Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
Subject: Re: [PATCH] MIPS: fix pmd_mkinvalid
To:     Jiaxun Yang <jiaxun.yang@flygoat.com>,
        Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
        "Matthew Wilcox (Oracle)" <willy@infradead.org>,
        Zhaolong Zhang <zhangzl2013@126.com>
Cc:     linux-mips@vger.kernel.org, linux-kernel@vger.kernel.org
References: <1653971769-14152-1-git-send-email-zhanghongchen@loongson.cn>
 <e8cca5c1-2db1-840e-609b-4ae98791ba80@flygoat.com>
From:   Hongchen Zhang <zhanghongchen@loongson.cn>
Message-ID: <a87c5693-a26c-a6a9-7149-74e94425703e@loongson.cn>
Date:   Wed, 1 Jun 2022 08:55:09 +0800
User-Agent: Mozilla/5.0 (X11; Linux loongarch64; rv:68.0) Gecko/20100101
 Thunderbird/68.7.0
MIME-Version: 1.0
In-Reply-To: <e8cca5c1-2db1-840e-609b-4ae98791ba80@flygoat.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 2022/5/31 下午11:40, Jiaxun Yang wrote:
> 
> 
> 在 2022/5/31 5:36, Hongchen Zhang 写道:
>> pmd_mkinvalid should not clear dirty flag,otherwise the pmd
>> dirty flag,it's ok to just clear _PAGE_VALID.
> Hi Hongcheng,
> 
> Sorry, what is meant by "otherwise the pmd dirty flag"? Could you please
> elaborate?
> In my humble opinion _PAGE_PRESENT must be cleared in mkinvalid as it
> nolonger present here. And thus it lose the dirty status.
> 
> Thanks
> - Jiaxun
> 
>> Signed-off-by: Hongchen Zhang <zhanghongchen@loongson.cn>
>> ---
>>   arch/mips/include/asm/pgtable.h | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/arch/mips/include/asm/pgtable.h 
>> b/arch/mips/include/asm/pgtable.h
>> index 374c632..d30a186 100644
>> --- a/arch/mips/include/asm/pgtable.h
>> +++ b/arch/mips/include/asm/pgtable.h
>> @@ -698,7 +698,7 @@ static inline pmd_t pmd_modify(pmd_t pmd, pgprot_t 
>> newprot)
>>   static inline pmd_t pmd_mkinvalid(pmd_t pmd)
>>   {
>> -    pmd_val(pmd) &= ~(_PAGE_PRESENT | _PAGE_VALID | _PAGE_DIRTY);
>> +    pmd_val(pmd) &= ~_PAGE_VALID;
>>       return pmd;
>>   }
Hi Jiaxun,

Refer to [ https://lkml.org/lkml/2018/10/17/231 ],Andrea said "
The clear of the real present bit during pmd (virtual) splitting is
done with pmdp_invalidate, that is created specifically to keeps
pmd_trans_huge=true, pmd_present=true despite the present bit is not
set. ". So it's wrong to clear _PAGE_PRESENT,because this will make 
pmd_present=false.

On the other hand, There is no need to clear dirty flag too.

Thanks.