From: kernel test robot <lkp@intel.com>
To: Jianxin Xiong <jianxin.xiong@intel.com>,
linux-rdma@vger.kernel.org, dri-devel@lists.freedesktop.org
Cc: kbuild-all@lists.01.org, Leon Romanovsky <leon@kernel.org>,
Jason Gunthorpe <jgg@ziepe.ca>,
Doug Ledford <dledford@redhat.com>,
Daniel Vetter <daniel.vetter@intel.com>,
Christian Koenig <christian.koenig@amd.com>,
Jianxin Xiong <jianxin.xiong@intel.com>
Subject: Re: [PATCH v15 1/4] RDMA/umem: Support importing dma-buf as user memory region
Date: Mon, 14 Dec 2020 14:43:34 +0800 [thread overview]
Message-ID: <202012141456.2b3HY1yU-lkp@intel.com> (raw)
In-Reply-To: <1607629058-141952-2-git-send-email-jianxin.xiong@intel.com>
[-- Attachment #1: Type: text/plain, Size: 7653 bytes --]
Hi Jianxin,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on rdma/for-next]
[also build test ERROR on next-20201211]
[cannot apply to tegra-drm/drm/tegra/for-next v5.10]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Jianxin-Xiong/RDMA-Add-dma-buf-support/20201211-032359
base: https://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma.git for-next
config: i386-randconfig-r025-20200629 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-15) 9.3.0
reproduce (this is a W=1 build):
# https://github.com/0day-ci/linux/commit/de7b8a9035076ce9b065dd12d421bab311c1cc43
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Jianxin-Xiong/RDMA-Add-dma-buf-support/20201211-032359
git checkout de7b8a9035076ce9b065dd12d421bab311c1cc43
# save the attached .config to linux build tree
make W=1 ARCH=i386
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
All errors (new ones prefixed by >>):
ld: drivers/infiniband/core/umem_dmabuf.o: in function `dma_fence_wait':
>> include/linux/dma-fence.h:581: undefined reference to `dma_fence_wait_timeout'
ld: drivers/infiniband/core/umem_dmabuf.o: in function `ib_umem_dmabuf_map_pages':
>> drivers/infiniband/core/umem_dmabuf.c:26: undefined reference to `dma_buf_map_attachment'
ld: drivers/infiniband/core/umem_dmabuf.o: in function `ib_umem_dmabuf_unmap_pages':
>> drivers/infiniband/core/umem_dmabuf.c:100: undefined reference to `dma_buf_unmap_attachment'
ld: drivers/infiniband/core/umem_dmabuf.o: in function `ib_umem_dmabuf_get':
>> drivers/infiniband/core/umem_dmabuf.c:124: undefined reference to `dma_buf_get'
>> ld: drivers/infiniband/core/umem_dmabuf.c:162: undefined reference to `dma_buf_put'
>> ld: drivers/infiniband/core/umem_dmabuf.c:147: undefined reference to `dma_buf_dynamic_attach'
ld: drivers/infiniband/core/umem_dmabuf.o: in function `ib_umem_dmabuf_release':
>> drivers/infiniband/core/umem_dmabuf.c:171: undefined reference to `dma_buf_detach'
ld: drivers/infiniband/core/umem_dmabuf.c:172: undefined reference to `dma_buf_put'
vim +26 drivers/infiniband/core/umem_dmabuf.c
11
12 int ib_umem_dmabuf_map_pages(struct ib_umem_dmabuf *umem_dmabuf)
13 {
14 struct sg_table *sgt;
15 struct scatterlist *sg;
16 struct dma_fence *fence;
17 unsigned long start, end, cur = 0;
18 unsigned int nmap = 0;
19 int i;
20
21 dma_resv_assert_held(umem_dmabuf->attach->dmabuf->resv);
22
23 if (umem_dmabuf->sgt)
24 goto wait_fence;
25
> 26 sgt = dma_buf_map_attachment(umem_dmabuf->attach, DMA_BIDIRECTIONAL);
27 if (IS_ERR(sgt))
28 return PTR_ERR(sgt);
29
30 /* modify the sg list in-place to match umem address and length */
31
32 start = ALIGN_DOWN(umem_dmabuf->umem.address, PAGE_SIZE);
33 end = ALIGN(umem_dmabuf->umem.address + umem_dmabuf->umem.length,
34 PAGE_SIZE);
35 for_each_sgtable_dma_sg(sgt, sg, i) {
36 if (start < cur + sg_dma_len(sg) && cur < end)
37 nmap++;
38 if (cur <= start && start < cur + sg_dma_len(sg)) {
39 unsigned long offset = start - cur;
40
41 umem_dmabuf->first_sg = sg;
42 umem_dmabuf->first_sg_offset = offset;
43 sg_dma_address(sg) += offset;
44 sg_dma_len(sg) -= offset;
45 cur += offset;
46 }
47 if (cur < end && end <= cur + sg_dma_len(sg)) {
48 unsigned long trim = cur + sg_dma_len(sg) - end;
49
50 umem_dmabuf->last_sg = sg;
51 umem_dmabuf->last_sg_trim = trim;
52 sg_dma_len(sg) -= trim;
53 break;
54 }
55 cur += sg_dma_len(sg);
56 }
57
58 umem_dmabuf->umem.sg_head.sgl = umem_dmabuf->first_sg;
59 umem_dmabuf->umem.sg_head.nents = nmap;
60 umem_dmabuf->umem.nmap = nmap;
61 umem_dmabuf->sgt = sgt;
62
63 wait_fence:
64 /*
65 * Although the sg list is valid now, the content of the pages
66 * may be not up-to-date. Wait for the exporter to finish
67 * the migration.
68 */
69 fence = dma_resv_get_excl(umem_dmabuf->attach->dmabuf->resv);
70 if (fence)
71 return dma_fence_wait(fence, false);
72
73 return 0;
74 }
75 EXPORT_SYMBOL(ib_umem_dmabuf_map_pages);
76
77 void ib_umem_dmabuf_unmap_pages(struct ib_umem_dmabuf *umem_dmabuf)
78 {
79 dma_resv_assert_held(umem_dmabuf->attach->dmabuf->resv);
80
81 if (!umem_dmabuf->sgt)
82 return;
83
84 /* retore the original sg list */
85 if (umem_dmabuf->first_sg) {
86 sg_dma_address(umem_dmabuf->first_sg) -=
87 umem_dmabuf->first_sg_offset;
88 sg_dma_len(umem_dmabuf->first_sg) +=
89 umem_dmabuf->first_sg_offset;
90 umem_dmabuf->first_sg = NULL;
91 umem_dmabuf->first_sg_offset = 0;
92 }
93 if (umem_dmabuf->last_sg) {
94 sg_dma_len(umem_dmabuf->last_sg) +=
95 umem_dmabuf->last_sg_trim;
96 umem_dmabuf->last_sg = NULL;
97 umem_dmabuf->last_sg_trim = 0;
98 }
99
> 100 dma_buf_unmap_attachment(umem_dmabuf->attach, umem_dmabuf->sgt,
101 DMA_BIDIRECTIONAL);
102
103 umem_dmabuf->sgt = NULL;
104 }
105 EXPORT_SYMBOL(ib_umem_dmabuf_unmap_pages);
106
107 struct ib_umem_dmabuf *ib_umem_dmabuf_get(struct ib_device *device,
108 unsigned long offset, size_t size,
109 int fd, int access,
110 const struct dma_buf_attach_ops *ops)
111 {
112 struct dma_buf *dmabuf;
113 struct ib_umem_dmabuf *umem_dmabuf;
114 struct ib_umem *umem;
115 unsigned long end;
116 struct ib_umem_dmabuf *ret = ERR_PTR(-EINVAL);
117
118 if (check_add_overflow(offset, (unsigned long)size, &end))
119 return ret;
120
121 if (unlikely(!ops || !ops->move_notify))
122 return ret;
123
> 124 dmabuf = dma_buf_get(fd);
125 if (IS_ERR(dmabuf))
126 return ERR_CAST(dmabuf);
127
128 if (dmabuf->size < end)
129 goto out_release_dmabuf;
130
131 umem_dmabuf = kzalloc(sizeof(*umem_dmabuf), GFP_KERNEL);
132 if (!umem_dmabuf) {
133 ret = ERR_PTR(-ENOMEM);
134 goto out_release_dmabuf;
135 }
136
137 umem = &umem_dmabuf->umem;
138 umem->ibdev = device;
139 umem->length = size;
140 umem->address = offset;
141 umem->writable = ib_access_writable(access);
142 umem->is_dmabuf = 1;
143
144 if (!ib_umem_num_pages(umem))
145 goto out_free_umem;
146
> 147 umem_dmabuf->attach = dma_buf_dynamic_attach(
148 dmabuf,
149 device->dma_device,
150 ops,
151 umem_dmabuf);
152 if (IS_ERR(umem_dmabuf->attach)) {
153 ret = ERR_CAST(umem_dmabuf->attach);
154 goto out_free_umem;
155 }
156 return umem_dmabuf;
157
158 out_free_umem:
159 kfree(umem_dmabuf);
160
161 out_release_dmabuf:
> 162 dma_buf_put(dmabuf);
163 return ret;
164 }
165 EXPORT_SYMBOL(ib_umem_dmabuf_get);
166
167 void ib_umem_dmabuf_release(struct ib_umem_dmabuf *umem_dmabuf)
168 {
169 struct dma_buf *dmabuf = umem_dmabuf->attach->dmabuf;
170
> 171 dma_buf_detach(dmabuf, umem_dmabuf->attach);
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 30708 bytes --]
[-- Attachment #3: Type: text/plain, Size: 160 bytes --]
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
next prev parent reply other threads:[~2020-12-14 6:43 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-10 19:37 [PATCH v15 0/4] RDMA: Add dma-buf support Jianxin Xiong
2020-12-10 19:37 ` [PATCH v15 1/4] RDMA/umem: Support importing dma-buf as user memory region Jianxin Xiong
2020-12-14 6:43 ` kernel test robot [this message]
2020-12-10 19:37 ` [PATCH v15 2/4] RDMA/core: Add device method for registering dma-buf based " Jianxin Xiong
2020-12-10 19:37 ` [PATCH v15 3/4] RDMA/uverbs: Add uverbs command for dma-buf based MR registration Jianxin Xiong
2020-12-10 19:37 ` [PATCH v15 4/4] RDMA/mlx5: Support dma-buf based userspace memory region Jianxin Xiong
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=202012141456.2b3HY1yU-lkp@intel.com \
--to=lkp@intel.com \
--cc=christian.koenig@amd.com \
--cc=daniel.vetter@intel.com \
--cc=dledford@redhat.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=jgg@ziepe.ca \
--cc=jianxin.xiong@intel.com \
--cc=kbuild-all@lists.01.org \
--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