From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:35344) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1a6kYJ-0005Ey-KV for qemu-devel@nongnu.org; Wed, 09 Dec 2015 14:37:28 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1a6kYG-0004w4-Ap for qemu-devel@nongnu.org; Wed, 09 Dec 2015 14:37:27 -0500 Received: from mx1.redhat.com ([209.132.183.28]:44045) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1a6kYG-0004vE-5j for qemu-devel@nongnu.org; Wed, 09 Dec 2015 14:37:24 -0500 Date: Wed, 9 Dec 2015 20:37:16 +0100 From: Igor Mammedov Message-ID: <20151209203716.478788f0@igors-macbook-pro.local> In-Reply-To: <56672516.3020809@gmail.com> References: <1449020831-8414-1-git-send-email-ehabkost@redhat.com> <1449020831-8414-7-git-send-email-ehabkost@redhat.com> <5665A831.2060504@gmail.com> <20151208175927.GN14490@thinpad.lan.raisama.net> <56672516.3020809@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH 06/16] acpi: Save PCMachineState on AcpiBuildState List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Marcel Apfelbaum Cc: marcel@redhat.com, qemu-devel@nongnu.org, Marcel Apfelbaum , Eduardo Habkost , "Michael S. Tsirkin" On Tue, 8 Dec 2015 20:44:38 +0200 Marcel Apfelbaum wrote: > On 12/08/2015 07:59 PM, Eduardo Habkost wrote: > > On Mon, Dec 07, 2015 at 05:39:29PM +0200, Marcel Apfelbaum wrote: > >> On 12/02/2015 03:47 AM, Eduardo Habkost wrote: > >>> PCMachineState will be used in some of the steps of ACPI table > >>> building. > >>> > >>> Signed-off-by: Eduardo Habkost > >>> --- > >>> hw/i386/acpi-build.c | 8 ++++---- > >>> 1 file changed, 4 insertions(+), 4 deletions(-) > >>> > >>> diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c > >>> index 85a5c53..ca11c88 100644 > >>> --- a/hw/i386/acpi-build.c > >>> +++ b/hw/i386/acpi-build.c > >>> @@ -1644,7 +1644,7 @@ struct AcpiBuildState { > >>> MemoryRegion *table_mr; > >>> /* Is table patched? */ > >>> uint8_t patched; > >>> - PcGuestInfo *guest_info; > >>> + PCMachineState *pcms; > >>> void *rsdp; > >>> MemoryRegion *rsdp_mr; > >>> MemoryRegion *linker_mr; > >>> @@ -1855,7 +1855,7 @@ static void acpi_build_update(void > >>> *build_opaque, uint32_t offset) > >>> > >>> acpi_build_tables_init(&tables); > >>> > >>> - acpi_build(build_state->guest_info, &tables); > >>> + acpi_build(&build_state->pcms->acpi_guest_info, &tables); > >>> > >>> acpi_ram_update(build_state->table_mr, tables.table_data); > >>> > >>> @@ -1916,12 +1916,12 @@ void acpi_setup(PCMachineState *pcms) > >>> > >>> build_state = g_malloc0(sizeof *build_state); > >>> > >>> - build_state->guest_info = guest_info; > >>> + build_state->pcms = pcms; > >> > >> I am not "sold" on keeping a reference to machine in the > >> build_state. We can always query current machine using > >> qdev_machine() or something. > >> > >> Keeping the "guest info" made sense since is used especially for > >> ACPI, however the machine has a wider scope. (And not having to > >> keep it around is a very good thing!) > > > > I wouldn't mind using qdev_get_machine() if preferred by the > > maintainer of that code, but I like to avoid it when possible. To > > me, qdev_get_machine() is just a global variable disguised as a > > harder-to-understand API. > > Really? Hmm, for me is looking like the other way around :) > I see it as "query the QOM tree", instead of "keep the reference > around" everywhere. But it may be just a personal preference. +1, It's not performance critical path so I'd prefer qdev_get_machine() instead of keeping extra reference/state. > > Thanks, > Marcel > > > > > >