Re: [Qemu-devel] [RESEND #3][PATCH] i386: Catch all non-present ptes in cpu_get_phys_page_debug

[Anthony Liguori]

Jan Kiszka wrote:
> It helps debugging guests when yet unmapped pages are correctly reported
> as, well, unmapped.
>
> Signed-off-by: Jan Kiszka <address@hidden>
> ---
>  target-i386/helper.c |    6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
>
> Index: b/target-i386/helper.c
> ===================================================================
> --- a/target-i386/helper.c
> +++ b/target-i386/helper.c
> @@ -1192,6 +1192,8 @@ target_phys_addr_t cpu_get_phys_page_deb
>              page_size = 4096;
>              pte = ldq_phys(pte_addr);
>          }
> +        if (!(pte & PG_PRESENT_MASK))
> +            return -1;
>      } else {
>          uint32_t pde;
>  
> @@ -1211,10 +1213,10 @@ target_phys_addr_t cpu_get_phys_page_deb
>                  /* page directory entry */
>                  pte_addr = ((pde & ~0xfff) + ((addr >> 10) & 0xffc)) & 
> env->a20_mask;
>                  pte = ldl_phys(pte_addr);
> -                if (!(pte & PG_PRESENT_MASK))
> -                    return -1;
>                  page_size = 4096;
>              }
> +            if (!(pte & PG_PRESENT_MASK))
> +                return -1;
>          }
>   

This looks redundant.  In the PSE=1 case, the check for pde & 
PG_PRESENT_MASK takes care of it.  In the PSE=0 clause, there is already 
a check for pte & PG_PRESENT_MASK.

The first chunk looks like a genuine fix though.

Regards,

Anthony Liguori

>          pte = pte & env->a20_mask;
>      }
>
>
>