[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20211011070405.GA173@LAPTOP-UKSR4ENP.internal.baidu.com>
Date: Mon, 11 Oct 2021 15:04:05 +0800
From: Cai Huoqing <caihuoqing@...du.com>
To: Christoph Hellwig <hch@...radead.org>
CC: Michael Cyr <mikecyr@...ux.ibm.com>,
"James E.J. Bottomley" <jejb@...ux.ibm.com>,
"Martin K. Petersen" <martin.petersen@...cle.com>,
<linux-scsi@...r.kernel.org>, <target-devel@...r.kernel.org>,
<linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] scsi: ibmvscsi_tgt: Use dma_alloc_coherent() instead of
get_zeroed_page/dma_map_single()
On 11 10月 21 07:42:28, Christoph Hellwig wrote:
> On Mon, Oct 11, 2021 at 12:00:53AM +0800, Cai Huoqing wrote:
> > Replacing get_zeroed_page/free_page/dma_map_single/dma_unmap_single()
> > with dma_alloc_coherent/dma_free_coherent() helps to reduce
> > code size, and simplify the code, and coherent DMA will not
> > clear the cache every time.
>
> This explanation does not make any sense whatsoever. Please explain
> why it would an show numbers.
Hi Christoph,
thanks for your feedback.
In this case, just simplify the code.
10 insertions(+), 31 deletions(-)
dma_sync_ API is not called, I think the hardware may keep cache coherent
directly or is a no cache system. No need to make perfermance compare.
thanks
Cai
Powered by blists - more mailing lists