From 0d6526f82c3cdefcca47f73f5fc08dc6f335eac6 Mon Sep 17 00:00:00 2001 From: Tim Chen Date: Mon, 21 Sep 2026 17:37:22 -0700 Subject: sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling bug alb_break_llc() decides whether to break LLC preference during active load balance. It does so by testing that every runnable fair task on the source rq prefers its LLC: env->src_rq->nr_pref_llc_running == env->src_rq->cfs.h_nr_runnable But the two counters cover different sets. nr_pref_llc_running is updated in account_llc_enqueue()/account_llc_dequeue(), next to cfs_rq->nr_queued, so it follows queued tasks. h_nr_runnable is updated in set_delayed()/ clear_delayed() and drops delay-dequeued tasks. So under DELAY_DEQUEUE, a preferring task that goes to sleep stays counted in nr_pref_llc_running while h_nr_runnable falls. The equality then breaks, alb_break_llc() returns false, and active balance is free to pull a task off its preferred LLC. Active balance only moves runnable tasks, and this is the only LLC check it consults: once the stopper runs, LBF_ACTIVE_LB skips the per-task test in can_migrate_task(). The runnable set is the one we want. Fix it on the counter side. A task should be counted in nr_pref_llc_running exactly while it is both queued on its preferred LLC (pref_llc_queued) and runnable (!sched_delayed). Define that membership once in task_pref_llc_runnable(), and adjust the counter only through pref_llc_running_inc()/pref_llc_running_dec() from the four sites that change either input: account_llc_enqueue(), account_llc_dequeue(), set_delayed() and clear_delayed(). Gating every update on the same predicate keeps the delay, wake and dequeue paths from double-counting or underflowing; see the comments at those sites for the ordering. nr_llc_running and sd->llc_counts are not touched and stay on queued semantics. Fixes: 714059f79ff0 ("sched/cache: Handle moving single tasks to/from their preferred LLC") Closes: https://lore.kernel.org/lkml/20260827135000.735138-1-zhanxusheng@xiaomi.com/ Reported-by: Zhan Xusheng Suggested-by: Chen Yu Signed-off-by: Tim Chen Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Ingo Molnar Reviewed-by: Kayra Cizmeci Cc: # v7.2.x Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com --- Documentation/userspace-api/lsm.rst | 73 +++++++++++++++++++++++++++++++++++++ 1 file changed, 73 insertions(+) create mode 100644 Documentation/userspace-api/lsm.rst (limited to 'Documentation/userspace-api/lsm.rst') diff --git a/Documentation/userspace-api/lsm.rst b/Documentation/userspace-api/lsm.rst new file mode 100644 index 000000000..a76da3738 --- /dev/null +++ b/Documentation/userspace-api/lsm.rst @@ -0,0 +1,73 @@ +.. SPDX-License-Identifier: GPL-2.0 +.. Copyright (C) 2022 Casey Schaufler +.. Copyright (C) 2022 Intel Corporation + +===================================== +Linux Security Modules +===================================== + +:Author: Casey Schaufler +:Date: July 2023 + +Linux security modules (LSM) provide a mechanism to implement +additional access controls to the Linux security policies. + +The various security modules may support any of these attributes: + +``LSM_ATTR_CURRENT`` is the current, active security context of the +process. +The proc filesystem provides this value in ``/proc/self/attr/current``. +This is supported by the SELinux, Smack and AppArmor security modules. +Smack also provides this value in ``/proc/self/attr/smack/current``. +AppArmor also provides this value in ``/proc/self/attr/apparmor/current``. + +``LSM_ATTR_EXEC`` is the security context of the process at the time the +current image was executed. +The proc filesystem provides this value in ``/proc/self/attr/exec``. +This is supported by the SELinux and AppArmor security modules. +AppArmor also provides this value in ``/proc/self/attr/apparmor/exec``. + +``LSM_ATTR_FSCREATE`` is the security context of the process used when +creating file system objects. +The proc filesystem provides this value in ``/proc/self/attr/fscreate``. +This is supported by the SELinux security module. + +``LSM_ATTR_KEYCREATE`` is the security context of the process used when +creating key objects. +The proc filesystem provides this value in ``/proc/self/attr/keycreate``. +This is supported by the SELinux security module. + +``LSM_ATTR_PREV`` is the security context of the process at the time the +current security context was set. +The proc filesystem provides this value in ``/proc/self/attr/prev``. +This is supported by the SELinux and AppArmor security modules. +AppArmor also provides this value in ``/proc/self/attr/apparmor/prev``. + +``LSM_ATTR_SOCKCREATE`` is the security context of the process used when +creating socket objects. +The proc filesystem provides this value in ``/proc/self/attr/sockcreate``. +This is supported by the SELinux security module. + +Kernel interface +================ + +Set a security attribute of the current process +----------------------------------------------- + +.. kernel-doc:: security/lsm_syscalls.c + :identifiers: sys_lsm_set_self_attr + +Get the specified security attributes of the current process +------------------------------------------------------------ + +.. kernel-doc:: security/lsm_syscalls.c + :identifiers: sys_lsm_get_self_attr + +.. kernel-doc:: security/lsm_syscalls.c + :identifiers: sys_lsm_list_modules + +Additional documentation +======================== + +* Documentation/security/lsm.rst +* Documentation/security/lsm-development.rst -- cgit v1.3.1