kernel-packages team mailing list archive

[dann frazier <dann.frazier@xxxxxxxxxxxxx>]

Public bug reported:

Two patches have landed upstream that improve thunderx nic performance:

commit b6e4038262bc933f2ef5427b6bcd2607d02ba4bb
Merge: 9bd160b 1d36879
Author: David S. Miller <davem@xxxxxxxxxxxxx>
Date:   Mon Mar 14 12:33:37 2016 -0400

    Merge branch 'thunderx-perf'
    
    Sunil Goutham says:
    
    ====================
    net: thunderx: Performance enhancement changes
    
    Below patches attempts to improve performance by reducing
    no of atomic operations while allocating new receive buffers
    and reducing cache misses by adjusting nicvf structure elements.
    
    Changes from v1:
     No changes, resubmitting a fresh as per David's suggestion.
    ====================
    
    Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>

commit 1d368790bc2d4111feae6fcc0b3b68d6ae92ff0f
Author: Sunil Goutham <sgoutham@xxxxxxxxxx>
Date:   Mon Mar 14 16:36:15 2016 +0530

    net: thunderx: Adjust nicvf structure to reduce cache misses
    
    Adjusted nicvf structure such that all elements used in hot
    path like napi, xmit e.t.c fall into same cache line. This reduced
    no of cache misses and resulted in ~2% increase in no of packets
    handled on a core.
    
    Also modified elements with :1 notation to boolean, to be
    consistent with other element definitions.
    
    Signed-off-by: Sunil Goutham <sgoutham@xxxxxxxxxx>
    Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>

commit 5c2e26f6f674ee93719769d024cc72fad3ba9134
Author: Sunil Goutham <sgoutham@xxxxxxxxxx>
Date:   Mon Mar 14 16:36:14 2016 +0530

    net: thunderx: Set recevie buffer page usage count in bulk
    
    Instead of calling get_page() for every receive buffer carved out
    of page, set page's usage count at the end, to reduce no of atomic
    calls.
    
    Signed-off-by: Sunil Goutham <sgoutham@xxxxxxxxxx>
    Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>

** Affects: linux (Ubuntu)
     Importance: Medium
     Assignee: dann frazier (dannf)
         Status: In Progress

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1567093

Title:
  thunderx nic performance improvements

Status in linux package in Ubuntu:
  In Progress

Bug description:
  Two patches have landed upstream that improve thunderx nic
  performance:

  commit b6e4038262bc933f2ef5427b6bcd2607d02ba4bb
  Merge: 9bd160b 1d36879
  Author: David S. Miller <davem@xxxxxxxxxxxxx>
  Date:   Mon Mar 14 12:33:37 2016 -0400

      Merge branch 'thunderx-perf'
      
      Sunil Goutham says:
      
      ====================
      net: thunderx: Performance enhancement changes
      
      Below patches attempts to improve performance by reducing
      no of atomic operations while allocating new receive buffers
      and reducing cache misses by adjusting nicvf structure elements.
      
      Changes from v1:
       No changes, resubmitting a fresh as per David's suggestion.
      ====================
      
      Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>

  commit 1d368790bc2d4111feae6fcc0b3b68d6ae92ff0f
  Author: Sunil Goutham <sgoutham@xxxxxxxxxx>
  Date:   Mon Mar 14 16:36:15 2016 +0530

      net: thunderx: Adjust nicvf structure to reduce cache misses
      
      Adjusted nicvf structure such that all elements used in hot
      path like napi, xmit e.t.c fall into same cache line. This reduced
      no of cache misses and resulted in ~2% increase in no of packets
      handled on a core.
      
      Also modified elements with :1 notation to boolean, to be
      consistent with other element definitions.
      
      Signed-off-by: Sunil Goutham <sgoutham@xxxxxxxxxx>
      Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>

  commit 5c2e26f6f674ee93719769d024cc72fad3ba9134
  Author: Sunil Goutham <sgoutham@xxxxxxxxxx>
  Date:   Mon Mar 14 16:36:14 2016 +0530

      net: thunderx: Set recevie buffer page usage count in bulk
      
      Instead of calling get_page() for every receive buffer carved out
      of page, set page's usage count at the end, to reduce no of atomic
      calls.
      
      Signed-off-by: Sunil Goutham <sgoutham@xxxxxxxxxx>
      Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1567093/+subscriptions