From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1MduGf-0001Wj-VE for qemu-devel@nongnu.org; Wed, 19 Aug 2009 19:08:33 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1MduGa-0001Ss-BC for qemu-devel@nongnu.org; Wed, 19 Aug 2009 19:08:32 -0400 Received: from [199.232.76.173] (port=49877 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1MduGa-0001Sn-9N for qemu-devel@nongnu.org; Wed, 19 Aug 2009 19:08:28 -0400 Received: from mx1.redhat.com ([209.132.183.28]:38953) by monty-python.gnu.org with esmtps (TLS-1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.60) (envelope-from ) id 1MduGZ-0005R3-QE for qemu-devel@nongnu.org; Wed, 19 Aug 2009 19:08:28 -0400 From: Luiz Capitulino Date: Wed, 19 Aug 2009 20:07:31 -0300 Message-Id: <1250723280-3509-1-git-send-email-lcapitulino@redhat.com> Subject: [Qemu-devel] [PATCH v1 00/29] QMonitor List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: aliguori@us.ibm.com, avi@redhat.com Hi there, This is a new version of my QMonitor series, it has some simplifications and changes suggested by Avi. Changelog and full introduction follow. changelog --------- V0 -> V1 - Make QType structs const - New macros: QOBJECT_INIT(), QDECREF()/QINCREF(), QType_HEAD - qobject_to_*() should return NULL if type doesn't match introduction ------------ Basically, this series introduces high-level data types (QInt, QString, QDict) and port all Monitor command handlers to use them to receive arguments. The following points should be observed: o Object life cycle management: I'm using refcouting and have adopted Python's approach, this means that you only do refcount when needed, more info here: http://docs.python.org/c-api/intro.html#objects-types-and-reference-counts This approach makes command handlers code simpler. o QString's get: QString exports a function called qstring_get_str(), this function returns a *pointer* to the stored string. A better API would return a copy instead, but it would be too much work and error-prone to go over all handlers and put a qemu_free() in the right place. Handlers only want quick and read-only access to strings anyway, so returning a pointer makes handlers' code a lot simpler. o unit-testing: I have written unit-tests for all the new code and have an off-tree suite for the Monitor's parsing code. I was in doubt if I should submit this work... I did, but it's in the end of the series, if people don't like this it can be easily dropped. Monitor's suite is not in the series because I couldn't make it build "in tree". While reviewing the series is important to bear in mind that all my design decisions were based on the need of the current and most important users of the API: monitor command handlers. Thanks for reading this all. :) Makefile | 8 + check-qdict.c | 347 ++++ check-qint.c | 124 ++ check-qstring.c | 100 + configure | 32 + console.h | 3 +- hw/pci-hotplug.c | 15 +- migration.c | 12 +- migration.h | 9 +- monitor.c | 463 +++--- monitor.h | 1 + net.c | 25 +- net.h | 13 +- qdict-test-data.txt | 4999 +++++++++++++++++++++++++++++++++++++++++++++++++++ qdict.c | 311 ++++ qdict.h | 40 + qemu-monitor.hx | 109 +- qint.c | 92 + qint.h | 20 + qobject.h | 113 ++ qstring.c | 73 + qstring.h | 16 + savevm.c | 13 +- sysemu.h | 18 +- vl.c | 13 +- 25 files changed, 6643 insertions(+), 326 deletions(-)