linux-kernel - Re: [PATCH 0/3] PCI: rockchip: assert PERST# in S3

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]

Date:   Fri, 13 Oct 2017 17:06:48 -0700
From:   Brian Norris <briannorris@...omium.org>
To:     Bjorn Helgaas <helgaas@...nel.org>
Cc:     Bjorn Helgaas <bhelgaas@...gle.com>,
        Mark Rutland <mark.rutland@....com>,
        devicetree@...r.kernel.org, Heiko Stuebner <heiko@...ech.de>,
        linux-pci@...r.kernel.org, Shawn Lin <shawn.lin@...k-chips.com>,
        linux-kernel@...r.kernel.org,
        "open list:ARM/Rockchip SoC..." <linux-rockchip@...ts.infradead.org>,
        Rob Herring <robh+dt@...nel.org>,
        Rajat Jain <rajatja@...gle.com>,
        Frank Rowand <frowand.list@...il.com>,
        linux-arm-kernel@...ts.infradead.org, furquan@...omium.org
Subject: Re: [PATCH 0/3] PCI: rockchip: assert PERST# in S3

Hi Bjorn,

One piece of information to add:

On Thu, Oct 12, 2017 at 11:27 PM, Brian Norris <briannorris@...omium.org> wrote:
> On Thu, Oct 12, 2017 at 10:15:23PM -0500, Bjorn Helgaas wrote:
> > Is this a hole in those specs?  Is this something that needs to be
> > clarified by the PCI-SIG to improve interoperability?
>
> After re-re-reading the specifications, I'm more convinced that the
> first Wifi vendor got it wrong. I also don't trust them to get many
> things right in general, so this is pretty much par for the course.
>
> The only things I can suggest:
> * "main power" is never defined, as far as I can tell. So "main" and
>   "auxiliary" power don't have much meaning for many M.2 cards, and so
>   I end up reading much of the spec with a grain of salt

This part is where the first vendor seems to disagree. They claim that
because power is not switched off, PERST# can remain asserted or
deasserted -- whichever leads to least power consumption. (i.e., they
don't consider they need to follow the parts that describe "when
removing main power" (e.g., in S3) we must assert PERST#.

> * S3 is never defined in the PCIe Card EM spec, but it's thrown around a
>   few times
> * if possible, the PCIe base spec should mention something about the
>   fundamental reset which is expected with an L2 transition. It's not
>   clear what to do if you don't want to switch off power completely (and
>   so enter L3), but you also don't have "auxiliary" power

Brian