Re: [libvirt] [PATCH v4 4/7] fdstream: Emit stream abort callback even if poll() doesnt.
On 02/06/2012 06:50 AM, Peter Krempa wrote:
This patch causes the fdstream driver to call the stream event
callback
if virStreamAbort() is issued on a stream using this driver. This
prohibited to abort streams from the daemon, as the daemon remote
handler installs a callback to watch for stream errors as the only mean
of detecting changes in the stream.
That sentence didn't parse well for me; are you trying to say:
A remote handler for a stream can only detect changes via stream events,
so this event callback is necessary in order to enable a daemon to abort
a stream in such a way that the client will see the change.
* src/fdstream.c:
- modify close function to call stream event callback
---
src/fdstream.c | 56 ++++++++++++++++++++++++++++++++++++++++++++++++++------
1 files changed, 50 insertions(+), 6 deletions(-)
diff --git a/src/fdstream.c b/src/fdstream.c
index 841f979..35f5135 100644
--- a/src/fdstream.c
+++ b/src/fdstream.c
@@ -1,5 +1,5 @@
/*
- * fdstream.h: generic streams impl for file descriptors
+ * fdstream.c: generic streams impl for file descriptors
Cute typo fix...
*
* Copyright (C) 2009-2011 Red Hat, Inc.
but it made me notice this. It's now 2012. If you use emacs, you can
add this to your ~/.emacs to auto-update copyright in any file you
touch; here's what I use:
(require 'copyright)
(defun my-copyright-update (&optional arg)
"My improvements to `copyright-update'."
(interactive "*P")
(and (not (eq major-mode 'fundamental-mode))
(copyright-update arg))
nil)
(add-hook 'before-save-hook 'my-copyright-update)
[hmm - should we follow the lead of GNU programs that just globally
update the copyright year on all files when a new year rolls around? but
that's a question to ask Red Hat legal]
@@ -120,6 +126,7 @@ static int virFDStreamUpdateCallback(virStreamPtr
stream, int events)
}
virEventUpdateHandle(fdst->watch, events);
+ fdst->events = events;
On update, you leave fdst->abortCallbackCalled unchanged,
ret = 0;
@@ -214,6 +221,8 @@ virFDStreamAddCallback(virStreamPtr st,
fdst->cb = cb;
fdst->opaque = opaque;
fdst->ff = ff;
+ fdst->events = events;
+ fdst->abortCallbackCalled = false;
but on Add, you always clear it. Any significance to this difference?
+ /* aborting the stream, ensure the callback is called if it's
+ * registered for stream error event */
+ if (streamAbort &&
+ fdst->cb &&
+ (fdst->events & (VIR_STREAM_EVENT_READABLE |
+ VIR_STREAM_EVENT_WRITABLE))) {
+ /* don't enter this function accidentaly from the callback again */
s/accidentaly/accidentally/
+ if (fdst->abortCallbackCalled) {
+ virMutexUnlock(&fdst->lock);
+ return 0;
+ }
+
+ fdst->abortCallbackCalled = true;
+ fdst->abortCallbackDispatching = true;
+ virMutexUnlock(&fdst->lock);
+
+ /* call failure callback, poll does report nothing on closed fd */
s/does report/reports/
+ (fdst->cb)(st, VIR_STREAM_EVENT_ERROR, fdst->opaque);
You need to cache fdst->cb and fdst->opaque before dropping locks, since
otherwise you could have a race with someone else updating the callback
to a different value.
I'd still feel more comfortable with a review from Dan, but since that
seems to be long in coming, you have my ACK if you can fix the problems
pointed out above.
--
Eric Blake eblake(a)redhat.com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
Attachments:
- signature.asc (application/pgp-signature — 620 bytes)