eventdev: negate maintenance capability flag
Checks
Commit Message
Replace RTE_EVENT_DEV_CAP_REQUIRES_MAINT, which signaled the need
for the application to call rte_event_maintain(), with
RTE_EVENT_DEV_CAP_MAINTENANCE_FREE, which does the opposite (i.e.,
signifies that the event device does not require maintenance).
This approach is more in line with how other eventdev hardware and/or
software limitations are handled in the Eventdev API.
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
---
doc/guides/eventdevs/dsw.rst | 4 ++--
drivers/event/cnxk/cnxk_eventdev.c | 3 ++-
drivers/event/dlb2/dlb2.c | 3 ++-
drivers/event/dpaa/dpaa_eventdev.c | 3 ++-
drivers/event/dpaa2/dpaa2_eventdev.c | 3 ++-
drivers/event/dsw/dsw_evdev.c | 3 +--
drivers/event/octeontx/ssovf_evdev.c | 3 ++-
drivers/event/octeontx2/otx2_evdev.c | 3 ++-
drivers/event/opdl/opdl_evdev.c | 1 +
drivers/event/skeleton/skeleton_eventdev.c | 3 ++-
drivers/event/sw/sw_evdev.c | 3 ++-
lib/eventdev/rte_eventdev.h | 19 ++++++++++---------
12 files changed, 30 insertions(+), 21 deletions(-)
Comments
> -----Original Message-----
> From: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
> Sent: Wednesday, November 10, 2021 9:29 AM
> To: jerinj@marvell.com; Jerin Jacob <jerinjacobk@gmail.com>; Sunil Kumar Kori
> <skori@marvell.com>
> Cc: dev@dpdk.org; Pavan Nikhilesh Bhagavatula <pbhagavatula@marvell.com>;
> Hemant Agrawal <hemant.agrawal@nxp.com>; Nipun Gupta
> <nipun.gupta@nxp.com>; Van Haaren, Harry <harry.van.haaren@intel.com>;
> Thomas Monjalon <thomas@monjalon.net>; David Marchand
> <david.marchand@redhat.com>; Gujjar, Abhinandan S
> <abhinandan.gujjar@intel.com>; Carrillo, Erik G <erik.g.carrillo@intel.com>;
> Jayatheerthan, Jay <jay.jayatheerthan@intel.com>; Yigit, Ferruh
> <ferruh.yigit@intel.com>; Akhil Goyal <gakhil@marvell.com>; mattias.ronnblom
> <mattias.ronnblom@ericsson.com>
> Subject: [PATCH] eventdev: negate maintenance capability flag
>
> Replace RTE_EVENT_DEV_CAP_REQUIRES_MAINT, which signaled the need
> for the application to call rte_event_maintain(), with
> RTE_EVENT_DEV_CAP_MAINTENANCE_FREE, which does the opposite (i.e.,
> signifies that the event device does not require maintenance).
>
> This approach is more in line with how other eventdev hardware and/or
> software limitations are handled in the Eventdev API.
>
> Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Generally patch looks fine to me, but on testing I got a strange warning from the
compiler on the OPDL driver:
In file included from ../drivers/event/opdl/opdl_evdev.h:8,
from ../drivers/event/opdl/opdl_evdev.c:15:
../drivers/event/opdl/opdl_evdev.c: In function ‘opdl_info_get’:
../lib/eventdev/rte_eventdev.h:302:44: warning: conversion from ‘long long unsigned int’ to ‘un
signed char’ changes value from ‘1024’ to ‘0’ [-Woverflow]
302 | #define RTE_EVENT_DEV_CAP_MAINTENANCE_FREE (1ULL << 10)
| ^
../drivers/event/opdl/opdl_evdev.c:379:34: note: in expansion of macro ‘RTE_EVENT_DEV_CAP_MAINT
ENANCE_FREE’
379 | RTE_EVENT_DEV_CAP_MAINTENANCE_FREE,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Investigating, RTE_EVENT_DEV_CAP_* are assigned to struct rte_event_dev_info::event_dev_cap,
which is a uint32_t. The error suggests that the variable being assigned to is a "unsigned char"?
<snip>
> static void
> diff --git a/drivers/event/opdl/opdl_evdev.c b/drivers/event/opdl/opdl_evdev.c
> index 5007e9a7bf..787ee4713d 100644
> --- a/drivers/event/opdl/opdl_evdev.c
> +++ b/drivers/event/opdl/opdl_evdev.c
> @@ -376,6 +376,7 @@ opdl_info_get(struct rte_eventdev *dev, struct
> rte_event_dev_info *info)
> .max_num_events = OPDL_INFLIGHT_EVENTS_TOTAL,
> .event_dev_cap = RTE_EVENT_DEV_CAP_BURST_MODE |
> RTE_EVENT_DEV_CAP_CARRY_FLOW_ID,
> + RTE_EVENT_DEV_CAP_MAINTENANCE_FREE,
> };
Aha! Replace:
> RTE_EVENT_DEV_CAP_CARRY_FLOW_ID,
with:
> RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
The "," character was causing the _FREE flag to not be combined with the previous
flags field, instead being its own (unnamed?) assignment to the struct rte_event_dev_info.
With that , to | change, all looks good to me. -Harry
On 2021-11-10 10:55, Van Haaren, Harry wrote:
>> -----Original Message-----
>> From: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
>> Sent: Wednesday, November 10, 2021 9:29 AM
>> To: jerinj@marvell.com; Jerin Jacob <jerinjacobk@gmail.com>; Sunil Kumar Kori
>> <skori@marvell.com>
>> Cc: dev@dpdk.org; Pavan Nikhilesh Bhagavatula <pbhagavatula@marvell.com>;
>> Hemant Agrawal <hemant.agrawal@nxp.com>; Nipun Gupta
>> <nipun.gupta@nxp.com>; Van Haaren, Harry <harry.van.haaren@intel.com>;
>> Thomas Monjalon <thomas@monjalon.net>; David Marchand
>> <david.marchand@redhat.com>; Gujjar, Abhinandan S
>> <abhinandan.gujjar@intel.com>; Carrillo, Erik G <erik.g.carrillo@intel.com>;
>> Jayatheerthan, Jay <jay.jayatheerthan@intel.com>; Yigit, Ferruh
>> <ferruh.yigit@intel.com>; Akhil Goyal <gakhil@marvell.com>; mattias.ronnblom
>> <mattias.ronnblom@ericsson.com>
>> Subject: [PATCH] eventdev: negate maintenance capability flag
>>
>> Replace RTE_EVENT_DEV_CAP_REQUIRES_MAINT, which signaled the need
>> for the application to call rte_event_maintain(), with
>> RTE_EVENT_DEV_CAP_MAINTENANCE_FREE, which does the opposite (i.e.,
>> signifies that the event device does not require maintenance).
>>
>> This approach is more in line with how other eventdev hardware and/or
>> software limitations are handled in the Eventdev API.
>>
>> Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
> Generally patch looks fine to me, but on testing I got a strange warning from the
> compiler on the OPDL driver:
>
> In file included from ../drivers/event/opdl/opdl_evdev.h:8,
> from ../drivers/event/opdl/opdl_evdev.c:15:
> ../drivers/event/opdl/opdl_evdev.c: In function ‘opdl_info_get’:
> ../lib/eventdev/rte_eventdev.h:302:44: warning: conversion from ‘long long unsigned int’ to ‘un
> signed char’ changes value from ‘1024’ to ‘0’ [-Woverflow]
> 302 | #define RTE_EVENT_DEV_CAP_MAINTENANCE_FREE (1ULL << 10)
> | ^
> ../drivers/event/opdl/opdl_evdev.c:379:34: note: in expansion of macro ‘RTE_EVENT_DEV_CAP_MAINT
> ENANCE_FREE’
> 379 | RTE_EVENT_DEV_CAP_MAINTENANCE_FREE,
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> Investigating, RTE_EVENT_DEV_CAP_* are assigned to struct rte_event_dev_info::event_dev_cap,
> which is a uint32_t. The error suggests that the variable being assigned to is a "unsigned char"?
>
> <snip>
>
>> static void
>> diff --git a/drivers/event/opdl/opdl_evdev.c b/drivers/event/opdl/opdl_evdev.c
>> index 5007e9a7bf..787ee4713d 100644
>> --- a/drivers/event/opdl/opdl_evdev.c
>> +++ b/drivers/event/opdl/opdl_evdev.c
>> @@ -376,6 +376,7 @@ opdl_info_get(struct rte_eventdev *dev, struct
>> rte_event_dev_info *info)
>> .max_num_events = OPDL_INFLIGHT_EVENTS_TOTAL,
>> .event_dev_cap = RTE_EVENT_DEV_CAP_BURST_MODE |
>> RTE_EVENT_DEV_CAP_CARRY_FLOW_ID,
>> + RTE_EVENT_DEV_CAP_MAINTENANCE_FREE,
>> };
> Aha! Replace:
>> RTE_EVENT_DEV_CAP_CARRY_FLOW_ID,
> with:
>> RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
> The "," character was causing the _FREE flag to not be combined with the previous
> flags field, instead being its own (unnamed?) assignment to the struct rte_event_dev_info.
>
> With that , to | change, all looks good to me. -Harry
>
>
I sent a v2. Thanks!
On Wed, Nov 10, 2021 at 3:37 PM Mattias Rönnblom
<mattias.ronnblom@ericsson.com> wrote:
>
> On 2021-11-10 10:55, Van Haaren, Harry wrote:
> >> -----Original Message-----
> >> From: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
> >> Sent: Wednesday, November 10, 2021 9:29 AM
> >> To: jerinj@marvell.com; Jerin Jacob <jerinjacobk@gmail.com>; Sunil Kumar Kori
> >> <skori@marvell.com>
> >> Cc: dev@dpdk.org; Pavan Nikhilesh Bhagavatula <pbhagavatula@marvell.com>;
> >> Hemant Agrawal <hemant.agrawal@nxp.com>; Nipun Gupta
> >> <nipun.gupta@nxp.com>; Van Haaren, Harry <harry.van.haaren@intel.com>;
> >> Thomas Monjalon <thomas@monjalon.net>; David Marchand
> >> <david.marchand@redhat.com>; Gujjar, Abhinandan S
> >> <abhinandan.gujjar@intel.com>; Carrillo, Erik G <erik.g.carrillo@intel.com>;
> >> Jayatheerthan, Jay <jay.jayatheerthan@intel.com>; Yigit, Ferruh
> >> <ferruh.yigit@intel.com>; Akhil Goyal <gakhil@marvell.com>; mattias.ronnblom
> >> <mattias.ronnblom@ericsson.com>
> >> Subject: [PATCH] eventdev: negate maintenance capability flag
> >>
> >> Replace RTE_EVENT_DEV_CAP_REQUIRES_MAINT, which signaled the need
> >> for the application to call rte_event_maintain(), with
> >> RTE_EVENT_DEV_CAP_MAINTENANCE_FREE, which does the opposite (i.e.,
> >> signifies that the event device does not require maintenance).
> >>
> >> This approach is more in line with how other eventdev hardware and/or
> >> software limitations are handled in the Eventdev API.
> >>
> >> Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
> > Generally patch looks fine to me, but on testing I got a strange warning from the
> > compiler on the OPDL driver:
> >
> > In file included from ../drivers/event/opdl/opdl_evdev.h:8,
> > from ../drivers/event/opdl/opdl_evdev.c:15:
> > ../drivers/event/opdl/opdl_evdev.c: In function ‘opdl_info_get’:
> > ../lib/eventdev/rte_eventdev.h:302:44: warning: conversion from ‘long long unsigned int’ to ‘un
> > signed char’ changes value from ‘1024’ to ‘0’ [-Woverflow]
> > 302 | #define RTE_EVENT_DEV_CAP_MAINTENANCE_FREE (1ULL << 10)
> > | ^
> > ../drivers/event/opdl/opdl_evdev.c:379:34: note: in expansion of macro ‘RTE_EVENT_DEV_CAP_MAINT
> > ENANCE_FREE’
> > 379 | RTE_EVENT_DEV_CAP_MAINTENANCE_FREE,
> > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >
> > Investigating, RTE_EVENT_DEV_CAP_* are assigned to struct rte_event_dev_info::event_dev_cap,
> > which is a uint32_t. The error suggests that the variable being assigned to is a "unsigned char"?
> >
> > <snip>
> >
> >> static void
> >> diff --git a/drivers/event/opdl/opdl_evdev.c b/drivers/event/opdl/opdl_evdev.c
> >> index 5007e9a7bf..787ee4713d 100644
> >> --- a/drivers/event/opdl/opdl_evdev.c
> >> +++ b/drivers/event/opdl/opdl_evdev.c
> >> @@ -376,6 +376,7 @@ opdl_info_get(struct rte_eventdev *dev, struct
> >> rte_event_dev_info *info)
> >> .max_num_events = OPDL_INFLIGHT_EVENTS_TOTAL,
> >> .event_dev_cap = RTE_EVENT_DEV_CAP_BURST_MODE |
> >> RTE_EVENT_DEV_CAP_CARRY_FLOW_ID,
> >> + RTE_EVENT_DEV_CAP_MAINTENANCE_FREE,
> >> };
> > Aha! Replace:
> >> RTE_EVENT_DEV_CAP_CARRY_FLOW_ID,
> > with:
> >> RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
> > The "," character was causing the _FREE flag to not be combined with the previous
> > flags field, instead being its own (unnamed?) assignment to the struct rte_event_dev_info.
> >
> > With that , to | change, all looks good to me. -Harry
> >
> >
> I sent a v2. Thanks!
v2 has doc build error. Please send v3
http://mails.dpdk.org/archives/test-report/2021-November/241649.html
>
@@ -44,8 +44,8 @@ Port Maintenance
~~~~~~~~~~~~~~~~
The distributed software eventdev uses an internal signaling scheme
-between the ports to achieve load balancing. Therefore, it sets the
-``RTE_EVENT_DEV_CAP_REQUIRES_MAINT`` flag.
+between the ports to achieve load balancing. Therefore, it does not
+set the ``RTE_EVENT_DEV_CAP_MAINTENANCE_FREE`` flag.
During periods when the application thread using a particular port is
neither attempting to enqueue nor to dequeue events, it must
@@ -119,7 +119,8 @@ cnxk_sso_info_get(struct cnxk_sso_evdev *dev,
RTE_EVENT_DEV_CAP_RUNTIME_PORT_LINK |
RTE_EVENT_DEV_CAP_MULTIPLE_QUEUE_PORT |
RTE_EVENT_DEV_CAP_NONSEQ_MODE |
- RTE_EVENT_DEV_CAP_CARRY_FLOW_ID;
+ RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
+ RTE_EVENT_DEV_CAP_MAINTENANCE_FREE;
}
int
@@ -66,7 +66,8 @@ static struct rte_event_dev_info evdev_dlb2_default_info = {
RTE_EVENT_DEV_CAP_BURST_MODE |
RTE_EVENT_DEV_CAP_DISTRIBUTED_SCHED |
RTE_EVENT_DEV_CAP_IMPLICIT_RELEASE_DISABLE |
- RTE_EVENT_DEV_CAP_QUEUE_ALL_TYPES),
+ RTE_EVENT_DEV_CAP_QUEUE_ALL_TYPES |
+ RTE_EVENT_DEV_CAP_MAINTENANCE_FREE),
};
struct process_local_port_data
@@ -356,7 +356,8 @@ dpaa_event_dev_info_get(struct rte_eventdev *dev,
RTE_EVENT_DEV_CAP_BURST_MODE |
RTE_EVENT_DEV_CAP_MULTIPLE_QUEUE_PORT |
RTE_EVENT_DEV_CAP_NONSEQ_MODE |
- RTE_EVENT_DEV_CAP_CARRY_FLOW_ID;
+ RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
+ RTE_EVENT_DEV_CAP_MAINTENANCE_FREE;
}
static int
@@ -408,7 +408,8 @@ dpaa2_eventdev_info_get(struct rte_eventdev *dev,
RTE_EVENT_DEV_CAP_MULTIPLE_QUEUE_PORT |
RTE_EVENT_DEV_CAP_NONSEQ_MODE |
RTE_EVENT_DEV_CAP_QUEUE_ALL_TYPES |
- RTE_EVENT_DEV_CAP_CARRY_FLOW_ID;
+ RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
+ RTE_EVENT_DEV_CAP_MAINTENANCE_FREE;
}
@@ -222,8 +222,7 @@ dsw_info_get(struct rte_eventdev *dev __rte_unused,
RTE_EVENT_DEV_CAP_DISTRIBUTED_SCHED|
RTE_EVENT_DEV_CAP_NONSEQ_MODE|
RTE_EVENT_DEV_CAP_MULTIPLE_QUEUE_PORT|
- RTE_EVENT_DEV_CAP_CARRY_FLOW_ID|
- RTE_EVENT_DEV_CAP_REQUIRES_MAINT
+ RTE_EVENT_DEV_CAP_CARRY_FLOW_ID
};
}
@@ -155,7 +155,8 @@ ssovf_info_get(struct rte_eventdev *dev, struct rte_event_dev_info *dev_info)
RTE_EVENT_DEV_CAP_RUNTIME_PORT_LINK |
RTE_EVENT_DEV_CAP_MULTIPLE_QUEUE_PORT |
RTE_EVENT_DEV_CAP_NONSEQ_MODE |
- RTE_EVENT_DEV_CAP_CARRY_FLOW_ID;
+ RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
+ RTE_EVENT_DEV_CAP_MAINTENANCE_FREE;
}
@@ -505,7 +505,8 @@ otx2_sso_info_get(struct rte_eventdev *event_dev,
RTE_EVENT_DEV_CAP_RUNTIME_PORT_LINK |
RTE_EVENT_DEV_CAP_MULTIPLE_QUEUE_PORT |
RTE_EVENT_DEV_CAP_NONSEQ_MODE |
- RTE_EVENT_DEV_CAP_CARRY_FLOW_ID;
+ RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
+ RTE_EVENT_DEV_CAP_MAINTENANCE_FREE;
}
static void
@@ -376,6 +376,7 @@ opdl_info_get(struct rte_eventdev *dev, struct rte_event_dev_info *info)
.max_num_events = OPDL_INFLIGHT_EVENTS_TOTAL,
.event_dev_cap = RTE_EVENT_DEV_CAP_BURST_MODE |
RTE_EVENT_DEV_CAP_CARRY_FLOW_ID,
+ RTE_EVENT_DEV_CAP_MAINTENANCE_FREE,
};
*info = evdev_opdl_info;
@@ -102,7 +102,8 @@ skeleton_eventdev_info_get(struct rte_eventdev *dev,
dev_info->event_dev_cap = RTE_EVENT_DEV_CAP_QUEUE_QOS |
RTE_EVENT_DEV_CAP_BURST_MODE |
RTE_EVENT_DEV_CAP_EVENT_QOS |
- RTE_EVENT_DEV_CAP_CARRY_FLOW_ID;
+ RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
+ RTE_EVENT_DEV_CAP_MAINTENANCE_FREE;
}
static int
@@ -609,7 +609,8 @@ sw_info_get(struct rte_eventdev *dev, struct rte_event_dev_info *info)
RTE_EVENT_DEV_CAP_RUNTIME_PORT_LINK |
RTE_EVENT_DEV_CAP_MULTIPLE_QUEUE_PORT |
RTE_EVENT_DEV_CAP_NONSEQ_MODE |
- RTE_EVENT_DEV_CAP_CARRY_FLOW_ID),
+ RTE_EVENT_DEV_CAP_CARRY_FLOW_ID |
+ RTE_EVENT_DEV_CAP_MAINTENANCE_FREE),
};
*info = evdev_sw_info;
@@ -299,13 +299,14 @@ struct rte_event;
* the content of this field is implementation dependent.
*/
-#define RTE_EVENT_DEV_CAP_REQUIRES_MAINT (1ULL << 10)
-/**< Event device requires calls to rte_event_maintain() during
- * periods when neither rte_event_dequeue_burst() nor
- * rte_event_enqueue_burst() are called on a port. This will allow the
- * event device to perform internal processing, such as flushing
- * buffered events, return credits to a global pool, or process
- * signaling related to load balancing.
+#define RTE_EVENT_DEV_CAP_MAINTENANCE_FREE (1ULL << 10)
+/**< Event device *does not* require calls to rte_event_maintain().
+ * An event device that does not set this flag requires calls to
+ * rte_event_maintain() during periods when neither
+ * rte_event_dequeue_burst() nor rte_event_enqueue_burst() are called
+ * on a port. This will allow the event device to perform internal
+ * processing, such as flushing buffered events, return credits to a
+ * global pool, or process signaling related to load balancing.
*/
/* Event device priority levels */
@@ -2082,8 +2083,8 @@ rte_event_dequeue_burst(uint8_t dev_id, uint8_t port_id, struct rte_event ev[],
/**
* Maintain an event device.
*
- * This function is only relevant for event devices which have the
- * @ref RTE_EVENT_DEV_CAP_REQUIRES_MAINT flag set. Such devices
+ * This function is only relevant for event devices which do not have
+ * the @ref RTE_EVENT_DEV_CAP_MAINTENANCE_FREE flag set. Such devices
* require an application thread using a particular port to
* periodically call rte_event_maintain() on that port during periods
* which it is neither attempting to enqueue events to nor dequeue