[systemd-bugs] [Bug 77005] New: bus-driverd racing external requests
bugzilla-daemon at freedesktop.org
bugzilla-daemon at freedesktop.org
Thu Apr 3 06:51:17 PDT 2014
https://bugs.freedesktop.org/show_bug.cgi?id=77005
Priority: medium
Bug ID: 77005
Assignee: systemd-bugs at lists.freedesktop.org
Summary: bus-driverd racing external requests
QA Contact: systemd-bugs at lists.freedesktop.org
Severity: normal
Classification: Unclassified
OS: All
Reporter: dh.herrmann at gmail.com
Hardware: Other
Status: NEW
Version: unspecified
Component: general
Product: systemd
Ryan just made me aware of this. The following sequence works reliably with
dbus1, but not with kdbus + bus-driverd:
ADD_MATCH to=org.freedesktop.DBus
METHOD_CALL to=someone.else
If both calls are async, with dbus1, the ADD_MATCH is guaranteed to be finished
before METHOD_CALL is forwarded to someone.else. With bus-driverd, there's no
such guarantee.
I'm not really sure how to fix it. Some ideas:
* make any org.freedesktop.DBus call synchronous in the kernel (ugh, ugly!)
* support org.freedesktop.DBus only when called through bus-proxyd and make it
synchronous inside of bus-proxyd (hm, still ugly?)
* make sd-bus inspect any outgoing messages and make them synchronous if
they're targeted as org.freedesktop.DBus
* document it, but it ignore it and make people wait for ADD_MATCH to actually
finish
With "synchronous" I obviously mean just delaying any other outgoing message
until we got the response. That "waiting" can still be async.. (like HELO)
I'm actually not sure whether _any_ of the bus-driverd calls is still needed
with kdbus, but if it is, we probably need to review the other ones, too.
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/systemd-bugs/attachments/20140403/867d633d/attachment.html>
More information about the systemd-bugs
mailing list