Thanks to all of you<br>
<br><br><div><span class="gmail_quote">On 3/27/06, <b class="gmail_sendername">Roland Dreier</b> <<a href="mailto:rdreier@cisco.com">rdreier@cisco.com</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Devesh> Hello all, Please any body explain me about the<br> Devesh> functionality of verbs ib_get_dma_mr()?<br><br>Actually, the responses you've gotten are not quite right.<br>ib_get_dma_mr() returns a memory region that can be used for any _bus_
<br>addresses. In other words, if an S/G entry is passed to the driver</blockquote><div><br>
S/G entry ? <br>
</div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">that uses the L_Key from ib_get_dma_mr() and an address of, say,<br>0xdeadbeef, then the RDMA device should use a bus address of
<br>0xdeadbeef to access that memory.</blockquote><div><br>
What is the size of this region ? is there any limitation in providing this size?<br>
</div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">The difference between bus addresses and physical addresses is<br>significant when IOMMUs are present.
<br><br>This is somewhat similar to the verbs extensions notion of "reserved<br>L_Key," except that it also provides an R_Key and the ability to<br>specify the access permissions of the region.</blockquote><div>
<br>
Finally you mean to say in the implementation of this
function providing a unique L_Key and R_Key is sufficient. Is it?<br>
</div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> - R.<br></blockquote></div><br>