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:   Wed, 20 Dec 2017 09:24:03 +0100
From:   Arnd Bergmann <>
To:     Joel Stanley <>
Cc:     Stephen Boyd <>,
        Michael Turquette <>,
        Rob Herring <>,
        C├ędric Le Goater <>,
        Benjamin Herrenschmidt <>,
        Jeremy Kerr <>,
        devicetree <>,
        Andrew Jeffery <>,
        Linux ARM <>,
        Linux Kernel Mailing List <>,,
        Patrick Venture <>, Xo Wang <>,
        Lei YU <>,
        Mark Rutland <>
Subject: Re: [PATCH v3 00/19] ARM: dts: aspeed: updates and new machines

On Wed, Dec 20, 2017 at 4:37 AM, Joel Stanley <> wrote:
> On Wed, Dec 20, 2017 at 1:53 PM, Joel Stanley <> wrote:
>> This series of device tree patches for the ASPEED BMC machines
>> moves all systems to use the soon to be merged clk driver, and
>> updates machines to use all of the drivers we have upstream.
>>  v3: Address review from Rob and Cedric
>>   - Move aspeed-gpio.h usage out into the patches where use of the GPIO
>>     is added
>>   - Clarify that the aspeed-clock.h patch will be merged as part of
>>     the device tree tree. This is to ensure we don't depend on the clk
>>     tree for building.
> Arnd, Michael, Stephen; how do we resolve this? We need the
> dt-bindings header to be present for both the clk driver and the
> device tree to build.
> The clk driver is not (yet - soon I hope?) merged by Michael and
> Stephen. I am about to commit the device tree changes that will go
> through the ARM SoC tree.

There are several options:

- avoid the use of the header and redefine the binding to have a
  clear mapping between hardware clock lines and the numeric
  representation. This works better for some SoCs than others,
- Don't use the constants in the dts files for now, but use the
  numbers directly, and update the dts files for the next merge window
- merge only one side for 4.16, either the driver or the dts files,
  and follow up with the other one in 4.17
- make one shared git branch that contains only the headers
  and base both the driver and the dts files on that branch so you
  get a single shared commit ID.

I'm fine with any of the above, please pick whatever suits you best.


Powered by blists - more mailing lists