linux-kernel - Re: [PATCH] mm, memory_hotplug: Don't bail out in do_migrate

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <616842fe-5c8d-0d51-cf7f-35b0bd0a1b34@redhat.com>
Date:   Tue, 11 Dec 2018 10:37:43 +0100
From:   David Hildenbrand <david@...hat.com>
To:     osalvador@...e.de, akpm@...ux-foundation.org
Cc:     mhocko@...e.com, pasha.tatashin@...een.com,
        dan.j.williams@...il.com, linux-kernel@...r.kernel.org,
        linux-mm@...ck.org, owner-linux-mm@...ck.org
Subject: Re: [PATCH] mm, memory_hotplug: Don't bail out in do_migrate_range
 prematurely

On 11.12.18 09:57, osalvador@...e.de wrote:
> On 2018-12-11 09:50, Oscar Salvador wrote:
> 
>> -		} else {
>> -			pr_warn("failed to isolate pfn %lx\n", pfn);
>> -			dump_page(page, "isolation failed");
>> -			put_page(page);
>> -			/* Because we don't have big zone->lock. we should
>> -			   check this again here. */
>> -			if (page_count(page)) {
>> -				not_managed++;
>> -				ret = -EBUSY;
>> -				break;
> 
> I forgot that here we should at least leave the put_page().
> But leave also the dump_page() and the pr_warn().
> 
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -1394,6 +1394,10 @@ do_migrate_range(unsigned long start_pfn, 
> unsigned long end_pfn)
>                                  inc_node_page_state(page, 
> NR_ISOLATED_ANON +
>                                                      
> page_is_file_cache(page));
> 
> +               } else {
> +                       pr_warn("failed to isolate pfn %lx\n", pfn);
> +                       dump_page(page, "isolation failed");
> +                       put_page(page);

When we're stuck with one problematic page, and we keep looping over
that function, won't that print out *way too much* messages? Shouldn't
that be rate limited somehow (same applies to other callers in this file)

>                  }
>          }
>          if (!list_empty(&source)) {
> 


-- 

Thanks,

David / dhildenb