From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:33186) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VBQNY-0005NL-V5 for qemu-devel@nongnu.org; Mon, 19 Aug 2013 10:24:27 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VBQNT-00030P-Fq for qemu-devel@nongnu.org; Mon, 19 Aug 2013 10:24:20 -0400 Message-ID: <52122A86.1050605@suse.de> Date: Mon, 19 Aug 2013 16:24:06 +0200 From: =?UTF-8?B?QW5kcmVhcyBGw6RyYmVy?= MIME-Version: 1.0 References: <1375447708-14545-1-git-send-email-chouteau@adacore.com> <1375447708-14545-2-git-send-email-chouteau@adacore.com> <5211D6C3.2000609@adacore.com> In-Reply-To: Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH 2/2] Add ARM registers definitions in Monitor commands List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Peter Maydell Cc: qemu-trivial@nongnu.org, qemu-devel@nongnu.org, Fabien Chouteau , lcapitulino@redhat.com Am 19.08.2013 10:31, schrieb Peter Maydell: > On 19 August 2013 09:26, Fabien Chouteau wrote: >> Any comments? >> >> Regards, >> >> On 08/02/2013 02:48 PM, Fabien Chouteau wrote: >>> Signed-off-by: Fabien Chouteau >>> --- >>> monitor.c | 17 +++++++++++++++++ >>> 1 file changed, 17 insertions(+) >>> >>> diff --git a/monitor.c b/monitor.c >>> index 5dc0aa9..78e93af 100644 >>> --- a/monitor.c >>> +++ b/monitor.c >>> @@ -3167,6 +3167,23 @@ static const MonitorDef monitor_defs[] =3D { >>> { "cleanwin", offsetof(CPUSPARCState, cleanwin) }, >>> { "fprs", offsetof(CPUSPARCState, fprs) }, >>> #endif >>> +#elif defined(TARGET_ARM) >>> + { "r0", offsetof(CPUARMState, regs[0]) }, >>> + { "r1", offsetof(CPUARMState, regs[1]) }, >=20 > Rather than adding yet another entry to this target-ifdef > ladder in common code, maybe we can abstract this out to be > a method/field of the CPU object? >=20 > Andreas can probably suggest the best approach. I had similar thoughts when I first saw this patch... I haven't looked closely into monitor yet though (other than seeing it has too much target dependencies). Either a MonitorDef* field (null-terminating the array then) or, depending on where/how this is used, a method in CPUClass might be options to better abstract this. Given that there is going to be a little bit of time before 1.7 now we could take some time to rethink our design. A related question that remained unanswered for the Program Counter in the gdbstub context was whether we want QOM properties for such registers, which would hint at further abstracting such a register list to reuse it outside monitor. Andreas --=20 SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 N=C3=BCrnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imend=C3=B6rffer; HRB 16746 AG N=C3=BC= rnberg