linux-kernel - Re: [PATCH v2 08/14] mm: copy_pte_range() use pte_offset_map_rw

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: <fe146c6c-60e6-4ff1-bb93-6c818f71b3de@bytedance.com>
Date: Fri, 30 Aug 2024 14:42:30 +0800
From: Qi Zheng <zhengqi.arch@...edance.com>
To: David Hildenbrand <david@...hat.com>
Cc: hughd@...gle.com, willy@...radead.org, muchun.song@...ux.dev,
 vbabka@...nel.org, akpm@...ux-foundation.org, rppt@...nel.org,
 vishal.moola@...il.com, peterx@...hat.com, ryan.roberts@....com,
 christophe.leroy2@...soprasteria.com, linux-kernel@...r.kernel.org,
 linux-mm@...ck.org, linux-arm-kernel@...ts.infradead.org,
 linuxppc-dev@...ts.ozlabs.org
Subject: Re: [PATCH v2 08/14] mm: copy_pte_range() use
 pte_offset_map_rw_nolock()



On 2024/8/29 23:36, David Hildenbrand wrote:
> On 22.08.24 09:13, Qi Zheng wrote:
>> In copy_pte_range(), we may modify the src_pte entry after holding the
>> src_ptl, so convert it to using pte_offset_map_rw_nolock(). But since we
>> already hold the write lock of mmap_lock, there is no need to get pmdval
>> to do pmd_same() check, just pass a dummy variable to it.
>>
>> Signed-off-by: Qi Zheng <zhengqi.arch@...edance.com>
>> ---
>>   mm/memory.c | 11 ++++++++++-
>>   1 file changed, 10 insertions(+), 1 deletion(-)
>>
>> diff --git a/mm/memory.c b/mm/memory.c
>> index 7b6071a0e21e2..30d98025b2a40 100644
>> --- a/mm/memory.c
>> +++ b/mm/memory.c
>> @@ -1083,6 +1083,7 @@ copy_pte_range(struct vm_area_struct *dst_vma, 
>> struct vm_area_struct *src_vma,
>>       struct mm_struct *src_mm = src_vma->vm_mm;
>>       pte_t *orig_src_pte, *orig_dst_pte;
>>       pte_t *src_pte, *dst_pte;
>> +    pmd_t dummy_pmdval;
>>       pte_t ptent;
>>       spinlock_t *src_ptl, *dst_ptl;
>>       int progress, max_nr, ret = 0;
>> @@ -1108,7 +1109,15 @@ copy_pte_range(struct vm_area_struct *dst_vma, 
>> struct vm_area_struct *src_vma,
>>           ret = -ENOMEM;
>>           goto out;
>>       }
>> -    src_pte = pte_offset_map_nolock(src_mm, src_pmd, addr, &src_ptl);
>> +
>> +    /*
>> +     * Use the maywrite version to indicate that dst_pte will be 
>> modified,
>> +     * but since we already hold the write lock of mmap_lock, there 
>> is no
>> +     * need to get pmdval to do pmd_same() check, just pass a dummy 
>> variable
>> +     * to it.
> 
> As we hold the mmap lock write lock, I assume it will prevent any page 
> table removal, because they need *at least* the mmap lock in read mode, 
> right?

Except for retract_page_tables(), all others hold the read lock of
mmap_lock.

> 
> We should probably document the rules for removing a page table -- which 
> locks must be held in which mode (if not already done).

Agree, I will document it in the v3.

>