From: Jason Gunthorpe <jgg@ziepe.ca>
To: "Xiong, Jianxin" <jianxin.xiong@intel.com>
Cc: "linux-rdma@vger.kernel.org" <linux-rdma@vger.kernel.org>,
"dri-devel@lists.freedesktop.org"
<dri-devel@lists.freedesktop.org>,
Doug Ledford <dledford@redhat.com>,
Leon Romanovsky <leon@kernel.org>,
Sumit Semwal <sumit.semwal@linaro.org>,
Christian Koenig <christian.koenig@amd.com>,
"Vetter, Daniel" <daniel.vetter@intel.com>
Subject: Re: [PATCH v6 4/4] RDMA/mlx5: Support dma-buf based userspace memory region
Date: Wed, 28 Oct 2020 13:35:46 -0300 [thread overview]
Message-ID: <20201028163546.GY36674@ziepe.ca> (raw)
In-Reply-To: <MW3PR11MB45559D700788EFFE08E9B639E5160@MW3PR11MB4555.namprd11.prod.outlook.com>
On Tue, Oct 27, 2020 at 08:33:52PM +0000, Xiong, Jianxin wrote:
> > > @@ -801,6 +816,52 @@ static int pagefault_implicit_mr(struct mlx5_ib_mr *imr,
> > > * Returns:
> > > * -EFAULT: The io_virt->bcnt is not within the MR, it covers pages that are
> > > * not accessible, or the MR is no longer valid.
> > > + * -EAGAIN: The operation should be retried
> > > + *
> > > + * >0: Number of pages mapped
> > > + */
> > > +static int pagefault_dmabuf_mr(struct mlx5_ib_mr *mr, struct ib_umem *umem,
> > > + u64 io_virt, size_t bcnt, u32 *bytes_mapped,
> > > + u32 flags)
> > > +{
> > > + struct ib_umem_dmabuf *umem_dmabuf = to_ib_umem_dmabuf(umem);
> > > + u64 user_va;
> > > + u64 end;
> > > + int npages;
> > > + int err;
> > > +
> > > + if (unlikely(io_virt < mr->mmkey.iova))
> > > + return -EFAULT;
> > > + if (check_add_overflow(io_virt - mr->mmkey.iova,
> > > + (u64)umem->address, &user_va))
> > > + return -EFAULT;
> > > + /* Overflow has alreddy been checked at the umem creation time */
> > > + end = umem->address + umem->length;
> > > + if (unlikely(user_va >= end || end - user_va < bcnt))
> > > + return -EFAULT;
> >
> > Why duplicate this sequence? Caller does it
>
> The sequence in the caller is for umem_odp only.
Nothing about umem_odp in this code though??
> > > /* prefetch with write-access must be supported by the MR */
> > > if (advice == IB_UVERBS_ADVISE_MR_ADVICE_PREFETCH_WRITE &&
> > > - !odp->umem.writable)
> > > + !mr->umem->writable)
> >
> > ??
> There is no need to use umem_odp here, mr->umem is the same as &odp->umem.
> This change makes the code works for both umem_odp and umem_dmabuf.
Ok
Can you please also think about how to test this? I very much prefer
to see new pyverbs tests for new APIs.
Distros are running the rdma-core test suite, if you want this to work
widely we need a public test for it.
Thanks,
Jason
next prev parent reply other threads:[~2020-10-29 0:55 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-23 16:39 [PATCH v6 0/4] RDMA: Add dma-buf support Jianxin Xiong
2020-10-23 16:39 ` [PATCH v6 1/4] RDMA/umem: Support importing dma-buf as user memory region Jianxin Xiong
2020-10-23 16:49 ` Daniel Vetter
2020-10-23 18:09 ` Xiong, Jianxin
2020-10-23 18:13 ` Daniel Vetter
2020-10-23 18:20 ` Jason Gunthorpe
2020-10-24 1:45 ` Daniel Vetter
2020-11-03 17:36 ` Xiong, Jianxin
2020-11-03 20:43 ` Daniel Vetter
2020-11-04 0:01 ` Xiong, Jianxin
2020-11-05 15:02 ` Jason Gunthorpe
2020-10-24 7:48 ` Christoph Hellwig
2020-10-26 12:26 ` Jason Gunthorpe
2020-10-27 8:08 ` Christoph Hellwig
2020-10-27 17:32 ` Xiong, Jianxin
2020-10-27 19:51 ` Jason Gunthorpe
2020-10-27 20:00 ` Jason Gunthorpe
2020-10-27 20:11 ` Xiong, Jianxin
2020-10-23 16:39 ` [PATCH v6 2/4] RDMA/core: Add device method for registering dma-buf base " Jianxin Xiong
2020-10-23 16:40 ` [PATCH v6 3/4] RDMA/uverbs: Add uverbs command for dma-buf based MR registration Jianxin Xiong
2020-10-23 16:40 ` [PATCH v6 4/4] RDMA/mlx5: Support dma-buf based userspace memory region Jianxin Xiong
2020-10-27 20:08 ` Jason Gunthorpe
2020-10-27 20:33 ` Xiong, Jianxin
2020-10-28 16:35 ` Jason Gunthorpe [this message]
2020-10-28 17:29 ` Xiong, Jianxin
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=20201028163546.GY36674@ziepe.ca \
--to=jgg@ziepe.ca \
--cc=christian.koenig@amd.com \
--cc=daniel.vetter@intel.com \
--cc=dledford@redhat.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=jianxin.xiong@intel.com \
--cc=leon@kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=sumit.semwal@linaro.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 NNTP newsgroup(s).