Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Subject: Re: [PATCH 25/30] nios2/mm: Enable ARCH_HAS_VM_GET_PAGE_PROT
To:     Dinh Nguyen <dinguyen@kernel.org>, linux-mm@kvack.org
Cc:     linux-kernel@vger.kernel.org,
        Christoph Hellwig <hch@infradead.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        linux-arch@vger.kernel.org
References: <1644805853-21338-1-git-send-email-anshuman.khandual@arm.com>
 <1644805853-21338-26-git-send-email-anshuman.khandual@arm.com>
 <50ac6dc2-7c71-2a8b-aa00-78926351b252@kernel.org>
 <637cfc45-60ad-3cd1-5127-76ecabb87def@arm.com>
 <7043506b-ad04-4572-316c-c5498873b8b1@kernel.org>
 <153130cc-e8d2-e65a-ff83-0a5ed243cc1c@kernel.org>
From:   Anshuman Khandual <anshuman.khandual@arm.com>
Message-ID: <bd239f45-3c9a-5faf-ba0e-610f2383720b@arm.com>
Date:   Mon, 28 Feb 2022 07:57:51 +0530
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.10.0
MIME-Version: 1.0
In-Reply-To: <153130cc-e8d2-e65a-ff83-0a5ed243cc1c@kernel.org>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk


On 2/25/22 8:08 PM, Dinh Nguyen wrote:
> 
> 
> On 2/25/22 08:29, Dinh Nguyen wrote:
>>
>>
>> On 2/25/22 02:52, Anshuman Khandual wrote:
>>>
>>>
>>> On 2/25/22 7:01 AM, Dinh Nguyen wrote:
>>>> Hi Anshuman,
>>>>
>>>> On 2/13/22 20:30, Anshuman Khandual wrote:
>>>>> This defines and exports a platform specific custom vm_get_page_prot() via
>>>>> subscribing ARCH_HAS_VM_GET_PAGE_PROT. Subsequently all __SXXX and __PXXX
>>>>> macros can be dropped which are no longer needed.
>>>>>
>>>>> Cc: Dinh Nguyen <dinguyen@kernel.org>
>>>>> Cc: linux-kernel@vger.kernel.org
>>>>> Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
>>>>> Acked-by: Dinh Nguyen <dinguyen@kernel.org>
>>>>> ---
>>>>>    arch/nios2/Kconfig               |  1 +
>>>>>    arch/nios2/include/asm/pgtable.h | 16 ------------
>>>>>    arch/nios2/mm/init.c             | 45 ++++++++++++++++++++++++++++++++
>>>>>    3 files changed, 46 insertions(+), 16 deletions(-)
>>>>>
>>>>> diff --git a/arch/nios2/Kconfig b/arch/nios2/Kconfig
>>>>> index 33fd06f5fa41..85a58a357a3b 100644
>>>>> --- a/arch/nios2/Kconfig
>>>>> +++ b/arch/nios2/Kconfig
>>>>> @@ -6,6 +6,7 @@ config NIOS2
>>>>>        select ARCH_HAS_SYNC_DMA_FOR_CPU
>>>>>        select ARCH_HAS_SYNC_DMA_FOR_DEVICE
>>>>>        select ARCH_HAS_DMA_SET_UNCACHED
>>>>> +    select ARCH_HAS_VM_GET_PAGE_PROT
>>>>>        select ARCH_NO_SWAP
>>>>>        select COMMON_CLK
>>>>>        select TIMER_OF
>>>>> diff --git a/arch/nios2/include/asm/pgtable.h b/arch/nios2/include/asm/pgtable.h
>>>>> index 4a995fa628ee..2678dad58a63 100644
>>>>> --- a/arch/nios2/include/asm/pgtable.h
>>>>> +++ b/arch/nios2/include/asm/pgtable.h
>>>>> @@ -40,24 +40,8 @@ struct mm_struct;
>>>>>     */
>>>>>      /* Remove W bit on private pages for COW support */
>>>>> -#define __P000    MKP(0, 0, 0)
>>>>> -#define __P001    MKP(0, 0, 1)
>>>>> -#define __P010    MKP(0, 0, 0)    /* COW */
>>>>> -#define __P011    MKP(0, 0, 1)    /* COW */
>>>>> -#define __P100    MKP(1, 0, 0)
>>>>> -#define __P101    MKP(1, 0, 1)
>>>>> -#define __P110    MKP(1, 0, 0)    /* COW */
>>>>> -#define __P111    MKP(1, 0, 1)    /* COW */
>>>>>      /* Shared pages can have exact HW mapping */
>>>>> -#define __S000    MKP(0, 0, 0)
>>>>> -#define __S001    MKP(0, 0, 1)
>>>>> -#define __S010    MKP(0, 1, 0)
>>>>> -#define __S011    MKP(0, 1, 1)
>>>>> -#define __S100    MKP(1, 0, 0)
>>>>> -#define __S101    MKP(1, 0, 1)
>>>>> -#define __S110    MKP(1, 1, 0)
>>>>> -#define __S111    MKP(1, 1, 1)
>>>>>      /* Used all over the kernel */
>>>>>    #define PAGE_KERNEL __pgprot(_PAGE_PRESENT | _PAGE_CACHED | _PAGE_READ | \
>>>>> diff --git a/arch/nios2/mm/init.c b/arch/nios2/mm/init.c
>>>>> index 613fcaa5988a..311b2146a248 100644
>>>>> --- a/arch/nios2/mm/init.c
>>>>> +++ b/arch/nios2/mm/init.c
>>>>> @@ -124,3 +124,48 @@ const char *arch_vma_name(struct vm_area_struct *vma)
>>>>>    {
>>>>>        return (vma->vm_start == KUSER_BASE) ? "[kuser]" : NULL;
>>>>>    }
>>>>> +
>>>>> +pgprot_t vm_get_page_prot(unsigned long vm_flags)
>>>>> +{
>>>>> +    switch (vm_flags & (VM_READ | VM_WRITE | VM_EXEC | VM_SHARED)) {
>>>>> +    case VM_NONE:
>>>>> +        return MKP(0, 0, 0);
>>>>> +    case VM_READ:
>>>>> +        return MKP(0, 0, 1);
>>>>> +    /* COW */
>>>>> +    case VM_WRITE:
>>>>> +        return MKP(0, 0, 0);
>>>>> +    /* COW */
>>>>> +    case VM_WRITE | VM_READ:
>>>>> +        return MKP(0, 0, 1);
>>>>> +    case VM_EXEC:
>>>>> +        return MKP(1, 0, 0);
>>>>> +    case VM_EXEC | VM_READ:
>>>>> +        return MKP(1, 0, 1);
>>>>> +    /* COW */
>>>>> +    case VM_EXEC | VM_WRITE:
>>>>> +        return MKP(1, 0, 0);
>>>>> +    /* COW */
>>>>> +    case VM_EXEC | VM_WRITE | VM_READ:
>>>>> +        return MKP(1, 0, 1);
>>>>> +    case VM_SHARED:
>>>>> +        return MKP(0, 0, 0);
>>>>> +    case VM_SHARED | VM_READ:
>>>>> +        return MKP(0, 0, 1);
>>>>> +    case VM_SHARED | VM_WRITE:
>>>>> +        return MKP(0, 1, 0);
>>>>> +    case VM_SHARED | VM_WRITE | VM_READ:
>>>>> +        return MKP(0, 1, 1);
>>>>> +    case VM_SHARED | VM_EXEC:
>>>>> +        return MKP(1, 0, 0);
>>>>> +    case VM_SHARED | VM_EXEC | VM_READ:
>>>>> +        return MKP(1, 0, 1);
>>>>> +    case VM_SHARED | VM_EXEC | VM_WRITE:
>>>>> +        return MKP(1, 1, 0);
>>>>> +    case VM_SHARED | VM_EXEC | VM_WRITE | VM_READ:
>>>>> +        return MKP(1, 1, 1);
>>>>> +    default:
>>>>> +        BUILD_BUG();
>>>>> +    }
>>>>> +}
>>>>> +EXPORT_SYMBOL(vm_get_page_prot);
>>>>
>>>> I'm getting this compile error after applying this patch when build NIOS2:
>>>
>>> Hmm, that is strange.
>>>
>>> Did you apply the entire series or atleast upto the nios2 patch ? Generic
>>> vm_get_page_prot() should not be called (which is build complaining here)
>>> when ARCH_HAS_VM_GET_PAGE_PROT is already enabled on nios2 platform.
>>>
>>> Ran a quick build test on nios2 for the entire series and also just upto
>>> this particular patch, build was successful.
>>>
>>
>> Ok, I did not apply the whole series, just this patch.
>>
> 
> 
> Is someone taking this whole series or should I just take this patch?

I expect the series (latest instead) will go via the mm tree, but will
really appreciate your tags if you find the series acceptable.