fix 3.3-testing c/s 18434
PATCH 1/1 fix 3.3-testing c/s 18434
c/s 18434 in xen-3.3-testing has a hunk that changes gatherDom() to _gatherDom() in tools/python/xen/xend/image.py. There is no such method, thus simple "xm list" fails
jfehlig4: # xm li
Error: _gatherDom
Usage: xm list [options] [Domain, ...]
..
Trivial fix attached. BTW, I don't see this patch in xen-unstable. Why is this only in xen3.3-testing?
Thanks,
Jim
tools/python/xen/xend/image.py +1 -1
@@ -502,7 +502,7 @@ class ImageHandler: if fifo_fd >= 0: self._openSentinel(sentinel_path_fifo) os.close(fifo_fd) - self.pid = self.vm._gatherDom(('image/device-model-pid', int)) + self.pid = self.vm.gatherDom(('image/device-model-pid', int)) log.debug("%s device model rediscovered, pid %s sentinel fifo %s", name, self.pid, sentinel_path_fifo) self.sentinel_thread = thread.start_new_thread(self._sentinel_watch,())
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel
Quoted 11 lines · Jim Fehlig
c/s 18434 in xen-3.3-testing has a hunk that changes
gatherDom()to_gatherDom()intools/python/xen/xend/image.py. There is no such method, thus simple "xm list" failsjfehlig4: # xm li
Error:_gatherDom
Usage: xm list [options] [Domain, ...]
..Trivial fix attached. BTW, I don't see this patch in xen-unstable. Why is this only in xen3.3-testing?
It is in xen-unstable, as c/s 18564 (which is part of xen-3.3-testing:18434). The bug is a stupid typo, and I will apply your patch to both trees.
Thanks,
Keir
Quoted 12 lines · Keir Fraser
Quoted 9 lines · Jim Fehlig
c/s 18434 in xen-3.3-testing has a hunk that changes
gatherDom()to_gatherDom()intools/python/xen/xend/image.py. There is no such method, thus simple "xm list" failsjfehlig4: # xm li
Error:_gatherDom
Usage: xm list [options] [Domain, ...]
..
Quoted 6 lines
Quoted 3 lines
Trivial fix attached. BTW, I don't see this patch in xen-unstable. Why is this only in xen3.3-testing?
It is in xen-unstable, as c/s 18564 (which is part of xen-3.3-testing:18434).
Missed that, sorry.
I've also found that this c/s causes /local/domain/0/backend/{vkbd, vfb, console} paths to leak :(. Patch attached.
Thanks,
Jim
tools/python/xen/xend/server/DevController.py +1 -1
@@ -257,7 +257,7 @@ class DevController: if force: frontpath = self.frontendPath(dev) - backpath = self.readVm(devid, "backend") + backpath = xstransact.Read(frontpath, "backend") if backpath: xstransact.Remove(backpath) xstransact.Remove(frontpath)
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel
Quoted 4 lines · Jim Fehlig
Missed that, sorry.
I've also found that this c/s causes /local/domain/0/backend/{vkbd, vfb, console} paths to leak :(. Patch attached.
A major point of the patch was to stop reading the backend link from the frontend directory. Is the problem that /vm/... is getting deleted too early? Could we delay it?
-- Keir
Quoted 11 lines · Keir Fraser
Quoted 5 lines · Jim Fehlig
Missed that, sorry.
I've also found that this c/s causes /local/domain/0/backend/{vkbd, vfb, console} paths to leak :(. Patch attached.
A major point of the patch was to stop reading the backend link from the frontend directory.
Ah, right. I should have looked at the patch more closely :-/
Quoted 3 lines
Is the problem that /vm/... is getting deleted too
early? Could we delay it?
Turns out that we were just constructing the path incorrectly. Patch attached.
Thanks,
Jim
tools/python/xen/xend/server/DevController.py +1 -1
@@ -257,7 +257,7 @@ class DevController: if force: frontpath = self.frontendPath(dev) - backpath = self.readVm(devid, "backend") + backpath = self.readVm(dev, "backend") if backpath: xstransact.Remove(backpath) xstransact.Remove(frontpath)
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel