From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:48786) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1apyZQ-0000bo-MA for qemu-devel@nongnu.org; Tue, 12 Apr 2016 09:41:33 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1apyZP-00019v-R6 for qemu-devel@nongnu.org; Tue, 12 Apr 2016 09:41:32 -0400 From: "Richard W.M. Jones" Date: Tue, 12 Apr 2016 14:41:12 +0100 Message-Id: <1460468474-29812-1-git-send-email-rjones@redhat.com> Subject: [Qemu-devel] [PATCH v2 REPOST 0/2] Add dynamic module loading for block drivers List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: kwolf@redhat.com Cc: qemu-devel@nongnu.org, qemu-block@nongnu.org, famz@redhat.com, den-lists@parallels.com This is a repost of the support for dynamically loaded block drivers. It is identical to how it was posted last summer, except that I have rebased it and checked that it still works. It was last posted here: https://lists.gnu.org/archive/html/qemu-devel/2015-09/threads.html#01995 Last time this was reviewed there were two strands of comments/ objections: (1) Fam Zheng objected to parsing C structs using the Python generator code and wanted special macros to be used instead: https://lists.gnu.org/archive/html/qemu-devel/2015-09/msg02201.html My objection to that is that it means the same information is defined in two places, with the usual opportunities for the information to get out of synch, which could cause crashes or modules not to be loaded. (2) Denis Lunev wanted module loading to work more like Linux modules: https://lists.gnu.org/archive/html/qemu-devel/2015-09/msg05331.html That makes the changes much larger. Rich.