Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756231AbZDRNYU (ORCPT ); Sat, 18 Apr 2009 09:24:20 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754511AbZDRNYK (ORCPT ); Sat, 18 Apr 2009 09:24:10 -0400 Received: from ogre.sisk.pl ([217.79.144.158]:44774 "EHLO ogre.sisk.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754599AbZDRNYI (ORCPT ); Sat, 18 Apr 2009 09:24:08 -0400 From: "Rafael J. Wysocki" To: Russell King , Len Brown Subject: Re: 900af0d breaks some embedded suspend/resume Date: Sat, 18 Apr 2009 15:23:35 +0200 User-Agent: KMail/1.11.2 (Linux/2.6.30-rc2-rjw; KDE/4.2.2; x86_64; ; ) Cc: Linux Kernel List , Linus Torvalds , pm list , ACPI Devel Maling List References: <20090417231009.GB6900@flint.arm.linux.org.uk> In-Reply-To: <20090417231009.GB6900@flint.arm.linux.org.uk> MIME-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200904181523.36010.rjw@sisk.pl> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 6163 Lines: 223 On Saturday 18 April 2009, Russell King wrote: > Some platforms need to talk via I2C to power control devices during > the suspend method. Currently, they do this via the platform PM ops > prepare callback, relying on the I2C driver being hooked into the > 'late' suspend method, and hence being shut down _after_ the prepare > callback. Well, I was not aware of this dependency and it seems quite unusual for a platform to depend on a driver like this. > However, as of the above commit, the ordering is changed such that > platforms don't get notified of suspends until after all devices are > well and truely shut down. > > This can't work, and actively breaks some platforms. Sorry for that. The patchset in question had been discussed quite extensively before it was merged and it's a pity none of the people caring for the affected platforms took part in those discussions. That would allow us to avoid the breakage. > Please come up with another solution for your PCI problems, I don't think this is possible, sorry. > or provide alternative equivalent functionality where the platform code is > notified of the PM event prior to the late suspend callback being issued. There is the .begin() callback that could be used, but if you need the platform to be notified _just_ prior to the late suspend callback, then the only thing I can think of at the moment is the appended patch. It shouldn't break anything in theory, because the majority of drivers put their devices into low power states in the "regular" suspend callbacks anyway. Len, what do you think? Thanks, Rafael --- kernel/power/disk.c | 43 +++++++++++++++++++++++-------------------- kernel/power/main.c | 19 +++++++++---------- 2 files changed, 32 insertions(+), 30 deletions(-) Index: linux-2.6/kernel/power/main.c =================================================================== --- linux-2.6.orig/kernel/power/main.c +++ linux-2.6/kernel/power/main.c @@ -291,21 +291,21 @@ static int suspend_enter(suspend_state_t device_pm_lock(); - error = device_power_down(PMSG_SUSPEND); - if (error) { - printk(KERN_ERR "PM: Some devices failed to power down\n"); - goto Done; - } - if (suspend_ops->prepare) { error = suspend_ops->prepare(); if (error) - goto Power_up_devices; + goto Done; } if (suspend_test(TEST_PLATFORM)) goto Platfrom_finish; + error = device_power_down(PMSG_SUSPEND); + if (error) { + printk(KERN_ERR "PM: Some devices failed to power down\n"); + goto Platfrom_finish; + } + error = disable_nonboot_cpus(); if (error || suspend_test(TEST_CPUS)) goto Enable_cpus; @@ -326,13 +326,12 @@ static int suspend_enter(suspend_state_t Enable_cpus: enable_nonboot_cpus(); + device_power_up(PMSG_RESUME); + Platfrom_finish: if (suspend_ops->finish) suspend_ops->finish(); - Power_up_devices: - device_power_up(PMSG_RESUME); - Done: device_pm_unlock(); Index: linux-2.6/kernel/power/disk.c =================================================================== --- linux-2.6.orig/kernel/power/disk.c +++ linux-2.6/kernel/power/disk.c @@ -217,6 +217,13 @@ static int create_image(int platform_mod device_pm_lock(); + error = platform_pre_snapshot(platform_mode); + if (error) + goto Unlock; + + if (hibernation_test(TEST_PLATFORM)) + goto Platform_finish; + /* At this point, device_suspend() has been called, but *not* * device_power_down(). We *must* call device_power_down() now. * Otherwise, drivers for some devices (e.g. interrupt controllers) @@ -227,12 +234,8 @@ static int create_image(int platform_mod if (error) { printk(KERN_ERR "PM: Some devices failed to power down, " "aborting hibernation\n"); - goto Unlock; - } - - error = platform_pre_snapshot(platform_mode); - if (error || hibernation_test(TEST_PLATFORM)) goto Platform_finish; + } error = disable_nonboot_cpus(); if (error || hibernation_test(TEST_CPUS) @@ -274,12 +277,12 @@ static int create_image(int platform_mod Enable_cpus: enable_nonboot_cpus(); - Platform_finish: - platform_finish(platform_mode); - device_power_up(in_suspend ? (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); + Platform_finish: + platform_finish(platform_mode); + Unlock: device_pm_unlock(); @@ -346,16 +349,16 @@ static int resume_target_kernel(bool pla device_pm_lock(); + error = platform_pre_restore(platform_mode); + if (error) + goto Unlock; + error = device_power_down(PMSG_QUIESCE); if (error) { printk(KERN_ERR "PM: Some devices failed to power down, " "aborting resume\n"); - goto Unlock; - } - - error = platform_pre_restore(platform_mode); - if (error) goto Cleanup; + } error = disable_nonboot_cpus(); if (error) @@ -398,11 +401,11 @@ static int resume_target_kernel(bool pla Enable_cpus: enable_nonboot_cpus(); + device_power_up(PMSG_RECOVER); + Cleanup: platform_restore_cleanup(platform_mode); - device_power_up(PMSG_RECOVER); - Unlock: device_pm_unlock(); @@ -466,17 +469,17 @@ int hibernation_platform_enter(void) device_pm_lock(); - error = device_power_down(PMSG_HIBERNATE); + error = hibernation_ops->prepare(); if (error) goto Unlock; - error = hibernation_ops->prepare(); + error = device_power_down(PMSG_HIBERNATE); if (error) goto Platofrm_finish; error = disable_nonboot_cpus(); if (error) - goto Platofrm_finish; + goto Power_up_devices; local_irq_disable(); sysdev_suspend(PMSG_HIBERNATE); @@ -484,6 +487,8 @@ int hibernation_platform_enter(void) /* We should never get here */ while (1); + Power_up_devices: + device_power_up(PMSG_RESTORE); /* * We don't need to reenable the nonboot CPUs or resume consoles, since * the system is going to be halted anyway. @@ -491,8 +496,6 @@ int hibernation_platform_enter(void) Platofrm_finish: hibernation_ops->finish(); - device_power_up(PMSG_RESTORE); - Unlock: device_pm_unlock(); -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/