[v4,14/14] eventdev: mark trace variables as internal
Checks
Commit Message
From: Pavan Nikhilesh <pbhagavatula@marvell.com>
Mark rte_trace global variables as internal i.e. remove them
from experimental section of version map.
Some of them are used in inline APIs, mark those as global.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Acked-by: Ray Kinsella <mdr@ashroe.eu>
---
doc/guides/rel_notes/release_21_11.rst | 12 +++++
lib/eventdev/version.map | 71 ++++++++++++--------------
2 files changed, 44 insertions(+), 39 deletions(-)
Comments
On Sat, Oct 16, 2021 at 12:34 AM <pbhagavatula@marvell.com> wrote:
>
> From: Pavan Nikhilesh <pbhagavatula@marvell.com>
>
> Mark rte_trace global variables as internal i.e. remove them
> from experimental section of version map.
> Some of them are used in inline APIs, mark those as global.
>
> Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
> Acked-by: Ray Kinsella <mdr@ashroe.eu>
> ---
> doc/guides/rel_notes/release_21_11.rst | 12 +++++
> lib/eventdev/version.map | 71 ++++++++++++--------------
> 2 files changed, 44 insertions(+), 39 deletions(-)
>
> diff --git a/doc/guides/rel_notes/release_21_11.rst b/doc/guides/rel_notes/release_21_11.rst
> index 38e601c236..5b4a05c3ae 100644
> --- a/doc/guides/rel_notes/release_21_11.rst
> +++ b/doc/guides/rel_notes/release_21_11.rst
> @@ -226,6 +226,9 @@ API Changes
> the crypto/security operation. This field will be used to communicate
> events such as soft expiry with IPsec in lookaside mode.
>
> +* eventdev: Event vector configuration APIs have been made stable.
> + Move memory used by timer adapters to hugepage. This will prevent TLB misses
> + if any and aligns to memory structure of other subsystems.
>
> ABI Changes
> -----------
> @@ -277,6 +280,15 @@ ABI Changes
> were added in structure ``rte_event_eth_rx_adapter_stats`` to get additional
> status.
>
> +* eventdev: A new structure ``rte_event_fp_ops`` has been added which is now used
> + by the fastpath inline functions. The structures ``rte_eventdev``,
> + ``rte_eventdev_data`` have been made internal. ``rte_eventdevs[]`` can't be
> + accessed directly by user any more. This change is transparent to both
> + applications and PMDs.
> +
> +* eventdev: Re-arrange fields in ``rte_event_timer`` to remove holes.
> + ``rte_event_timer_adapter_pmd.h`` has been made internal.
Looks good. Please fix the following, If there are no objections, I
will merge the next version.
1) Please move the doc update to respective patches
2) Following checkpath issue
[for-main]dell[dpdk-next-eventdev] $ ./devtools/checkpatches.sh -n 14
### eventdev: move inline APIs into separate structure
INFO: symbol event_dev_fp_ops_reset has been added to the INTERNAL
section of the version map
INFO: symbol event_dev_fp_ops_set has been added to the INTERNAL
section of the version map
INFO: symbol event_dev_probing_finish has been added to the INTERNAL
section of the version map
ERROR: symbol rte_event_fp_ops is added in the DPDK_22 section, but is
expected to be added in the EXPERIMENTAL section of the version map
>On Sat, Oct 16, 2021 at 12:34 AM <pbhagavatula@marvell.com> wrote:
>>
>> From: Pavan Nikhilesh <pbhagavatula@marvell.com>
>>
>> Mark rte_trace global variables as internal i.e. remove them
>> from experimental section of version map.
>> Some of them are used in inline APIs, mark those as global.
>>
>> Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
>> Acked-by: Ray Kinsella <mdr@ashroe.eu>
>> ---
>> doc/guides/rel_notes/release_21_11.rst | 12 +++++
>> lib/eventdev/version.map | 71 ++++++++++++--------------
>> 2 files changed, 44 insertions(+), 39 deletions(-)
>>
>> diff --git a/doc/guides/rel_notes/release_21_11.rst
>b/doc/guides/rel_notes/release_21_11.rst
>> index 38e601c236..5b4a05c3ae 100644
>> --- a/doc/guides/rel_notes/release_21_11.rst
>> +++ b/doc/guides/rel_notes/release_21_11.rst
>> @@ -226,6 +226,9 @@ API Changes
>> the crypto/security operation. This field will be used to communicate
>> events such as soft expiry with IPsec in lookaside mode.
>>
>> +* eventdev: Event vector configuration APIs have been made stable.
>> + Move memory used by timer adapters to hugepage. This will
>prevent TLB misses
>> + if any and aligns to memory structure of other subsystems.
>>
>> ABI Changes
>> -----------
>> @@ -277,6 +280,15 @@ ABI Changes
>> were added in structure ``rte_event_eth_rx_adapter_stats`` to get
>additional
>> status.
>>
>> +* eventdev: A new structure ``rte_event_fp_ops`` has been added
>which is now used
>> + by the fastpath inline functions. The structures ``rte_eventdev``,
>> + ``rte_eventdev_data`` have been made internal.
>``rte_eventdevs[]`` can't be
>> + accessed directly by user any more. This change is transparent to
>both
>> + applications and PMDs.
>> +
>> +* eventdev: Re-arrange fields in ``rte_event_timer`` to remove
>holes.
>> + ``rte_event_timer_adapter_pmd.h`` has been made internal.
>
>Looks good. Please fix the following, If there are no objections, I
>will merge the next version.
>
>1) Please move the doc update to respective patches
Ack, will move in next version.
>2) Following checkpath issue
>[for-main]dell[dpdk-next-eventdev] $ ./devtools/checkpatches.sh -n
>14
>
>### eventdev: move inline APIs into separate structure
>
>INFO: symbol event_dev_fp_ops_reset has been added to the
>INTERNAL
>section of the version map
>INFO: symbol event_dev_fp_ops_set has been added to the INTERNAL
>section of the version map
>INFO: symbol event_dev_probing_finish has been added to the
>INTERNAL
>section of the version map
These can be ignored as they are internal
>ERROR: symbol rte_event_fp_ops is added in the DPDK_22 section, but
>is
>expected to be added in the EXPERIMENTAL section of the version map
This is a replacement for rte_eventdevs, ethdev rework also doesn’t mark
it as experimental. @David Marchand @Ray Kinsella any opinions?
Hello Pavan,
On Mon, Oct 18, 2021 at 5:07 PM Pavan Nikhilesh Bhagavatula
<pbhagavatula@marvell.com> wrote:
> >[for-main]dell[dpdk-next-eventdev] $ ./devtools/checkpatches.sh -n
> >14
> >
> >### eventdev: move inline APIs into separate structure
> >
> >INFO: symbol event_dev_fp_ops_reset has been added to the
> >INTERNAL
> >section of the version map
> >INFO: symbol event_dev_fp_ops_set has been added to the INTERNAL
> >section of the version map
> >INFO: symbol event_dev_probing_finish has been added to the
> >INTERNAL
> >section of the version map
>
> These can be ignored as they are internal
Those first warnings are informational.
>
> >ERROR: symbol rte_event_fp_ops is added in the DPDK_22 section, but
> >is
> >expected to be added in the EXPERIMENTAL section of the version map
>
> This is a replacement for rte_eventdevs, ethdev rework also doesn’t mark
> it as experimental. @David Marchand @Ray Kinsella any opinions?
This check is there to ensure that added symbols first go through a
period in experimental status.
Same as for ethdev, the use of inlines in stable API directly exposes
a new symbol to applications.
With this implementation, this check can be waived and the symbol can
go directly to stable status.
This symbol being exposed as stable, it will be frozen in ABI until
next breakage.
I see you reserved 6 spots for new ops, so it looks ok.
@@ -226,6 +226,9 @@ API Changes
the crypto/security operation. This field will be used to communicate
events such as soft expiry with IPsec in lookaside mode.
+* eventdev: Event vector configuration APIs have been made stable.
+ Move memory used by timer adapters to hugepage. This will prevent TLB misses
+ if any and aligns to memory structure of other subsystems.
ABI Changes
-----------
@@ -277,6 +280,15 @@ ABI Changes
were added in structure ``rte_event_eth_rx_adapter_stats`` to get additional
status.
+* eventdev: A new structure ``rte_event_fp_ops`` has been added which is now used
+ by the fastpath inline functions. The structures ``rte_eventdev``,
+ ``rte_eventdev_data`` have been made internal. ``rte_eventdevs[]`` can't be
+ accessed directly by user any more. This change is transparent to both
+ applications and PMDs.
+
+* eventdev: Re-arrange fields in ``rte_event_timer`` to remove holes.
+ ``rte_event_timer_adapter_pmd.h`` has been made internal.
+
Known Issues
------------
@@ -1,6 +1,13 @@
DPDK_22 {
global:
+ __rte_eventdev_trace_crypto_adapter_enqueue;
+ __rte_eventdev_trace_deq_burst;
+ __rte_eventdev_trace_enq_burst;
+ __rte_eventdev_trace_eth_tx_adapter_enqueue;
+ __rte_eventdev_trace_timer_arm_burst;
+ __rte_eventdev_trace_timer_arm_tmo_tick_burst;
+ __rte_eventdev_trace_timer_cancel_burst;
rte_event_crypto_adapter_caps_get;
rte_event_crypto_adapter_create;
rte_event_crypto_adapter_create_ext;
@@ -42,8 +49,8 @@ DPDK_22 {
rte_event_eth_rx_adapter_start;
rte_event_eth_rx_adapter_stats_get;
rte_event_eth_rx_adapter_stats_reset;
- rte_event_eth_rx_adapter_vector_limits_get;
rte_event_eth_rx_adapter_stop;
+ rte_event_eth_rx_adapter_vector_limits_get;
rte_event_eth_tx_adapter_caps_get;
rte_event_eth_tx_adapter_create;
rte_event_eth_tx_adapter_create_ext;
@@ -56,6 +63,7 @@ DPDK_22 {
rte_event_eth_tx_adapter_stats_get;
rte_event_eth_tx_adapter_stats_reset;
rte_event_eth_tx_adapter_stop;
+ rte_event_fp_ops;
rte_event_port_attr_get;
rte_event_port_default_conf_get;
rte_event_port_link;
@@ -86,25 +94,28 @@ DPDK_22 {
rte_event_timer_cancel_burst;
rte_event_vector_pool_create;
- #added in 21.11
- rte_event_fp_ops;
-
local: *;
};
EXPERIMENTAL {
global:
- # added in 20.05
- __rte_eventdev_trace_configure;
- __rte_eventdev_trace_queue_setup;
- __rte_eventdev_trace_port_link;
- __rte_eventdev_trace_port_unlink;
- __rte_eventdev_trace_start;
- __rte_eventdev_trace_stop;
+ # added in 21.11
+ rte_event_eth_rx_adapter_create_with_params;
+ rte_event_eth_rx_adapter_queue_conf_get;
+};
+
+INTERNAL {
+ global:
+
__rte_eventdev_trace_close;
- __rte_eventdev_trace_deq_burst;
- __rte_eventdev_trace_enq_burst;
+ __rte_eventdev_trace_configure;
+ __rte_eventdev_trace_crypto_adapter_create;
+ __rte_eventdev_trace_crypto_adapter_free;
+ __rte_eventdev_trace_crypto_adapter_queue_pair_add;
+ __rte_eventdev_trace_crypto_adapter_queue_pair_del;
+ __rte_eventdev_trace_crypto_adapter_start;
+ __rte_eventdev_trace_crypto_adapter_stop;
__rte_eventdev_trace_eth_rx_adapter_create;
__rte_eventdev_trace_eth_rx_adapter_free;
__rte_eventdev_trace_eth_rx_adapter_queue_add;
@@ -117,38 +128,19 @@ EXPERIMENTAL {
__rte_eventdev_trace_eth_tx_adapter_queue_del;
__rte_eventdev_trace_eth_tx_adapter_start;
__rte_eventdev_trace_eth_tx_adapter_stop;
- __rte_eventdev_trace_eth_tx_adapter_enqueue;
+ __rte_eventdev_trace_port_link;
+ __rte_eventdev_trace_port_setup;
+ __rte_eventdev_trace_port_unlink;
+ __rte_eventdev_trace_queue_setup;
+ __rte_eventdev_trace_start;
+ __rte_eventdev_trace_stop;
__rte_eventdev_trace_timer_adapter_create;
+ __rte_eventdev_trace_timer_adapter_free;
__rte_eventdev_trace_timer_adapter_start;
__rte_eventdev_trace_timer_adapter_stop;
- __rte_eventdev_trace_timer_adapter_free;
- __rte_eventdev_trace_timer_arm_burst;
- __rte_eventdev_trace_timer_arm_tmo_tick_burst;
- __rte_eventdev_trace_timer_cancel_burst;
- __rte_eventdev_trace_crypto_adapter_create;
- __rte_eventdev_trace_crypto_adapter_free;
- __rte_eventdev_trace_crypto_adapter_queue_pair_add;
- __rte_eventdev_trace_crypto_adapter_queue_pair_del;
- __rte_eventdev_trace_crypto_adapter_start;
- __rte_eventdev_trace_crypto_adapter_stop;
-
- # changed in 20.11
- __rte_eventdev_trace_port_setup;
- # added in 21.11
- rte_event_eth_rx_adapter_create_with_params;
-
- #added in 21.05
- __rte_eventdev_trace_crypto_adapter_enqueue;
- rte_event_eth_rx_adapter_queue_conf_get;
-};
-
-INTERNAL {
- global:
-
event_dev_fp_ops_reset;
event_dev_fp_ops_set;
event_dev_probing_finish;
- rte_event_pmd_selftest_seqn_dynfield_offset;
rte_event_pmd_allocate;
rte_event_pmd_get_named_dev;
rte_event_pmd_is_valid_dev;
@@ -156,6 +148,7 @@ INTERNAL {
rte_event_pmd_pci_probe_named;
rte_event_pmd_pci_remove;
rte_event_pmd_release;
+ rte_event_pmd_selftest_seqn_dynfield_offset;
rte_event_pmd_vdev_init;
rte_event_pmd_vdev_uninit;
rte_eventdevs;