[LU-9909] memory leak in lnet_discover()

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.11.0
• Affects Version/s: Lustre 2.11.0
• Assignee: Amir Shehata (Inactive)
• Reporter: John Hammond

Description

In lnet_discover() the buffer allocated by LIBCFS_ALLOC(buf, n_ids * sizeof(*buf)) is never freed.

Issue Links

• is related to: LU-9914 Dynamic Discovery - discovery hangs if max_interfaces is changed from 200->16

Comments

[Peter Jones]

Landed for 2.11

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/28702/
Subject: LU-9909 lnet: fix memory leak and lnet_interfaces_max
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 81d4f7a253193ebfe559f675d3c0975c0899d592

[Gerrit Updater]

Amir Shehata ([e-mail]) uploaded a new patch: https://review.whamcloud.com/28702
Subject: LU-9909 lnet: fix memory leak and lnet_interfaces_max
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 52ea572d5e828584ba49f703f0161407384323a2

[James Nunez (Inactive)]

I created LU-9913 to capture the lnet_peer memory leak error.

[Olaf Weber (Inactive)]

If the leak is an lnet_peer then it differs from the problem for which this LU was opened.

[John Hammond]

I bisected this locally by running conf-sanity 35a. This was introduced by commit 0f1aaad4c1b4447ee5097b8bb79a49d09eaa23c2 https://review.whamcloud.com/25789 LU-9480 lnet: implement Peer Discovery. Unfortunately leak finder doesn't work for LNet allocations. But the leak is most likely an LNet peer:

m:lustre-release# gdb lnet/lnet/lnet.ko
...
(gdb) p sizeof(struct lnet_peer)
$1 = 184

[James Nunez (Inactive)]

So far, the earliest date I see the 'Portals memory leaked' error is on August 22. So, yes, they started after the last batch of patch landings to master.

[Amir Shehata (Inactive)]

Did that just start happening after the latest landing? or has it been happening for a while?

[James Nunez (Inactive)]

I'm seeing several of our tests failing with

[13455.203067] LNetError: 14901:0:(module.c:689:libcfs_exit()) Portals memory leaked: 184 bytes
mv: cannot stat '/tmp/debug': No such file or directory
Memory leaks detected

Is this the same issue?

Logs for this failure can be found at
https://testing.hpdd.intel.com/test_sets/972a44c6-87e0-11e7-b4b0-5254006e85c2

[Amir Shehata (Inactive)]

thanks for catching that. Will address it.