<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body>
    Am 31.12.21 um 09:37 schrieb 温志伟:<br>
    <blockquote type="cite" cite="mid:xfsceflzgr-xftmc911ak@nsmail6.0">
      
      <p><br>
      </p>
      <div class="viewer_part">
        <div>   <br>
          <p>Then the judgment in line 944 is needed. Are the
            possibilities of these two positions consistent.</p>
          <p>Why keep this judgment?</p>
        </div>
      </div>
    </blockquote>
    <br>
    Because when mem is non-NULL bo_va->bo can't be NULL either.<br>
    <br>
    Regards,<br>
    Christian.<br>
    <br>
    <blockquote type="cite" cite="mid:xfsceflzgr-xftmc911ak@nsmail6.0">
      <div class="viewer_part">
        <div>
          <p><br>
          </p>
          <p>Regards,</p>
          <p>Wen Zhiwei.</p>
          <p><br>
          </p>
             
          <p>----</p>
          <p><br>
          </p>
          <div id="re" style="margin-left:0.5em;padding-left:0.5em;border-left:1px
            solid green;"><br>
            <br>
            <br>
                   
            <div style="background-color:#f5f7fa"><b>主 题:</b><span id="subject">Re: [PATCH] gpu/drm/radeon:Fix null pointer
                risk</span>            <br>
              <b>日 期:</b><span id="date">2021-12-31 00:36</span>        
                 <br>
              <b>发件人:</b><span id="from">Christian König</span>        
                 <br>
              <b>收件人:</b><span id="to">Wen
<a class="moz-txt-link-abbreviated" href="mailto:Zhiweialexander.deucher@amd.comXinhui.Pan@amd.comairlied@linux.iedaniel@ffwll.ch">Zhiweialexander.deucher@amd.comXinhui.Pan@amd.comairlied@linux.iedaniel@ffwll.ch</a></span>
                                 </div>
            <br>
                   
            <div id="content">
              <div class="viewer_part">
                <div>Am 28.12.21 um 08:31 schrieb Wen Zhiwei:<br>
                  > If the null pointer is not judged in advance,<br>
                  > there is a risk that the pointer will cross<br>
                  > the boundary<br>
                  <br>
                  As far as I can see that case is impossible, why do
                  you want to add a <br>
                  check for it?<br>
                  <br>
                  Regards,<br>
                  Christian.<br>
                  <br>
                  ><br>
                  > Signed-off-by: Wen Zhiwei<wenzhiwei@kylinos.cn><br>
                    > ---<br>
                    >   drivers/gpu/drm/radeon/radeon_vm.c | 2 +-<br>
                    >   1 file changed, 1 insertion(+), 1 deletion(-)<br>
                    ><br>
                    > diff --git a/drivers/gpu/drm/radeon/radeon_vm.c
                    b/drivers/gpu/drm/radeon/radeon_vm.c<br>
                    > index bb53016f3138..d3d342041adf 100644<br>
                    > --- a/drivers/gpu/drm/radeon/radeon_vm.c<br>
                    > +++ b/drivers/gpu/drm/radeon/radeon_vm.c<br>
                    > @@ -951,7 +951,7 @@ int
                    radeon_vm_bo_update(struct radeon_device *rdev,<br>
                    >   <br>
                    >   if (mem->mem_type == TTM_PL_TT) {<br>
                    >   bo_va->flags |= RADEON_VM_PAGE_SYSTEM;<br>
                    > - if (!(bo_va->bo->flags &
                    (RADEON_GEM_GTT_WC | RADEON_GEM_GTT_UC)))<br>
                    > + if (bo_va->bo &&
                    !(bo_va->bo->flags & (RADEON_GEM_GTT_WC |
                    RADEON_GEM_GTT_UC)))<br>
                    >   bo_va->flags |= RADEON_VM_PAGE_SNOOPED;<br>
                    >   <br>
                    >   } else {<br>
                    <br>
                  </wenzhiwei@kylinos.cn></div>
              </div>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
  </body>
</html>