x86/resctrl: Do not reconfigure exiting tasks
authorXiaochen Shen <xiaochen.shen@intel.com>
Wed, 8 Jan 2020 16:13:46 +0000 (00:13 +0800)
committerBorislav Petkov <bp@suse.de>
Mon, 13 Jan 2020 13:10:21 +0000 (14:10 +0100)
commitdc433797c6f639e46824585bbf943578f13d54bf
tree5ddd9ea8026bfac1bfba437197c71b632d987506
parentb3a987b0264d3ddbb24293ebff10eddfc472f653
x86/resctrl: Do not reconfigure exiting tasks

When writing a pid to file "tasks", a callback function move_myself() is
queued to this task to be called when the task returns from kernel mode
or exits. The purpose of move_myself() is to activate the newly assigned
closid and/or rmid associated with this task. This activation is done by
calling resctrl_sched_in() from move_myself(), the same function that is
called when switching to this task.

If this work is successfully queued but then the task enters PF_EXITING
status (e.g., receiving signal SIGKILL, SIGTERM) prior to the
execution of the callback move_myself(), move_myself() still calls
resctrl_sched_in() since the task status is not currently considered.

When a task is exiting, the data structure of the task itself will
be freed soon. Calling resctrl_sched_in() to write the register that
controls the task's resources is unnecessary and it implies extra
performance overhead.

Add check on task status in move_myself() and return immediately if the
task is PF_EXITING.

 [ bp: Massage. ]

Signed-off-by: Xiaochen Shen <xiaochen.shen@intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Reinette Chatre <reinette.chatre@intel.com>
Link: https://lkml.kernel.org/r/1578500026-21152-1-git-send-email-xiaochen.shen@intel.com
arch/x86/kernel/cpu/resctrl/rdtgroup.c