linux-kernel - Re: [PATCH] mm: define pte_add

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]

Date:   Wed, 1 Jul 2020 10:11:13 +0800
From:   Wei Yang <richard.weiyang@...ux.alibaba.com>
To:     David Hildenbrand <david@...hat.com>
Cc:     Wei Yang <richard.weiyang@...ux.alibaba.com>,
        dave.hansen@...ux.intel.com, luto@...nel.org, peterz@...radead.org,
        tglx@...utronix.de, mingo@...hat.com, bp@...en8.de,
        akpm@...ux-foundation.org, x86@...nel.org,
        linux-kernel@...r.kernel.org, kasan-dev@...glegroups.com,
        linux-mm@...ck.org
Subject: Re: [PATCH] mm: define pte_add_end for consistency

On Tue, Jun 30, 2020 at 02:44:00PM +0200, David Hildenbrand wrote:
>On 30.06.20 05:18, Wei Yang wrote:
>> When walking page tables, we define several helpers to get the address of
>> the next boundary. But we don't have one for pte level.
>> 
>> Let's define it and consolidate the code in several places.
>> 
>> Signed-off-by: Wei Yang <richard.weiyang@...ux.alibaba.com>
>> ---
>>  arch/x86/mm/init_64.c   | 6 ++----
>>  include/linux/pgtable.h | 7 +++++++
>>  mm/kasan/init.c         | 4 +---
>>  3 files changed, 10 insertions(+), 7 deletions(-)
>> 
>> diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
>> index dbae185511cd..f902fbd17f27 100644
>> --- a/arch/x86/mm/init_64.c
>> +++ b/arch/x86/mm/init_64.c
>> @@ -973,9 +973,7 @@ remove_pte_table(pte_t *pte_start, unsigned long addr, unsigned long end,
>>  
>>  	pte = pte_start + pte_index(addr);
>>  	for (; addr < end; addr = next, pte++) {
>> -		next = (addr + PAGE_SIZE) & PAGE_MASK;
>> -		if (next > end)
>> -			next = end;
>> +		next = pte_addr_end(addr, end);
>>  
>>  		if (!pte_present(*pte))
>>  			continue;
>> @@ -1558,7 +1556,7 @@ void register_page_bootmem_memmap(unsigned long section_nr,
>>  		get_page_bootmem(section_nr, pud_page(*pud), MIX_SECTION_INFO);
>>  
>>  		if (!boot_cpu_has(X86_FEATURE_PSE)) {
>> -			next = (addr + PAGE_SIZE) & PAGE_MASK;
>> +			next = pte_addr_end(addr, end);
>>  			pmd = pmd_offset(pud, addr);
>>  			if (pmd_none(*pmd))
>>  				continue;
>> diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
>> index 32b6c52d41b9..0de09c6c89d2 100644
>> --- a/include/linux/pgtable.h
>> +++ b/include/linux/pgtable.h
>> @@ -706,6 +706,13 @@ static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot)
>>  })
>>  #endif
>>  
>> +#ifndef pte_addr_end
>> +#define pte_addr_end(addr, end)						\
>> +({	unsigned long __boundary = ((addr) + PAGE_SIZE) & PAGE_MASK;	\
>> +	(__boundary - 1 < (end) - 1) ? __boundary : (end);		\
>> +})
>> +#endif
>> +
>>  /*
>>   * When walking page tables, we usually want to skip any p?d_none entries;
>>   * and any p?d_bad entries - reporting the error before resetting to none.
>> diff --git a/mm/kasan/init.c b/mm/kasan/init.c
>> index fe6be0be1f76..89f748601f74 100644
>> --- a/mm/kasan/init.c
>> +++ b/mm/kasan/init.c
>> @@ -349,9 +349,7 @@ static void kasan_remove_pte_table(pte_t *pte, unsigned long addr,
>>  	unsigned long next;
>>  
>>  	for (; addr < end; addr = next, pte++) {
>> -		next = (addr + PAGE_SIZE) & PAGE_MASK;
>> -		if (next > end)
>> -			next = end;
>> +		next = pte_addr_end(addr, end);
>>  
>>  		if (!pte_present(*pte))
>>  			continue;
>> 
>
>I'm not really a friend of this I have to say. We're simply iterating
>over single pages, not much magic ....

Hmm... yes, we are iterating on Page boundary, while we many have the case
when addr or end is not PAGE_ALIGN.

>
>What would definitely make sense is replacing (addr + PAGE_SIZE) &
>PAGE_MASK; by PAGE_ALIGN() ...
>

No, PAGE_ALIGN() is expanded to be 

	(addr + PAGE_SIZE - 1) & PAGE_MASK;

If we change the code to PAGE_ALIGN(), we would end up with infinite loop.

>-- 
>Thanks,
>
>David / dhildenb

-- 
Wei Yang
Help you, Help me