ip_append_page and the socket send buffer

I wrote a function that is equivalent to udp_sendmsg, but uses ip_append_page to attach data to an skb. The function is implemented as follows:

1. Allocate a page and copy the given data to that page
2. Set up routing and cork the socket
3. Call ip_append_data to create an initial skb (with data length set to 0)
4. Call ip_append_page with the allocated page
5. Call udp_push_pending_frames to send the packet

The function works correctly. Packets are generated and sent as expected: this was verified by looking at the packet contents on the receiving machine.
However, under load, there is a significant difference in the behaviour of udp_sendmsg, compared with my function. The problem is that the socket send buffer (wmem_alloc) quickly grows beyond its upper limit (which is 131071 by default). This results in numerous failures of ip_append_data with EAGAIN, degrading performance considerably. udp_sendmsg, on the other hand, keeps wmem_alloc in a much smaller range under the same load.

Two notes:
1. Modifying the upper limit to 524287 solved the problem completely (regardless of the load)
2. The same thing happens with multiple calls to ip_append_data (e.g., if I want to copy the data in two sections), so it is not a problem with ip_append_page. This leads me o believe that the problem lies with Scatter/Gather I/O.

Any thoughts?

Elad

Why?

Where is the code?

James Morris
<[email protected]>

This scheme allows you to send the same data to multiple recipients. The user's data is copied only once into a set of pages, which are then shared among multiple skbuff objects.

The code is probably too long to be attached to a mailing list message. It pretty much follows the steps I specified in my original e-mail. I can provide more details, if required.

Elad