Re: [Qemu-devel] [PATCH] walk_pml4e(): fix abort on bad PML4E/PDPTE/PDE/PTE addresses

[Andreas Färber]

Am 30.05.2013 15:16, schrieb Luiz Capitulino:
> On Thu, 30 May 2013 15:16:18 +0200
> Laszlo Ersek <address@hidden> wrote:
> 
>> On 05/30/13 14:59, Luiz Capitulino wrote:
>>> On Tue, 28 May 2013 14:19:22 -0400
>>> Luiz Capitulino <address@hidden> wrote:
>>>
>>>> The code used to walk IA-32e page-tables, and possibly PAE page-tables,
>>>> uses the bit mask ~0xfff to get the next PML4E/PDPTE/PDE/PTE address.
>>>>
>>>> However, as we use a uint64_t to store the resulting address, that mask
>>>> gets expanded to 0xfffffffffffff000 which not only ends up selecting
>>>> reserved bits but also selects the XD bit (execute-disable) which
>>>> happens to be enabled by Windows 8, causing qemu_get_ram_ptr() to abort.
>>>>
>>>> This commit fixes that problem by replacing ~0xfff by a correct mask
>>>> that only selects the address bit range (ie. bits 51:12).
>>>>
>>>> Signed-off-by: Luiz Capitulino <address@hidden>
>>>
>>> Ping? Wen?
>>>
>>> Would be nice get a Reviewed-by before merging...
>>
>> I didn't miss your submission and did find it OK, I just felt unsure
>> about stating so, because "simple" patches like this are prime territory
>> to burn someone's R-b's worth (ie. to expose a reviewer's lack of
>> information / experience). But hey, what can I lose? The patch does look
>> good to me, so
> 
> Thank you Laszlo! It's also new territory for me, that's why I'm asking
> for reviews (otherwise I'd just sneak it in some pull request :-)

Luiz, you aware aware that I have another fix by Nuohan queued that
seemed orthogonal? If someone reviews my refactoring series (which
resent that patch) I would like to send out a PULL for that rather soon,
since it blocks further CPU work. I would then include your fix as well
to avoid merge conflicts.

Andreas

> 
>>
>> Reviewed-by: Laszlo Ersek <address@hidden>
>>
>>>
>>>> ---
>>>>
>>>> PS: I (obviously) don't any more core dumps with this patch applied, but
>>>>     I couldn't check if the Windows dump is correct (does anyone know
>>>>     how to do this?). I did quickly check on Linux though.
>>>>
>>>>  target-i386/arch_memory_mapping.c | 10 ++++++----
>>>>  1 file changed, 6 insertions(+), 4 deletions(-)
>>>>
>>>> diff --git a/target-i386/arch_memory_mapping.c 
>>>> b/target-i386/arch_memory_mapping.c
>>>> index 844893f..24884bd 100644
>>>> --- a/target-i386/arch_memory_mapping.c
>>>> +++ b/target-i386/arch_memory_mapping.c
>>>> @@ -75,6 +75,8 @@ static void walk_pte2(MemoryMappingList *list,
>>>>  }
>>>>  
>>>>  /* PAE Paging or IA-32e Paging */
>>>> +#define PLM4_ADDR_MASK 0xffffffffff000 /* selects bits 51:12 */
>>>> +
>>>>  static void walk_pde(MemoryMappingList *list, hwaddr pde_start_addr,
>>>>                       int32_t a20_mask, target_ulong start_line_addr)
>>>>  {
>>>> @@ -105,7 +107,7 @@ static void walk_pde(MemoryMappingList *list, hwaddr 
>>>> pde_start_addr,
>>>>              continue;
>>>>          }
>>>>  
>>>> -        pte_start_addr = (pde & ~0xfff) & a20_mask;
>>>> +        pte_start_addr = (pde & PLM4_ADDR_MASK) & a20_mask;
>>>>          walk_pte(list, pte_start_addr, a20_mask, line_addr);
>>>>      }
>>>>  }
>>>> @@ -208,7 +210,7 @@ static void walk_pdpe(MemoryMappingList *list,
>>>>              continue;
>>>>          }
>>>>  
>>>> -        pde_start_addr = (pdpe & ~0xfff) & a20_mask;
>>>> +        pde_start_addr = (pdpe & PLM4_ADDR_MASK) & a20_mask;
>>>>          walk_pde(list, pde_start_addr, a20_mask, line_addr);
>>>>      }
>>>>  }
>>>> @@ -231,7 +233,7 @@ static void walk_pml4e(MemoryMappingList *list,
>>>>          }
>>>>  
>>>>          line_addr = ((i & 0x1ffULL) << 39) | (0xffffULL << 48);
>>>> -        pdpe_start_addr = (pml4e & ~0xfff) & a20_mask;
>>>> +        pdpe_start_addr = (pml4e & PLM4_ADDR_MASK) & a20_mask;
>>>>          walk_pdpe(list, pdpe_start_addr, a20_mask, line_addr);
>>>>      }
>>>>  }
>>>> @@ -249,7 +251,7 @@ int cpu_get_memory_mapping(MemoryMappingList *list, 
>>>> CPUArchState *env)
>>>>          if (env->hflags & HF_LMA_MASK) {
>>>>              hwaddr pml4e_addr;
>>>>  
>>>> -            pml4e_addr = (env->cr[3] & ~0xfff) & env->a20_mask;
>>>> +            pml4e_addr = (env->cr[3] & PLM4_ADDR_MASK) & env->a20_mask;
>>>>              walk_pml4e(list, pml4e_addr, env->a20_mask);
>>>>          } else
>>>>  #endif
>>>
>>>
>>
> 


-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg