[LU-8841] Replace l_wait_info calls that use zeroed info parameters Created: 16/Nov/16 Updated: 06/Jul/21 Resolved: 06/Jul/21 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.9.0 |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Minor |
| Reporter: | Steve Guminski (Inactive) | Assignee: | WC Triage |
| Resolution: | Duplicate | Votes: | 0 |
| Labels: | None | ||
| Issue Links: |
|
||||||||
| Severity: | 3 | ||||||||
| Rank (Obsolete): | 9223372036854775807 | ||||||||
| Description |
|
There are many cases where l_wait_event is used with an l_wait_info parameter that is completely zeroed. This is done to wait for new kernel threads to be started (or exit due to a failure). These cases should be replaced with the native wait_event or wait_event_timeout as appropriate. An example of the zeroed l_wait_info is in sptlrpc_gc_init: struct l_wait_info lwi = { 0 };
...
task = kthread_run(...);
...
l_wait_event(sec_gc_thread.t_ctl_waitq, thread_is_running(&sec_gc_thread), &lwi);
This issue was uncovered during the code review at http://review.whamcloud.com/#/c/23754/ |