[prev in list] [next in list] [prev in thread] [next in thread] 

List:       wireshark-users
Subject:    Re: [Wireshark-users] Windows 7 Ignores TCP MSS?
From:       Hansang Bae <for_list_hbae () nyc ! rr ! com>
Date:       2011-04-07 1:14:59
Message-ID: 4D9D1013.9010507 () nyc ! rr ! com
[Download RAW message or body]

[Attachment #2 (multipart/alternative)]


On 3/23/2011 9:11 PM, Robinson, Eric wrote:
> If the NIC has offloading enabled, I guess the stack knows this and 
> sends it bigger frames, knowing it will packetize them smaller? 
> Because I went into the adapter properties and turned off all the 
> offloading features, and now it is being reported as expected in 
> WireShark. So I guess when the NIC's offload capabilities are 
> disabled, the stack detects this and only sends it frames that are 
> appropriate for the MTU?

(wow..sorry for the late reply - it's been hard to keep up on the list!)

Yes, I don't know how it breaks it, but TCP cheats by looking at what 
the MTU is and subtracts the header appropriately to come up with MSS.  
Theoretically, IP is supposed to do that, but I suppose it's one of 
those things that started and no one ever deviated from it.  I do recall 
seeing a post by Rich Seifert on Usenet (when it actually was useful) 
comp.dcom.lans.ethernet explaining how this came to be.  But I don't 
recall the gory details.

hsb

[Attachment #5 (text/html)]

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
    <title></title>
  </head>
  <body text="#000000" bgcolor="#ffffff">
    On 3/23/2011 9:11 PM, Robinson, Eric wrote:
    <blockquote
      cite="mid:768FEA3F936926468F594E3CA7741FE802C74614@ipa-vault.nvipa.com"
      type="cite">
      <div dir="ltr" align="left"><span class="980250801-24032011"><font
            color="#0000ff" face="Verdana" size="2">&nbsp;</font></span></div>
      <div dir="ltr" align="left"><span class="980250801-24032011"><font
            color="#0000ff" face="Verdana" size="2">If the NIC has
            offloading enabled, I guess the stack knows this and sends
            it bigger frames, knowing it will packetize them&nbsp;smaller?
            Because I went into the adapter properties and turned off
            all the offloading features, and now it is being reported as
            expected in WireShark. So I guess when the NIC's offload
            capabilities are disabled, the stack&nbsp;detects this and only
            sends it frames that are appropriate for the MTU?</font></span><br>
      </div>
    </blockquote>
    <font color="#0000ff"><font size="2"><font face="Verdana"><br>
          (wow..sorry for the late reply - it's been hard to keep up on
          the list!)<br>
          <br>
          Yes, I don't know how it breaks it, but TCP cheats by looking
          at what the MTU is and subtracts the header appropriately to
          come up with MSS.&nbsp; Theoretically, IP is supposed to do that,
          but I suppose it's one of those things that started and no one
          ever deviated from it.&nbsp; I do recall seeing a post by Rich
          Seifert on Usenet (when it actually was useful)
          comp.dcom.lans.ethernet explaining how this came to be.&nbsp; But I
          don't recall the gory details.<br>
          <br>
          hsb<br>
        </font></font></font>
  </body>
</html>


___________________________________________________________________________
Sent via:    Wireshark-users mailing list <wireshark-users@wireshark.org>
Archives:    http://www.wireshark.org/lists/wireshark-users
Unsubscribe: https://wireshark.org/mailman/options/wireshark-users
             mailto:wireshark-users-request@wireshark.org?subject=unsubscribe

[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic