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;
MIME-Version: 1.0
References: <20210929041905.126454-1-mie@igel.co.jp> <20210929041905.126454-3-mie@igel.co.jp>
 <YVXMkSDXybju88TU@phenom.ffwll.local> <CANXvt5rD82Lvvag_k9k+XE-Sj1S6Qwp5uf+-feUTvez1-t4xUA@mail.gmail.com>
 <YWbGFkzkFRHmBcpa@phenom.ffwll.local>
In-Reply-To: <YWbGFkzkFRHmBcpa@phenom.ffwll.local>
From:   Shunsuke Mie <mie@igel.co.jp>
Date:   Thu, 14 Oct 2021 08:10:22 +0900
Message-ID: <CANXvt5p3Oo7VBq1DSDaPH8QexAus9+Q4DuzFo0JO583C7ZS9Gw@mail.gmail.com>
Subject: Re: [RFC PATCH v2 2/2] RDMA/rxe: Add dma-buf support
To:     Shunsuke Mie <mie@igel.co.jp>, Zhu Yanjun <zyjzyj2000@gmail.com>,
        =?UTF-8?Q?Christian_K=C3=B6nig?= <christian.koenig@amd.com>,
        Alex Deucher <alexander.deucher@amd.com>,
        Doug Ledford <dledford@redhat.com>,
        Jason Gunthorpe <jgg@ziepe.ca>,
        Jianxin Xiong <jianxin.xiong@intel.com>,
        Leon Romanovsky <leon@kernel.org>,
        Maor Gottlieb <maorg@nvidia.com>,
        Sean Hefty <sean.hefty@intel.com>,
        Sumit Semwal <sumit.semwal@linaro.org>,
        dri-devel@lists.freedesktop.org, linaro-mm-sig@lists.linaro.org,
        linux-media@vger.kernel.org,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        linux-rdma <linux-rdma@vger.kernel.org>,
        Damian Hobson-Garcia <dhobsong@igel.co.jp>,
        Takanari Hayama <taki@igel.co.jp>,
        Tomohito Esaki <etom@igel.co.jp>
Cc:     Daniel Vetter <daniel.vetter@ffwll.ch>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

2021=E5=B9=B410=E6=9C=8813=E6=97=A5(=E6=B0=B4) 20:42 Daniel Vetter <daniel@=
ffwll.ch>:
>
> On Fri, Oct 01, 2021 at 12:56:48PM +0900, Shunsuke Mie wrote:
> > 2021=E5=B9=B49=E6=9C=8830=E6=97=A5(=E6=9C=A8) 23:41 Daniel Vetter <dani=
el@ffwll.ch>:
> > >
> > > On Wed, Sep 29, 2021 at 01:19:05PM +0900, Shunsuke Mie wrote:
> > > > Implement a ib device operation =E2=80=98reg_user_mr_dmabuf=E2=80=
=99. Generate a
> > > > rxe_map from the memory space linked the passed dma-buf.
> > > >
> > > > Signed-off-by: Shunsuke Mie <mie@igel.co.jp>
> > > > ---
> > > >  drivers/infiniband/sw/rxe/rxe_loc.h   |   2 +
> > > >  drivers/infiniband/sw/rxe/rxe_mr.c    | 118 ++++++++++++++++++++++=
++++
> > > >  drivers/infiniband/sw/rxe/rxe_verbs.c |  34 ++++++++
> > > >  drivers/infiniband/sw/rxe/rxe_verbs.h |   2 +
> > > >  4 files changed, 156 insertions(+)
> > > >
> > > > diff --git a/drivers/infiniband/sw/rxe/rxe_loc.h b/drivers/infiniba=
nd/sw/rxe/rxe_loc.h
> > > > index 1ca43b859d80..8bc19ea1a376 100644
> > > > --- a/drivers/infiniband/sw/rxe/rxe_loc.h
> > > > +++ b/drivers/infiniband/sw/rxe/rxe_loc.h
> > > > @@ -75,6 +75,8 @@ u8 rxe_get_next_key(u32 last_key);
> > > >  void rxe_mr_init_dma(struct rxe_pd *pd, int access, struct rxe_mr =
*mr);
> > > >  int rxe_mr_init_user(struct rxe_pd *pd, u64 start, u64 length, u64=
 iova,
> > > >                    int access, struct rxe_mr *mr);
> > > > +int rxe_mr_dmabuf_init_user(struct rxe_pd *pd, int fd, u64 start, =
u64 length,
> > > > +                         u64 iova, int access, struct rxe_mr *mr);
> > > >  int rxe_mr_init_fast(struct rxe_pd *pd, int max_pages, struct rxe_=
mr *mr);
> > > >  int rxe_mr_copy(struct rxe_mr *mr, u64 iova, void *addr, int lengt=
h,
> > > >               enum rxe_mr_copy_dir dir);
> > > > diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniban=
d/sw/rxe/rxe_mr.c
> > > > index 53271df10e47..af6ef671c3a5 100644
> > > > --- a/drivers/infiniband/sw/rxe/rxe_mr.c
> > > > +++ b/drivers/infiniband/sw/rxe/rxe_mr.c
> > > > @@ -4,6 +4,7 @@
> > > >   * Copyright (c) 2015 System Fabric Works, Inc. All rights reserve=
d.
> > > >   */
> > > >
> > > > +#include <linux/dma-buf.h>
> > > >  #include "rxe.h"
> > > >  #include "rxe_loc.h"
> > > >
> > > > @@ -245,6 +246,120 @@ int rxe_mr_init_user(struct rxe_pd *pd, u64 s=
tart, u64 length, u64 iova,
> > > >       return err;
> > > >  }
> > > >
> > > > +static int rxe_map_dmabuf_mr(struct rxe_mr *mr,
> > > > +                          struct ib_umem_dmabuf *umem_dmabuf)
> > > > +{
> > > > +     struct rxe_map_set *set;
> > > > +     struct rxe_phys_buf *buf =3D NULL;
> > > > +     struct rxe_map **map;
> > > > +     void *vaddr, *vaddr_end;
> > > > +     int num_buf =3D 0;
> > > > +     int err;
> > > > +     size_t remain;
> > > > +
> > > > +     mr->dmabuf_map =3D kzalloc(sizeof &mr->dmabuf_map, GFP_KERNEL=
);
> > >
> > > dmabuf_maps are just tagged pointers (and we could shrink them to act=
ually
> > > just a tagged pointer if anyone cares about the overhead of the separ=
ate
> > > bool), allocating them seperately is overkill.
> >
> > I agree with you. However, I think it is needed to unmap by
> > dma_buf_vunmap(). If there is another simple way to unmap it. It is not
> > needed I think. What do you think about it?
>
> dma_buf_vunmap does not kfree the dma_buf_map argument, so that's no
> reason to allocate it separately. Or I'm confused.
I had a misunderstood. Yes, It is not needed to allocate an object.
Actually some
implementations don't alloc/free the argument.
e.g. gpu/drm/drm_gem_cma_helper.c
I'll fix it.

> Also apologies, I'm way behind on mails.
No problem. Thank you for your answer.

Thanks,
Shunsuke

> -Daniel
>
> >
> > > > +     if (!mr->dmabuf_map) {
> > > > +             err =3D -ENOMEM;
> > > > +             goto err_out;
> > > > +     }
> > > > +
> > > > +     err =3D dma_buf_vmap(umem_dmabuf->dmabuf, mr->dmabuf_map);
> > > > +     if (err)
> > > > +             goto err_free_dmabuf_map;
> > > > +
> > > > +     set =3D mr->cur_map_set;
> > > > +     set->page_shift =3D PAGE_SHIFT;
> > > > +     set->page_mask =3D PAGE_SIZE - 1;
> > > > +
> > > > +     map =3D set->map;
> > > > +     buf =3D map[0]->buf;
> > > > +
> > > > +     vaddr =3D mr->dmabuf_map->vaddr;
> > >
> > > dma_buf_map can be an __iomem too, you shouldn't dig around in this, =
but
> > > use the dma-buf-map.h helpers instead. On x86 (and I think also on mo=
st
> > > arm) it doesn't matter, but it's kinda not very nice in a pure softwa=
re
> > > driver.
> > >
> > > If anything is missing in dma-buf-map.h wrappers just add more.
> > >
> > > Or alternatively you need to fail the import if you can't handle __io=
mem.
> > >
> > > Aside from these I think the dma-buf side here for cpu access looks
> > > reasonable now.
> > > -Daniel
> > I'll see the dma-buf-map.h and consider the error handling that you sug=
gested.
> > I appreciate your support.
> >
> > Thanks a lot,
> > Shunsuke.
> >
> > > > +     vaddr_end =3D vaddr + umem_dmabuf->dmabuf->size;
> > > > +     remain =3D umem_dmabuf->dmabuf->size;
> > > > +
> > > > +     for (; remain; vaddr +=3D PAGE_SIZE) {
> > > > +             if (num_buf >=3D RXE_BUF_PER_MAP) {
> > > > +                     map++;
> > > > +                     buf =3D map[0]->buf;
> > > > +                     num_buf =3D 0;
> > > > +             }
> > > > +
> > > > +             buf->addr =3D (uintptr_t)vaddr;
> > > > +             if (remain >=3D PAGE_SIZE)
> > > > +                     buf->size =3D PAGE_SIZE;
> > > > +             else
> > > > +                     buf->size =3D remain;
> > > > +             remain -=3D buf->size;
> > > > +
> > > > +             num_buf++;
> > > > +             buf++;
> > > > +     }
> > > > +
> > > > +     return 0;
> > > > +
> > > > +err_free_dmabuf_map:
> > > > +     kfree(mr->dmabuf_map);
> > > > +err_out:
> > > > +     return err;
> > > > +}
> > > > +
> > > > +static void rxe_unmap_dmabuf_mr(struct rxe_mr *mr)
> > > > +{
> > > > +     struct ib_umem_dmabuf *umem_dmabuf =3D to_ib_umem_dmabuf(mr->=
umem);
> > > > +
> > > > +     dma_buf_vunmap(umem_dmabuf->dmabuf, mr->dmabuf_map);
> > > > +     kfree(mr->dmabuf_map);
> > > > +}
> > > > +
> > > > +int rxe_mr_dmabuf_init_user(struct rxe_pd *pd, int fd, u64 start, =
u64 length,
> > > > +                         u64 iova, int access, struct rxe_mr *mr)
> > > > +{
> > > > +     struct ib_umem_dmabuf *umem_dmabuf;
> > > > +     struct rxe_map_set *set;
> > > > +     int err;
> > > > +
> > > > +     umem_dmabuf =3D ib_umem_dmabuf_get(pd->ibpd.device, start, le=
ngth, fd,
> > > > +                                      access, NULL);
> > > > +     if (IS_ERR(umem_dmabuf)) {
> > > > +             err =3D PTR_ERR(umem_dmabuf);
> > > > +             goto err_out;
> > > > +     }
> > > > +
> > > > +     rxe_mr_init(access, mr);
> > > > +
> > > > +     err =3D rxe_mr_alloc(mr, ib_umem_num_pages(&umem_dmabuf->umem=
), 0);
> > > > +     if (err) {
> > > > +             pr_warn("%s: Unable to allocate memory for map\n", __=
func__);
> > > > +             goto err_release_umem;
> > > > +     }
> > > > +
> > > > +     mr->ibmr.pd =3D &pd->ibpd;
> > > > +     mr->umem =3D &umem_dmabuf->umem;
> > > > +     mr->access =3D access;
> > > > +     mr->state =3D RXE_MR_STATE_VALID;
> > > > +     mr->type =3D IB_MR_TYPE_USER;
> > > > +
> > > > +     set =3D mr->cur_map_set;
> > > > +     set->length =3D length;
> > > > +     set->iova =3D iova;
> > > > +     set->va =3D start;
> > > > +     set->offset =3D ib_umem_offset(mr->umem);
> > > > +
> > > > +     err =3D rxe_map_dmabuf_mr(mr, umem_dmabuf);
> > > > +     if (err)
> > > > +             goto err_free_map_set;
> > > > +
> > > > +     return 0;
> > > > +
> > > > +err_free_map_set:
> > > > +     rxe_mr_free_map_set(mr->num_map, mr->cur_map_set);
> > > > +err_release_umem:
> > > > +     ib_umem_release(&umem_dmabuf->umem);
> > > > +err_out:
> > > > +     return err;
> > > > +}
> > > > +
> > > >  int rxe_mr_init_fast(struct rxe_pd *pd, int max_pages, struct rxe_=
mr *mr)
> > > >  {
> > > >       int err;
> > > > @@ -703,6 +818,9 @@ void rxe_mr_cleanup(struct rxe_pool_entry *arg)
> > > >  {
> > > >       struct rxe_mr *mr =3D container_of(arg, typeof(*mr), pelem);
> > > >
> > > > +     if (mr->umem && mr->umem->is_dmabuf)
> > > > +             rxe_unmap_dmabuf_mr(mr);
> > > > +
> > > >       ib_umem_release(mr->umem);
> > > >
> > > >       if (mr->cur_map_set)
> > > > diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.c b/drivers/infini=
band/sw/rxe/rxe_verbs.c
> > > > index 9d0bb9aa7514..6191bb4f434d 100644
> > > > --- a/drivers/infiniband/sw/rxe/rxe_verbs.c
> > > > +++ b/drivers/infiniband/sw/rxe/rxe_verbs.c
> > > > @@ -916,6 +916,39 @@ static struct ib_mr *rxe_reg_user_mr(struct ib=
_pd *ibpd,
> > > >       return ERR_PTR(err);
> > > >  }
> > > >
> > > > +static struct ib_mr *rxe_reg_user_mr_dmabuf(struct ib_pd *ibpd, u6=
4 start,
> > > > +                                         u64 length, u64 iova, int=
 fd,
> > > > +                                         int access, struct ib_uda=
ta *udata)
> > > > +{
> > > > +     int err;
> > > > +     struct rxe_dev *rxe =3D to_rdev(ibpd->device);
> > > > +     struct rxe_pd *pd =3D to_rpd(ibpd);
> > > > +     struct rxe_mr *mr;
> > > > +
> > > > +     mr =3D rxe_alloc(&rxe->mr_pool);
> > > > +     if (!mr) {
> > > > +             err =3D -ENOMEM;
> > > > +             goto err2;
> > > > +     }
> > > > +
> > > > +     rxe_add_index(mr);
> > > > +
> > > > +     rxe_add_ref(pd);
> > > > +
> > > > +     err =3D rxe_mr_dmabuf_init_user(pd, fd, start, length, iova, =
access, mr);
> > > > +     if (err)
> > > > +             goto err3;
> > > > +
> > > > +     return &mr->ibmr;
> > > > +
> > > > +err3:
> > > > +     rxe_drop_ref(pd);
> > > > +     rxe_drop_index(mr);
> > > > +     rxe_drop_ref(mr);
> > > > +err2:
> > > > +     return ERR_PTR(err);
> > > > +}
> > > > +
> > > >  static struct ib_mr *rxe_alloc_mr(struct ib_pd *ibpd, enum ib_mr_t=
ype mr_type,
> > > >                                 u32 max_num_sg)
> > > >  {
> > > > @@ -1081,6 +1114,7 @@ static const struct ib_device_ops rxe_dev_ops=
 =3D {
> > > >       .query_qp =3D rxe_query_qp,
> > > >       .query_srq =3D rxe_query_srq,
> > > >       .reg_user_mr =3D rxe_reg_user_mr,
> > > > +     .reg_user_mr_dmabuf =3D rxe_reg_user_mr_dmabuf,
> > > >       .req_notify_cq =3D rxe_req_notify_cq,
> > > >       .resize_cq =3D rxe_resize_cq,
> > > >
> > > > diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.h b/drivers/infini=
band/sw/rxe/rxe_verbs.h
> > > > index c807639435eb..0aa95ab06b6e 100644
> > > > --- a/drivers/infiniband/sw/rxe/rxe_verbs.h
> > > > +++ b/drivers/infiniband/sw/rxe/rxe_verbs.h
> > > > @@ -334,6 +334,8 @@ struct rxe_mr {
> > > >
> > > >       struct rxe_map_set      *cur_map_set;
> > > >       struct rxe_map_set      *next_map_set;
> > > > +
> > > > +     struct dma_buf_map *dmabuf_map;
> > > >  };
> > > >
> > > >  enum rxe_mw_state {
> > > > --
> > > > 2.17.1
> > > >
> > >
> > > --
> > > Daniel Vetter
> > > Software Engineer, Intel Corporation
> > > http://blog.ffwll.ch
>
> --
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch