Skip to contentSkip to conversation

lockdep: improve comments in wait-type checks

RFC PATCH Superseded Zhouyi Zhou 3 participants 4 messages Aug 10, 2021 – Aug 24, 2021

Subsystems (2)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 lockdep: improve comments in wait-type checks

Hi Peter,
I think comments in wait-type checks may benifit from some comments
improvement.
I recklessly add Paul to signed-off-by, because I asked his opinion
about comment in rcu_read_lock_bh part.

Thanks a lot
Zhouyi

---
 include/linux/lockdep_types.h | 2 +-
 kernel/locking/lockdep.c      | 2 +-
 kernel/rcu/update.c           | 4 ++--
 3 files changed, 4 insertions(+), 4 deletions(-)
include/linux/lockdep_types.h +1 -1
index 3e726ace5c62..d22430840b53 100644
@@ -21,7 +21,7 @@ enum lockdep_wait_type {
 	LD_WAIT_SPIN,		/* spin loops, raw_spinlock_t etc.. */
 
 #ifdef CONFIG_PROVE_RAW_LOCK_NESTING
-	LD_WAIT_CONFIG,		/* CONFIG_PREEMPT_LOCK, spinlock_t etc.. */
+	LD_WAIT_CONFIG,		/* preemptible in PREEMPT_RT, spinlock_t etc.. */
 #else
 	LD_WAIT_CONFIG = LD_WAIT_SPIN,
 #endif
kernel/locking/lockdep.c +1 -1
index bf1c00c881e4..952d0ccf8776 100644
@@ -4671,7 +4671,7 @@ print_lock_invalid_wait_context(struct task_struct *curr,
 /*
  * Verify the wait_type context.
  *
- * This check validates we takes locks in the right wait-type order; that is it
+ * This check validates we take locks in the right wait-type order; that is it
  * ensures that we do not take mutexes inside spinlocks and do not attempt to
  * acquire spinlocks inside raw_spinlocks and the sort.
  *
kernel/rcu/update.c +2 -2
index c21b38cc25e9..690b0cec7459 100644
@@ -247,7 +247,7 @@ struct lockdep_map rcu_lock_map = {
 	.name = "rcu_read_lock",
 	.key = &rcu_lock_key,
 	.wait_type_outer = LD_WAIT_FREE,
-	.wait_type_inner = LD_WAIT_CONFIG, /* XXX PREEMPT_RCU ? */
+	.wait_type_inner = LD_WAIT_CONFIG, /* PREEMPT_RT implies PREEMPT_RCU */
 };
 EXPORT_SYMBOL_GPL(rcu_lock_map);
 
@@ -256,7 +256,7 @@ struct lockdep_map rcu_bh_lock_map = {
 	.name = "rcu_read_lock_bh",
 	.key = &rcu_bh_lock_key,
 	.wait_type_outer = LD_WAIT_FREE,
-	.wait_type_inner = LD_WAIT_CONFIG, /* PREEMPT_LOCK also makes BH preemptible */
+	.wait_type_inner = LD_WAIT_CONFIG, /* PREEMPT_RT makes BH preemptible. */
 };
 EXPORT_SYMBOL_GPL(rcu_bh_lock_map);
 
Signature · 1 line
2.25.1
#1 lore raw

Hi,

Quoted 3 lines · Zhouyi Zhou

Hi Peter,
I think comments in wait-type checks may benifit from some comments
improvement.

Thanks for looking into this!

Even for an RFC patch, it's better that you write a proper commit log, and note that you can put all this background information right after the "---" line, if you think that will help people review it.

Quoted 3 lines

I recklessly add Paul to signed-off-by, because I asked his opinion
about comment in rcu_read_lock_bh part.

Probably a Suggested-by will suffice, but I leave that to Paul ;-)

Quoted 5 lines

Thanks a lot
Zhouyi

Please make you the last one in the "Signed-off-by" list, see Documentation/process/submitting-patches.rst:

"Notably, the last Signed-off-by: must always be that of the developer submitting the patch."

Regards,
Boqun

Quoted 57 lines
---
 include/linux/lockdep_types.h | 2 +-
 kernel/locking/lockdep.c      | 2 +-
 kernel/rcu/update.c           | 4 ++--
 3 files changed, 4 insertions(+), 4 deletions(-)
include/linux/lockdep_types.h +1 -1
index 3e726ace5c62..d22430840b53 100644
@@ -21,7 +21,7 @@ enum lockdep_wait_type {
 	LD_WAIT_SPIN,		/* spin loops, raw_spinlock_t etc.. */
 
 #ifdef CONFIG_PROVE_RAW_LOCK_NESTING
-	LD_WAIT_CONFIG,		/* CONFIG_PREEMPT_LOCK, spinlock_t etc.. */
+	LD_WAIT_CONFIG,		/* preemptible in PREEMPT_RT, spinlock_t etc.. */
 #else
 	LD_WAIT_CONFIG = LD_WAIT_SPIN,
 #endif
kernel/locking/lockdep.c +1 -1
index bf1c00c881e4..952d0ccf8776 100644
@@ -4671,7 +4671,7 @@ print_lock_invalid_wait_context(struct task_struct *curr,
 /*
  * Verify the wait_type context.
  *
- * This check validates we takes locks in the right wait-type order; that is it
+ * This check validates we take locks in the right wait-type order; that is it
  * ensures that we do not take mutexes inside spinlocks and do not attempt to
  * acquire spinlocks inside raw_spinlocks and the sort.
  *
kernel/rcu/update.c +2 -2
index c21b38cc25e9..690b0cec7459 100644
@@ -247,7 +247,7 @@ struct lockdep_map rcu_lock_map = {
 	.name = "rcu_read_lock",
 	.key = &rcu_lock_key,
 	.wait_type_outer = LD_WAIT_FREE,
-	.wait_type_inner = LD_WAIT_CONFIG, /* XXX PREEMPT_RCU ? */
+	.wait_type_inner = LD_WAIT_CONFIG, /* PREEMPT_RT implies PREEMPT_RCU */
 };
 EXPORT_SYMBOL_GPL(rcu_lock_map);
 
@@ -256,7 +256,7 @@ struct lockdep_map rcu_bh_lock_map = {
 	.name = "rcu_read_lock_bh",
 	.key = &rcu_bh_lock_key,
 	.wait_type_outer = LD_WAIT_FREE,
-	.wait_type_inner = LD_WAIT_CONFIG, /* PREEMPT_LOCK also makes BH preemptible */
+	.wait_type_inner = LD_WAIT_CONFIG, /* PREEMPT_RT makes BH preemptible. */
 };
 EXPORT_SYMBOL_GPL(rcu_bh_lock_map);
 
Signature · 1 line
2.25.1
maintainer
#2 lore raw
Quoted 13 lines · Boqun Feng

Hi,

Quoted 3 lines · Zhouyi Zhou

Hi Peter,
I think comments in wait-type checks may benifit from some comments
improvement.

Thanks for looking into this!

Even for an RFC patch, it's better that you write a proper commit log, and note that you can put all this background information right after the "---" line, if you think that will help people review it.

Quoted 5 lines
Quoted 3 lines

I recklessly add Paul to signed-off-by, because I asked his opinion
about comment in rcu_read_lock_bh part.

Probably a Suggested-by will suffice, but I leave that to Paul ;-)

This, also since Paul never actually touched the patch.

Quoted 11 lines
Quoted 5 lines

Thanks a lot
Zhouyi

Please make you the last one in the "Signed-off-by" list, see Documentation/process/submitting-patches.rst:

"Notably, the last Signed-off-by: must always be that of the developer submitting the patch."

The first sob should be that of the author, and given that's the same person as the submitter in this case, he's got a problem at his hands ;-)

#3 lore raw

Thank both of you for correcting my mistakes.
I do benefit a lot from your emails.

Cheers
Zhouyi

Quoted 39 lines · Peter Zijlstra
Quoted 13 lines · Boqun Feng

Hi,

Quoted 3 lines · Zhouyi Zhou

Hi Peter,
I think comments in wait-type checks may benifit from some comments
improvement.

Thanks for looking into this!

Even for an RFC patch, it's better that you write a proper commit log, and note that you can put all this background information right after the "---" line, if you think that will help people review it.

Quoted 5 lines
Quoted 3 lines

I recklessly add Paul to signed-off-by, because I asked his opinion
about comment in rcu_read_lock_bh part.

Probably a Suggested-by will suffice, but I leave that to Paul ;-)

This, also since Paul never actually touched the patch.

Quoted 11 lines
Quoted 5 lines

Thanks a lot
Zhouyi

Please make you the last one in the "Signed-off-by" list, see Documentation/process/submitting-patches.rst:

"Notably, the last Signed-off-by: must always be that of the developer submitting the patch."

The first sob should be that of the author, and given that's the same person as the submitter in this case, he's got a problem at his hands ;-)