[PATCH] target/m68k: fix physical address translation in m68k_cpu_get_phys_page_debug()

[PATCH] target/m68k: fix physical address translation in m68k_cpu_get_phys_page_debug()

[Mark Cave-Ayland]

The result of the get_physical_address() function should be combined with the
offset of the original page access before being returned. Otherwise the
m68k_cpu_get_phys_page_debug() function can round to the wrong page causing
incorrect lookups in gdbstub and various "Disassembler disagrees with
translator over instruction decoding" warnings to appear at translation time.

Fixes: 88b2fef6c3 ("target/m68k: add MC68040 MMU")
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
---
 target/m68k/helper.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/target/m68k/helper.c b/target/m68k/helper.c
index 79b0b10ea9..631eab7774 100644
--- a/target/m68k/helper.c
+++ b/target/m68k/helper.c
@@ -820,10 +820,14 @@ hwaddr m68k_cpu_get_phys_page_debug(CPUState *cs, vaddr addr)
     if (env->sr & SR_S) {
         access_type |= ACCESS_SUPER;
     }
+
     if (get_physical_address(env, &phys_addr, &prot,
                              addr, access_type, &page_size) != 0) {
         return -1;
     }
+
+    addr &= TARGET_PAGE_MASK;
+    phys_addr += addr & (page_size - 1);
     return phys_addr;
 }
 
-- 
2.20.1

Re: [PATCH] target/m68k: fix physical address translation in m68k_cpu_get_phys_page_debug()

[Philippe Mathieu-Daudé]

On 6/29/20 6:26 PM, Mark Cave-Ayland wrote:
> The result of the get_physical_address() function should be combined with the
> offset of the original page access before being returned. Otherwise the
> m68k_cpu_get_phys_page_debug() function can round to the wrong page causing
> incorrect lookups in gdbstub and various "Disassembler disagrees with
> translator over instruction decoding" warnings to appear at translation time.
> 
> Fixes: 88b2fef6c3 ("target/m68k: add MC68040 MMU")
> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
> ---
>  target/m68k/helper.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/target/m68k/helper.c b/target/m68k/helper.c
> index 79b0b10ea9..631eab7774 100644
> --- a/target/m68k/helper.c
> +++ b/target/m68k/helper.c
> @@ -820,10 +820,14 @@ hwaddr m68k_cpu_get_phys_page_debug(CPUState *cs, vaddr addr)
>      if (env->sr & SR_S) {
>          access_type |= ACCESS_SUPER;
>      }
> +
>      if (get_physical_address(env, &phys_addr, &prot,
>                               addr, access_type, &page_size) != 0) {
>          return -1;
>      }
> +
> +    addr &= TARGET_PAGE_MASK;
> +    phys_addr += addr & (page_size - 1);

Correct but all the callers do the same, maybe this can somehow
be moved to the callee? Anyway:

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>

>      return phys_addr;
>  }
>  
> 

Re: [PATCH] target/m68k: fix physical address translation in m68k_cpu_get_phys_page_debug()

[Laurent Vivier]

Le 29/06/2020 à 19:56, Philippe Mathieu-Daudé a écrit :
> On 6/29/20 6:26 PM, Mark Cave-Ayland wrote:
>> The result of the get_physical_address() function should be combined with the
>> offset of the original page access before being returned. Otherwise the
>> m68k_cpu_get_phys_page_debug() function can round to the wrong page causing
>> incorrect lookups in gdbstub and various "Disassembler disagrees with
>> translator over instruction decoding" warnings to appear at translation time.
>>
>> Fixes: 88b2fef6c3 ("target/m68k: add MC68040 MMU")
>> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
>> ---
>>  target/m68k/helper.c | 4 ++++
>>  1 file changed, 4 insertions(+)
>>
>> diff --git a/target/m68k/helper.c b/target/m68k/helper.c
>> index 79b0b10ea9..631eab7774 100644
>> --- a/target/m68k/helper.c
>> +++ b/target/m68k/helper.c
>> @@ -820,10 +820,14 @@ hwaddr m68k_cpu_get_phys_page_debug(CPUState *cs, vaddr addr)
>>      if (env->sr & SR_S) {
>>          access_type |= ACCESS_SUPER;
>>      }
>> +
>>      if (get_physical_address(env, &phys_addr, &prot,
>>                               addr, access_type, &page_size) != 0) {
>>          return -1;
>>      }
>> +
>> +    addr &= TARGET_PAGE_MASK;
>> +    phys_addr += addr & (page_size - 1);
> 
> Correct but all the callers do the same, maybe this can somehow
> be moved to the callee? Anyway:

I agree.

Anyway:
Reviewed-by: Laurent Vivier <laurent@vivier.eu>

Thanks,
Laurent