<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=gb2312">
<STYLE type=text/css>.style3 {
        COLOR: #0072a8
}
</STYLE>
<META content="MSHTML 6.00.6000.16809" name=GENERATOR></HEAD>
<BODY bottomMargin=5 bgColor=#ffffff leftMargin=5 topMargin=5 rightMargin=5
marginheight="0" marginwidth="0">
<DIV><FONT face="Courier New" size=2></FONT> </DIV>
<BLOCKQUOTE
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
<DIV style="FONT: 10pt arial">----- Original Message ----- </DIV>
<DIV
style="BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: black"><B>From:</B>
<A title=jli@altobeam.com href="mailto:jli@altobeam.com">jli@altobeam.com</A>
</DIV>
<DIV style="FONT: 10pt arial"><B>To:</B> <A title=winpcap-users@winpcap.org
href="mailto:winpcap-users@winpcap.org">winpcap-users@winpcap.org</A> </DIV>
<DIV style="FONT: 10pt arial"><B>Sent:</B> Sunday, March 15, 2009 8:39
PM</DIV>
<DIV style="FONT: 10pt arial"><B>Subject:</B> [Winpcap-users] packet lost at
480Mbps</DIV>
<DIV><BR></DIV>
<DIV><BR><BR>Hi guys:<BR> Last mail has some format problems,and
i'm sorry for that .<BR><!-- --> I use wireshark to capture data
at 480Mbps and found packets lost when wireshark indicate that "Drop" is
zero and "Receive but no buffer" is not zero. <BR><!-- -->So I learned the
sourcecode of wireshark and winpcap, found that "Receive but no buffer" in
wireshark come from PacketRequest() in packet32.dll using DeviceIoControl()
and the function code is "OID_GEN_RCV_NO_BUFFER", which MSDN explains as
follows:<BR><BR>As a query, the OID_GEN_RCV_NO_BUFFER OID specifies the number
of frames that the NIC cannot receive due to lack of NIC receive buffer space.
Some<BR>NICs do not provide the exact number of missed frames; they provide
only the number of times at least one frame is missed.<BR><BR>Is that means my
NIC hardware buffer is too small?</DIV>
<DIV> </DIV></BLOCKQUOTE>
<DIV><FONT face="Courier New" size=2>Yeah. Actually "Receive buffer space" is a
generic term. Most probably the DMA receive buffers (that are in the RAM of your
PC) are not big enough, or the PCI/PCIe/USB bus on which the NIC is installed is
not fast enough to transfer all the data from the device to the
host.</FONT></DIV>
<BLOCKQUOTE
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"><FONT
face="Courier New" size=2></FONT>
<DIV><BR>Can i improve reading speed by recompiling npf.sys to fix this
problem?</DIV>
<DIV> </DIV></BLOCKQUOTE>
<DIV><FONT face="Courier New" size=2>No</FONT></DIV>
<BLOCKQUOTE
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"><FONT
face="Courier New" size=2></FONT>
<DIV><BR><!-- -->If i use packet.lib, what should i do to make read operation
more fast besides decrease value of mintocopy ?<BR></DIV><!-- --></BLOCKQUOTE>
<DIV><FONT face="Courier New" size=2>Decreasing the mintocopy can actually
impact the perfomance in the opposite way: if mintocopy is low, the number of
user/kernel </FONT><FONT face="Courier New" size=2>transitions (i.e. system
calls) to move packets from kernel mode to user mode gets higher and can impact
the CPU load on the system.</FONT></DIV>
<DIV><FONT face="Courier New" size=2></FONT> </DIV>
<DIV><FONT face="Courier New" size=2>Have a nice day</FONT></DIV>
<DIV><FONT face="Courier New" size=2>GV</FONT></DIV>
<BLOCKQUOTE
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
<DIV> </DIV><FONT face="Courier New" size=2></FONT></BLOCKQUOTE>
<BLOCKQUOTE
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
<DIV><!-- -->Here are some information of my laptop: WinXp sp3, NDIS5.1, dual
CPU at 2GHz, 2G memory, Realtek RTL8168C(P)/8111C(P) PCI-E Gagibit Ethernet
NIC<BR></DIV><!-- -->
<P>
<HR>
<P></P>_______________________________________________<BR>Winpcap-users
mailing
list<BR>Winpcap-users@winpcap.org<BR>https://www.winpcap.org/mailman/listinfo/winpcap-users<BR></BLOCKQUOTE></BODY></HTML>