linux-kernel - Re: [PATCH] PCI: ACPI: Fix ThunderX PEM initialization

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: <20170315113323.GA28027@localhost.localdomain>
Date:   Wed, 15 Mar 2017 04:33:23 -0700
From:   Vadim Lomovtsev <Vadim.Lomovtsev@...iumnetworks.com>
To:     Jon Masters <jcm@...masters.org>
Cc:     Bjorn Helgaas <helgaas@...nel.org>, David.Daney@...ium.com,
        tn@...ihalf.com, linux-kernel@...r.kernel.org,
        stemerkhanov@...IUMNETWORKS.onmicrosoft.com,
        linux-pci@...r.kernel.org, bhelgaas@...gle.com,
        linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH] PCI: ACPI: Fix ThunderX PEM initialization

Hi Jon,

On Wed, Mar 15, 2017 at 07:14:38AM -0400, Jon Masters wrote:
> Hi Bjorn, Vadim,
> 
> Following up to this old thread...
> 
> On 02/01/2017 10:18 AM, Bjorn Helgaas wrote:
> > On Wed, Feb 01, 2017 at 04:53:25AM -0800, Vadim Lomovtsev wrote:
> 
> >>>> Because there is no such ACPI ID as "THRX0002" registered
> >>>> (http://www.uefi.org/acpi_id_list).
> 
> There is still no "THRX" prefix registered with UEFI as of this morning.
> 
> >>> To be pedantically correct, I think you want "THRX" registered.  Then
> >>> you can manage the "0002" part internally without registering each
> >>> individual device.
> 
> The upstream Linux kernel contains a quirk matching entry that looks for
> "THRX". Therefore, you have already agreed (as of at least January) that
> this is the prefix that you will use in any firmware updates to support
> the latest upstream Linux kernel. Please register this prefix promptly.

And from what I know for now - we wont going to register this
since we have already regsitered "CAV" prefix for that. And this was the part
of our discussion also.

We had a bit long review of proper implementation of legacy firmware support,
so my apologise on that.

Please take a look at link to the patchset posted by Tomasz.
https://www.spinics.net/lists/arm-kernel/msg568741.html

> 
> >> Not sure if it would be registered that way, because (AFAIK)
> >> it expected to be string constructed from Vendor ID (not the Product ID) plus
> >> four hex digit manged internaly. So we suggest to change it to 177DXXXX
> >> which corresponds to Cavium PCI ID https://pci-ids.ucw.cz/pci.ids.
> >> It's also possible to use the 3-digit PNP ID, "CAV", to construct these
> >> _HID/_CID/_SUB values (http://www.uefi.org/pnp_id_list).
> > 
> > My point was that you only need to register the prefix ("CAV" or
> > "THRX") of the PNP or ACPI ID.  Then you manage the suffixes
> > internally.  You as long as you register "CAV" or "THRX", you can
> > assign and use "THRX0002" yourself without registering that
> > specifically.
> >

And my reply here was :
"Yes, exactly. And the "CAV" perfix is already registered.
And I think will'll use it to keep things aligned to specs & rules."

> >> So the FW will be updated accordingly.
> 
> Indeed.

Yes, it is now contains "CAVxxx" as _HID for device config object.

> 
> The version Bjorn merged looks for "THRX". This is the version that you will
> use, and you will promptly register that prefix with UEFI and provide fixes
> for existing firmware to correctly use the solution that is upstream.

Cavium FW is updated accordingly to use already registered prefix.
For existent FW legacy support is posted by Tomasz.

> 
> Thanks,
> 
> Jon.
> 


WBR,
Vadim