Hello:

I'm looking for a coding example to better understand how to use PACKET_MMAP for packet reception into user space. I'm trying to read raw L2 packets bound to an ethernet interface and get them into the shared memory ring buffer to be used for user space applications to manipulate.

I have seen the documentation at:

Documentation/networking/packet_mmap.txt

http://wiki.ipxwarzone.com/index.php...ux_packet_mmap

Note has a decent example of transmission, but I could really use an rx example.

This one has potential but I still just don't quite get it:

http://www.scaramanga.co.uk/code-fu/lincap.c

http://yusufonlinux.blogspot.com/201...zero-copy.html

but its still unclear to me how to read the memory mapped raw packets from a buffer in user space. A working example would be great for self study...

Thanks!

The idea is simple: you allocate a buffer in an area of memory that has been made directly accessible both to your process and to the kernel. You are obliged to initialize that area of memory in the form of a linked list of equal-sized slots. Now, when you access the socket (which you do using the same "poll()" based logic as before), the kernel moves the data directly into that agreed-upon area of memory, thereby occupying zero-or-more of those slots. Your process wakes up to (maybe...) discover that zero or more of those slots are now "magically" filled with data. It processes the content of the slots in the required way, then goes to sleep again.

The advantages of this arrangement are that (a) the data only has to be copied once (by the kernel, directly to/from a location that is agreeable to you both); and (b) more than one piece of data can be exchanged between your process and the kernel using only one "wakeup call." The number of trips through the system dispatcher and the number of system-calls in general is greatly reduced.

The source-code in lincap.c illustrates the procedure. The user-land process allocates the buffer, initializes it, announces it to the kernel when opening the socket, then enters an endless loop which consists first of processing the zero-or-more chunks of data that it might find there, then waiting for more. The loop is constructed in the manner that it is (with the poll call at the bottom) in recognition of the fact that the buffer might be filled with some data when the socket is successfully opened.

Thank you for the reply. I'm trying to build the lincap.c example but I'm failing with the problem:

gcc lincap.c
lincap.c:29:24: fatal error: asm/system.h: No such file or directory
compilation terminated.

Some of my system info:

uname -a
Linux swamp 3.3.4-3.fc16.x86_64 #1 SMP Thu May 3 14:46:44 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

ls /usr/src/kernels/3.3.4-3.fc16.x86_64/include/asm-generic/system.h
/usr/src/kernels/3.3.4-3.fc16.x86_64/include/asm-generic/system.h

So I change the include path from <asm/system.h> to <asm-generic/system.h>:

gcc lincap.c
lincap.c:29:32: fatal error: asm-generic/system.h: No such file or directory
compilation terminated.

So I then brute force the whole path in there:

#include "/usr/src/kernels/3.3.4-3.fc16.x86_64/include/asm-generic/system.h"

but I get:

gcc lincap.c
In file included from lincap.c:30:0:
/usr/src/kernels/3.3.4-3.fc16.x86_64/include/asm-generic/system.h:20:28: fatal error: linux/irqflags.h: No such file or directory
compilation terminated.

I don't think I'm going about this correctly... it seems like system.h is needed for a function called mb(). Any help with the build would be appreciated.

Hi

Did you manage to build lincap.c?

I also face the issue that lincap.c uses mb(), which is defined in asm/system.h. But the code in asm/system.h is embraced by #ifdef __KERNEL__. How is this supposed to work if lincap is a user space application?

BR

David

Have you work it out? I've got the same problem as you. When I removed "mb()", I compiled the source and I could access first frame in buffer. But when I was trying to write "tp_status" to buffer it gives me "Segmentation fault".