[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220204233858.GA1469@paulmck-ThinkPad-P17-Gen-1>
Date: Fri, 4 Feb 2022 15:38:58 -0800
From: "Paul E. McKenney" <paulmck@...nel.org>
To: rcu@...r.kernel.org
Cc: linux-kernel@...r.kernel.org, kernel-team@...com,
rostedt@...dmis.org
Subject: [PATCH rcu 0/19] SRCU updates for v5.18
Hello!
This series provides SRCU updates, perhaps most notably some changes
that greatly reduce the size of the srcu_struct structure in kernels
built with large values for the CONFIG_NR_CPUS Kconfig option.
1. Tighten cleanup_srcu_struct() GP checks.
2. Fix s/is/if/ typo in srcu_node comment.
3. Make srcu_funnel_gp_start() cache ->mynode in snp_leaf.
4. Dynamically allocate srcu_node array.
5. Make Tree SRCU able to operate without snp_node array.
6. Add size-state transitioning code.
7. Make rcutorture dump the SRCU size state.
8. Add boot-time control over srcu_node array allocation.
9. Use export for srcu_struct defined by DEFINE_STATIC_SRCU(),
courtesy of Alexander Aring.
10. Compute snp_seq earlier in srcu_funnel_gp_start().
11. Use invalid initial value for srcu_node GP sequence numbers.
12. Avoid NULL dereference in srcu_torture_stats_print().
13. Prevent cleanup_srcu_struct() from freeing non-dynamic ->sda.
14. Explain srcu_funnel_gp_start() call to list_add() is safe.
15. Create concurrency-safe helper for initiating size transition.
16. Add contention-triggered addition of srcu_node tree.
17. Make srcu_size_state_name static, courtesy of Jiapeng Chong.
18. Automatically determine size-transition strategy at boot.
19. Add contention check to call_srcu() srcu_data ->lock acquisition.
Thanx, Paul
------------------------------------------------------------------------
Documentation/admin-guide/kernel-parameters.txt | 9
b/Documentation/admin-guide/kernel-parameters.txt | 22
b/include/linux/srcutree.h | 8
b/kernel/rcu/srcutree.c | 6
include/linux/srcutree.h | 20
kernel/rcu/srcutree.c | 644 +++++++++++++++-------
6 files changed, 500 insertions(+), 209 deletions(-)
Powered by blists - more mailing lists