Could we add a grep for this to the make?
Ethan
On Fri, Dec 6, 2013 at 3:57 PM, Joe Stringer wrote:
> Acked-by: Joe Stringer
>
> On 20 November 2013 16:50, Ben Pfaff wrote:
>> Putting "static inline" on a function definition in a .c file does not help
>> the compiler and does suppress warnings
I'm not planning to review this patch, but could you please reformat
the commit message in a more standard way? I.E. something like:
poll-loop: \n
\n
The current formatting doesn't play nicely with git send email.
Ethan
On Fri, Dec 6, 2013 at 4:41 PM, Linda Sun wrote:
> Can you elaborate whi
Can you elaborate which cases have duplicate fds?
I've already used closesocket instead of close handle for sockets. I'll send a
separate patch with the stream related changes. E.g. Replace read/write on
socket with send/recv.
Linda
- Original Message -
From: Eitan Eliahu
To: Ben Pfaf
Ben and Linda,
It seems that the first two comments could be addressed (by removing NULL
events and duplicates handles.
In regards to the third comments, if these handles are Sockets (which I believe
they are) these handles should never be closed by calling CloseHandle(). For
these socket hand
On Fri, Dec 06, 2013 at 02:47:29PM -0800, Linda Sun wrote:
> Signed-off-by: Linda Sun
Hi Linda. Thanks for sending this along so that we can start figuring
out how to work Windows into the poll loop.
Looking at the documentation for WaitForMultipleObjects, I see a few
possible issues.
One is t
Acked-by: Joe Stringer
On 20 November 2013 16:50, Ben Pfaff wrote:
> Putting "static inline" on a function definition in a .c file does not help
> the compiler and does suppress warnings for unused functions that one would
> want, should the function ever become unused.
>
> Signed-off-by: Ben Pf
> On Dec 6, 2013, at 3:42 PM, Ben Pfaff wrote:
>
>> On Thu, Dec 05, 2013 at 04:27:24PM -0800, Jarno Rajahalme wrote:
>> Ofproto always provides a valid pointer to a mask, but the mask length
>> is zero when megaflows are disabled. Use match_wc_init() to get an
>> appropriate exact match mask wh
On Thu, Dec 05, 2013 at 04:27:24PM -0800, Jarno Rajahalme wrote:
> Ofproto always provides a valid pointer to a mask, but the mask length
> is zero when megaflows are disabled. Use match_wc_init() to get an
> appropriate exact match mask when there is no mask.
>
> Signed-off-by: Jarno Rajahalme
On Fri, Dec 06, 2013 at 01:26:16PM -0800, Jarno Rajahalme wrote:
>
> On Dec 6, 2013, at 11:02 AM, Ben Pfaff wrote:
>
> > On Thu, Dec 05, 2013 at 04:27:26PM -0800, Jarno Rajahalme wrote:
> >> Add a prefix tree (trie) structure for tracking the used address
> >> space, enabling skipping classifier
Looks good to me.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Pushed with changes suggested. Thanks a lot for the review.
On Fri, Dec 6, 2013 at 2:46 PM, Joe Stringer wrote:
> On 6 December 2013 14:27, Andy Zhou wrote:
> > diff --git a/vswitchd/bridge.c b/vswitchd/bridge.c
> > index 0f4cf2c..d64de65 100644
> > --- a/vswitchd/bridge.c
> > +++ b/vswitchd/b
On Fri, Dec 06, 2013 at 12:54:40PM -0800, Jarno Rajahalme wrote:
> I had forgot the line in there, was testing for alignment just for
> my own understanding. flow_compose() would be little simpler if we
> had the 32-bit aligned header definitions available. I felt hesitant
> about using the Linux d
Signed-off-by: Linda Sun
---
lib/dpif-linux.c |2 +-
lib/fatal-signal.c |2 +-
lib/latch.c | 37 -
lib/latch.h |3 +++
lib/netdev-linux.c |2 +-
lib/netlink-socket.c |2 +-
lib/poll-loop.c | 18
On 6 December 2013 14:27, Andy Zhou wrote:
> diff --git a/vswitchd/bridge.c b/vswitchd/bridge.c
> index 0f4cf2c..d64de65 100644
> --- a/vswitchd/bridge.c
> +++ b/vswitchd/bridge.c
> @@ -21,7 +21,7 @@
> #include "async-append.h"
> #include "bfd.h"
> #include "bitmap.h"
> -#include "bond.h"
> +#i
Ah, I forgot. Will send it out again. Thanks.
On Fri, Dec 6, 2013 at 2:09 PM, Joe Stringer wrote:
> On 6 December 2013 14:02, Andy Zhou wrote:
> > This are no functional changes.
> > ...
> > 5 files changed, 3 insertions(+), 1662 deletions(-)
>
> Uh, did you add the new files to the patch?
On 6 December 2013 14:02, Andy Zhou wrote:
> This are no functional changes.
> ...
> 5 files changed, 3 insertions(+), 1662 deletions(-)
Uh, did you add the new files to the patch?
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman
Relocating bond.[ch] to allow bond.c to make ofproto calls.
This is needed for upcoming patches that enable megaflow support
for bond ports.
This are no functional changes.
Signed-off-by: Andy Zhou
---
lib/automake.mk |2 -
lib/bond.c | 1560
On 5 December 2013 15:42, Joe Stringer wrote:
> + * Tries not to count hyper-threads, but may be inaccurate - particularly on
> + * platforms that do not provide /proc/cpuinfo, but also if /proc/cpuinfo is
> + * formatted different to the layout that parse_cpuinfo() expects. */
Just highlighting
On Fri, Dec 6, 2013 at 1:45 PM, Pravin Shelar wrote:
> On Fri, Dec 6, 2013 at 1:19 PM, Jesse Gross wrote:
>> Per-CPU flow stats are now allocated separately (if necessary) so
>> there's no longer any need to allocate space for them directly in
>> the base flow.
>>
>> Signed-off-by: Jesse Gross
>
On Fri, Dec 6, 2013 at 1:19 PM, Jesse Gross wrote:
> Per-CPU flow stats are now allocated separately (if necessary) so
> there's no longer any need to allocate space for them directly in
> the base flow.
>
> Signed-off-by: Jesse Gross
Looks good.
Acked-by: Pravin B Shelar
> ---
> datapath/flo
On Dec 6, 2013, at 1:26 PM, Jarno Rajahalme wrote:
>
> On Dec 6, 2013, at 11:02 AM, Ben Pfaff wrote:
>>
>> Clang reports:
>>../lib/classifier.c:1570:40: error: bad constant expression
>> referring to this line in trie_remove():
>>struct trie_node *nodes[trie->field->n_bytes];
>>
>
Pushed as-is, leaving Ben’s preference as an opportunity for future improvement.
Jarno
On Dec 6, 2013, at 9:24 AM, Ben Pfaff wrote:
> On Thu, Dec 05, 2013 at 04:36:26PM -0800, Jarno Rajahalme wrote:
>> Inline, use another well-known algorithm for 64-bit builds, and use
>> builtins when they a
On Dec 6, 2013, at 11:02 AM, Ben Pfaff wrote:
> On Thu, Dec 05, 2013 at 04:27:26PM -0800, Jarno Rajahalme wrote:
>> Add a prefix tree (trie) structure for tracking the used address
>> space, enabling skipping classifier tables containing longer masks
>> than necessary for an address field value
Per-CPU flow stats are now allocated separately (if necessary) so
there's no longer any need to allocate space for them directly in
the base flow.
Signed-off-by: Jesse Gross
---
datapath/flow_table.c | 8 ++--
1 file changed, 2 insertions(+), 6 deletions(-)
diff --git a/datapath/flow_table.
On Dec 6, 2013, at 12:53 PM, Ben Pfaff wrote:
> On Fri, Dec 06, 2013 at 11:48:58AM -0800, Jarno Rajahalme wrote:
>>
>> On Dec 6, 2013, at 9:40 AM, Ben Pfaff wrote:
>>
>>> On Thu, Dec 05, 2013 at 04:27:24PM -0800, Jarno Rajahalme wrote:
Ofproto always provides a valid pointer to a mask, b
On Fri, Dec 06, 2013 at 12:51:25PM -0800, Ben Pfaff wrote:
> On Fri, Dec 06, 2013 at 07:35:45PM +, Alin Serdean wrote:
> > Signed-off-by: Alin Serdean
> > ---
> > lib/string.h |4
> > 1 files changed, 4 insertions(+), 0 deletions(-)
> >
> > diff --git a/lib/string.h b/lib/string.h
>
Ben,
I pushed this after I removed the offending assert and changed to use normal
hton macros.
I had forgot the line in there, was testing for alignment just for my own
understanding. flow_compose() would be little simpler if we had the 32-bit
aligned header definitions available. I felt hesit
On Fri, Dec 06, 2013 at 11:48:58AM -0800, Jarno Rajahalme wrote:
>
> On Dec 6, 2013, at 9:40 AM, Ben Pfaff wrote:
>
> > On Thu, Dec 05, 2013 at 04:27:24PM -0800, Jarno Rajahalme wrote:
> >> Ofproto always provides a valid pointer to a mask, but the mask length
> >> is zero when megaflows are dis
On Fri, Dec 06, 2013 at 07:35:45PM +, Alin Serdean wrote:
> Signed-off-by: Alin Serdean
> ---
> lib/string.h |4
> 1 files changed, 4 insertions(+), 0 deletions(-)
>
> diff --git a/lib/string.h b/lib/string.h
> index 2b7b454..6981742 100644
> --- a/lib/string.h
> +++ b/lib/string.h
In the case where we have not yet sent a control packet for a bfd
connection, and we receive a control packet from the remote host,
bfd->next_tx is updated to an unusual value. This causes the logging to
incorrectly report that there has been long delays (in the order of
weeks) since the last bfd t
Sounds good, I'll have joe or alex look at it.
Ethan
On Thu, Dec 5, 2013 at 4:42 PM, Jarno Rajahalme wrote:
>
> On Dec 5, 2013, at 1:02 PM, Ethan Jackson wrote:
>
>> I'm working on a patch to disable wildcard calculation when we're
>> doing things like pushing stats and don't actually need to k
On Fri, Dec 6, 2013 at 12:03 PM, Alin Serdean
wrote:
> I think we should use _WIN32 (it is predefined for both Win32 and Win64) :).
>
> http://msdn.microsoft.com/en-us/library/b0084kay.aspx
I see what you mean. You are correct. I was referring to the WIN32
defined through config.h.
So your change
I think we should use _WIN32 (it is predefined for both Win32 and Win64) :).
http://msdn.microsoft.com/en-us/library/b0084kay.aspx
Kind Regards,
Alin.
From: Gurucharan Shetty [shet...@nicira.com]
Sent: Friday, December 06, 2013 9:47 PM
To: Alin Serdean
Cc:
On Fri, Dec 6, 2013 at 11:43 AM, Pravin B Shelar wrote:
> ipv6_addr_hash() is not available on kernel 3.5, Use compat version.
>
> Signed-off-by: Pravin B Shelar
Acked-by: Jesse Gross
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mai
On Dec 6, 2013, at 9:40 AM, Ben Pfaff wrote:
> On Thu, Dec 05, 2013 at 04:27:24PM -0800, Jarno Rajahalme wrote:
>> Ofproto always provides a valid pointer to a mask, but the mask length
>> is zero when megaflows are disabled. Use match_wc_init() to get an
>> appropriate exact match mask when th
On Fri, Dec 6, 2013 at 11:35 AM, Alin Serdean
wrote:
> Signed-off-by: Alin Serdean
> ---
> lib/string.h |4
> 1 files changed, 4 insertions(+), 0 deletions(-)
>
> diff --git a/lib/string.h b/lib/string.h
> index 2b7b454..6981742 100644
> --- a/lib/string.h
> +++ b/lib/string.h
> @@ -17,
ipv6_addr_hash() is not available on kernel 3.5, Use compat version.
Signed-off-by: Pravin B Shelar
---
datapath/linux/compat/include/net/ipv6.h |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/datapath/linux/compat/include/net/ipv6.h
b/datapath/linux/compat/include/net
Signed-off-by: Alin Serdean
---
lib/string.h |4
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/lib/string.h b/lib/string.h
index 2b7b454..6981742 100644
--- a/lib/string.h
+++ b/lib/string.h
@@ -17,7 +17,11 @@
#ifndef STRING_WRAPPER_H
#define STRING_WRAPPER_H 1
+#ifde
On Fri, Dec 6, 2013 at 9:42 AM, Alin Serdean
wrote:
> Guru can you do a quick review over the following:
Can you please send as a inline mail with a proper commit message (you
can look at SubmittingPatches or dev archives for examples) and a
Signed-off-by.
> diff --git a/lib/string.h b/lib/string
I pushed patch to master.
Thanks.
On Fri, Dec 6, 2013 at 11:05 AM, Jesse Gross wrote:
> On Fri, Dec 6, 2013 at 10:43 AM, Pravin B Shelar wrote:
>> Kernel 3.3 to 3.8 has defined `struct flow_keys` but does not
>> contains flow_keys.thoff field. Therefore we need to use
>> compat definition for
On Fri, Dec 06, 2013 at 08:24:03AM -0800, Jarno Rajahalme wrote:
>
> On Dec 6, 2013, at 1:59 AM, Helmut Schaa wrote:
>
> > On Wed, Nov 27, 2013 at 10:35 AM, Helmut Schaa
> > wrote:
> >> Currently openvswitch builds all libraries static only. However,
> >> libopenvswitch is linked into nearly al
On Fri, Dec 6, 2013 at 10:43 AM, Pravin B Shelar wrote:
> Kernel 3.3 to 3.8 has defined `struct flow_keys` but does not
> contains flow_keys.thoff field. Therefore we need to use
> compat definition for flow_keys struct.
>
> Signed-off-by: Pravin B Shelar
Acked-by: Jesse Gross
On Thu, Dec 05, 2013 at 04:27:26PM -0800, Jarno Rajahalme wrote:
> Add a prefix tree (trie) structure for tracking the used address
> space, enabling skipping classifier tables containing longer masks
> than necessary for an address field value in a packet header being
> classified. This enables l
On Wed, Nov 27, 2013 at 10:35:39AM +0100, Helmut Schaa wrote:
> Currently openvswitch builds all libraries static only. However,
> libopenvswitch is linked into nearly all openvswitch executables
> making it hardly possible to run openvswitch on embedded devices
> (for example running OpenWrt).
>
Kernel 3.3 to 3.8 has defined `struct flow_keys` but does not
contains flow_keys.thoff field. Therefore we need to use
compat definition for flow_keys struct.
Signed-off-by: Pravin B Shelar
---
datapath/linux/compat/include/net/flow_keys.h |2 +-
1 files changed, 1 insertions(+), 1 deletion
On Fri, Dec 06, 2013 at 09:46:29AM -0800, Ben Pfaff wrote:
> On Thu, Dec 05, 2013 at 04:27:25PM -0800, Jarno Rajahalme wrote:
> > Add the missing code for generating IPv6 packets for testing purposes.
> >
> > Also make flow_compose() set the l4 and l7 pointers more consistently
> > with flow_extra
I applied this patch to master. Henry, thanks for the review.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Thu, Dec 05, 2013 at 05:02:00PM -0800, Ben Pfaff wrote:
> Anything inside vlog_init__() that tried to log a message was going to
> deadlock, since it would hit pthread_once() recursively and wait for the
> previous call to complete. Unfortunately, there was a VLOG_ERR call inside
> vlog_init__(
Anything inside vlog_init__() that tried to log a message was going to
deadlock, since it would hit pthread_once() recursively and wait for the
previous call to complete. Unfortunately, there was a VLOG_ERR call inside
vlog_init__(), only called in the corner case where the system's clock was
wron
On Thu, Dec 05, 2013 at 04:27:25PM -0800, Jarno Rajahalme wrote:
> Add the missing code for generating IPv6 packets for testing purposes.
>
> Also make flow_compose() set the l4 and l7 pointers more consistently
> with flow_extract().
>
> Signed-off-by: Jarno Rajahalme
Would you please use uint
Guru can you do a quick review over the following:
diff --git a/lib/string.h b/lib/string.h
index 2b7b454..6981742 100644
--- a/lib/string.h
+++ b/lib/string.h
@@ -17,7 +17,11 @@
#ifndef STRING_WRAPPER_H
#define STRING_WRAPPER_H 1
+#ifdef _WIN32
+#include <../include/string.h>
+#else
#include_n
On Thu, Dec 05, 2013 at 04:27:24PM -0800, Jarno Rajahalme wrote:
> Ofproto always provides a valid pointer to a mask, but the mask length
> is zero when megaflows are disabled. Use match_wc_init() to get an
> appropriate exact match mask when there is no mask.
I'm having a little trouble understa
On Fri, Dec 6, 2013 at 8:39 AM, Jesse Gross wrote:
> On Thu, Dec 5, 2013 at 7:48 PM, Pravin B Shelar wrote:
>> Earlier RHEL6 kernel do not have linux/atomic.h header
>> file. Use types.h to get atomic_t definition.
>>
>> Signed-off-by: Pravin B Shelar
>
> Acked-by: Jesse Gross
I pushed it to
On Thu, Dec 05, 2013 at 04:36:26PM -0800, Jarno Rajahalme wrote:
> Inline, use another well-known algorithm for 64-bit builds, and use
> builtins when they are known to be fast at compile time. A 32-bit
> version of the alternate algorithm is slower than the existing
> implementation, so the old o
On Thu, Dec 5, 2013 at 7:48 PM, Pravin B Shelar wrote:
> Earlier RHEL6 kernel do not have linux/atomic.h header
> file. Use types.h to get atomic_t definition.
>
> Signed-off-by: Pravin B Shelar
Acked-by: Jesse Gross
___
dev mailing list
dev@openvswi
Pushed, thanks!
Jarno
On Dec 5, 2013, at 5:11 PM, Ben Pfaff wrote:
> On Thu, Dec 05, 2013 at 04:27:23PM -0800, Jarno Rajahalme wrote:
>> parse_l2_5_onward() previously used out of range 'expected_bit' when
>> it was called with no attrs, which happens when parsing a mask with
>> zero length.
On Dec 6, 2013, at 1:59 AM, Helmut Schaa wrote:
> On Wed, Nov 27, 2013 at 10:35 AM, Helmut Schaa
> wrote:
>> Currently openvswitch builds all libraries static only. However,
>> libopenvswitch is linked into nearly all openvswitch executables
>> making it hardly possible to run openvswitch on em
On Dec 6, 2013, at 1:18 AM, Alexander Wu wrote:
> Hi Jarno,
>
> I've read your patch "better count1_bits", and I test the gcc
> builtins separately.
>
> Call __builtin_popcount|__builtin_popcountl|__builtin_popcountll 10 million
> times
> --
>suse-kvm-o
On Dec 6, 2013, at 7:36 AM, Jarno Rajahalme wrote:
> On Dec 5, 2013, at 4:44 PM, Ben Pfaff wrote:
>> How did you measure the benefit of inlining?
>
> With a standalone C program running different variants (original, inlined,
> different algorithm) over
I should mention that I linked with a st
On Dec 5, 2013, at 4:44 PM, Ben Pfaff wrote:
> On Thu, Dec 05, 2013 at 04:36:26PM -0800, Jarno Rajahalme wrote:
>> Inline, use another well-known algorithm for 64-bit builds, and use
>> builtins when they are known to be fast at compile time. A 32-bit
>> version of the alternate algorithm is sl
When adding a physical port to the main bridge the mac address
of the bridge is updated. We can do the same for fake bridges by
copying the mac address of the parent bridge.
There exists only one fake bridge per vlan, hence it is safe
to copy the mac address of the parent bridge.
Signed-off-by: H
On Wed, Nov 27, 2013 at 10:35 AM, Helmut Schaa
wrote:
> Currently openvswitch builds all libraries static only. However,
> libopenvswitch is linked into nearly all openvswitch executables
> making it hardly possible to run openvswitch on embedded devices
> (for example running OpenWrt).
>
> Conver
v4:
No change.
v3:
No change.
v2:
No change
v1:
Now the cli we implement is very crude. Maybe it could
display better.
Signed-off-by: Alexander Wu
Reviewed-by: Simon Horman
---
utilities/ovs-ofctl.c | 17 +
1 files changed, 17 insertions(+), 0 deletions(-)
diff -
v4:
1. update abstract table-features to bitmap
2. update ofproto, reduce copy when reply table features.
3. Move max_table_feature to oftable.
v3:
1. Delete variable OFPIT13_END, use OFPIT13_METER directly.
2. Fix implement, restructure with macro.
3. Add comments that describe table-
v4.1:
Update bitmap_count1 function: call count_1bits to improve
performance.
v4:
Add bitmap_count1 function to count all 1 bits.
---
lib/bitmap.c | 15 +++
lib/bitmap.h |1 +
2 files changed, 16 insertions(+), 0 deletions(-)
diff --git a/lib/bitmap.c b/lib/bitmap.c
index
v4:
1. Delete duplication code.
2. Add new acts in *.def.
3. Update abstract table-features to bitmap.
4. Make decode_openflow13_props more general.
5. Update tests. Correct the wriable oxms.
v3:
1. Update names of functions/macros to make them meaningful.
2. Fix codingstyle.
3. Re
v4.1:
Update bitmap_count1 implemnt to call count1_bits directly.
Bitmap is a separate subsystem, so I think bitmap_count1 is needed.
v4:
Implement abstract bitmap for table feature properties.
Make prop functions more general.
Remove duplicated code, use existed implement.
Squash some
v4:
Add enums of ofp13_action_type and make them work.
v3:
Add func to get instruction/action name by OpenFlow instruction type directly.
---
include/openflow/openflow-1.3.h | 20 ++--
lib/ofp-actions.c | 19 +++
lib/ofp-actions.h
Hi Jarno,
I've read your patch "better count1_bits", and I test the gcc
builtins separately.
Call __builtin_popcount|__builtin_popcountl|__builtin_popcountll 10 million
times
--
suse-kvm-of13:/test # time ./bit4
real0m0.034s
user0m0.032s
69 matches
Mail list logo
