lkml.org 
[lkml]   [2021]   [Sep]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH 1/2] ARM: Add per-cpu variable holding cpu number
Date
To help move thread_info into task_struct, stop using the cpu number
contained in the thread_info block in C code and use a per-cpu
variable instead. This value will be initialized long before the
task_struct cpu value for the various idle threads are set, which
avoids ordering issues during CPU startup.

Signed-off-by: Keith Packard <keithpac@amazon.com>
---
arch/arm/include/asm/smp.h | 5 ++++-
arch/arm/kernel/smp.c | 14 ++++++++++++++
2 files changed, 18 insertions(+), 1 deletion(-)

diff --git a/arch/arm/include/asm/smp.h b/arch/arm/include/asm/smp.h
index 5d508f5d56c4..3aca2c2089bc 100644
--- a/arch/arm/include/asm/smp.h
+++ b/arch/arm/include/asm/smp.h
@@ -15,7 +15,10 @@
# error "<asm/smp.h> included in non-SMP build"
#endif

-#define raw_smp_processor_id() (current_thread_info()->cpu)
+#define raw_smp_processor_id() this_cpu_read(cpu_number)
+#define __smp_processor_id() __this_cpu_read(cpu_number)
+
+DECLARE_PER_CPU_READ_MOSTLY(unsigned int, cpu_number);

struct seq_file;

diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c
index 74679240a9d8..0457e25109c6 100644
--- a/arch/arm/kernel/smp.c
+++ b/arch/arm/kernel/smp.c
@@ -51,6 +51,9 @@
#define CREATE_TRACE_POINTS
#include <trace/events/ipi.h>

+DEFINE_PER_CPU_READ_MOSTLY(unsigned int, cpu_number);
+EXPORT_PER_CPU_SYMBOL(cpu_number);
+
/*
* as from 2.5, kernels no longer have an init_tasks structure
* so we need some other way of telling a new secondary core
@@ -495,6 +498,7 @@ void __init smp_prepare_boot_cpu(void)
void __init smp_prepare_cpus(unsigned int max_cpus)
{
unsigned int ncores = num_possible_cpus();
+ unsigned int cpu;

init_cpu_topology();

@@ -505,6 +509,16 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
*/
if (max_cpus > ncores)
max_cpus = ncores;
+
+ /*
+ * Initialize the cpu_number value for each cpu before we
+ * start it. This ensures that the value is valid during cpu
+ * initialization, even before the idle task_struct cpu member
+ * is set
+ */
+ for_each_possible_cpu(cpu)
+ per_cpu(cpu_number, cpu) = cpu;
+
if (ncores > 1 && max_cpus) {
/*
* Initialise the present map, which describes the set of CPUs
--
2.33.0
\
 
 \ /
  Last update: 2021-09-02 17:55    [W:0.182 / U:0.176 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site