Skip to contentSkip to conversation

doc: update atomic operation deprecation

PATCH Superseded Joyce Kong 3 participants 6 messages Jul 12, 2021 – Aug 2, 2021

lore
Share & tools
lore

Joyce Kong

author
#0 lore raw

PATCH 1/1 doc: update atomic operation deprecation

Update the incorrect description about atomic operations with provided wrappers in deprecation doc[1].

[1]https://mails.dpdk.org/archives/dev/2021-July/213333.html

Fixes 7518c5c4ae6a ("doc: announce adoption of C11 atomic operations semantics")
---
 doc/guides/rel_notes/deprecation.rst | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
doc/guides/rel_notes/deprecation.rst +8 -8
index 9584d6bfd7..4142315842 100644
@@ -19,16 +19,16 @@ Deprecation Notices
 
 * rte_atomicNN_xxx: These APIs do not take memory order parameter. This does
   not allow for writing optimized code for all the CPU architectures supported
-  in DPDK. DPDK will adopt C11 atomic operations semantics and provide wrappers
-  using C11 atomic built-ins. These wrappers must be used for patches that
-  need to be merged in 20.08 onwards. This change will not introduce any
-  performance degradation.
+  in DPDK. DPDK has adopted atomic operations semantics. GCC atomic built-ins
+  must be used for patches that need to be merged in 20.08 onwards. This change
+  will not introduce any performance degradation.
 
 * rte_smp_*mb: These APIs provide full barrier functionality. However, many
-  use cases do not require full barriers. To support such use cases, DPDK will
-  adopt C11 barrier semantics and provide wrappers using C11 atomic built-ins.
-  These wrappers must be used for patches that need to be merged in 20.08
-  onwards. This change will not introduce any performance degradation.
+  use cases do not require full barriers. To support such use cases, DPDK has
+  adopted atomic barrier semantics. GCC atomic built-ins and a new wrapper
+  ``rte_atomic_thread_fence`` instead of ``__atomic_thread_fence`` must be
+  used for patches that need to be merged in 20.08 onwards. This change will
+  not introduce any performance degradation.
 
 * lib: will fix extending some enum/define breaking the ABI. There are multiple
   samples in DPDK that enum/define terminated with a ``.*MAX.*`` value which is
Signature · 1 line
2.17.1
#1 lore raw

<snip>

Quoted 28 lines

Update the incorrect description about atomic operations with provided wrappers in deprecation doc[1].

[1]https://mails.dpdk.org/archives/dev/2021-July/213333.html

Fixes 7518c5c4ae6a ("doc: announce adoption of C11 atomic operations

semantics")
Cc: [email protected]

---
 doc/guides/rel_notes/deprecation.rst | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
doc/guides/rel_notes/deprecation.rst +0 -0
index 9584d6bfd7..4142315842 100644
@@ -19,16 +19,16 @@ Deprecation Notices
 
 * rte_atomicNN_xxx: These APIs do not take memory order parameter. This

does
not allow for writing optimized code for all the CPU architectures supported
- in DPDK. DPDK will adopt C11 atomic operations semantics and provide wrappers

doc/guides/rel_notes/deprecation.rst Deprecation Notices

Quoted 6 lines
doc/guides/rel_notes/deprecation.rst +3 -3
-  using C11 atomic built-ins. These wrappers must be used for patches that
-  need to be merged in 20.08 onwards. This change will not introduce any
-  performance degradation.
+  in DPDK. DPDK has adopted atomic operations semantics. GCC atomic
+ built-ins  must be used for patches that need to be merged in 20.08
+ onwards. This change  will not introduce any performance degradation.

Since there have been objections to the language used to refer to GCC C11 atomic built-ins, may be we add a reference to the GCC pages?

DPDK has adopted the atomic operations from https://gcc.gnu.org/onlinedocs/gcc/_005f_005fatomic-Builtins.html. These operations must be used for patches that need to be merged in 20.08 onwards. This change will not introduce any performance degradation.

Quoted 12 lines

* rte_smp_*mb: These APIs provide full barrier functionality. However, many

doc/guides/rel_notes/deprecation.rst +0 -2
-  use cases do not require full barriers. To support such use cases, DPDK will
-  adopt C11 barrier semantics and provide wrappers using C11 atomic built-

ins.

doc/guides/rel_notes/deprecation.rst +5 -2
-  These wrappers must be used for patches that need to be merged in 20.08
-  onwards. This change will not introduce any performance degradation.
+  use cases do not require full barriers. To support such use cases,
+ DPDK has  adopted atomic barrier semantics. GCC atomic built-ins and a
+ new wrapper  ``rte_atomic_thread_fence`` instead of
+ ``__atomic_thread_fence`` must be  used for patches that need to be
+ merged in 20.08 onwards. This change will  not introduce any performance
Quoted 1 line

degradation.

Same here.
To support such use cases, DPDK has adopted atomic operations from https://gcc.gnu.org/onlinedocs/gcc/_005f_005fatomic-Builtins.html. A new wrapper ``rte_atomic_thread_fence`` instead of ``__atomic_thread_fence`` must be used for patches that need to be merged in 20.08 onwards. This change will not introduce any performance degradation.

Quoted 7 lines

* lib: will fix extending some enum/define breaking the ABI. There are
multiple
samples in DPDK that enum/define terminated with a ``.*MAX.*`` value
which is
--
2.17.1

author
#2 lore raw

PATCH v2 1/1 doc: update atomic operation deprecation

In reply to Joyce Kong

Update the incorrect description about atomic operations with provided wrappers in deprecation doc[1].

[1]https://mails.dpdk.org/archives/dev/2021-July/213333.html

Fixes 7518c5c4ae6a ("doc: announce adoption of C11 atomic operations semantics")
---
 doc/guides/rel_notes/deprecation.rst | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)
doc/guides/rel_notes/deprecation.rst +10 -8
index 9584d6bfd7..a4f350fa09 100644
@@ -19,16 +19,18 @@ Deprecation Notices
 
 * rte_atomicNN_xxx: These APIs do not take memory order parameter. This does
   not allow for writing optimized code for all the CPU architectures supported
-  in DPDK. DPDK will adopt C11 atomic operations semantics and provide wrappers
-  using C11 atomic built-ins. These wrappers must be used for patches that
-  need to be merged in 20.08 onwards. This change will not introduce any
-  performance degradation.
+  in DPDK. DPDK has adopted the atomic operations from
+  https://gcc.gnu.org/onlinedocs/gcc/_005f_005fatomic-Builtins.html. These
+  operations must be used for patches that need to be merged in 20.08 onwards.
+  This change will not introduce any performance degradation.
 
 * rte_smp_*mb: These APIs provide full barrier functionality. However, many
-  use cases do not require full barriers. To support such use cases, DPDK will
-  adopt C11 barrier semantics and provide wrappers using C11 atomic built-ins.
-  These wrappers must be used for patches that need to be merged in 20.08
-  onwards. This change will not introduce any performance degradation.
+  use cases do not require full barriers. To support such use cases, DPDK has
+  adopted atomic operations from
+  https://gcc.gnu.org/onlinedocs/gcc/_005f_005fatomic-Builtins.html. These
+  operations and a new wrapper ``rte_atomic_thread_fence`` instead of
+  ``__atomic_thread_fence`` must be used for patches that need to be merged in
+  20.08 onwards. This change will not introduce any performance degradation.
 
 * lib: will fix extending some enum/define breaking the ABI. There are multiple
   samples in DPDK that enum/define terminated with a ``.*MAX.*`` value which is
Signature · 1 line
2.17.1
#3 lore raw

23/07/2021 11:49, Joyce Kong:

Quoted 10 lines

Update the incorrect description about atomic operations with provided wrappers in deprecation doc[1].

[1]https://mails.dpdk.org/archives/dev/2021-July/213333.html

Fixes 7518c5c4ae6a ("doc: announce adoption of C11 atomic operations semantics")

Applied, thanks.

Considering all the questions regarding usage of atomics, I still think we need a documentation about their correct use.

#4 lore raw
Quoted 21 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Saturday, July 31, 2021 3:27 PM
To: Honnappa Nagarahalli <[email protected]>; Ruifeng Wang <[email protected]>; Joyce Kong <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH v2] doc: update atomic operation deprecation

23/07/2021 11:49, Joyce Kong:

Quoted 9 lines

Update the incorrect description about atomic operations with provided wrappers in deprecation doc[1].

[1]https://mails.dpdk.org/archives/dev/2021-July/213333.html

Fixes 7518c5c4ae6a ("doc: announce adoption of C11 atomic operations

semantics")
Cc: [email protected]

Quoted 7 lines
Quoted 2 lines

Applied, thanks.

Considering all the questions regarding usage of atomics, I still think we need a documentation about their correct use.

I think few things to document are:
1) Use of __atomic_add_fetch vs __atomic_fetch_add
2) Using __ATOMIC_RELAXED for statistics
3) Using rte_atomic_thread_fence API

Anything else?

Quoted 1 line
#5 lore raw

02/08/2021 19:22, Honnappa Nagarahalli:

Quoted 9 lines

From: Thomas Monjalon <[email protected]>

Quoted 3 lines

23/07/2021 11:49, Joyce Kong:
Considering all the questions regarding usage of atomics, I still think we need a documentation about their correct use.

I think few things to document are:
1) Use of __atomic_add_fetch vs __atomic_fetch_add
2) Using __ATOMIC_RELAXED for statistics
3) Using rte_atomic_thread_fence API

That's a good list.
Let's start with that please.

I would like an explanation about when full barrier may be used.