linux-kernel - Re: [PATCH 1/4] mm/page_alloc.c: extract commom part to check page

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: <20200120003359.GB26292@richard>
Date:   Mon, 20 Jan 2020 08:33:59 +0800
From:   Wei Yang <richardw.yang@...ux.intel.com>
To:     David Rientjes <rientjes@...gle.com>
Cc:     Wei Yang <richardw.yang@...ux.intel.com>,
        akpm@...ux-foundation.org, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/4] mm/page_alloc.c: extract commom part to check page

On Sun, Jan 19, 2020 at 02:06:18PM -0800, David Rientjes wrote:
>On Sun, 19 Jan 2020, Wei Yang wrote:
>
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index d047bf7d8fd4..8cd06729169f 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -1025,13 +1025,9 @@ static inline bool page_expected_state(struct page *page,
>>  	return true;
>>  }
>>  
>> -static void free_pages_check_bad(struct page *page)
>> +static inline const char *__check_page(struct page *page)
>>  {
>> -	const char *bad_reason;
>> -	unsigned long bad_flags;
>> -
>> -	bad_reason = NULL;
>> -	bad_flags = 0;
>> +	const char *bad_reason = NULL;
>>  
>>  	if (unlikely(atomic_read(&page->_mapcount) != -1))
>>  		bad_reason = "nonzero mapcount";
>> @@ -1039,14 +1035,23 @@ static void free_pages_check_bad(struct page *page)
>>  		bad_reason = "non-NULL mapping";
>>  	if (unlikely(page_ref_count(page) != 0))
>>  		bad_reason = "nonzero _refcount";
>> -	if (unlikely(page->flags & PAGE_FLAGS_CHECK_AT_FREE)) {
>> -		bad_reason = "PAGE_FLAGS_CHECK_AT_FREE flag(s) set";
>> -		bad_flags = PAGE_FLAGS_CHECK_AT_FREE;
>> -	}
>>  #ifdef CONFIG_MEMCG
>>  	if (unlikely(page->mem_cgroup))
>>  		bad_reason = "page still charged to cgroup";
>>  #endif
>> +	return bad_reason;
>> +}
>> +
>> +static void free_pages_check_bad(struct page *page)
>> +{
>> +	const char *bad_reason = NULL;
>> +	unsigned long bad_flags = 0;
>> +
>> +	bad_reason = __check_page(page);
>> +	if (unlikely(page->flags & PAGE_FLAGS_CHECK_AT_FREE)) {
>> +		bad_reason = "PAGE_FLAGS_CHECK_AT_FREE flag(s) set";
>> +		bad_flags = PAGE_FLAGS_CHECK_AT_FREE;
>> +	}
>>  	bad_page(page, bad_reason, bad_flags);
>>  }
>>  
>> @@ -2044,12 +2049,7 @@ static void check_new_page_bad(struct page *page)
>>  	const char *bad_reason = NULL;
>>  	unsigned long bad_flags = 0;
>>  
>> -	if (unlikely(atomic_read(&page->_mapcount) != -1))
>> -		bad_reason = "nonzero mapcount";
>> -	if (unlikely(page->mapping != NULL))
>> -		bad_reason = "non-NULL mapping";
>> -	if (unlikely(page_ref_count(page) != 0))
>> -		bad_reason = "nonzero _refcount";
>> +	bad_reason = __check_page(page);
>>  	if (unlikely(page->flags & __PG_HWPOISON)) {
>>  		bad_reason = "HWPoisoned (hardware-corrupted)";
>>  		bad_flags = __PG_HWPOISON;
>> @@ -2061,10 +2061,6 @@ static void check_new_page_bad(struct page *page)
>>  		bad_reason = "PAGE_FLAGS_CHECK_AT_PREP flag set";
>>  		bad_flags = PAGE_FLAGS_CHECK_AT_PREP;
>>  	}
>> -#ifdef CONFIG_MEMCG
>> -	if (unlikely(page->mem_cgroup))
>> -		bad_reason = "page still charged to cgroup";
>> -#endif
>>  	bad_page(page, bad_reason, bad_flags);
>>  }
>>  
>
>I think this is compounding a previous problem in these functions: these 
>are all "if" clauses, not "else if" clauses so they are presumably ordered 
>based on least significant to most significant (we only see the last 
>bad_reason that we find).  For the page->mem_cgroup check, this leaves 
>bad_flags set but it doesn't match bad_reason.
>

I have thought about this. And curious about the order of those reasons.

>Could you instead fix the problem with these functions so that we actually 
>list *all* the problems with the page rather than only the last 
>conditional that is true?

Sure, thanks for the suggestion.

-- 
Wei Yang
Help you, Help me