[U-Boot] [RFC PATCH v2 13/20] fastboot: Merge reboot-bootloader handling

[Alex Kiernan]

On Wed, May 2, 2018 at 6:51 AM Jocelyn Bohr <bohr at verily.com> wrote:



> On Tue, May 1, 2018 at 10:46 PM Jocelyn Bohr <bohr at google.com> wrote:



>> On Tue, May 1, 2018 at 1:21 AM Alex Kiernan <alex.kiernan at gmail.com>
wrote:

>>> On Tue, May 1, 2018 at 8:22 AM Alex Kiernan <alex.kiernan at gmail.com>
wrote:


>>> > On Tue, May 1, 2018 at 7:50 AM Jocelyn Bohr <bohr at google.com> wrote:



>>> > > On Mon, Apr 30, 2018 at 1:33 AM Alex Kiernan <alex.kiernan at gmail.com

>>> > wrote:


...

>>> > >> diff --git a/drivers/fastboot/fb_common.c
>>> b/drivers/fastboot/fb_common.c
>>> > >> index 8b3627b..36ef669 100644
>>> > >> --- a/drivers/fastboot/fb_common.c
>>> > >> +++ b/drivers/fastboot/fb_common.c
>>> > >> @@ -102,3 +102,8 @@ int fastboot_lookup_command(const char
*cmd_string)

>>> > >>          return -1;
>>> > >>   }
>>> > >> +
>>> > >> +int __weak fb_set_reboot_flag(void)
>>> > >> +{
>>> > >> +       return -1;
>>> > >> +}

>>> > > Instead could this write the string "reboot-bootloader" to
>>> > CONFIG_FASTBOOT_BUF_ADDR?


>>> > I guess the thing you lose is you don't then get a default "I don't
know
>>> > how to do that" response. Am I right in thinking you're only using
this
>>> > behaviour on Raspberry Pi, or is it broader?


>>> Thinking about this some more... if we moved that string write into the
>>> default fb_set_reboot_flag implementation and then guarded the
>>> reboot-bootloader command with a new Kconfig symbol, you could get the
"I
>>> don't know how to do that" behaviour by disabling the command?


>> The "reboot-bootloader" behavior isn't RPi specific, it's part of the
fastboot
>> protocol and is required by all Android bootloaders that support
fastboot.

>> https://android.googlesource.com/platform/system/core/+/master/fastboot/

>> So my thinking was that since it's part of the protocol, the weak
function could
>> have some default implementation. I can't think of a reason for a device
to
>> support fastboot but not the "reboot-bootloader" command.


> I think I'm also fine with adding a default implementation later, when
> common/android-bootloader.c from AOSP gets merged. That would probably
make
> more sense, because currently nothing in the upstream code reads the
> "reboot-bootloader" flag. But I prefer to not add a new Kconfig to guard
the
> reboot-bootloader command.


Thanks, I'll leave it like it is...

Whenever code to honour it gets added, the various other mechanisms that
already exist for fb_set_reboot_flag() want looking at too as there's a mix
of approaches today (my least favourite of which is
arch/arm/mach-omap2/boot-common.c... which the board I'm interested in ends
up using).

-- 
Alex Kiernan