From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751426AbbCQC7Z (ORCPT ); Mon, 16 Mar 2015 22:59:25 -0400 Received: from v094114.home.net.pl ([79.96.170.134]:58563 "HELO v094114.home.net.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1751069AbbCQC7T convert rfc822-to-8bit (ORCPT ); Mon, 16 Mar 2015 22:59:19 -0400 From: "Rafael J. Wysocki" To: Hanjun Guo Cc: Hanjun Guo , Catalin Marinas , Will Deacon , Olof Johansson , Grant Likely , Lorenzo Pieralisi , Arnd Bergmann , Mark Rutland , Graeme Gregory , Sudeep Holla , Jon Masters , Marc Zyngier , Mark Brown , Robert Richter , Timur Tabi , Ashwin Chaugule , suravee.suthikulpanit@amd.com, linux-acpi@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, linaro-acpi@lists.linaro.org, Tomasz Nowicki , Zhangdianfang Subject: Re: [update][PATCH v10 06/21] ACPI / sleep: Introduce CONFIG_ACPI_GENERIC_SLEEP Date: Tue, 17 Mar 2015 04:23:11 +0100 Message-ID: <10995588.scpoK3KRg2@vostro.rjw.lan> User-Agent: KMail/4.11.5 (Linux/3.19.0+; KDE/4.11.5; x86_64; ; ) In-Reply-To: <5507933F.2020203@huawei.com> References: <1426234469-6434-1-git-send-email-hanjun.guo@linaro.org> <3737028.Ib0fKlgSkd@vostro.rjw.lan> <5507933F.2020203@huawei.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8BIT Content-Type: text/plain; charset="utf-8" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tuesday, March 17, 2015 10:36:47 AM Hanjun Guo wrote: > On 2015/3/17 10:28, Rafael J. Wysocki wrote: > > On Tuesday, March 17, 2015 09:08:45 AM Hanjun Guo wrote: > >> On 2015/3/17 7:15, Rafael J. Wysocki wrote: > >>> On Monday, March 16, 2015 08:14:52 PM Hanjun Guo wrote: > >>>> On 2015年03月14日 05:49, Rafael J. Wysocki wrote: > >>>>> On Friday, March 13, 2015 04:14:29 PM Hanjun Guo wrote: > >> [...] > >>>>>> diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig > >>>>>> index 074e52b..e8728d7 100644 > >>>>>> --- a/arch/ia64/Kconfig > >>>>>> +++ b/arch/ia64/Kconfig > >>>>>> @@ -10,6 +10,7 @@ config IA64 > >>>>>> select ARCH_MIGHT_HAVE_PC_SERIO > >>>>>> select PCI if (!IA64_HP_SIM) > >>>>>> select ACPI if (!IA64_HP_SIM) > >>>>>> + select ACPI_GENERIC_SLEEP if ACPI > >>>>>> select ARCH_MIGHT_HAVE_ACPI_PDC if ACPI > >>>>>> select HAVE_UNSTABLE_SCHED_CLOCK > >>>>>> select HAVE_IDE > >>>>>> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig > >>>>>> index b7d31ca..9804431 100644 > >>>>>> --- a/arch/x86/Kconfig > >>>>>> +++ b/arch/x86/Kconfig > >>>>>> @@ -22,6 +22,7 @@ config X86_64 > >>>>>> ### Arch settings > >>>>>> config X86 > >>>>>> def_bool y > >>>>>> + select ACPI_GENERIC_SLEEP if ACPI > >>>>> One more nit. If you did > >>>>> > >>>>> + select ACPI_GENERIC_SLEEP if ACPI_SLEEP > >>>>> > >>>>> here (and above for ia64), you'd avoid having to make ACPI_SLEEP > >>>>> depend on ACPI_GENERIC_SLEEP which goes somewhat backwards. > >>>> In sleep.c, > >>>> > >>>> #ifdef CONFIG_ACPI_SLEEP > >>>> acpi_target_system_state() > >>>> { > >>>> } > >>>> #endif > >>>> > >>>> and CONFIG_ACPI_SLEEP depends on SUSPEND || HIBERNATION, > >>>> which one of them will be enabled on ARM64 so ACPI_SLEEP > >>>> will also enabled too. > >>>> > >>>> So if we > >>>> > >>>> +select ACPI_GENERIC_SLEEP if ACPI_SLEEP > >>>> > >>>> and > >>>> > >>>> +acpi-$(CONFIG_ACPI_GENERIC_SLEEP) += sleep.o > >>>> > >>>> it will lead to errors for acpi_target_system_state() that > >>>> is declared but not defined, so I will keep the code as > >>>> it is, what do you think? > >>> No, we need to hash this out. Having two different Kconfig options meaning > >>> almost the same thing (ACPI_SLEEP and ACPI_GENERIC_SLEEP) is beyond ugly. > >>> > >>> Do you need ACPI_SLEEP on ARM64 at all? > >> No, at least for now we don't need it, the spec for sleep is not ready for > >> ARM64 arch, so ACPI_SLEEP will not work at all on ARM64. > > Well, so what about selecting ACPI_SLEEP from the architectures that use it? > > Do you mean remove CONFIG_ACPI_GENERIC_SLEEP and > > +acpi-$(CONFIG_ACPI_SLEEP) += sleep.o > > as well (also need to remove duplicate #ifdef CONFIG_ACPI_SLEEP in sleep.c if > we doing so)? Well, almost. There is one problem with that, becuase sleep.c contains code outside of the ACPI_SLEEP-dependent blocks. That code is used for powering off ACPI platforms. I guess you don't want that code on ARM too, right? Perhaps we can use ACPI_REDUCED_HARDWARE_ONLY for that? ARM64 will be the only arch setting it at least for the time being, is that correct? -- I speak only for myself. Rafael J. Wysocki, Intel Open Source Technology Center.