On 30/04/21 2:12 pm, Christophe Leroy wrote:
>
>
> Le 30/04/2021 à 09:55, Sandipan Das a écrit :
>> Trace memory is cleared and the corresponding dcache lines
>> are flushed after allocation. However, this should not be
>> done using the PFN. This adds the missing __va() conversion.
>>
>> Fixes: 2ac02e5ecec0 ("powerpc/mm: Remove dcache flush from memory remove.")
>> Signed-off-by: Sandipan Das <sandi...@linux.ibm.com>
>> ---
>> arch/powerpc/platforms/powernv/memtrace.c | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/powerpc/platforms/powernv/memtrace.c
>> b/arch/powerpc/platforms/powernv/memtrace.c
>> index 71c1262589fe..a31f13814f2e 100644
>> --- a/arch/powerpc/platforms/powernv/memtrace.c
>> +++ b/arch/powerpc/platforms/powernv/memtrace.c
>> @@ -104,8 +104,8 @@ static void memtrace_clear_range(unsigned long start_pfn,
>> * Before we go ahead and use this range as cache inhibited range
>> * flush the cache.
>> */
>> - flush_dcache_range_chunked(PFN_PHYS(start_pfn),
>> - PFN_PHYS(start_pfn + nr_pages),
>> + flush_dcache_range_chunked((unsigned long)__va(PFN_PHYS(start_pfn)),
>> + (unsigned long)__va(PFN_PHYS(start_pfn + nr_pages)),
>
> Can you use pfn_to_virt() instead ?
>
>> FLUSH_CHUNK_SIZE);
>> }
>>
Sure, that looks cleaner. Thanks for the suggestion.
- Sandipan
