DISPATCH_LEVEL. But
> in many places of the code, while holding a spin lock, there
> are usel
Acquiring a spin lock raises the IRQL to DISPATCH_LEVEL. But
in many places of the code, while holding a spin lock, there
are useless checks for the current IRQL against DISPATCH_LEVEL.
Also, the dispatch flag is not correctly set when calling
NdisAcquireRWLockXXX() functions, which causes an extra
Acked-by: Eitan Eliahu
Thanks!
-Original Message-
From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Sorin Vinturis
Sent: Monday, October 06, 2014 6:19 AM
To: dev@openvswitch.org
Subject: [ovs-dev] [PATCH v3] datapath-windows: Incorrect assumption of the IRQL
Acquiring a spin
Acquiring a spin lock raises the IRQL to DISPATCH_LEVEL. But
in many places of the code, while holding a spin lock, there
are useless checks for the current IRQL against DISPATCH_LEVEL.
Also, the dispatch flag is not correctly set when calling
NdisAcquireRWLockXXX() functions, which causes an extra
Acquiring a spin lock raises the IRQL to DISPATCH_LEVEL. But
in many places of the code, while holding a spin lock, there
are useless checks for the current IRQL against DISPATCH_LEVEL.
Also, the dispatch flag is not correctly set when calling
NdisAcquireRWLockXXX() functions, which causes an extra
Acked-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Ankur Sharma
Trimis: Saturday, October 4, 2014 1:54 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH v3 1/5] datapath-windows: Netlink additional APIs.
In this patch we have
Acked-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Ankur Sharma
Trimis: Saturday, October 4, 2014 1:54 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH v3 2/5] datapath-windows: Added the API for getting
unused space in nlbuf
Acked-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Ankur Sharma
Trimis: Saturday, October 4, 2014 1:54 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH v3 3/5] datapath-windows: Flow Dump handler
In this patch we have added b
Acked-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Ankur Sharma
Trimis: Saturday, October 4, 2014 1:54 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH v3 4/5] datapath-windows: Flow dump implementation.
In this patch we have
Acked-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Ankur Sharma
Trimis: Saturday, October 4, 2014 1:54 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH v3 5/5] datapath-windows: Flow get implementation.
In this patch we have
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Nithin Raju
Trimis: Sunday, October 5, 2014 1:10 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH 6/7 v2] datapath-windows: Some fixes in vpor
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Nithin Raju
Trimis: Sunday, October 5, 2014 1:10 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH 5/7 v2] datapath-windows: Add stub handlers
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Nithin Raju
Trimis: Sunday, October 5, 2014 1:10 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH 4/7 v2] lib/netdev-windows.c: add code to qu
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Nithin Raju
Trimis: Sunday, October 5, 2014 1:10 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH 7/7 v2] datapath-windows: implementation of
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Nithin Raju
Trimis: Sunday, October 5, 2014 1:10 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH 3/7 v2] add lib/netdev-windows.c
In this pa
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Nithin Raju
Trimis: Sunday, October 5, 2014 1:10 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH 1/7 v2] OvsDpInterfaceExt.h: add support for
Acked-by: Alin Gabriel Serdean
Tested-by: Alin Gabriel Serdean
-Mesaj original-
De la: dev [mailto:dev-boun...@openvswitch.org] În numele Nithin Raju
Trimis: Sunday, October 5, 2014 1:10 AM
Către: dev@openvswitch.org
Subiect: [ovs-dev] [PATCH 2/7 v2] lib/netlink-socket.c: add support f
I dropped this patch for the moment, it'll be easier to take a fresh look
once the bulk of the changes are in.
On 30 September 2014 16:21, Joe Stringer wrote:
> On 30 September 2014 10:27, Ben Pfaff wrote:
>
>> On Fri, Sep 26, 2014 at 09:28:18PM +1200, Joe Stringer wrote:
>> > Previously we sto
On 7 October 2014 00:23, Joe Stringer wrote:
> This series is also made available here to assist review:
> https://github.com/joestringer/openvswitch/tree/submit/ufid_v6
This should be
https://github.com/joestringer/openvswitch/tree/submit/ufid_v7.
__
One of the limiting factors on the number of flows that can be supported
in the datapath is the overhead of assembling flow dump messages in the
datapath. This patch modifies the flow_dump interface to allow
revalidators to skip dumping the key, mask and actions from the
datapath, by making use of
If a datapath is created with the flag OVS_DP_F_INDEX_BY_UID, then an
additional table_instance is added to the flow_table, which is indexed
by unique identifiers ("UID"). Userspace implementations can specify a
UID of up to 128 bits along with a flow operation as shorthand for the
key. This allows
This patch shifts the responsibility for determining the hash for a flow
from the revalidation logic down to the dpif layer. This assists in
handling backward-compatibility for revalidation with the upcoming
unique identifier "UID" patches, and allows ovs-dpctl to automatically
determine flow hashe
This patch modifies the dpif interface to allow flows to be manipulated
using a 128-bit identifier. This allows revalidator threads to perform
datapath operations faster, as they do not need to serialise the entire
flow key for operations like flow_get and flow_delete. In conjunction
with a future
Add the 128-bit murmurhash by Austin Appleby, r150 from:
http://code.google.com/p/smhasher/source/browse/trunk/MurmurHash3.cpp
Signed-off-by: Joe Stringer
Acked-by: Ben Pfaff
---
v7: Replace getblock*() with get_unaligned_u*().
Add ovs_u128_equal() function rather than using memcmp.
Only
Currently, when a revalidator thread first dumps a flow, it creates a
'udpif_key' object and caches a copy of a kernel flow key. This allows
us to perform lookups in the classifier to attribute stats and validate
the correctness of the datapath flow.
This patch sets up this cache from the handler
This allows us to ignore most fields of a flow_dump, requiring only the
flow key for looking up the ukey. Fetching flows can also be avoided in
the corner case where a flow is missed from a dump but revalidation is
required.
A future patch will modify the datapath interface to allow datapaths to
s
Future patches will make use of the 'struct dump_op' in a broader sense,
so this patch renames it to make things a bit clearer.
Signed-off-by: Joe Stringer
Acked-by: Ben Pfaff
---
v7: Rebase.
v4-v6: No change.
v3: Rebase.
v2: No change.
RFC: First post.
---
ofproto/ofproto-dpif-upcall.c | 97
1 - 100 of 109 matches
Mail list logo
