Some randome fixes and cleanups to swapfile

[PATCH 3/4] mm: swap: fix potensial buffer overflow in setup_clusters()

Posted by Kemeng Shi 6 months, 3 weeks ago

In setup_swap_map(), we only ensure badpages are in range (0, last_page].
As maxpages might be < last_page, setup_clusters() will encounter a
buffer overflow when a badpage is >= maxpages.
Only call inc_cluster_info_page() for badpage which is < maxpages to
fix the issue.

Fixes: b843786b0bd01 ("mm: swapfile: fix SSD detection with swapfile on btrfs")
Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
---
 mm/swapfile.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/mm/swapfile.c b/mm/swapfile.c
index a82f4ebefca3..63ab9f14b2c6 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -3208,9 +3208,13 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
 	 * and the EOF part of the last cluster.
 	 */
 	inc_cluster_info_page(si, cluster_info, 0);
-	for (i = 0; i < swap_header->info.nr_badpages; i++)
-		inc_cluster_info_page(si, cluster_info,
-				      swap_header->info.badpages[i]);
+	for (i = 0; i < swap_header->info.nr_badpages; i++) {
+		unsigned int page_nr = swap_header->info.badpages[i];
+
+		if (page_nr >= maxpages)
+			continue;
+		inc_cluster_info_page(si, cluster_info, page_nr);
+	}
 	for (i = maxpages; i < round_up(maxpages, SWAPFILE_CLUSTER); i++)
 		inc_cluster_info_page(si, cluster_info, i);
 
-- 
2.30.0

Re: [PATCH 3/4] mm: swap: fix potensial buffer overflow in setup_clusters()

Posted by Baoquan He 6 months, 2 weeks ago

On 05/22/25 at 08:25pm, Kemeng Shi wrote:
> In setup_swap_map(), we only ensure badpages are in range (0, last_page].
> As maxpages might be < last_page, setup_clusters() will encounter a
> buffer overflow when a badpage is >= maxpages.
> Only call inc_cluster_info_page() for badpage which is < maxpages to
> fix the issue.
> 
> Fixes: b843786b0bd01 ("mm: swapfile: fix SSD detection with swapfile on btrfs")
> Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
> ---
>  mm/swapfile.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)

Reviewed-by: Baoquan He <bhe@redhat.com>

> 
> diff --git a/mm/swapfile.c b/mm/swapfile.c
> index a82f4ebefca3..63ab9f14b2c6 100644
> --- a/mm/swapfile.c
> +++ b/mm/swapfile.c
> @@ -3208,9 +3208,13 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
>  	 * and the EOF part of the last cluster.
>  	 */
>  	inc_cluster_info_page(si, cluster_info, 0);
> -	for (i = 0; i < swap_header->info.nr_badpages; i++)
> -		inc_cluster_info_page(si, cluster_info,
> -				      swap_header->info.badpages[i]);
> +	for (i = 0; i < swap_header->info.nr_badpages; i++) {
> +		unsigned int page_nr = swap_header->info.badpages[i];
> +
> +		if (page_nr >= maxpages)
> +			continue;
> +		inc_cluster_info_page(si, cluster_info, page_nr);
> +	}
>  	for (i = maxpages; i < round_up(maxpages, SWAPFILE_CLUSTER); i++)
>  		inc_cluster_info_page(si, cluster_info, i);
>  
> -- 
> 2.30.0
>

Re: [PATCH 3/4] mm: swap: fix potensial buffer overflow in setup_clusters()

Posted by Kairui Song 6 months, 3 weeks ago

On Thu, May 22, 2025 at 11:32 AM Kemeng Shi <shikemeng@huaweicloud.com> wrote:
>
> In setup_swap_map(), we only ensure badpages are in range (0, last_page].
> As maxpages might be < last_page, setup_clusters() will encounter a
> buffer overflow when a badpage is >= maxpages.
> Only call inc_cluster_info_page() for badpage which is < maxpages to
> fix the issue.
>
> Fixes: b843786b0bd01 ("mm: swapfile: fix SSD detection with swapfile on btrfs")
> Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
> ---
>  mm/swapfile.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
>
> diff --git a/mm/swapfile.c b/mm/swapfile.c
> index a82f4ebefca3..63ab9f14b2c6 100644
> --- a/mm/swapfile.c
> +++ b/mm/swapfile.c
> @@ -3208,9 +3208,13 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
>          * and the EOF part of the last cluster.
>          */
>         inc_cluster_info_page(si, cluster_info, 0);
> -       for (i = 0; i < swap_header->info.nr_badpages; i++)
> -               inc_cluster_info_page(si, cluster_info,
> -                                     swap_header->info.badpages[i]);
> +       for (i = 0; i < swap_header->info.nr_badpages; i++) {
> +               unsigned int page_nr = swap_header->info.badpages[i];
> +
> +               if (page_nr >= maxpages)
> +                       continue;
> +               inc_cluster_info_page(si, cluster_info, page_nr);

I think we might need a pr_err or pr_warn here, this means mkswap
marked the wrong region as a bad block? Or some fs side things went
wrong.


> +       }
>         for (i = maxpages; i < round_up(maxpages, SWAPFILE_CLUSTER); i++)
>                 inc_cluster_info_page(si, cluster_info, i);
>
> --
> 2.30.0
>
>

Re: [PATCH 3/4] mm: swap: fix potensial buffer overflow in setup_clusters()

Posted by Kemeng Shi 6 months ago


on 5/26/2025 2:44 AM, Kairui Song wrote:
> On Thu, May 22, 2025 at 11:32 AM Kemeng Shi <shikemeng@huaweicloud.com> wrote:
>>
>> In setup_swap_map(), we only ensure badpages are in range (0, last_page].
>> As maxpages might be < last_page, setup_clusters() will encounter a
>> buffer overflow when a badpage is >= maxpages.
>> Only call inc_cluster_info_page() for badpage which is < maxpages to
>> fix the issue.
>>
>> Fixes: b843786b0bd01 ("mm: swapfile: fix SSD detection with swapfile on btrfs")
>> Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
>> ---
>>  mm/swapfile.c | 10 +++++++---
>>  1 file changed, 7 insertions(+), 3 deletions(-)
>>
>> diff --git a/mm/swapfile.c b/mm/swapfile.c
>> index a82f4ebefca3..63ab9f14b2c6 100644
>> --- a/mm/swapfile.c
>> +++ b/mm/swapfile.c
>> @@ -3208,9 +3208,13 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
>>          * and the EOF part of the last cluster.
>>          */
>>         inc_cluster_info_page(si, cluster_info, 0);
>> -       for (i = 0; i < swap_header->info.nr_badpages; i++)
>> -               inc_cluster_info_page(si, cluster_info,
>> -                                     swap_header->info.badpages[i]);
>> +       for (i = 0; i < swap_header->info.nr_badpages; i++) {
>> +               unsigned int page_nr = swap_header->info.badpages[i];
>> +
>> +               if (page_nr >= maxpages)
>> +                       continue;
>> +               inc_cluster_info_page(si, cluster_info, page_nr);
> 
> I think we might need a pr_err or pr_warn here, this means mkswap
> marked the wrong region as a bad block? Or some fs side things went
> wrong.
As Baoquan metioned that there is already warning in read_swap_header().
Besides, I think last_page in swap_header already indicates the range
which is acceptable to fs and we only need to explicitly handle page_nr
which is > last_page.
> 
> 
>> +       }
>>         for (i = maxpages; i < round_up(maxpages, SWAPFILE_CLUSTER); i++)
>>                 inc_cluster_info_page(si, cluster_info, i);
>>
>> --
>> 2.30.0
>>
>>
>

Re: [PATCH 3/4] mm: swap: fix potensial buffer overflow in setup_clusters()

Posted by Baoquan He 6 months, 2 weeks ago

On 05/26/25 at 02:44am, Kairui Song wrote:
> On Thu, May 22, 2025 at 11:32 AM Kemeng Shi <shikemeng@huaweicloud.com> wrote:
> >
> > In setup_swap_map(), we only ensure badpages are in range (0, last_page].
> > As maxpages might be < last_page, setup_clusters() will encounter a
> > buffer overflow when a badpage is >= maxpages.
> > Only call inc_cluster_info_page() for badpage which is < maxpages to
> > fix the issue.
> >
> > Fixes: b843786b0bd01 ("mm: swapfile: fix SSD detection with swapfile on btrfs")
> > Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
> > ---
> >  mm/swapfile.c | 10 +++++++---
> >  1 file changed, 7 insertions(+), 3 deletions(-)
> >
> > diff --git a/mm/swapfile.c b/mm/swapfile.c
> > index a82f4ebefca3..63ab9f14b2c6 100644
> > --- a/mm/swapfile.c
> > +++ b/mm/swapfile.c
> > @@ -3208,9 +3208,13 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
> >          * and the EOF part of the last cluster.
> >          */
> >         inc_cluster_info_page(si, cluster_info, 0);
> > -       for (i = 0; i < swap_header->info.nr_badpages; i++)
> > -               inc_cluster_info_page(si, cluster_info,
> > -                                     swap_header->info.badpages[i]);
> > +       for (i = 0; i < swap_header->info.nr_badpages; i++) {
> > +               unsigned int page_nr = swap_header->info.badpages[i];
> > +
> > +               if (page_nr >= maxpages)
> > +                       continue;
> > +               inc_cluster_info_page(si, cluster_info, page_nr);
> 
> I think we might need a pr_err or pr_warn here, this means mkswap
> marked the wrong region as a bad block? Or some fs side things went
> wrong.

There's aready warning in read_swap_header():
static unsigned long read_swap_header(struct swap_info_struct *si,
                                        union swap_header *swap_header,
                                        struct inode *inode)
{
	......
	if (last_page > maxpages) {
                pr_warn("Truncating oversized swap area, only using %luk out of %luk\n",
                        K(maxpages), K(last_page));
        }
	...
}

And if we add pr_err|warn here, we also need add it in setup_swap_map()
when filling swap_map.

> 
> 
> > +       }
> >         for (i = maxpages; i < round_up(maxpages, SWAPFILE_CLUSTER); i++)
> >                 inc_cluster_info_page(si, cluster_info, i);
> >
> > --
> > 2.30.0
> >
> >
>