[LU-4011] problems with upstream lustre client code

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Environment: fc19, 3.11 kernel
• Severity: 3
• Assignee: Bob Glossman (Inactive)
• Reporter: Bob Glossman (Inactive)

Description

This ticket is to track issues with the upstream lustre client code that is part of the 3.11 kernel source in fc19.

Making a separate ticket as suggested by Andreas in https://jira.hpdd.intel.com/browse/LU-3974?focusedCommentId=67574&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-67574

Issue Links

• is related to: LU-4451 Kernel Oops with NFS reexport using mainline 3.12 client
• is related to: LU-6204 modinfo data is stale, and would be nice if they included the lustre version
• is related to: LU-6285 Assert fails in staging client module crashes kernel if CPUMASK_OFFSTACK set
• is related to: LU-7747 sanity test_56w: dataversion changed during copy, migration aborted
• is related to: LU-6209 remove old LNDs - ralnd, mxlnd
• is related to: LU-2355 orph_index_delete()) ASSERTION(obj->mod_flags & ORPHAN_OBJ)
• is related to: LU-4530 Mainline kernel client (3.12-3.14): lfs quota -u ... -> Permission denied
• is related to: LU-3974 Support for linux 3.11 kernel

Comments

[James A Simmons]

Bob can you close this ticket.

[Bob Glossman (Inactive)]

James,
I think this ticket was to track only old fc19 upstream client.
Would suggest a new distinct ticket for upstream clients in 4.12 kernels.

[James A Simmons]

With lastest 4.12-rc5 upstream client as of today in my testing we fail the following sanity test. Besides these test the patch LU-8680 needs to be applied to make the lustre client stable.

sanity 27z, 27D, 29, 77c, 101g, 102a, 102b, 102n, 103a, 125, 133h, 154B, 154a, 154g, 160a, 160c, 160e, 161c, 161d, 162a, 205, 215, 226a, 242, 251, 405, 900

[James Nunez (Inactive)]

No, but found these failures in our regular autotest runs on master.

[James A Simmons]

Numez are you testing the upstream client?

[James Nunez (Inactive)]

Another runtests test 1 failure in review-dne-part-2:
2015-07-27 11:06:17 - https://testing.hpdd.intel.com/test_sets/558dd67a-3497-11e5-a9b3-5254006e85c2

[James Nunez (Inactive)]

We're seeing a similar (same?) failure with runtests test_1 again at:
2015-05-29 16:08:46 - https://testing.hpdd.intel.com/test_sets/21fd7836-0668-11e5-bf9f-5254006e85c2

[Dmitry Eremin (Inactive)]

The situation on master even worse. It's not even compiled. I have submit http://review.whamcloud.com/#/c/8853/. And after this fix I observe other crash LU-4489.

[Peng Tao]

Dmitry, thanks for digging the patch. Your patch also applies to lustre master. Do you see the same crash with master?

[Dmitry Eremin (Inactive)]

This LBUG was fixed by the following small patch:

diff --git a/lustre/obdclass/cl_lock.c b/lustre/obdclass/cl_lock.c
index d440da9..2544053 100644
--- a/lustre/obdclass/cl_lock.c
+++ b/lustre/obdclass/cl_lock.c
@@ -2053,8 +2053,8 @@ void cl_lock_hold_add(const struct lu_env *env, struct cl_
lock *lock,
         LASSERT(lock->cll_state != CLS_FREEING);

         ENTRY;
-        cl_lock_hold_mod(env, lock, +1);
         cl_lock_get(lock);
+        cl_lock_hold_mod(env, lock, +1);
         lu_ref_add(&lock->cll_holders, scope, source);
         lu_ref_add(&lock->cll_reference, scope, source);
         EXIT;

and also it happens only if CONFIG_LUSTRE_DEBUG_EXPENSIVE_CHECK=y is set.