Re: [PATCH 01/10] ehci early_printk: split ehci debug driver from early_printk.c

[Jason Wessel <jason.wessel@windriver.com>]

Jason Wessel wrote:
> Yinghai Lu wrote:
>   
>> Jason Wessel wrote:
>>   
>>     
>>> Move the dbgp early printk driver in advance of refactoring and adding
>>> new code, so the changes to this code are tracked separately from the
>>> move of the code.
>>>
>>> The drivers/usb/early directory will be the location of the current
>>> and future early usb code for driving usb devices prior initializing
>>> the standard interrupt driven USB drivers.
>>>
>>> Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
>>> Cc: Greg KH <gregkh@suse.de>
>>> Cc: Ingo Molnar <mingo@elte.hu>
>>> Cc: Andrew Morton <akpm@linux-foundation.org>
>>> Cc: Yinghai Lu <yinghai@kernel.org>
>>> Cc: "Eric W. Biederman" <ebiederm@xmission.com>
>>> ---
>>>  arch/x86/kernel/early_printk.c |  715 ---------------------------------------
>>>  drivers/usb/Makefile           |    1 +
>>>  drivers/usb/early/Makefile     |    5 +
>>>  drivers/usb/early/ehci-dbgp.c  |  723 ++++++++++++++++++++++++++++++++++++++++
>>>  include/linux/usb/ehci_def.h   |    6 +
>>>  5 files changed, 735 insertions(+), 715 deletions(-)
>>>  create mode 100644 drivers/usb/early/Makefile
>>>  create mode 100644 drivers/usb/early/ehci-dbgp.c
>>>     
>>>       
>> before this patch, we could use dbgp without compiling USB susbsystem in...
>>
>>   
>>     
>
> And you shouldn't need it today.  You pointed out an error in the patch
> which I will correct.  The drivers/Makefile needs to be changed, instead
> of the drivers/usb/Makefile.
>
> The dbgp code was changed in the patch series such that it works with or
> with out the usb subsystem and with or without the earlyconsole=keep,...
>
> All the code in the drivers/usb/early would work this way as well.  As
> an example, I have an experimental UHCI serial driver that can be used
> via the earlyprintk.
>
>
>   
So I checked into this further because I was curious how all the tests
passed. 

The drivers/usb/Makefile is read because of the line in drivers/Makefile

     67 obj-$(CONFIG_PCI)               += usb/

That of course will cause the drivers/usb/early/Makefile to get read. 
This explains why all the testing passed.  For now it might be ok, but
it will have to get changed to compile the dbgp driver when/if the dbgp
gets changed to work with a non-pci EHCI controller.

Jason.