Method of handling network traffic through optimization of receive side scaling

• US 8,842,562 B2
• Filed: 10/25/2011
• Issued: 09/23/2014
• Est. Priority Date: 10/25/2011
• Status: Active Grant

• Current Assignee: Dell Products LP (Dell Technologies Inc.)
• Original Assignee: Dell Products LP (Dell Technologies Inc.)
• Inventors: Domsch, Matthew L., Hernandez, Hendrich M., Winter, Robert L., Dube, Shawn J.
• Primary Examiner(s): PHAN, MAN U
• Application Number:
  US13/281,192
• Publication Number:
  US 20130103871A1
• Time in Patent Office: 1,064 Days
• Field of Search: 370235-252, 370389-412, 709201-230
• US Class Current: 370/252
• CPC Class Codes: G06F 13/24 using interrupt G06F13/32 takes precedence
• CPC Class Codes: G06F 9/5005 to service a request
• CPC Class Codes: G06F 9/54 Interprogram communication
• CPC Class Codes: H04L 69/32 Architecture of open systems interconnection [OSI] 7-layer type protocol stacks, e.g. the interfaces between the data link level and the physical level

Abstract

An information handling system includes a plurality of processors that each includes a cache memory, and a receive side scaling (RSS) indirection table with a plurality of pointers that each points to one of the processors. A network data packet received by the information handling system determines a pointer to a first processor. In response to determining the pointer, information associated with the network data packet is transferred to the cache memory of the first processor, The information handling system also includes a process scheduler that moves a process associated with the network data packet from a second processor to the first processor, and an RSS module that directs the process scheduler to move the process and associates the first pointer with the processor in response to directing the process scheduler.

Claims

1. An information handling system comprising:

a plurality of processors, each processor including a cache memory;

a receive side scaling (RSS) indirection table including a plurality of pointers, each pointer identifying a respective processor, wherein;

a first network data packet received by the information handling system determines a first one of the pointers, wherein the first pointer identifies a first one of the processors; and

in response to determining the first pointer, first information associated with the first network data packet is transferred to the cache memory of the first processor;

a process scheduler that moves a process associated with the first network data packet from a second one of the processors to the first processor;

an RSS module that;

directs the process scheduler to move the process; and

associates the first pointer with the first processor in response to directing the process scheduler; and

a message signaled interrupt (MSI-X) table including a plurality of interrupts, wherein each interrupt;

is associated with a respective pointer; and

identifies the respective processor, such that when an associated pointer is determined, the MSI-X table provides an interrupt to the respective processor;

wherein the RSS module further associates a first interrupt with the first processor in response to directing the process scheduler.

2. The information handling system of claim 1, further comprising:

a hash type module that implements a plurality of hash type operations, each hash type operation being operable to select a field of the first network data packet;

wherein;

determining the first pointer is in response to the hash type module implementing a first one of the hash type operations; and

the RSS module further selects the first hash type operation.

3. The information handling system of claim 2, wherein the RSS module selects the first hash type operation in response to determining a load on the first processor.

4. The information handling system of claim 1, further comprising:

a hash function module that implements a plurality of hash function operations, each hash function operation being operable to select a hash function to perform on the first network data packet;

wherein;

determining the first pointer is in response to the hash function module implementing a first one of the hash function operations; and

the RSS module further selects the first hash function operation.

5. The information handling system of claim 4, wherein the RSS module selects the first hash function in response to determining a load on the first processor.

6. The information handling system of claim 1, wherein the process scheduler:

moves the process from the first processor to a third one of the processors; and

directs the RSS module to associate the first pointer with the third processor in response to moving the process to the third processor.

7. The information handling system of claim 1, wherein:

a second network data packet received by the information handling system determines a second pointer that identifies a third one of the processors; and

in response to determining the second pointer, second information associated with the second network data packet is transferred to the cache memory of the third processor.

8. The information handling system of claim 1, wherein transferring the first information includes:

transferring the first information to a receive memory buffer of the information handling system; and

transferring the first information from the receive memory buffer to the first cache memory of the first processor.

9. A network interface device comprising:

a receive side scaling (RSS) indirection table including a plurality of pointers, each pointer identifying a respective processor, wherein;

a first network data packet received by the network interface device determines a first one of the pointers, wherein the first pointer identifies a first one of the processors; and

in response to determining the first pointer, first information associated with the first network data packet is transferred to the cache memory of the first processor;

an RSS module that;

directs a process scheduler to move a process associated with the first network data packet from a second one of the processors to the first processor; and

associates the first pointer with the first processor in response to directing the process scheduler; and

a message signaled interrupt (MSI-X) table including a plurality of interrupts, wherein each interrupt;

is associated with a respective pointer; and

identities the respective processor, such that when an associated pointer is determined, the MSI-X table provides an interrupt to the respective processor;

wherein the RSS module further associates a first interrupt with the first processor espouse to directing the process scheduler.

10. The network interface device of claim 9, further comprising:

a hash type module that implements a plurality of hash type operations, each hash type operation being operable to select a field of the first network data packet;

wherein;

determining the first pointer is in response to the hash type module implementing a first one of the hash type operations; and

the RSS module further selects the first hash type operation.

11. The network interface device of claim 9, further comprising:

a hash function module that implements a plurality of hash function operations, each hash function operation being operable to select a hash function to perform on the first network data packet;

wherein;

determining the first pointer is in response to the hash function module implementing a first one of the hash function operations; and

the RSS module further selects the first hash function operation.

12. The network interface of claim 11, wherein the RSS module selects the first hash function operation in response to determining a load on the first processor.

13. The network interface device of claim 9, wherein the RSS module associates the first pointer with a third one of the processors in response to the process scheduler moving the process from the first processor to the third processor.

14. The network interface device of claim 9, wherein:

a second network data packet received by the network interface device determines a second pointer that identifies a third one of the processors; and

in response to determining the second pointer, second information associated with the second network data packet is transferred to the cache memory of the third processor.

15. A method comprising:

retrieving a process scheduling policy from a process scheduler of an information handling system;

retrieving a receive side scaling (RSS) indirection table of the information handling system;

determining an optimal processor of the information handling system on which to run a process, wherein the determining is in response to retrieving the process scheduling policy and to retrieving the RSS indirection table;

in response to determining the optimal processor;

scheduling by the process scheduler the process on the optimal processor; and

setting an RSS pointer of the RSS indirection table to point to the optimal processor;

retrieving a message signaled interrupt table of the information handling system; and

setting MSI-X pointer of the MSI-X table to provide an interrupt to the optimal processor in further response to determining the optimal processor;

wherein the determining is in further response to retrieving the MSI-X table.

16. The method of claim 15, further comprising:

retrieving a hash type capability of the information handling system; and

setting a hash type in further response to determining the optimal processor;

wherein the determining is in further response to retrieving the hash type capability.

17. The method of claim 15, further comprising:

retrieving a hash function capability of the information handling system; and

setting a hash function in further response to determining the optimal processor;

wherein the determining is in further response to retrieving the hash function capability.