Re: [PATCH] mm/cma: print total and used pages in cma_alloc()

[Giorgi Tchankvetadze <giorgitchankvetadze1997@gmail.com>]

What about using tracepoints?
Add a trace_cma_alloc() event that only runs when tracing is enabled. No 
lock unless someone is actively monitoring

On 8/16/2025 10:27 AM, David Hildenbrand wrote:
> On 16.08.25 06:28, Xiang Gao wrote:
>> From: gaoxiang17 <gaoxiang17@xiaomi.com>
>>
>> This makes cma info more intuitive during debugging.
>>
>> before:
>> [   24.407814] cma: cma_alloc(cma (____ptrval____), name: reserved, 
>> count 1, align 0)
>> [   24.413397] cma: cma_alloc(cma (____ptrval____), name: reserved, 
>> count 1, align 0)
>> [   24.415886] cma: cma_alloc(cma (____ptrval____), name: reserved, 
>> count 1, align 0)
>>
>> after:
>> [   24.069738] cma: cma_alloc(cma (____ptrval____), name: reserved, 
>> total pages: 16384, used pages: 64, request pages: 1, align 0)
>> [   24.075317] cma: cma_alloc(cma (____ptrval____), name: reserved, 
>> total pages: 16384, used pages: 65, request pages: 1, align 0)
>> [   24.078455] cma: cma_alloc(cma (____ptrval____), name: reserved, 
>> total pages: 16384, used pages: 66, request pages: 1, align 0)
>>
>> Signed-off-by: gaoxiang17 <gaoxiang17@xiaomi.com>
>> ---
>>   mm/cma.c | 15 +++++++++++++--
>>   1 file changed, 13 insertions(+), 2 deletions(-)
>>
>> diff --git a/mm/cma.c b/mm/cma.c
>> index 2ffa4befb99a..46cc98e7f587 100644
>> --- a/mm/cma.c
>> +++ b/mm/cma.c
>> @@ -776,6 +776,17 @@ static void cma_debug_show_areas(struct cma *cma)
>>       spin_unlock_irq(&cma->lock);
>>   }
>> +static unsigned long cma_get_used_pages(struct cma *cma)
>> +{
>> +    unsigned long used;
>> +
>> +    spin_lock_irq(&cma->lock);
>> +    used = bitmap_weight(cma->bitmap, (int)cma_bitmap_maxno(cma));
>> +    spin_unlock_irq(&cma->lock);
>> +
>> +    return used << cma->order_per_bit;
>> +}
>> +
>>   static int cma_range_alloc(struct cma *cma, struct cma_memrange *cmr,
>>                   unsigned long count, unsigned int align,
>>                   struct page **pagep, gfp_t gfp)
>> @@ -858,8 +869,8 @@ static struct page *__cma_alloc(struct cma *cma, 
>> unsigned long count,
>>       if (!cma || !cma->count)
>>           return page;
>> -    pr_debug("%s(cma %p, name: %s, count %lu, align %d)\n", __func__,
>> -        (void *)cma, cma->name, count, align);
>> +    pr_debug("%s(cma %p, name: %s, total pages: %lu, used pages: %lu, 
>> request pages: %lu, align %d)\n",
>> +        __func__, (void *)cma, cma->name, cma->count, 
>> cma_get_used_pages(cma), count, align);
> 
>          ^ one space missing for proper indentation.
> 
> But doing another spinlock cycle just for debugging purposes? That does 
> not feel right, sorry.
>