linux-kernel - Re: [PATCH 3/3] ACPI / sleep: Simplify suspend-to-idle event processing loop

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <9231606.5a3ByxjlJX@aspire.rjw.lan>
Date:   Sun, 14 May 2017 13:17:51 +0200
From:   "Rafael J. Wysocki" <rjw@...ysocki.net>
To:     Linux PM <linux-pm@...r.kernel.org>
Cc:     Linux ACPI <linux-acpi@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Mika Westerberg <mika.westerberg@...ux.intel.com>,
        Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>,
        David Box <david.e.box@...ux.intel.com>
Subject: Re: [PATCH 3/3] ACPI / sleep: Simplify suspend-to-idle event processing loop

On Sunday, May 14, 2017 01:57:13 AM Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> 
> Commit eed4d47efe95 (ACPI / sleep: Ignore spurious SCI wakeups from
> suspend-to-idle) modified the core suspend-to-idle code to filter
> out spurious SCI interrupts received while suspended, but it
> implemented that by resuming the system partially every time an
> SCI triggers wakeup, so that the SCI handler can run (and possibly
> further event handlers invoked by it can run too), which requires the
> "noirq" phase of device resume to be carried out.  One drawback of
> that implementation is that PCI devices are put into the full-power
> state (D0) during the "noirq" resume phase and they need to be put
> into low-power states again in case the wakeup event turns out to be
> spurious, which may cause some unpleasant power fluctuations in the
> system to happen among other things.
> 
> However, that can be avoided by using the observation that it is
> generally possible to call the SCI handler directly from the ACPI
> suspend-to-idle ->wake callback and the processing initiated by it
> can be carried out before the "noirq" phase of device resume starts,
> so if the SCI interrupt turns out to be non-wakeup, the system can
> go back to sleep immediately.
> 
> Implement that idea, drop the suspend-to-idle ->sync callback added
> by commit eed4d47efe95 as it is not necessary any more and simplify
> the suspend-to-idle event processing loop in the core code.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@...el.com>

Scratch this one.

It is not sufficient to run the SCI handler alone, the SCI IRQ has to be
re-armed properly for system wakeup too, so resume_device_irqs()
and suspend_device_irqs() need to run for that.

Thanks,
Rafael