Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH 6/9] dma-remap: support DMA_ATTR_NO_KERNEL_MAPPING
To:     Christoph Hellwig <hch@lst.de>, iommu@lists.linux-foundation.org
Cc:     Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will.deacon@arm.com>, Guo Ren <ren_guo@c-sky.com>,
        Laura Abbott <labbott@redhat.com>,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org
References: <20181105121931.13481-1-hch@lst.de>
 <20181105121931.13481-7-hch@lst.de>
From:   Robin Murphy <robin.murphy@arm.com>
Message-ID: <e993440c-53bf-8071-b1b6-06d022d39e4e@arm.com>
Date:   Fri, 30 Nov 2018 19:05:40 +0000
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.2.1
MIME-Version: 1.0
In-Reply-To: <20181105121931.13481-7-hch@lst.de>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-GB
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 05/11/2018 12:19, Christoph Hellwig wrote:
> Do not waste vmalloc space on allocations that do not require a mapping
> into the kernel address space.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>   kernel/dma/remap.c | 11 +++++++++--
>   1 file changed, 9 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/dma/remap.c b/kernel/dma/remap.c
> index 8f1fca34b894..10a545126b0b 100644
> --- a/kernel/dma/remap.c
> +++ b/kernel/dma/remap.c
> @@ -200,7 +200,8 @@ void *arch_dma_alloc(struct device *dev, size_t size, dma_addr_t *dma_handle,
>   
>   	size = PAGE_ALIGN(size);
>   
> -	if (!gfpflags_allow_blocking(flags)) {
> +	if (!gfpflags_allow_blocking(flags) &&
> +	    !(attrs & DMA_ATTR_NO_KERNEL_MAPPING)) {
>   		ret = dma_alloc_from_pool(size, &page, flags);
>   		if (!ret)
>   			return NULL;
> @@ -215,6 +216,9 @@ void *arch_dma_alloc(struct device *dev, size_t size, dma_addr_t *dma_handle,
>   	/* remove any dirty cache lines on the kernel alias */
>   	arch_dma_prep_coherent(page, size);
>   
> +	if (attrs & DMA_ATTR_NO_KERNEL_MAPPING)
> +		return page; /* opaqueue cookie */

Best to preempt the inevitable patch fixing that typo in 3 months' time. 
Otherwise,

Reviewed-by: Robin Murphy <robin.murphy@arm.com>

> +
>   	/* create a coherent mapping */
>   	ret = dma_common_contiguous_remap(page, size, VM_USERMAP,
>   			arch_dma_mmap_pgprot(dev, PAGE_KERNEL, attrs),
> @@ -227,7 +231,10 @@ void *arch_dma_alloc(struct device *dev, size_t size, dma_addr_t *dma_handle,
>   void arch_dma_free(struct device *dev, size_t size, void *vaddr,
>   		dma_addr_t dma_handle, unsigned long attrs)
>   {
> -	if (!dma_free_from_pool(vaddr, PAGE_ALIGN(size))) {
> +	if (attrs & DMA_ATTR_NO_KERNEL_MAPPING) {
> +		/* vaddr is a struct page cookie, not a kernel address */
> +		__dma_direct_free_pages(dev, size, vaddr);
> +	} else if (!dma_free_from_pool(vaddr, PAGE_ALIGN(size))) {
>   		phys_addr_t phys = dma_to_phys(dev, dma_handle);
>   		struct page *page = pfn_to_page(__phys_to_pfn(phys));
>   
>