[PATCH] powerpc: Add support for early debugging via Serial 16550 console

[Christophe Leroy]

Le 20/08/2022 à 00:38, Pali Rohár a écrit :
> On Friday 19 August 2022 15:34:14 Randy Dunlap wrote:
>> Hi--
>>
>> On 8/19/22 14:12, Pali Rohár wrote:
>>> Currently powerpc early debugging contains lot of platform specific
>>> options, but does not support standard UART / serial 16550 console.
>>>
>>> Later legacy_serial.c code supports registering UART as early debug console
>>> from device tree but it is not early during booting, but rather later after
>>> machine description code finishes.
>>>
>>> So for real early debugging via UART is current code unsuitable.
>>>
>>> Add support for new early debugging option CONFIG_PPC_EARLY_DEBUG_16550
>>> which enable Serial 16550 console on address defined by new option
>>> CONFIG_PPC_EARLY_DEBUG_16550_PHYSADDR and by stride by option
>>> CONFIG_PPC_EARLY_DEBUG_16550_STRIDE.
>>>
>>> With this change it is possible to debug powerpc machine descriptor code.
>>> For example this early debugging code can print on serial console also
>>> "No suitable machine description found" error which is done before
>>> legacy_serial.c code.
>>>
>>> Signed-off-by: Pali Rohár <pali at kernel.org>
>>> ---
>>> Tested on P2020 board. It allowed me do debug and implement this patch series:
>>> https://lore.kernel.org/linuxppc-dev/20220819191557.28116-1-pali@kernel.org/
>>> ---
>>>   arch/powerpc/Kconfig.debug       | 14 ++++++++++++++
>>>   arch/powerpc/include/asm/udbg.h  |  1 +
>>>   arch/powerpc/kernel/udbg.c       |  2 ++
>>>   arch/powerpc/kernel/udbg_16550.c | 33 ++++++++++++++++++++++++++++++++
>>>   4 files changed, 50 insertions(+)
>>>
>>> diff --git a/arch/powerpc/Kconfig.debug b/arch/powerpc/Kconfig.debug
>>> index 9f363c143d86..a4e7d90a45d2 100644
>>> --- a/arch/powerpc/Kconfig.debug
>>> +++ b/arch/powerpc/Kconfig.debug
>>> @@ -276,6 +276,11 @@ config PPC_EARLY_DEBUG_OPAL_HVSI
>>>   	  Select this to enable early debugging for the PowerNV platform
>>>   	  using an "hvsi" console
>>>   
>>> +config PPC_EARLY_DEBUG_16550
>>> +	bool "Serial 16550"
>>> +	help
>>> +	  Select this to enable early debugging via Serial 16550 console
>>> +
>>>   config PPC_EARLY_DEBUG_MEMCONS
>>>   	bool "In memory console"
>>>   	help
>>> @@ -355,6 +360,15 @@ config PPC_EARLY_DEBUG_CPM_ADDR
>>>   	  platform probing is done, all platforms selected must
>>>   	  share the same address.
>>>   
>>> +config PPC_EARLY_DEBUG_16550_PHYSADDR
>>> +	hex "Early debug Serial 16550 physical address"
>>> +	depends on PPC_EARLY_DEBUG_16550
>>
>> Is there any chance that you could provide a default value here
>> so that 'make olddefconfig' does not end up like it does
>> without having a default value?
> 
> No. Because there is not any default value. Physical address of 16550 is
> hardly HW dependent. Possibly also bootloader dependent (if it can remap
> peripherals to different physical addresses).

The exact same problem exists with EARLY_DEBUG_CPM. Nevertheless there 
is a default CONFIG_PPC_EARLY_DEBUG_CPM_ADDR. This value is wrong for 
some people, but at least it avoids prompt during the build of existing 
defconfig.

Other solution is to update the defconfigs with the good value to avoid 
getting an empty value.

> 
> User _has to_ specify correct physical address if want to use early
> debug 16550 console.

By the wait the build robot doesn't mind.

> 
>> CONFIG_PPC_EARLY_DEBUG_16550=y
>> # CONFIG_PPC_EARLY_DEBUG_MEMCONS is not set
>> CONFIG_PPC_EARLY_DEBUG_16550_PHYSADDR=
>> CONFIG_PPC_EARLY_DEBUG_16550_STRIDE=1
>>
>> which then causes a kconfig prompt when starting
>> the build...
> 
> Cannot we set somehow that PPC_EARLY_DEBUG_16550 would be disabled by
> default when upgrading defconfig?

The only solution I see is to put it at the end of the list, so that the 
previous default value which is PPC_EARLY_DEBUG_MEMCONS gets selected.


> 
>>> +
>>> +config PPC_EARLY_DEBUG_16550_STRIDE
>>> +	int "Early debug Serial 16550 stride"
>>> +	depends on PPC_EARLY_DEBUG_16550
>>> +	default 1
>>> +
>>>   config FAIL_IOMMU
>>>   	bool "Fault-injection capability for IOMMU"
>>>   	depends on FAULT_INJECTION
>>
>> Thanks.
>> -- 
>> ~Randy