[Hyper-V] hv_netvsc: Eliminate memory allocation in the packet send path

This kernel patch applies to 15.10, 15.04, 14.10, 14.04, and 14.04 HWE

This bug is missing log files that will aid in diagnosing the problem. From a terminal window please run:

apport-collect 1452074

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

These are upstream commits, not a apport-based report.

The importance for this update, and 1445195, has become critical, due to customer concerns here https://social.technet.microsoft.com/Forums/windowsserver/en-US/8807f61c-565e-45bc-abc4-af09abf59de2/ubuntu-14042-lts-generation-2-scsi-errors-on-vss-based-backups

This needs to go to 15.10, 15.04, 14.10, 14.04, and 14.04 HWE.

I was able to build a test kernel with the three requested commits. The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1452074/

However, the three requested commits required three additional commits as prerequisites. Here are the details:

Prereq for hv_netvsc_packet->xmit_more:
7c3877fhv_netvsc: Implement batching in send buffer

Requested:
cbacec7 hv_netvsc: Cleanup the test for freeing skb when we use sendbuf mechanism
Requested:
b08cc79 hv_netvsc: Eliminate memory allocation in the packet send path

Prereq for definition of netvsc_xmit_completion:
ee90b81 hv_netvsc: Fix the packet free when it is in skb headroom

Prereq for implementing partial copies.
aa0a34b hv_netvsc: Implement partial copy into send buffer

Requested:
b56fc3c hv_netvsc: Fix a bug in netvsc_start_xmit()

Can you see if the test kernel resolves the bug?

Hrm, I had thought those other commits had been requested previously, sorry I missed them in the roundup. We'll look at testing the kernel right away. There is another skb-related fix that is still in review upstream and is not yet committed as well.

@Joshua, just curious if you had a chance to have the test kernel tested?

@jsalisbury We encountered unexpected performance regressions which we were not able to diagnose (a 10% decrease in throughput). Is it possible to rebase on 3.19.0-20 and release a new test kernel?

I built a new test kernel, which was rebased to 3.19.0-20. The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1452074/

This kernel looks good.

Are the issues caused by this bug the same reliability issues we were talking about in bug 1445195 and bug 1454758 ?

This patch affects networking and not storage, so it should not have an effect on storage reliability. It may be more appropriate to open a new issue to investigate the remaining instability.

The new bug is bug 1470250 . Let's now use that bug to focus on the VSS Backup issues.

The patch requested in bug 1454892 has all the commits in this bug as pre-reqs. Is it okay to mark this bug as a duplicate of bug 1454892 so all these commits can be sent together as one SRU request?

It looks like testing is good for both bugs in Vivid. I'll work on building test kernels for Utopic, Trusty and Precise.

Thanks in advance!

Yes, It is okay to mark this as a duplicate since the other request encapsulates these commits.