When a packet_in message results from a packet_out message
there is no particular relationship between the buffer_ids of
the two messages.
I noticed this when using Open vSwitch's "make ryu-check".
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Wireshark may misunderstand a MAC frame (including ARP) whose length is 64 bytes or more, as a frame with the FCS field.
This patch fixes the issue by modifying lengths of pad fields for ARP.
Reported-by: Arne Goetje <arne_goetje@accton.com>
Signed-off-by: Yuichi Ito <ito.yuichi0@gmail.com>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
IEEE 802.3-2012 (3.2.3 Address fields) says:
b) The first bit (LSB) shall be used in the Destination Address field as an address type designation bit
to identify the Destination Address either as an individual or as a group address. If this bit is 0, it
shall indicate that the address field contains an individual address. If this bit is 1, it shall indicate that
the address field contains a group address that identifies none, one or more, or all of the stations
connected to the LAN. In the Source Address field, the first bit is reserved and set to 0.
c) The second bit shall be used to distinguish between locally or globally administered addresses. For
globally administered (or U, universal) addresses, the bit is set to 0. If an address is to be assigned
locally, this bit shall be set to 1. Note that for the broadcast address, this bit is also a 1.
This patch fixes MAC addresses used in tests as follows:
- set the first bit to 0
- set the second bit to 1
Reported-by: Arne Goetje <arne_goetje@accton.com>
Signed-off-by: Yuichi Ito <ito.yuichi0@gmail.com>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
For a goto-table instruction use a table_id greater than the table in which
a flow will be inserted. OpenFlow 1.2 section 5.6 stipulates that this is
required.
I noticed this when using Open vSwitch's "make ryu-check" as
Open vSwitch enforces the above stipulation.
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Use actual table ids rather than OFPTT_ALL for
flow_mod modify tests as OpenFlow 1.2 section A3.4.1
stipulates that OFPTT_ALL is only valid for flow_mod delete requests.
I noticed this when using Open vSwitch's "make ryu-check" as
Open vSwitch enforces the above stipulation.
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
RROR 2014-04-23 11:54:59,239 hub 58 hub: uncaught exception: Traceback (most recent c\
all last):
File "/Users/fujita/git/ryu/ryu/lib/hub.py", line 50, in _launch
func(*args, **kwargs)
File "/Users/fujita/git/ryu/ryu/services/protocols/bgp/base.py", line 239, in start
self._run(*args, **kwargs)
File "/Users/fujita/git/ryu/ryu/services/protocols/bgp/peer.py", line 448, in _run
self._connect_loop(client_factory)
File "/Users/fujita/git/ryu/ryu/services/protocols/bgp/peer.py", line 820, in _conne\
ct_loop
bind_address=bind_addr)
File "/Users/fujita/git/ryu/ryu/services/protocols/bgp/base.py", line 354, in _conne\
ct_tcp
return sock
UnboundLocalError: local variable 'sock' referenced before assignment
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
RFC2402 says:
All IPv6 extension headers, as per RFC 1883, encode the "Hdr Ext Len" field by first
subtracting 1 (64-bit word) from the header length (measured in 64-bit words).
AH is an IPv6 extension header. However, since its length is measured in 32-bit words,
the "Payload Length" is calculated by subtracting 2 (32 bit words).
This patch fixes as follows:
return (int(size) - 1) * 8 ->
return (int(size) + 2) * 4
^ ^ ^
And, this patch also fixes a default argument of length.
Signed-off-by: TAKAHASHI Minoru <takahashi.minoru7@gmail.com>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
this new command takes a sub-command name as its first positional argument
and passes the rest of arguments to the corresponding sub-module.
for example, 'run' sub-command is an equivalent to bin/ryu-manager.
this commit does not replace the existing commands like ryu-manager.
this merely adds new ways to invoke the functionalities.
Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
specifically we want their commit de803cdf
(Change-Id: If712e1d51c1e1329093acb16ed6a84e9f0baf8c1)
for stable order of cli options.
the following sub-command commits rely on it.
Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
this will be used by the comming 'ryu' command.
Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Mask lengths without byte boundary are fixed as follows:
- VLAN_VID, from 16 bits (2 bytes) to 12 + 1 bits
- IPV6_EXTHDR, from 16 bits (2 bytes) to 9 bits
Reported-by: Arne Goetje <arne_goetje@accton.com>
Signed-off-by: Yuichi Ito <ito.yuichi0@gmail.com>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
This makes the GitHub interface aware of the contribution guidelines,
so it will be displayed to contributors when they submit issues or
pull requests.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
OF 1.3.3 spec (3 Glossary) says:
Flow Entry: an element in a flow table used to match and process packets. It contains a set of
match fields for matching packets, a priority for matching precedence, a set of counters to track
packets, and a set of instructions to apply (see 5.2).
According to OF spec, the match fields are treated as the set.
This means that the order of matches which flow_stats message returns is inconsistant.
This patch corrects the comparison method of matches by sorting in a particular order before comparison.
Signed-off-by: Yuichi Ito <ito.yuichi0@gmail.com>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
OF 1.3.3 spec (7.2.3.5 Flow Match Field Masking) says:
An all-zero-bits oxm_mask is equivalent to omitting the OXM TLV entirely. An all-one-bits oxm_mask
is equivalent to specifying 0 for oxm_hasmask and omitting oxm_mask.
This patch fixes to compare OFPMatch ignoring fields that masks are all zero bits.
Signed-off-by: Yuichi Ito <ito.yuichi0@gmail.com>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
OF 1.3.3 spec (7.2.3.5 Flow Match Field Masking) says:
An all-zero-bits oxm_mask is equivalent to omitting the OXM TLV entirely. An all-one-bits oxm_mask
is equivalent to specifying 0 for oxm_hasmask and omitting oxm_mask.
This patch fixes to compare OFPMatch ignoring masks that are all one bits.
Reported-by: Arne Goetje <arne_goetje@accton.com>
Signed-off-by: Yuichi Ito <ito.yuichi0@gmail.com>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
i-tagged packets will have ETH_TYPE=0x88E7. The tests are
rewritten in the same way with MPLS.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
this will be used by ofctl app which registers handlers dynamically.
when unobserving an event cls dynamically, there might be events of
the cls left in app's event queue already.
Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
