From: Boshi Yu <boshiyu@linux.alibaba.com>
To: Leon Romanovsky <leon@kernel.org>
Cc: jgg@ziepe.ca, linux-rdma@vger.kernel.org,
chengyou@linux.alibaba.com, KaiShen@linux.alibaba.com
Subject: Re: [PATCH for-next 3/3] RDMA/erdma: Remove unnecessary __GFP_ZERO flag
Date: Wed, 13 Mar 2024 10:40:06 +0800 [thread overview]
Message-ID: <ZfESBtLmXfs27Ya-@xy-macbook.local> (raw)
In-Reply-To: <20240312105305.GR12921@unreal>
On Tue, Mar 12, 2024 at 12:53:05PM +0200, Leon Romanovsky wrote:
> On Mon, Mar 11, 2024 at 07:38:21PM +0800, Boshi Yu wrote:
> > From: Boshi Yu <boshiyu@linux.alibaba.com>
> >
> > The dma_alloc_coherent() interface automatically zero the memory returned.
>
> Can you please point to the DMA code which does that?
We have noticed a patchset which ensures that dma_alloc_coherent() always
returns zeroed memory. The url of this patchset is listed as below:
https://lore.kernel.org/all/20181214082515.14835-1-hch@lst.de/T/#m70c723c646004445713f31b7837f7e9d910c06f5
Besides, you may refer to commit 518a2f1925c3 ("dma-mapping: zero memory returned from dma_alloc_*")
for details. This commit zeros memory by passing __GFP_ZERO flag or
calling memset internally. For example, the dma_alloc_direct() interface
calls memset() to zero the allocated memory.
Thanks,
Boshi Yu
>
> > Thus, we do not need to specify the __GFP_ZERO flag explicitly when we call
> > dma_alloc_coherent().
> >
> > Reviewed-by: Cheng Xu <chengyou@linux.alibaba.com>
> > Signed-off-by: Boshi Yu <boshiyu@linux.alibaba.com>
> > ---
> > drivers/infiniband/hw/erdma/erdma_cmdq.c | 6 ++----
> > drivers/infiniband/hw/erdma/erdma_eq.c | 6 ++----
> > 2 files changed, 4 insertions(+), 8 deletions(-)
> >
> > diff --git a/drivers/infiniband/hw/erdma/erdma_cmdq.c b/drivers/infiniband/hw/erdma/erdma_cmdq.c
> > index 0ac2683cfccf..43ff40b5a09d 100644
> > --- a/drivers/infiniband/hw/erdma/erdma_cmdq.c
> > +++ b/drivers/infiniband/hw/erdma/erdma_cmdq.c
> > @@ -127,8 +127,7 @@ static int erdma_cmdq_cq_init(struct erdma_dev *dev)
> >
> > cq->depth = cmdq->sq.depth;
> > cq->qbuf = dma_alloc_coherent(&dev->pdev->dev, cq->depth << CQE_SHIFT,
> > - &cq->qbuf_dma_addr,
> > - GFP_KERNEL | __GFP_ZERO);
> > + &cq->qbuf_dma_addr, GFP_KERNEL);
> > if (!cq->qbuf)
> > return -ENOMEM;
> >
> > @@ -162,8 +161,7 @@ static int erdma_cmdq_eq_init(struct erdma_dev *dev)
> >
> > eq->depth = cmdq->max_outstandings;
> > eq->qbuf = dma_alloc_coherent(&dev->pdev->dev, eq->depth << EQE_SHIFT,
> > - &eq->qbuf_dma_addr,
> > - GFP_KERNEL | __GFP_ZERO);
> > + &eq->qbuf_dma_addr, GFP_KERNEL);
> > if (!eq->qbuf)
> > return -ENOMEM;
> >
> > diff --git a/drivers/infiniband/hw/erdma/erdma_eq.c b/drivers/infiniband/hw/erdma/erdma_eq.c
> > index 0a4746e6d05c..84ccdd8144c9 100644
> > --- a/drivers/infiniband/hw/erdma/erdma_eq.c
> > +++ b/drivers/infiniband/hw/erdma/erdma_eq.c
> > @@ -87,8 +87,7 @@ int erdma_aeq_init(struct erdma_dev *dev)
> > eq->depth = ERDMA_DEFAULT_EQ_DEPTH;
> >
> > eq->qbuf = dma_alloc_coherent(&dev->pdev->dev, eq->depth << EQE_SHIFT,
> > - &eq->qbuf_dma_addr,
> > - GFP_KERNEL | __GFP_ZERO);
> > + &eq->qbuf_dma_addr, GFP_KERNEL);
> > if (!eq->qbuf)
> > return -ENOMEM;
> >
> > @@ -237,8 +236,7 @@ static int erdma_ceq_init_one(struct erdma_dev *dev, u16 ceqn)
> >
> > eq->depth = ERDMA_DEFAULT_EQ_DEPTH;
> > eq->qbuf = dma_alloc_coherent(&dev->pdev->dev, eq->depth << EQE_SHIFT,
> > - &eq->qbuf_dma_addr,
> > - GFP_KERNEL | __GFP_ZERO);
> > + &eq->qbuf_dma_addr, GFP_KERNEL);
> > if (!eq->qbuf)
> > return -ENOMEM;
> >
> > --
> > 2.39.3
> >
> >
next prev parent reply other threads:[~2024-03-13 2:40 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-03-11 11:38 [PATCH for-next 0/3] RDMA/erdma: A series of fixes for the erdma driver Boshi Yu
2024-03-11 11:38 ` [PATCH for-next 1/3] RDMA/erdma: Allocate doorbell records from dma pool Boshi Yu
2024-03-13 16:20 ` Zhu Yanjun
2024-03-14 10:23 ` Zhu Yanjun
2024-03-11 11:38 ` [PATCH for-next 2/3] RDMA/erdma: Unify the names related to doorbell records Boshi Yu
2024-03-11 11:38 ` [PATCH for-next 3/3] RDMA/erdma: Remove unnecessary __GFP_ZERO flag Boshi Yu
2024-03-12 10:53 ` Leon Romanovsky
2024-03-13 2:40 ` Boshi Yu [this message]
2024-03-13 9:31 ` Leon Romanovsky
2024-03-12 10:54 ` [PATCH for-next 0/3] RDMA/erdma: A series of fixes for the erdma driver Leon Romanovsky
2024-04-01 11:46 ` Leon Romanovsky
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ZfESBtLmXfs27Ya-@xy-macbook.local \
--to=boshiyu@linux.alibaba.com \
--cc=KaiShen@linux.alibaba.com \
--cc=chengyou@linux.alibaba.com \
--cc=jgg@ziepe.ca \
--cc=leon@kernel.org \
--cc=linux-rdma@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).