mm/mm_init.c: add debug messsge for dma zone

[PATCH] mm/mm_init.c: add debug messsge for dma zone

Posted by Haifeng Xu 2 years, 8 months ago

If freesize is less than dma_reserve, print warning message to report
this case.

Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
---
 mm/mm_init.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/mm/mm_init.c b/mm/mm_init.c
index 232efac9a929..9a9d6a52471c 100644
--- a/mm/mm_init.c
+++ b/mm/mm_init.c
@@ -1561,9 +1561,14 @@ static void __init free_area_init_core(struct pglist_data *pgdat)
 		}
 
 		/* Account for reserved pages */
-		if (j == 0 && freesize > dma_reserve) {
-			freesize -= dma_reserve;
-			pr_debug("  %s zone: %lu pages reserved\n", zone_names[0], dma_reserve);
+		if (j == 0) {
+			if (freesize >= dma_reserve) {
+				freesize -= dma_reserve;
+				pr_debug("  %s zone: %lu pages reserved\n",
+					 zone_names[0], dma_reserve);
+			} else
+				pr_warn("  %s zone: %lu reserved pages exceeds freesize %lu\n",
+					zone_names[0], dma_reserve, freesize);
 		}
 
 		if (!is_highmem_idx(j))
-- 
2.25.1

Re: [PATCH] mm/mm_init.c: add debug messsge for dma zone

Posted by Michal Hocko 2 years, 8 months ago

On Wed 07-06-23 09:07:34, Haifeng Xu wrote:
> If freesize is less than dma_reserve, print warning message to report
> this case.

Why?

> Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
> ---
>  mm/mm_init.c | 11 ++++++++---
>  1 file changed, 8 insertions(+), 3 deletions(-)
> 
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index 232efac9a929..9a9d6a52471c 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -1561,9 +1561,14 @@ static void __init free_area_init_core(struct pglist_data *pgdat)
>  		}
>  
>  		/* Account for reserved pages */
> -		if (j == 0 && freesize > dma_reserve) {
> -			freesize -= dma_reserve;
> -			pr_debug("  %s zone: %lu pages reserved\n", zone_names[0], dma_reserve);
> +		if (j == 0) {
> +			if (freesize >= dma_reserve) {
> +				freesize -= dma_reserve;
> +				pr_debug("  %s zone: %lu pages reserved\n",
> +					 zone_names[0], dma_reserve);
> +			} else
> +				pr_warn("  %s zone: %lu reserved pages exceeds freesize %lu\n",
> +					zone_names[0], dma_reserve, freesize);
>  		}
>  
>  		if (!is_highmem_idx(j))
> -- 
> 2.25.1

-- 
Michal Hocko
SUSE Labs

Re: [PATCH] mm/mm_init.c: add debug messsge for dma zone

Posted by David Hildenbrand 2 years, 8 months ago

On 07.06.23 12:16, Michal Hocko wrote:
> On Wed 07-06-23 09:07:34, Haifeng Xu wrote:
>> If freesize is less than dma_reserve, print warning message to report
>> this case.
> 
> Why?

I'd like to second that question, and add

a) Did you run into that scenario?
b) What can an admin do in that case with that error messages?

If it reveals a buggy situation, maybe a WARN_ON_ONCE() is warranted ... 
but maybe only if anybody actually ran into that issue.

-- 
Cheers,

David / dhildenb

Re: [PATCH] mm/mm_init.c: add debug messsge for dma zone

Posted by Haifeng Xu 2 years, 8 months ago


On 2023/6/7 18:22, David Hildenbrand wrote:
> On 07.06.23 12:16, Michal Hocko wrote:
>> On Wed 07-06-23 09:07:34, Haifeng Xu wrote:
>>> If freesize is less than dma_reserve, print warning message to report
>>> this case.
>>
>> Why?
> 
> I'd like to second that question, and add
> 
> a) Did you run into that scenario?
> b) What can an admin do in that case with that error messages?

In theory，dma_reserve shouldn't exceed freesize, so the error messages can remind us
to verify whether the configuration of reserved memory is correct.

> 
> If it reveals a buggy situation, maybe a WARN_ON_ONCE() is warranted ... but maybe only if anybody actually ran into that issue.
>

Re: [PATCH] mm/mm_init.c: add debug messsge for dma zone

Posted by Michal Hocko 2 years, 8 months ago

On Thu 08-06-23 15:38:48, Haifeng Xu wrote:
> 
> 
> On 2023/6/7 18:22, David Hildenbrand wrote:
> > On 07.06.23 12:16, Michal Hocko wrote:
> >> On Wed 07-06-23 09:07:34, Haifeng Xu wrote:
> >>> If freesize is less than dma_reserve, print warning message to report
> >>> this case.
> >>
> >> Why?
> > 
> > I'd like to second that question, and add
> > 
> > a) Did you run into that scenario?
> > b) What can an admin do in that case with that error messages?
> 
> In theory，dma_reserve shouldn't exceed freesize, so the error messages can remind us
> to verify whether the configuration of reserved memory is correct.

I am not really convinced this is worth touching the code TBH.

-- 
Michal Hocko
SUSE Labs

Re: [PATCH] mm/mm_init.c: add debug messsge for dma zone

Posted by Mike Rapoport 2 years, 8 months ago

On Thu, Jun 08, 2023 at 11:18:02AM +0200, Michal Hocko wrote:
> On Thu 08-06-23 15:38:48, Haifeng Xu wrote:
> > 
> > 
> > On 2023/6/7 18:22, David Hildenbrand wrote:
> > > On 07.06.23 12:16, Michal Hocko wrote:
> > >> On Wed 07-06-23 09:07:34, Haifeng Xu wrote:
> > >>> If freesize is less than dma_reserve, print warning message to report
> > >>> this case.
> > >>
> > >> Why?
> > > 
> > > I'd like to second that question, and add
> > > 
> > > a) Did you run into that scenario?
> > > b) What can an admin do in that case with that error messages?
> > 
> > In theory，dma_reserve shouldn't exceed freesize, so the error messages can remind us
> > to verify whether the configuration of reserved memory is correct.
> 
> I am not really convinced this is worth touching the code TBH.

The only architecture that sets the dma_reserve is x86_64 and it sets it to
the number of reserved pages in DMA zone. There is no way freesize will be
less than dma_reserve.

I'm not sure that in general dma_reserve has some value now, but that's a
completely different story.

> -- 
> Michal Hocko
> SUSE Labs

-- 
Sincerely yours,
Mike.

Re: [PATCH] mm/mm_init.c: add debug messsge for dma zone

Posted by Haifeng Xu 2 years, 8 months ago


On 2023/6/8 18:13, Mike Rapoport wrote:
> On Thu, Jun 08, 2023 at 11:18:02AM +0200, Michal Hocko wrote:
>> On Thu 08-06-23 15:38:48, Haifeng Xu wrote:
>>>
>>>
>>> On 2023/6/7 18:22, David Hildenbrand wrote:
>>>> On 07.06.23 12:16, Michal Hocko wrote:
>>>>> On Wed 07-06-23 09:07:34, Haifeng Xu wrote:
>>>>>> If freesize is less than dma_reserve, print warning message to report
>>>>>> this case.
>>>>>
>>>>> Why?
>>>>
>>>> I'd like to second that question, and add
>>>>
>>>> a) Did you run into that scenario?
>>>> b) What can an admin do in that case with that error messages?
>>>
>>> In theory，dma_reserve shouldn't exceed freesize, so the error messages can remind us
>>> to verify whether the configuration of reserved memory is correct.
>>
>> I am not really convinced this is worth touching the code TBH.
> 
> The only architecture that sets the dma_reserve is x86_64 and it sets it to
> the number of reserved pages in DMA zone. There is no way freesize will be
> less than dma_reserve.

Yes. From the comments, x86_64 calculates the dma_reserve in order to set zone watermarks more
accurately. But berfore init_per_zone_wmark_min(), memblock_free_all() has already recalculated
the managed pages. It seems that the dma_reserve is not really helpful to this.

> 
> I'm not sure that in general dma_reserve has some value now, but that's a
> completely different story.
> 
>> -- 
>> Michal Hocko
>> SUSE Labs
>

Re: [PATCH] mm/mm_init.c: add debug messsge for dma zone

Posted by Haifeng Xu 2 years, 8 months ago

On 2023/6/7 18:22, David Hildenbrand wrote:
> On 07.06.23 12:16, Michal Hocko wrote:
>> On Wed 07-06-23 09:07:34, Haifeng Xu wrote:
>>> If freesize is less than dma_reserve, print warning message to report
>>> this case.
>>
>> Why?
> 
> I'd like to second that question, and add
> 
> a) Did you run into that scenario?
> b) What can an admin do in that case with that error messages?
> 
> If it reveals a buggy situation, maybe a WARN_ON_ONCE() is warranted ... but maybe only if anybody actually ran into that issue.
> 

I didn't run into that scenario, just from code review. I think the account for reserved pages is similar to memmap pages, if freesize
is less than memmap pages, it print a warning message, so for dma_reserve, it can also does the same thing.