linux-kernel - Re: [PATCH 1/1] ACPI: Document _OSI and

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: <20161122100258.GA1552@wunner.de>
Date:   Tue, 22 Nov 2016 11:02:58 +0100
From:   Lukas Wunner <lukas@...ner.de>
To:     Len Brown <lenb@...nel.org>
Cc:     linux-acpi@...r.kernel.org, linux-kernel@...r.kernel.org,
        Len Brown <len.brown@...el.com>
Subject: Re: [PATCH 1/1] ACPI: Document _OSI and _REV for Linux BIOS writers

On Tue, Nov 22, 2016 at 12:29:46AM -0500, Len Brown wrote:
> +On a platform tasked with running multiple versions of Windows,
> +the BIOS could use _OS to enable/disable devices that and OS
                                                           ^
							   delete

> +_OSI("Darwin")
> +--------------
> +
> +No description of _OSI would be complete without mentioning _OSI("Darwin").
> +This string is mis-used by Apple on Mac platforms, just as Linux
> +mis-used _OSI("Linux").
> +
> +Linux did not originally return TRUE to _OSI("Darwin"),
> +on the assumption that on Apple hardware, Linux was
> +probably best pretending to be Windows,
> +rather than pretending to be OSX.
> +
> +But _OSI("Darwin") was required to make Thunderbolt work,
> +and so Linux changed to enable it by default.  Making matters worse,
> +that change did not include the ability to disable _OSI("Darwin")
> +on the cmdline with acpi_OSI=!Darwin, and so the resulting
> +regressions had no workaround.
> +
> +The cmdline has since been fixed to enable !Darwin.
> +Today "Darwin" enables Thunderbolt, but significantly
> +hurts battery life, and so the default for _OSI("Darwin")
> +is still under debate.

So first of all, I'm currently upstreaming a series which enables
runtime PM for Thunderbolt on Macs.  Two series with prep work went
into 4.9 and 4.10.  The final bits are slated for 4.11.

IOW the whole "answering affirmative to Darwin hurts battery life"
argument will be nonsensical and moot from 4.11 onwards.

Link to the series:
https://github.com/l1k/linux/commits/thunderbolt_runpm_v3

All patches up to 3890c45 ("mfd: intel-lpss: Avoid resuming
runtime-suspended lpss unnecessarily") are either already in
mainline or queued up for it.

Second, your description is incorrect in that Thunderbolt is only
powered down in the non-Darwin case on Macs introduced before 2015.
That's because they didn't support Thunderbolt on Windows (with
BootCamp) on the older machines, so they powered the controller down
and made it invisible to the OS.

However on Macs introduced 2015+, they now seem to support Thunderbolt
on Windows.  They seem to switch the controller between SMM-managed
versus OS-managed mode based on the _OSI string.  In any case they no
longer just power it down.

Third, Apple uses _OSI("Darwin") not just for Thunderbolt, so the
above description is incomplete.  They use it *generally* to
distinguish between macOS versus Windows code paths in AML.
(These are the only two OSes they care about.)

E.g. on the 12" MacBook, which doesn't even have Thunderbolt, the
_PS0 and _PS3 methods become no-ops in the non-Darwin case.

Usually Apple only supports a subset of the hardware on Windows,
so we have to pretend being Darwin to be afforded full access to
the machine.

One noteworthy issue with _OSI("Darwin") is that we have to respond
negatively to all other vendor strings once it's encountered.
See here for details:  http://mjg59.dreamwidth.org/29954.html

Thanks,

Lukas