> There is a Kernel API called dev_queue_xmit. Actually > this function calls the xmit function of the device. > > The layer upper to the driver is the network layer. So > each type of these fills the sk_buff, before giving > to the driver. > Eg:- IPv4, IPv6,etc... (it will be there in the > protocol specific codes). > ---- net/ipv4/arp.c: dev_queue_xmit(skb); > > Then calls the dev_queue_xmit. And I think dev_queue_xmit > handles the network QOS. > I've not gone deep into it. But went through the > code in brief. hi there. today i was making some progress because of the help from you. i was able to send a bootp_pkt like one function is doing in ipconfig.c. when i changed the code a bit to make the same with tcp/ip packets, it still sends the packets but tcpdump for example, displays the packets in a raw form. it does neither recognize the packet as a tcp nor as a valid one. i built the packet similar to the bootp_pkt in ipconfig.c and added some things like checksum calculation over pseudo tcp header and the like. same that is required in userspace to send raw packets. could it be something with the tcp options set, or may it more be related to a misfunctioning or not correctly initialized buffer that is used to transmit the packet? honestly i dont know if still problem still belongs in here. thanks best regards, threadhead ______________________________________________________________________________ Erdbeben im Iran: Zehntausende Kinder brauchen Hilfe. UNICEF hilft den Kindern - helfen Sie mit! https://www.unicef.de/spe/spe_03.php -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
