On 2019/11/3 12:30 上午, Borislav Petkov wrote:
Resending again because your mail has the craziest header:
Reply-To: 20191101162109.GN20975@paulmck-ThinkPad-P72
and hitting Reply-to-all creates only confusion. WTF?
Sorry, I send the emails via git send-email with the wrong
argument "--reply-to", it should have been "--in-reply-to".
On Sat, Nov 02, 2019 at 12:45:59PM +0000, Lai Jiangshan wrote:
Convert x86 to use a per-cpu rcu_preempt_depth. The reason for doing so
is that accessing per-cpu variables is a lot cheaper than accessing
task_struct or thread_info variables.
We need to save/restore the actual rcu_preempt_depth when switch.
We also place the per-cpu rcu_preempt_depth close to __preempt_count
and current_task variable.
Using the idea of per-cpu __preempt_count.
No function call when using rcu_read_[un]lock().
Single instruction for rcu_read_lock().
2 instructions for fast path of rcu_read_unlock().
CC: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Signed-off-by: Lai Jiangshan <laijs@xxxxxxxxxxxxxxxxx>
---
arch/x86/Kconfig | 2 +
arch/x86/include/asm/rcu_preempt_depth.h | 87 ++++++++++++++++++++++++
arch/x86/kernel/cpu/common.c | 7 ++
arch/x86/kernel/process_32.c | 2 +
arch/x86/kernel/process_64.c | 2 +
include/linux/rcupdate.h | 24 +++++++
init/init_task.c | 2 +-
kernel/fork.c | 2 +-
kernel/rcu/Kconfig | 3 +
kernel/rcu/tree_exp.h | 2 +
kernel/rcu/tree_plugin.h | 37 +++++++---
11 files changed, 158 insertions(+), 12 deletions(-)
create mode 100644 arch/x86/include/asm/rcu_preempt_depth.h
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index d6e1faa28c58..af9fedc0fdc4 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -18,6 +18,7 @@ config X86_32
select MODULES_USE_ELF_REL
select OLD_SIGACTION
select GENERIC_VDSO_32
+ select ARCH_HAVE_RCU_PREEMPT_DEEPTH
WTF is a "DEEPTH"?
Sorry, bad English.
