Duplicate Information in NDIS_PACKET?

NTDEV
1

Hi All,

While going through NDIS_PACKET_EXTENSION, I found that
NDIS_TCP_IP_CHECKSUM_PACKET_INFO provides the checksum information about the
packet. But the same information is present in the contents of the packet
(buffers data) in standard tcp, udp and ip headers.
Is there any special use of this information?

Another thing of confusion is that according to documentation: fields like
NdisPacketTcpChecksumFailed are set by miniport ? Why are we making lower
layers to be worried about upper layers? Are not we violating the
architecture of OSI?

Amit Manocha

2

Hi Amit,

Good Luck,

 

From: “Amit Manocha”

>Reply-To: “NT Developers Interest List”
>To: “NT Developers Interest List”
>Subject: [ntdev] Duplicate Information in NDIS_PACKET?
>Date: Fri, 9 May 2003 12:09:25 +0530
>
>Hi All,
>
>While going through NDIS_PACKET_EXTENSION, I found that
>NDIS_TCP_IP_CHECKSUM_PACKET_INFO provides the checksum information about the
>packet. But the same information is present in the contents of the packet
>(buffers data) in standard tcp, udp and ip headers.


>Is there any special use of this information?



[Yogi]



Indexes a pointer that points to or contains an NDIS_TCP_IP_CHECKSUM_PACKET_INFO structure. This structure specifies per-packet information for checksum operations offloaded from the TCP/IP transport to a NIC.



It is releted to the offloading.



>
>Another thing of confusion is that according to documentation: fields like
>NdisPacketTcpChecksumFailed are set by miniport ? Why are we making lower
>layers to be worried about upper layers? Are not we violating the


>architecture of OSI?



[Yogi] Just read comments given for NDIS_TCP_IP_CHECKSUM_PACKET_INFO, this tells that it will be set by the NIC while the checksum is offloaded.



>
>Amit Manocha
>
>
>
>—
>You are currently subscribed to ntdev as: xxxxx@hotmail.com
>To unsubscribe send a blank email to xxxxx@lists.osr.com

Win a PC. Every 15 days! Just take this quiz.

3

> NDIS_TCP_IP_CHECKSUM_PACKET_INFO provides the checksum information
about
the

packet. But the same information is present in the contents of the
packet
(buffers data) in standard tcp, udp and ip headers.
Is there any special use of this information?

Yes, this is the hardware-calculated checksum of the real data.

layers to be worried about upper layers? Are not we violating the
architecture of OSI?

Yes, for speed.

Max