Re: [PATCH] xen/arm: fix unmapped access trapping on GICv2 hardware

[Stewart Hildebrand <stewart.hildebrand@xxxxxxx>]

Gentle ping.

To be clear, domU interrupts with vGICv2 are currently broken when
XEN_DOMCTL_CDF_trap_unmapped_accesses is not set.

On 2/5/26 14:01, Stewart Hildebrand wrote:
> Since 4dbcb0653621, the vGICv2 CPU interface is mapped in a deferred
> manner. On domains with XEN_DOMCTL_CDF_trap_unmapped_accesses unset on
> GICv2 hardware, the vGICv2 CPU interface fails to be mapped. A visible
> symptom is that a domU gets stuck at:
> 
>   [    0.177983] smp: Bringing up secondary CPUs ...
> 
> Move the 2nd check_p2m earlier so it's prioritized over try_handle_mmio.
> 
> Fixes: 980aff4e8fcd ("xen/arm: Add way to disable traps on accesses to 
> unmapped addresses")
> Signed-off-by: Stewart Hildebrand <stewart.hildebrand@xxxxxxx>
> ---
> This should be backported to 4.21.
> 
> Pipeline: 
> https://gitlab.com/xen-project/people/stewarthildebrand/xen/-/pipelines/2010469665
> ---
>  xen/arch/arm/traps.c | 16 ++++++++--------
>  1 file changed, 8 insertions(+), 8 deletions(-)
> 
> diff --git a/xen/arch/arm/traps.c b/xen/arch/arm/traps.c
> index 040c0f2e0db1..0c01f37ad6b4 100644
> --- a/xen/arch/arm/traps.c
> +++ b/xen/arch/arm/traps.c
> @@ -1915,6 +1915,14 @@ static void do_trap_stage2_abort_guest(struct 
> cpu_user_regs *regs,
>          if ( info.dabt_instr.state == INSTR_ERROR )
>              goto inject_abt;
>  
> +        /*
> +         * If the instruction syndrome was invalid, then we already checked 
> if
> +         * this was due to a P2M fault. So no point to check again as the 
> result
> +         * will be the same.
> +         */
> +        if ( (info.dabt_instr.state == INSTR_VALID) && check_p2m(is_data, 
> gpa) )
> +            return;
> +
>          state = try_handle_mmio(regs, &info);
>  
>          switch ( state )
> @@ -1939,14 +1947,6 @@ static void do_trap_stage2_abort_guest(struct 
> cpu_user_regs *regs,
>                  break;
>          }
>  
> -        /*
> -         * If the instruction syndrome was invalid, then we already checked 
> if
> -         * this was due to a P2M fault. So no point to check again as the 
> result
> -         * will be the same.
> -         */
> -        if ( (info.dabt_instr.state == INSTR_VALID) && check_p2m(is_data, 
> gpa) )
> -            return;
> -
>          break;
>      }
>      default: