From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sudeep Holla Subject: Re: [PATCH v5 3/5] ACPI / processor: Introduce map_gic_id() to get apic id from MADT or _MAT method Date: Wed, 19 Feb 2014 14:33:40 +0000 Message-ID: <5304C0C4.3080601@arm.com> References: <1392740638-2479-1-git-send-email-hanjun.guo@linaro.org> <1392740638-2479-4-git-send-email-hanjun.guo@linaro.org> Mime-Version: 1.0 Content-Type: text/plain; charset=WINDOWS-1252 Content-Transfer-Encoding: 8BIT Return-path: Received: from service87.mimecast.com ([91.220.42.44]:38010 "EHLO service87.mimecast.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753885AbaBSOdQ convert rfc822-to-8bit (ORCPT ); Wed, 19 Feb 2014 09:33:16 -0500 In-Reply-To: <1392740638-2479-4-git-send-email-hanjun.guo@linaro.org> Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: Hanjun Guo , "Rafael J. Wysocki" Cc: Sudeep.Holla@arm.com, Lan Tianyu , "patches@linaro.org" , "linaro-acpi@lists.linaro.org" , "linux-kernel@vger.kernel.org" , "linux-acpi@vger.kernel.org" , Joe Perches , "linux-arm-kernel@lists.infradead.org" Hi Hanjun, On 18/02/14 16:23, Hanjun Guo wrote: > Get apic id from MADT or _MAT method is not implemented on arm/arm64, > and ACPI 5.0 introduces GIC Structure for it, so this patch introduces > map_gic_id() to get apic id followed the ACPI 5.0 spec. > > Signed-off-by: Hanjun Guo > --- > drivers/acpi/processor_core.c | 26 ++++++++++++++++++++++++++ > 1 file changed, 26 insertions(+) > > diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c > index 4dcf776..d316d9b 100644 > --- a/drivers/acpi/processor_core.c > +++ b/drivers/acpi/processor_core.c > @@ -71,6 +71,27 @@ static int map_lsapic_id(struct acpi_subtable_header *entry, > return 0; > } > > +static int map_gic_id(struct acpi_subtable_header *entry, > + int device_declaration, u32 acpi_id, int *apic_id) > +{ > + struct acpi_madt_generic_interrupt *gic = > + (struct acpi_madt_generic_interrupt *)entry; > + > + if (!(gic->flags & ACPI_MADT_ENABLED)) > + return -ENODEV; > + > + /* In the GIC interrupt model, logical processors are > + * required to have a Processor Device object in the DSDT, > + * so we should check device_declaration here > + */ > + if (device_declaration && (gic->uid == acpi_id)) { > + *apic_id = gic->gic_id; I have mentioned this earlier, it's not clear yet to me how does this work ? It needs more clarity in the form of comment here at-least as the ACPIv5.0 is also not so clear or explicit on how to handle this. Here you are expecting gic->uid = acpi_id which is fine, while acpi_map_cpuid matches apic_id with cpu_physical_id(which must be MPIDR in ARM{32,64}). The latter imposes restriction that gic->gic_id has to be MPIDR. Does that mean we are imposing restriction on GIC ID to be MPIDR ? If so please document it here and please explain the reason behind that choice. I would expect _UID to be MPIDR rather than GIC ID but you may have some reasons for this choice. Regards, Sudeep From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sudeep.Holla@arm.com (Sudeep Holla) Date: Wed, 19 Feb 2014 14:33:40 +0000 Subject: [PATCH v5 3/5] ACPI / processor: Introduce map_gic_id() to get apic id from MADT or _MAT method In-Reply-To: <1392740638-2479-4-git-send-email-hanjun.guo@linaro.org> References: <1392740638-2479-1-git-send-email-hanjun.guo@linaro.org> <1392740638-2479-4-git-send-email-hanjun.guo@linaro.org> Message-ID: <5304C0C4.3080601@arm.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi Hanjun, On 18/02/14 16:23, Hanjun Guo wrote: > Get apic id from MADT or _MAT method is not implemented on arm/arm64, > and ACPI 5.0 introduces GIC Structure for it, so this patch introduces > map_gic_id() to get apic id followed the ACPI 5.0 spec. > > Signed-off-by: Hanjun Guo > --- > drivers/acpi/processor_core.c | 26 ++++++++++++++++++++++++++ > 1 file changed, 26 insertions(+) > > diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c > index 4dcf776..d316d9b 100644 > --- a/drivers/acpi/processor_core.c > +++ b/drivers/acpi/processor_core.c > @@ -71,6 +71,27 @@ static int map_lsapic_id(struct acpi_subtable_header *entry, > return 0; > } > > +static int map_gic_id(struct acpi_subtable_header *entry, > + int device_declaration, u32 acpi_id, int *apic_id) > +{ > + struct acpi_madt_generic_interrupt *gic = > + (struct acpi_madt_generic_interrupt *)entry; > + > + if (!(gic->flags & ACPI_MADT_ENABLED)) > + return -ENODEV; > + > + /* In the GIC interrupt model, logical processors are > + * required to have a Processor Device object in the DSDT, > + * so we should check device_declaration here > + */ > + if (device_declaration && (gic->uid == acpi_id)) { > + *apic_id = gic->gic_id; I have mentioned this earlier, it's not clear yet to me how does this work ? It needs more clarity in the form of comment here at-least as the ACPIv5.0 is also not so clear or explicit on how to handle this. Here you are expecting gic->uid = acpi_id which is fine, while acpi_map_cpuid matches apic_id with cpu_physical_id(which must be MPIDR in ARM{32,64}). The latter imposes restriction that gic->gic_id has to be MPIDR. Does that mean we are imposing restriction on GIC ID to be MPIDR ? If so please document it here and please explain the reason behind that choice. I would expect _UID to be MPIDR rather than GIC ID but you may have some reasons for this choice. Regards, Sudeep