Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 12 Oct 2021 06:20:32 +0100
From:   Christoph Hellwig <hch@infradead.org>
To:     Cai Huoqing <caihuoqing@baidu.com>
Cc:     hch@infradead.org, Michael Cyr <mikecyr@linux.ibm.com>,
        "James E.J. Bottomley" <jejb@linux.ibm.com>,
        "Martin K. Petersen" <martin.petersen@oracle.com>,
        linux-scsi@vger.kernel.org, target-devel@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] scsi: ibmvscsi_tgt: Use dma_alloc_noncoherent()
 instead of get_zeroed_page/dma_map_single()
Message-ID: <YWUbIBna0NAOJu62@infradead.org>
References: <20211012032110.2224-1-caihuoqing@baidu.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20211012032110.2224-1-caihuoqing@baidu.com>
Precedence: bulk

On Tue, Oct 12, 2021 at 11:21:09AM +0800, Cai Huoqing wrote:
> Replacing get_zeroed_page/free_page/dma_map_single/dma_unmap_single()
> with dma_alloc_noncoherent/dma_free_noncoherent() helps to reduce
> code size, and simplify the code, and the hardware keep DMA coherent
> itself.

It might also be worth to mention that this also avoids potential
bounce buffering.  Although for pseries vio devices this probably can't
happen anyway.

> +	vscsi->cmd_q.base_addr = dma_alloc_noncoherent(&vdev->dev, PAGE_SIZE,
> +						       &vscsi->cmd_q.crq_token,
> +						       DMA_BIDIRECTIONAL, GFP_KERNEL);

Please avoid the overly long line.

The same comments also apply to the other patch.