Re: [PATCH v2] of: Support CONFIG_CMDLINE_EXTEND config option

[Doug Anderson]

John,

On Thu, Jul 10, 2014 at 12:33 AM, John Stultz <john.stultz@xxxxxxxxxx> wrote:
> On Fri, Jan 13, 2012 at 11:15 PM, Doug Anderson <dianders@xxxxxxxxxxxx> wrote:
>> The old logic assumes CMDLINE_FROM_BOOTLOADER vs. CMDLINE_FORCE and
>> ignores CMDLINE_EXTEND.  Here's the old logic:
>>
>> - CONFIG_CMDLINE_FORCE=true
>>     CONFIG_CMDLINE
>> - dt bootargs=non-empty:
>>     dt bootargs
>> - dt bootargs=empty, @data is non-empty string
>>     @data is left unchanged
>> - dt bootargs=empty, @data is empty string
>>     CONFIG_CMDLINE (or "" if that's not defined)
>>
>> The new logic is now documented in of_fdt.h and is copied here for
>> reference:
>>
>> - CONFIG_CMDLINE_FORCE=true
>>     CONFIG_CMDLINE
>> - CONFIG_CMDLINE_EXTEND=true, @data is non-empty string
>>     @data + dt bootargs (even if dt bootargs are empty)
>> - CONFIG_CMDLINE_EXTEND=true, @data is empty string
>>     CONFIG_CMDLINE + dt bootargs (even if dt bootargs are empty)
>> - CMDLINE_FROM_BOOTLOADER=true, dt bootargs=non-empty:
>>     dt bootargs
>> - CMDLINE_FROM_BOOTLOADER=true, dt bootargs=empty, @data is non-empty string
>>     @data is left unchanged
>> - CMDLINE_FROM_BOOTLOADER=true, dt bootargs=empty, @data is empty string
>>     CONFIG_CMDLINE (or "" if that's not defined)
>>
>> Signed-off-by: Doug Anderson <dianders@xxxxxxxxxxxx>
>> CC: devicetree-discuss@xxxxxxxxxxxxxxxx
>> CC: Grant Likely <grant.likely@xxxxxxxxxxxx>
>> CC: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
>> CC: Rob Herring <rob.herring@xxxxxxxxxxx>
>> ---
>> I didn't hear a response from Ben about whether he was OK with
>> my proposal for how CONFIG_CMDLINE_EXTEND should behave with
>> a non-empty @data, so sending out code that implements my
>> interpretation.  Hopefully this looks OK.
>>
>> Changes in v2:
>> - Removed broken fix for the case when there is no "chosen" attribute.
>> - Changed CONFIG_CMDLINE_EXTEND behavior when @data is a non-empty string.
>
> In reviewing some of the random patches in the Android AOSP common.git
> tree, I noticed this one, and found this discussion which petered out
> without resolution.  Were there any objections still, or did this just
> slip through the cracks?

I gave up and decided not to keep pushing and solved the problem
another way myself.  ...but I still agree that it's a reasonable thing
to do.  I haven't really thought about this in several years now...

I do remember seeing
<https://android-review.googlesource.com/#/c/57349/1/drivers/of/fdt.c>
where Colin looked to be making a bunch of changes to the patch before
merging, but I guess he ended up splitting that into a cleanup patch.
I think I pinged him to ask if he wanted me to get involved in the
review but I got no response.


> Note, there's also this follow-on fix from Colin, which probably
> should be included if the parent patch is merged:
> https://android.googlesource.com/kernel/common.git/+/a2acc58bc7460a03627f20842a7169d7a1328299

OK.  Seems reasonable to me.  I haven't done a full review of all of
the boundary conditions here, but otherwise it looks fine.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/