From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:60956) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TjSLa-00009G-FU for qemu-devel@nongnu.org; Fri, 14 Dec 2012 05:18:27 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1TjSLX-0004jx-T0 for qemu-devel@nongnu.org; Fri, 14 Dec 2012 05:18:26 -0500 Received: from mx1.redhat.com ([209.132.183.28]:53076) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TjSLX-0004jn-LI for qemu-devel@nongnu.org; Fri, 14 Dec 2012 05:18:23 -0500 Date: Fri, 14 Dec 2012 11:18:18 +0100 From: Igor Mammedov Message-ID: <20121214111818.25ad95d2@thinkpad.mammed.net> In-Reply-To: <1355336183-18858-3-git-send-email-ehabkost@redhat.com> References: <1355336183-18858-1-git-send-email-ehabkost@redhat.com> <1355336183-18858-3-git-send-email-ehabkost@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH 2/2] target-i386: make cpu_x86_create() get Error argument List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eduardo Habkost Cc: qemu-devel@nongnu.org, Andreas =?UTF-8?B?RsOkcmJlcg==?= On Wed, 12 Dec 2012 16:16:23 -0200 Eduardo Habkost wrote: > Instead of forcing the caller to guess what went wrong while creating > the CPU object, return error information in a Error argument. > > Also, as cpu_x86_create() won't print error messages itself anymore, > change cpu_x86_init() to print any error returned by cpu_x86_create() > or cpu_x86_realize(). > > Signed-off-by: Eduardo Habkost > --- > target-i386/cpu.c | 8 ++++++-- > target-i386/cpu.h | 2 +- > target-i386/helper.c | 21 ++++++++++++++------- > 3 files changed, 21 insertions(+), 10 deletions(-) > > diff --git a/target-i386/cpu.c b/target-i386/cpu.c > index b242bf1..fba872d 100644 > --- a/target-i386/cpu.c > +++ b/target-i386/cpu.c > @@ -1542,7 +1542,7 @@ static void filter_features_for_kvm(X86CPU *cpu) > /* Create and initialize a X86CPU object, based on the full CPU model string > * (that may include "+feature,-feature,feature=xxx" feature strings) > */ > -X86CPU *cpu_x86_create(const char *cpu_model) > +X86CPU *cpu_x86_create(const char *cpu_model, Error **errp) > { > X86CPU *cpu; > CPUX86State *env; > @@ -1559,12 +1559,14 @@ X86CPU *cpu_x86_create(const char *cpu_model) > > model_pieces = g_strsplit(cpu_model, ",", 2); > if (!model_pieces[0]) { > + error_setg(errp, "invalid CPU model string: %s", cpu_model); > goto error; > } > name = model_pieces[0]; > features = model_pieces[1]; > > if (cpu_x86_find_by_name(def, name) < 0) { > + error_setg(errp, "CPU model not found: %s", name); > goto error; > } > > @@ -1575,13 +1577,15 @@ X86CPU *cpu_x86_create(const char *cpu_model) > &def->svm_features, &def->cpuid_7_0_ebx_features); > > if (cpu_x86_parse_featurestr(def, features) < 0) { > + error_setg(errp, "Error parsing feature string: %s", > + features ? features : "(none)"); It could be simplified, it shouldn't get here if features == NULL > goto error; > } > > cpudef_2_x86_cpu(cpu, def, &error); > > if (error) { > - fprintf(stderr, "%s\n", error_get_pretty(error)); > + error_propagate(errp, error); Why do it here but not above? > error_free(error); > goto error; > } > diff --git a/target-i386/cpu.h b/target-i386/cpu.h [...] > > X86CPU *cpu_x86_init(const char *cpu_model) > { > - X86CPU *cpu; > + X86CPU *cpu = NULL; > Error *error = NULL; > > - cpu = cpu_x86_create(cpu_model); > - if (!cpu) { > - return NULL; > + cpu = cpu_x86_create(cpu_model, &error); > + if (error) { > + goto error; > } > > x86_cpu_realize(OBJECT(cpu), &error); if x86_cpu_realize() behave as visit* functions, i.e. return early if error has been already set, error check & goto could be removed here and above and consolidated at function exit. > if (error) { > - error_free(error); > - object_delete(OBJECT(cpu)); > - return NULL; > + goto error; > } > return cpu; > + > +error: > + if (cpu) { > + object_delete(OBJECT(cpu)); > + } > + error_report("%s", error_get_pretty(error)); > + error_free(error); > + return NULL; > } > > #if !defined(CONFIG_USER_ONLY) > -- > 1.7.11.7 > -- Regards, Igor