Re: [PATCH v2 2/2] xen/console: unify printout behavior for UART emulators

[Jan Beulich <jbeulich@xxxxxxxx>]

On 06.06.2025 21:47, dmkhn@xxxxxxxxx wrote:
> On Fri, Jun 06, 2025 at 09:12:06AM +0200, Jan Beulich wrote:
>> On 06.06.2025 09:06, dmkhn@xxxxxxxxx wrote:
>>> On Thu, Jun 05, 2025 at 08:18:34AM +0200, Jan Beulich wrote:
>>>> On 05.06.2025 02:46, dmkhn@xxxxxxxxx wrote:
>>>>> From: Denis Mukhin <dmukhin@xxxxxxxx>
>>>>>
>>>>> If virtual UART from domain X prints on the physical console, the 
>>>>> behavior is
>>>>> updated to (see [1]):
>>>>> - console focus in domain X: do not prefix messages;
>>>>> - no console focus in domain X: prefix all messages with "(dX)".
>>>>>
>>>>> Use guest_printk() without rate-limiting in all current in-hypervisor UART
>>>>> emulators. That aligns the behavior with debug I/O port 0xe9 handler on 
>>>>> x86 and
>>>>> slightly improves the logging since guest_printk() already prints the 
>>>>> domain
>>>>> ID. guest_printk() was modified to account for console focus ownership.
>>>>>
>>>>> Modify guest_console_write() for hardware domain case by adding domain ID 
>>>>> to
>>>>> the message when hwdom does not have console focus.
>>>>>
>>>>> [1] 
>>>>> https://lore.kernel.org/xen-devel/alpine.DEB.2.22.394.2412121655360.463523@ubuntu-linux-20-04-desktop/
>>>>> Signed-off-by: Denis Mukhin <dmukhin@xxxxxxxx>
>>>>> ---
>>>>> Changes since v1:
>>>>> - dropped change for debug port and for HYPERVISOR_console_io hypercall
>>>>
>>>> Yet then what about ...
>>>>
>>>>> --- a/xen/arch/arm/vuart.c
>>>>> +++ b/xen/arch/arm/vuart.c
>>>>> @@ -89,7 +89,7 @@ static void vuart_print_char(struct vcpu *v, char c)
>>>>>          if ( c != '\n' )
>>>>>              uart->buf[uart->idx++] = '\n';
>>>>>          uart->buf[uart->idx] = '\0';
>>>>> -        printk(XENLOG_G_DEBUG "DOM%u: %s", d->domain_id, uart->buf);
>>>>> +        guest_printk(d, "%s", uart->buf);
>>>>>          uart->idx = 0;
>>>>>      }
>>>>>      spin_unlock(&uart->lock);
>>>>
>>>> ... this dropping of XENLOG_G_DEBUG? In fact I'd have expected such to
>>>> be _added_ where presently missing.
>>>
>>> vUART is a debugging facility. This flavor of UART is specifically for 
>>> guest OS
>>> early boot debugging.
>>> I think it is not desirable to potentially lose guest messages while doing 
>>> such
>>> early guest OS boot debugging.
>>
>> That is the host admin's decision, not a policy we should enforce.
> 
> re: policy: agreed, I will drop that hunk.
> 
> I think for the policy control, there can be a compile time setting (separate
> patch) which enables/disables the debug output rate-limiting - and that 
> setting
> applies to:
>   - vUARTs (currently vpl011 and "vuart", later ns16550 (x86) and upcoming
>     emulator for RISC-V)
>   - debug port on x86
>   - HYPERVISOR_console_io
> 
> What do you think?

I'm not convinced, but much would depend on the justification for such a change.

Jan