From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:47023) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YDE8g-0006uX-WB for qemu-devel@nongnu.org; Mon, 19 Jan 2015 10:21:15 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1YDE8c-0006jX-0x for qemu-devel@nongnu.org; Mon, 19 Jan 2015 10:21:14 -0500 Received: from mx1.redhat.com ([209.132.183.28]:50324) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YDE8b-0006jL-Oh for qemu-devel@nongnu.org; Mon, 19 Jan 2015 10:21:09 -0500 From: Markus Armbruster References: <1414481739-19939-1-git-send-email-armbru@redhat.com> <545244CA.5010809@suse.de> <87egtk3f5i.fsf@blackfin.pond.sub.org> Date: Mon, 19 Jan 2015 16:21:00 +0100 In-Reply-To: <87egtk3f5i.fsf@blackfin.pond.sub.org> (Markus Armbruster's message of "Mon, 03 Nov 2014 08:40:57 +0100") Message-ID: <87d26akd9f.fsf@blackfin.pond.sub.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH RFC 00/10] pci: Partial conversion to realize List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Andreas =?utf-8?Q?F=C3=A4rber?= Cc: pbonzini@redhat.com, arei.gonglei@huawei.com, mst@redhat.com, qemu-devel@nongnu.org, kraxel@redhat.com Markus Armbruster writes: > Andreas F=C3=A4rber writes: > >> Hi Markus, >> >> Am 28.10.2014 um 08:35 schrieb Markus Armbruster: >>> While discussing Gonglei's "[PATCH v2 00/19] usb: convert device init >>> to realize", Paolo called the PCI conversion job "Gargantuan". This >>> series attempts to crack it into manageable jobs. >> >> Thanks for giving this a stab! What kept me from diving into the PCI >> converstion was that I first invested into qtests for the non-default >> PCI devices. How many of the converted devices are actually covered in >> qtest? > > Can't tell offhand, but I can find out. The vast majority of my conversions are utterly trivial [PATCH 03]: change return type to void, rename, put into ->realize instead of ->init. I could enumerate the devices so changed and look for qtests, but I feel it's rather pointless busywork. But if you should insist... The remaining conversions are all very, very simple: * PATCH 05: "pcnet" Utterly trivial after trivial PATCH 04 changed a helper's return type to void. There's pcnet-test.c, which basically tests "-device pcnet doesn't explode right away". * PATCH 06: "pci-serial", "pci-serial-2x", "pci-serial-4x" Two error paths trivially converted from qerror_report_err() to error_propagate(). Not covered in qtest as far as I can tell. * PATCH 07: "ich9-ahci" One pci_add_capability() replaced by pci_add_capability2(). The former is a wrapper around the latter which additionally passes any error to error_report(), then frees it. Straightforward conversion of the error path to error_setg(). Not covered in qtest as far as I can tell. * PATCH 08: "cirrus-vga" One error path trivially converted from error_report() to error_setg(). There's display-vga-test.c, which tests "-device cirrus-vga doesn't explode right away". * PATCH 09: "qxl-vga", "qxl" Two error paths trivially converted from error_report() to error_setg(). Not covered in qtest as far as I can tell. * PATCH 10: "kvm-pci-assign" Two error paths trivially converted from qerror_report_err() to error_propagate(). Not covered in qtest as far as I can tell. I'm prepared to drop conversions you consider risky without test coverage (although I have a hard time seeing risks, considering how silly-simple my conversions are). But I'd really like to get the core changes plus some examples in.