Now ovs-vswitchd fills in the Interface ofport and Bridge datapath_id
fields when it reconfigures.
The existing Bridge datapath_id and hwaddr columns, which had surprising
meanings, have been banished to a new other_config column.
make_unix_socket() can return EAGAIN in rare circumstances, e.g. when the
server's socket listen queue is full. A lot of OVS callers interpret
EAGAIN as a "try again" error code, but in this case it means that the
attempt to create the socket failed. So munge EAGAIN into another error
code to prevent that misinterpretation.
If the transaction modified a row and then deleted it, the IDL would
instead mistakenly leave the row entirely untouched.
This commit fixes this bug. It needs a regression test, but this commit
does not add one.
When the IDL was used to insert a row, but all of the new row's columns
were left at the default values, then the IDL would not insert the row at
all.
When the IDL was used to delete one or more rows, and the transaction did
not include any update or insertion operations, the transaction was dropped
entirely.
This commit fixes these two bugs. It needs a regression test, but this
commit does not add one.
Some programs require a "Open_vSwitch" table to exist in the
configuration DB. This change adds that table when a new DB is created.
Also, don't try to create the DB unless it doesn't exist.
With ovs-vswitchd using the config DB, it is no longer necessary to tell
it to reload its configuration file. This removes references to the
need for reloading.
It also cleans up some messages placed on the console during boot up.
This change makes xsconsole able to reconfigure the management interface
on some XenServer hosts. It's not clear why it is needed--apparently
xapi is creating the bridge behind our back.
Tested by starting and stopping a VM that has a single VIF and verifying
that the VM could get an IP address via DHCP and access the network and
that the VIF's external-ids were set to plausible values.
Now a XenServer can boot OK with the database at least in simple cases.
interface-reconfigure is slow because there is a 5-second sleep in place of
a proper way to wait for vswitchd reconfiguration to complete.
This is an extremely lightly tested attempt at switching ovs-brcompatd
from using the config file to the new config db. There are a lot of
shortcomings in this cut, but we need to make progress on the XenServer
integration, so it's going in now. Expect changes in the near future.
This passes at least one test (the one named "add-br a"). It probably
doesn't pass any more than that.
This is *way* not up to my quality standards, but we are in a super hurry
so I'm pushing it anyhow.
Until now, the "set" functions generated by the IDL updated the data in the
database (during commit) but not the data exposed by the IDL in its data
structures. This was just an oversight, so this commit causes the data
exposed by IDL to be updated also.
bitmap_scan() can be optimized significantly for the case of a sparse
bitmap but it doesn't seem worth the additional overhead of writing a test
unless and until we show that it's a useful optimization in practice.
This is useful in cases where one might want to know whether an hmap_node
is actually part of an hmap, without using a separate variable to indicate
it.
