Patchwork Fix mask of pte index

login
register
mail settings
Submitter Qiao Nuohan
Date May 21, 2013, 3 a.m.
Message ID <1369105221-26307-1-git-send-email-qiaonuohan@cn.fujitsu.com>
Download mbox | patch
Permalink /patch/245166/
State New
Headers show

Comments

Qiao Nuohan - May 21, 2013, 3 a.m.
Function walk_pte needs pte index to calculate virtual address. However, pte
index of PAE paging or IA-32e paging is 9 bit, so the mask should be 0x1ff.
---
 target-i386/arch_memory_mapping.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
Jesse Larrew - May 21, 2013, 6:16 a.m.
On 05/20/2013 10:00 PM, Qiao Nuohan wrote:
> Function walk_pte needs pte index to calculate virtual address. However, pte
> index of PAE paging or IA-32e paging is 9 bit, so the mask should be 0x1ff.
> ---
>  target-i386/arch_memory_mapping.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/target-i386/arch_memory_mapping.c b/target-i386/arch_memory_mapping.c
> index 844893f..a2eb7e7 100644
> --- a/target-i386/arch_memory_mapping.c
> +++ b/target-i386/arch_memory_mapping.c
> @@ -38,7 +38,7 @@ static void walk_pte(MemoryMappingList *list, hwaddr pte_start_addr,
>              continue;
>          }
> 
> -        start_vaddr = start_line_addr | ((i & 0x1fff) << 12);
> +        start_vaddr = start_line_addr | ((i & 0x1ff) << 12);
>          memory_mapping_list_add_merge_sorted(list, start_paddr,
>                                               start_vaddr, 1 << 12);
>      }
> 

Verified that this agrees with the spec for IA-32e/PAE paging. Note that
walk_pte2() is correct; only walk_pte() has the typo.

Reviewed by: Jesse Larrew <jlarrew@linux.vnet.ibm.com>

Jesse Larrew
Software Engineer, KVM Team
IBM Linux Technology Center
Phone: (512) 973-2052 (T/L: 363-2052)
jlarrew@linux.vnet.ibm.com
Andreas Färber - May 21, 2013, 11:32 a.m.
Nuohan,

Am 21.05.2013 08:16, schrieb Jesse Larrew:
> On 05/20/2013 10:00 PM, Qiao Nuohan wrote:
>> Function walk_pte needs pte index to calculate virtual address. However, pte
>> index of PAE paging or IA-32e paging is 9 bit, so the mask should be 0x1ff.
>> ---
>>  target-i386/arch_memory_mapping.c |    2 +-
>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/target-i386/arch_memory_mapping.c b/target-i386/arch_memory_mapping.c
>> index 844893f..a2eb7e7 100644
>> --- a/target-i386/arch_memory_mapping.c
>> +++ b/target-i386/arch_memory_mapping.c
>> @@ -38,7 +38,7 @@ static void walk_pte(MemoryMappingList *list, hwaddr pte_start_addr,
>>              continue;
>>          }
>>
>> -        start_vaddr = start_line_addr | ((i & 0x1fff) << 12);
>> +        start_vaddr = start_line_addr | ((i & 0x1ff) << 12);
>>          memory_mapping_list_add_merge_sorted(list, start_paddr,
>>                                               start_vaddr, 1 << 12);
>>      }
>>
> 
> Verified that this agrees with the spec for IA-32e/PAE paging. Note that
> walk_pte2() is correct; only walk_pte() has the typo.
> 
> Reviewed by: Jesse Larrew <jlarrew@linux.vnet.ibm.com>

This patch is missing a Signed-off-by!

If you can reply with one, I'd fix up the subject for you (it should
indicate where you are fixing it) and queue it together with my CPU'ish
memory_mapping refactorings, if no one objects.

Regards,
Andreas
Qiao Nuohan - May 21, 2013, 12:05 p.m.
On 05/21/2013 07:32 PM, Andreas Färber wrote:
> This patch is missing a Signed-off-by!

Ha! I forgot it.

>
> If you can reply with one, I'd fix up the subject for you (it should
> indicate where you are fixing it) and queue it together with my CPU'ish
> memory_mapping refactorings, if no one objects.

Well, I have got no reason to object. And thanks.
Andreas Färber - May 21, 2013, 4:53 p.m.
Am 21.05.2013 14:05, schrieb Qiao Nuohan:
> On 05/21/2013 07:32 PM, Andreas Färber wrote:
>> This patch is missing a Signed-off-by!
> 
> Ha! I forgot it.
> 
>>
>> If you can reply with one, I'd fix up the subject for you (it should
>> indicate where you are fixing it) and queue it together with my CPU'ish
>> memory_mapping refactorings, if no one objects.
> 
> Well, I have got no reason to object. And thanks.

Still no Signed-off-by line in your reply nor a resend on the list, so I
can't do anything yet - did you forget? Or do you intend to send a v2?

Regards,
Andreas

Patch

diff --git a/target-i386/arch_memory_mapping.c b/target-i386/arch_memory_mapping.c
index 844893f..a2eb7e7 100644
--- a/target-i386/arch_memory_mapping.c
+++ b/target-i386/arch_memory_mapping.c
@@ -38,7 +38,7 @@  static void walk_pte(MemoryMappingList *list, hwaddr pte_start_addr,
             continue;
         }
 
-        start_vaddr = start_line_addr | ((i & 0x1fff) << 12);
+        start_vaddr = start_line_addr | ((i & 0x1ff) << 12);
         memory_mapping_list_add_merge_sorted(list, start_paddr,
                                              start_vaddr, 1 << 12);
     }