tunnel not checksum offloading properly.
Fixes: 85bcbbed839a ("userspace: Enable tunnel tests with TSO.")
Reported-by: David Marchand
Signed-off-by: Mike Pattrick
---
nb: The first patch was removed from the v2 version of this series,
which was not explicitly a bug fix.
---
lib/netd
On Wed, Feb 14, 2024 at 12:09 PM David Marchand
wrote:
>
> Hello Mike,
>
> On Mon, Feb 12, 2024 at 8:50 PM Mike Pattrick wrote:
> >
> > Previously if an OVS configuration nested multiple layers of UDP tunnels
> > like VXLAN or GENEVE on top of each other throu
Note, this failed github CI with the following error message:
../../tests/testsuite: line 4034: ./atconfig: No such file or directory
This appears to be a false negative.
-M
On Mon, Feb 12, 2024 at 2:50 PM Mike Pattrick wrote:
>
> Previously if an OVS configuration nested multiple
right after its generation. So, there is no point in
> creating it with restricted permissions in the first place.
>
> Fixes: 99e5e05db37a ("ovs-pki: Create private keys with restricted
> permissions.")
> Signed-off-by: Ilya Maximets
Acked-by: Mike Pattrick
___
On Tue, Feb 13, 2024 at 2:42 PM Ilya Maximets wrote:
>
> It's not an executable file.
>
> Signed-off-by: Ilya Maximets
Acked-by: Mike Pattrick
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On Mon, Feb 12, 2024 at 10:04 AM Ilya Maximets wrote:
>
> On 2/12/24 15:19, Mike Pattrick wrote:
> > On Mon, Feb 12, 2024 at 7:52 AM Ilya Maximets wrote:
> >>
> >> On 2/12/24 09:13, Mike Pattrick wrote:
> >>> When sending packets that are flagged as re
tunnel not checksum offloading properly.
Fixes: 3337e6d91c5b ("userspace: Enable L4 checksum offloading by default.")
Reported-by: David Marchand
Signed-off-by: Mike Pattrick
---
lib/netdev-linux.c | 15 ---
1 file changed, 12 insertions(+), 3 deletions(-)
diff --git a/
an appropriate check for ncat support. That check was
added to these tests.
Signed-off-by: Mike Pattrick
---
v2: Start the ncat listener before pings, so the socket has a better
chance of being ready to accept connections when the ncat client starts.
Signed-off-by: Mike Pattrick
---
tests/system
t;)
Signed-off-by: Mike Pattrick
---
lib/netdev-linux.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/netdev-linux.c b/lib/netdev-linux.c
index 7a156cc28..51517854b 100644
--- a/lib/netdev-linux.c
+++ b/lib/netdev-linux.c
@@ -7191,8 +7191,8 @@ netdev_linux_prepen
When sending packets that are flagged as requiring segmentation to an
interface that doens't support this feature, send the packet to the TSO
software fallback instead of dropping it.
Signed-off-by: Mike Pattrick
---
v2: Start the ncat listener before pings, so the socket has a better
chance
On Mon, Feb 12, 2024 at 7:52 AM Ilya Maximets wrote:
>
> On 2/12/24 09:13, Mike Pattrick wrote:
> > When sending packets that are flagged as requiring segmentation to an
> > interface that doens't support this feature, send the packet to the TSO
> > software fall
Previously a gap existed in the tunnel system tests where only ICMP and
TCP traffic was tested. However, the code paths using for UDP traffic is
different then either of those and should also be tested.
Signed-off-by: Mike Pattrick
---
tests/system-traffic.at | 50
t;)
Signed-off-by: Mike Pattrick
---
lib/netdev-linux.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/netdev-linux.c b/lib/netdev-linux.c
index 7a156cc28..51517854b 100644
--- a/lib/netdev-linux.c
+++ b/lib/netdev-linux.c
@@ -7191,8 +7191,8 @@ netdev_linux_prepen
tunnel not checksum offloading properly.
Fixes: 3337e6d91c5b ("userspace: Enable L4 checksum offloading by default.")
Reported-by: David Marchand
Signed-off-by: Mike Pattrick
---
lib/netdev-linux.c | 15 ---
1 file changed, 12 insertions(+), 3 deletions(-)
diff --git a/
When sending packets that are flagged as requiring segmentation to an
interface that doens't support this feature, send the packet to the TSO
software fallback instead of dropping it.
Signed-off-by: Mike Pattrick
---
Note: Previously this patch failed gitlab ci, however, I was not able
Currently the monitor will call dp_packet_clear() on the dp_packet that
is shared amongst BFD, LLDP, and CFM. However, all of these packets are
created with eth_compose(), which already calls dp_packet_clear().
Reviewed-by: David Marchand
Signed-off-by: Mike Pattrick
---
ofproto/ofproto-dpif
Include inner offsets in functions where l3 and l4 offsets are either
modified or checked.
Fixes: 084c8087292c ("userspace: Support VXLAN and GENEVE TSO.")
Signed-off-by: Mike Pattrick
---
v2:
- Prints out new offsets in autovalidator
- Extends resize_l2 change to avx512
v3:
-
t;bfd: Implement Bidirectional Forwarding Detection.")
Signed-off-by: Mike Pattrick
---
v2: Corrected formatting, and just calculate checksum up front
v3: Extended patch comment
---
lib/bfd.c | 23 +++
1 file changed, 11 insertions(+), 12 deletions(-)
diff --git a/lib/bfd.c b
as long as
the flags of the packet are correct.
Fixes: 85bcbbed839a ("userspace: Enable tunnel tests with TSO.")
Reviewed-by: David Marchand
Signed-off-by: Mike Pattrick
---
v2: Corrected patch subject
---
lib/dp-packet.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --gi
as long as
the flags of the packet are correct.
Fixes: 85bcbbed839a ("userspace: Enable tunnel tests with TSO.")
Reviewed-by: David Marchand
Signed-off-by: Mike Pattrick
---
v2: Corrected patch subject
---
lib/dp-packet.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --gi
t;bfd: Implement Bidirectional Forwarding Detection.")
Signed-off-by: Mike Pattrick
---
v2: Corrected formatting, and just calculate checksum up front
v3: Extended patch comment
---
lib/bfd.c | 23 +++
1 file changed, 11 insertions(+), 12 deletions(-)
diff --git a/lib/bfd.c b
Currently the monitor will call dp_packet_clear() on the dp_packet that
is shared amongst BFD, LLDP, and CFM. However, all of these packets are
created with eth_compose(), which already calls dp_packet_clear().
Reviewed-by: David Marchand
Signed-off-by: Mike Pattrick
---
ofproto/ofproto-dpif
Include inner offsets in functions where l3 and l4 offsets are either
modified or checked.
Fixes: 084c8087292c ("userspace: Support VXLAN and GENEVE TSO.")
Signed-off-by: Mike Pattrick
---
v2:
- Prints out new offsets in autovalidator
- Extends resize_l2 change to avx512
v3:
-
, returning immediately. One notable change in
functionality is not performing the RBAC check in this condition, as no
mutation actually takes place.
Reported-by: Terry Wilson
Reported-at: https://issues.redhat.com/browse/FDP-359
Signed-off-by: Mike Pattrick
---
ovsdb/execution.c | 9
On Wed, Jan 31, 2024 at 10:04 AM David Marchand
wrote:
>
> On Tue, Jan 30, 2024 at 11:15 PM Mike Pattrick wrote:
> >
> > Include inner offsets in functions where l3 and l4 offsets are either
> > modified or checked.
> >
> > Fixes: 084c8087292c (&quo
Include inner offsets in functions where l3 and l4 offsets are either
modified or checked.
Fixes: 084c8087292c ("userspace: Support VXLAN and GENEVE TSO.")
Signed-off-by: Mike Pattrick
---
v2:
- Prints out new offsets in autovalidator
- Extends resize_l2 change to avx512
Signed-of
Currently the monitor will call dp_packet_clear() on the dp_packet that
is shared amongst BFD, LLDP, and CFM. However, all of these packets are
created with eth_compose(), which already calls dp_packet_clear().
Signed-off-by: Mike Pattrick
---
ofproto/ofproto-dpif-monitor.c | 3 ---
1 file
Previously the BFD packet creation code did not appropriately set
offsets or flags. This contributed to issues involving encapsulation and
the TSO code.
Fixes: ccc096898c46 ("bfd: Implement Bidirectional Forwarding Detection.")
Signed-off-by: Mike Pattrick
---
v2: Corrected formatting
as long as
the flags of the packet are correct.
Fixes: 85bcbbed839a ("userspace: Enable tunnel tests with TSO.")
Signed-off-by: Mike Pattrick
---
v2: Corrected patch subject
---
lib/dp-packet.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/dp-packet.h b/lib/d
-at: https://issues.redhat.com/browse/FDP-300
Signed-off-by: Mike Pattrick
---
lib/dp-packet.h | 3 +++
lib/packets.c | 3 ---
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/lib/dp-packet.h b/lib/dp-packet.h
index 939bec5c8..dceb701e8 100644
--- a/lib/dp-packet.h
+++ b/lib/dp-packet.h
as long as
the flags of the packet are correct.
Fixes: 85bcbbed839a ("userspace: Enable tunnel tests with TSO.")
Signed-off-by: Mike Pattrick
---
lib/dp-packet.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/dp-packet.h b/lib/dp-packet.h
index dceb701e8..802d3f
Currently the monitor will call dp_packet_clear() on the dp_packet that
is shared amongst BFD, LLDP, and CFM. However, all of these packets are
created with eth_compose(), which already calls dp_packet_clear().
Signed-off-by: Mike Pattrick
---
ofproto/ofproto-dpif-monitor.c | 3 ---
1 file
Include inner offsets in functions where l3 and l4 offsets are either
modified or checked.
Signed-off-by: Mike Pattrick
Fixes: 084c8087292c ("userspace: Support VXLAN and GENEVE TSO.")
---
lib/dp-packet.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/lib/dp-
Previously the BFD packet creation code did not appropriately set
offsets or flags. This contributed to issues involving encapsulation and
the TSO code.
Signed-off-by: Mike Pattrick
Fixes: ccc096898c46 ("bfd: Implement Bidirectional Forwarding Detection.")
Signed-off-by: Mike Pattrick
On Wed, Jan 24, 2024 at 1:14 PM Mike Pattrick wrote:
>
> When sending packets that are flagged as requiring segmentation to an
> interface that doens't support this feature, send the packet to the TSO
> software fallback instead of dropping it.
>
> Signed-off-by: Mike Pattrick
When sending packets that are flagged as requiring segmentation to an
interface that doens't support this feature, send the packet to the TSO
software fallback instead of dropping it.
Signed-off-by: Mike Pattrick
---
lib/dp-packet-gso.c | 73 +
lib/dp
On Tue, Jan 23, 2024 at 7:41 AM Ilya Maximets wrote:
>
> On 1/22/24 23:24, Mike Pattrick wrote:
> > On Mon, Jan 22, 2024 at 4:10 PM Ilya Maximets wrote:
> >>
> >> On 1/22/24 21:33, Mike Pattrick wrote:
> >>> On Mon, Jan 22, 2024 at 1:47 PM Ilya Maxim
("userspace: Add Generic Segmentation Offloading.")
Fixes: 85bcbbed839a ("userspace: Enable tunnel tests with TSO.")
Reported-by: Dumitru Ceara
Reported-at: https://issues.redhat.com/browse/FDP-300
Signed-off-by: Mike Pattrick
---
lib/dp-packet.h | 4 +++-
1 file changed
On Mon, Jan 22, 2024 at 4:10 PM Ilya Maximets wrote:
>
> On 1/22/24 21:33, Mike Pattrick wrote:
> > On Mon, Jan 22, 2024 at 1:47 PM Ilya Maximets wrote:
> >>
> >> On 1/22/24 18:51, Mike Pattrick wrote:
> >>> The OVN test suite identified a bug in dp_pac
On Mon, Jan 22, 2024 at 1:47 PM Ilya Maximets wrote:
>
> On 1/22/24 18:51, Mike Pattrick wrote:
> > The OVN test suite identified a bug in dp_packet_ol_send_prepare() where
> > a double encapsulated BFD packet would trigger a seg fault. This
> > happened because w
Signed-off-by: Mike Pattrick
---
lib/dp-packet.h | 10 --
lib/packets.c | 6 +++---
2 files changed, 11 insertions(+), 5 deletions(-)
diff --git a/lib/dp-packet.h b/lib/dp-packet.h
index 939bec5c8..81b01f331 100644
--- a/lib/dp-packet.h
+++ b/lib/dp-packet.h
@@ -537,7 +537,7 @@ dp_packe
=),protoinfo=(state=)
+tcp,orig=(src=10.1.1.4,dst=20.22.166.15,sport=,dport=),reply=(src=20.22.166.15,dst=10.1.1.4,sport=,dport=),protoinfo=(state=)
Those are github IP's, don't know how github traffic got into the tests.
-M
On Fri, Jan 19, 2024 at 11:13 AM Mike Pattrick wrote:
>
> This
in the single function that is causing all of
the errors.
Signed-off-by: Mike Pattrick
---
include/openvswitch/compiler.h | 11 +++
lib/ovs-rcu.c | 1 +
2 files changed, 12 insertions(+)
diff --git a/include/openvswitch/compiler.h b/include/openvswitch/compiler.h
index 878c5c6a7
n inline function and calling into the full send prepare if it
passes, but that change didn't yield any noticeable performance
improvement over the current solution.
Acked-by: Mike Pattrick
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On Wed, Jan 17, 2024 at 5:09 AM Dexia Li wrote:
>
>
>
> On 1/14/24 00:34, Ilya Maximets wrote:
> > On 12/27/23 19:55, Mike Pattrick wrote:
> >> From: Dexia Li
> >>
> >> For userspace datapath, this patch provides vxlan and geneve tunnel tso.
>
and geneve tso,packets will drop. Front-end devices can close offload
features by ethtool also.
Signed-off-by: Dexia Li
Co-authored-by: Mike Pattrick
Signed-off-by: Mike Pattrick
---
v9: Rebased patch
v12:
- Sent in by Dexia
v13:
- Corrected formatting and comment/function naming consistency
as the traffic produced by this test gives
incorrect values in the vnet header.
Signed-off-by: Mike Pattrick
---
v10:
- Software TCP checksums now support encapsulated TSO case
- Redundant inner offset code was removed
v11:
- Corrected logic in netdev-linux vnet prepend code
v12:
- Submitted
On Sat, Jan 13, 2024 at 6:51 PM Ilya Maximets wrote:
>
> On 12/27/23 19:55, Mike Pattrick wrote:
> > This patch enables most of the tunnel tests in the testsuite, and adds a
> > large TCP transfer to a vxlan and geneve test to verify TSO
> > functionality. Some addition
explicitly sets the appropriate flags when the segmentation
flag is removed, to guarantee that packets always end up with correct
checksums.
Signed-off-by: Mike Pattrick
Fixes: 8b5fe2dc6080 ("userspace: Add Generic Segmentation Offloading.")
---
v3: Moved logic from ofproto-dpif-upcall to
Test that netdev-dummy is able to send and receive segment offloaded
packets.
Signed-off-by: Mike Pattrick
---
v2: Fix clang build error: mutex needed to access netdev_dummy members
v3: Remove use of tcpdump, hexdump, and otherwise clean up test
v4: Wrapped long lines in test
v5:
- Implemented
Correct the length field in IPv6 packets when applying software fallback
GSO. Previosuly the field retained an IPv4 header size, which was
incorrect.
Signed-off-by: Mike Pattrick
Fixes: 8b5fe2dc6080 ("userspace: Add Generic Segmentation Offloading.")
---
v6: include ip6 extensions
-
Correct the length field in IPv6 packets when applying software fallback
GSO. Previosuly the field retained an IPv4 header size, which was
incorrect.
Signed-off-by: Mike Pattrick
---
lib/dp-packet-gso.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/dp-packet-gso.c b
Test that netdev-dummy is able to send and receive segment offloaded
packets.
Signed-off-by: Mike Pattrick
---
v2: Fix clang build error: mutex needed to access netdev_dummy members
v3: Remove use of tcpdump, hexdump, and otherwise clean up test
v4: Wrapped long lines in test
v5:
- Implemented
explicitly sets the appropriate flags when the segmentation
flag is removed, to guarantee that packets always end up with correct
checksums.
Signed-off-by: Mike Pattrick
---
v3: Moved logic from ofproto-dpif-upcall to dp-packet
v5: Corrected indectionation, and removed a redundant flag from the mask
On Sat, Jan 13, 2024 at 4:47 PM Ilya Maximets wrote:
>
> On 12/19/23 15:47, Mike Pattrick wrote:
> > Test that netdev-dummy is able to send and receive segment offloaded
> > packets.
> >
> > Signed-off-by: Mike Pattrick
> > ---
> > v2: Fix clang build
On Sat, Jan 13, 2024 at 4:21 PM Ilya Maximets wrote:
>
> On 12/19/23 15:47, Mike Pattrick wrote:
> > When OVS needs to fallback on the software TSO implementation to segment
> > a packet, it currently doesn't guarantee that IP and TCP checksum
> > offload flags are set.
Ilya Maximets
Thanks for the updated comment.
Acked-by: Mike Pattrick
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
te a new structure. This
> is done in order for different callers to have their own
> default values and only update them with what is provided
> by the user explicitly. For example, replication and relay
> have different default probe intervals.
>
> Signed-off-by: Ily
-by: Frode Nordahl
That looks correct to me.
Acked-by: Mike Pattrick
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On Fri, Jan 5, 2024 at 9:40 AM Ilya Maximets wrote:
>
> On 1/1/24 07:37, Mike Pattrick wrote:
> > On Sun, Dec 17, 2023 at 9:03 PM Ilya Maximets wrote:
> >>
> >> In case the difference between 'old' and 'new' rows is readily
> >> available, it can be
On Fri, Jan 5, 2024 at 9:30 AM Ilya Maximets wrote:
>
> On 12/29/23 17:28, Mike Pattrick wrote:
> > On Sun, Dec 17, 2023 at 9:03 PM Ilya Maximets wrote:
> >>
> >> While reassessing weak references the code attempts to collect added
> >> and removed
a database itself. Will be used
> later for running active and backup databases in a single process.
>
> Marking the _Server database as read only is not necessary, because
> modifications of internal databases are not allowed anyway, but it
> doesn't hurt.
>
> Signed-off-by:
On Wed, Dec 13, 2023 at 8:05 PM Ilya Maximets wrote:
>
> It's not a big problem, but it would be nice to ensure that
> the backup database cannot be locally converted.
>
> Fixes: e51879e99b3e ("ovsdb: Make OVSDB backup sever read only")
> Signed-off-by: Ilya Maxime
d-off-by: Ilya Maximets
Acked-by: Mike Pattrick
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
adic test failures.
> Should not affect performance in real-world cases as the actual
> outgoing message is still not sorted.
>
> Signed-off-by: Ilya Maximets
Acked-by: Mike Pattrick
___
dev mailing list
d...@openvswitch.org
https://mail.openvs
On Wed, Dec 13, 2023 at 8:05 PM Ilya Maximets wrote:
>
> It's easier to analyze failures when the lines that are different
> are shown next to each other.
>
> Signed-off-by: Ilya Maximets
Acked-by: Mike Pattrick
___
dev
On Wed, Dec 13, 2023 at 8:05 PM Ilya Maximets wrote:
>
> It's almost impossible to debug test failures without them.
>
> Signed-off-by: Ilya Maximets
Acked-by: Mike Pattrick
___
dev mailing list
d...@openvswitch.org
https://mail.ope
On Sun, Dec 17, 2023 at 9:03 PM Ilya Maximets wrote:
>
> In case the difference between 'old' and 'new' rows is readily
> available, it can be used to construct added/removed datums
> instead. Diffs are typically much smaller than the column
> itself. This change more than doubles the
long, we keep them around only while transaction
> is alive, so should not affect memory consumption.
>
> Users for this data will be added in later commits.
>
> Signed-off-by: Ilya Maximets
Acked-by: Mike Pattrick
___
dev mailin
index],
> - datum, >type);
> -} else {
> -ovsdb_datum_init_empty();
> -ovsdb_datum_clone(, datum);
> +ovsdb_datum_init_empty();
> +ovsdb_datum_init_empty();
Nit: Why init these here if they will be overwrit
uot;)
> Fixes: b2712d026eae ("ovsdb: transaction: Use diffs for strong reference
> counting.")
> Signed-off-by: Ilya Maximets
Acked-by: Mike Pattrick
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
citly
> unshared.
>
> Also making ovsdb_datum_subtract() call conditional, so the
> issue can be hit by existing unit tests.
>
> Fixes: 485ac63d10f8 ("ovsdb: Add lazy-copy support for ovsdb_datum objects.")
> Signed-off-by: Il
as the traffic produced by this test gives
incorrect values in the vnet header.
Signed-off-by: Mike Pattrick
---
v11:
- Corrected logic in netdev-linux vnet prepend code
v10:
- Software TCP checksums now support encapsulated TSO case
- Redundant inner offset code was removed
Signed-off
and geneve tso,packets will drop. Front-end devices can close offload
features by ethtool also.
Signed-off-by: Dexia Li
Co-authored-by: Mike Pattrick
Signed-off-by: Mike Pattrick
---
v9: Rebased patch
---
lib/dp-packet.c | 41 +++-
lib/dp-packet.h | 216
Test that netdev-dummy is able to send and receive segment offloaded
packets.
Signed-off-by: Mike Pattrick
---
v2: Fix clang build error: mutex needed to access netdev_dummy members
v3: Remove use of tcpdump, hexdump, and otherwise clean up test
v4: Wrapped long lines in test
---
lib/netdev
explicitly sets the appropriate flags when the segmentation
flag is removed, to guarantee that packets always end up with correct
checksums.
Signed-off-by: Mike Pattrick
---
v3: Moved logic from ofproto-dpif-upcall to dp-packet
Signed-off-by: Mike Pattrick
---
lib/dp-packet.h | 16
explicitly sets the appropriate flags when the segmentation
flag is removed, to guarantee that packets always end up with correct
checksums.
Signed-off-by: Mike Pattrick
---
v3: Moved logic from ofproto-dpif-upcall to dp-packet
Signed-off-by: Mike Pattrick
---
lib/dp-packet.h | 16
Test that netdev-dummy is able to send and receive segment offloaded
packets.
Signed-off-by: Mike Pattrick
---
v2: Fix clang build error: mutex needed to access netdev_dummy members
v3: Remove use of tcpdump, hexdump, and otherwise clean up test
Note, somehow the v3 patch reported as corrupted
These patches apply cleanly on the current master branch for me. I
will try to resubmit.
-M
On Mon, Dec 18, 2023 at 4:21 PM 0-day Robot wrote:
>
> Bleep bloop. Greetings Mike Pattrick, I am a robot and I have tried out your
> patch.
> Thanks for your contribution.
>
>
Test that netdev-dummy is able to send and receive segment offloaded
packets.
Signed-off-by: Mike Pattrick
---
v2: Fix clang build error: mutex needed to access netdev_dummy members
v3: Remove use of tcpdump, hexdump, and otherwise clean up test
---
lib/netdev-dummy.c | 32
explicitly sets the appropriate flags when the segmentation
flag is removed, to guarantee that packets always end up with correct
checksums.
Signed-off-by: Mike Pattrick
---
v3: Moved logic from ofproto-dpif-upcall to dp-packet
---
lib/dp-packet.h | 16 ++--
1 file changed, 14
On Fri, Dec 15, 2023 at 8:05 PM Ilya Maximets wrote:
>
> On 12/11/23 16:39, Mike Pattrick wrote:
> > Test that netdev-dummy is able to send and recieve segment offloaded
> > packets.
> >
> > Signed-off-by: Mike Pattrick
> >
> > ---
> > v
On Mon, Dec 18, 2023 at 5:04 AM Eelco Chaudron wrote:
>
>
>
> On 15 Dec 2023, at 19:22, Mike Pattrick wrote:
>
> > On Fri, Dec 15, 2023 at 8:01 AM Eelco Chaudron wrote:
> >>
> >>
> >>
> >> On 11 Dec 2023, at 16:39, Mike Pattrick wrote:
&
as the traffic produced by this test gives
incorrect values in the vnet header.
Signed-off-by: Mike Pattrick
---
v10:
- Software TCP checksums now support encapsulated TSO case
- Redundant inner offset code was removed
---
lib/dp-packet.h | 49
and geneve tso,packets will drop. Front-end devices can close offload
features by ethtool also.
Signed-off-by: Dexia Li
Co-authored-by: Mike Pattrick
Signed-off-by: Mike Pattrick
---
v9: Rebased patch
---
lib/dp-packet.c | 41 +++-
lib/dp-packet.h | 216
On Fri, Dec 15, 2023 at 8:01 AM Eelco Chaudron wrote:
>
>
>
> On 11 Dec 2023, at 16:39, Mike Pattrick wrote:
>
> > Currently dp_netdev_upcall() resolves checksums on all packets, but this
> > isn't strictly needed. The checksums will be resolved before
> >
as the traffic produced by this test gives
incorrect values in the vnet header.
Signed-off-by: Mike Pattrick
---
lib/dp-packet.h | 39 ---
lib/dpif-netdev-extract-avx512.c | 8 ++
lib/flow.c | 12 +++--
lib/netdev-linux.c
and geneve tso,packets will drop. Front-end devices can close offload
features by ethtool also.
Signed-off-by: Dexia Li
Signed-off-by: Mike Pattrick
---
lib/dp-packet.c | 41 +++-
lib/dp-packet.h | 216
lib/dpif-netdev.c | 4
serve references for deleted rows.")
Reported-at: https://issues.redhat.com/browse/FDP-193
Signed-off-by: Mike Pattrick
---
v2: Removed a redundant case.
Signed-off-by: Mike Pattrick
---
lib/ovsdb-idl.c | 11 +--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/lib/ovsdb-i
From: Dumitru Ceara
Add a check for inserting and then deleting rows referenced by weak
references by the same idl client. Verify that insertion and deletion
events are recorded.
Reported-at: https://issues.redhat.com/browse/FDP-193
Signed-off-by: Dumitru Ceara
Co-authored-by: Mike Pattrick
On Tue, Dec 12, 2023 at 7:45 AM Dumitru Ceara wrote:
>
> On 12/12/23 05:04, Mike Pattrick wrote:
> > In the case of a weak reference, clearing all change_seqno's can delete
> > useful information. Instead of clearing all seqno's when removing
> > track_node, only cle
From: Dumitru Ceara
Add a check for inserting and then deleting rows referenced by weak
references by the same idl client. Verify that insertion and deletion
events are recorded.
Reported-at: https://issues.redhat.com/browse/FDP-193
Signed-off-by: Dumitru Ceara
Co-authored-by: Mike Pattrick
serve references for deleted rows.")
Reported-at: https://issues.redhat.com/browse/FDP-193
Signed-off-by: Mike Pattrick
---
lib/ovsdb-idl.c | 13 +++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/lib/ovsdb-idl.c b/lib/ovsdb-idl.c
index 634fbb56d..77f1a6db4 100644
--- a
Test that netdev-dummy is able to send and recieve segment offloaded
packets.
Signed-off-by: Mike Pattrick
---
v2: Fix clang build error: mutex needed to access netdev_dummy members
Signed-off-by: Mike Pattrick
---
lib/netdev-dummy.c | 32 +++-
tests/dpif-netdev.at
Currently dp_netdev_upcall() resolves checksums on all packets, but this
isn't strictly needed. The checksums will be resolved before
transmission. However, we do have to resolve the checksums before
sending a packet to the controller as offload flags aren't retained.
Signed-off-by: Mike Pattrick
Test that netdev-dummy is able to send and recieve segment offloaded
packets.
Signed-off-by: Mike Pattrick
---
lib/netdev-dummy.c | 28 -
tests/dpif-netdev.at | 50
2 files changed, 77 insertions(+), 1 deletion(-)
diff
Currently dp_netdev_upcall() resolves checksums on all packets, but this
isn't strictly needed. The checksums will be resolved before
transmission. However, we do have to resolve the checksums before
sending a packet to the controller as offload flags aren't retained.
Signed-off-by: Mike Pattrick
On Fri, Dec 1, 2023 at 7:39 PM Ilya Maximets wrote:
>
> On 11/22/23 11:16, Simon Horman wrote:
> > On Tue, Nov 21, 2023 at 02:26:52PM -0500, Mike Pattrick wrote:
> >> From: Flavio Leitner
> >>
> >> This provides a software implementation in the case
On Fri, Nov 3, 2023 at 12:01 AM Dexia Li via dev
wrote:
>
> For userspace datapath, this patch provides vxlan and geneve tunnel tso.
> Only support userspace vxlan or geneve tunnel, meanwhile support
> tunnel outter and inner csum offload. If netdev do not support offload
> features, there is
is sent in order to the netdev.
Signed-off-by: Flavio Leitner
Co-authored-by: Mike Pattrick
Signed-off-by: Mike Pattrick
---
v4:
- Various formatting changes
- Fixed memory leak in soft-gso code if packet is flagged
for GSO but incorrectly lacks segment size.
v5:
- Corrected indentation
101 - 200 of 664 matches
Mail list logo
