[dpdk-dev,1/7] ethdev: remove unused flag from header
Checks
Commit Message
remove RTE_ETHDEV_HAS_LRO_SUPPORT flag from header.
Flag seems added with the patch that adds LRO support, and intention
looks like giving a pointer to application that library supports LRO.
Fixes: 8eecb3295aed ("ixgbe: add LRO support")
Cc: vladz@cloudius-systems.com
Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
lib/librte_ether/rte_ethdev.h | 3 ---
1 file changed, 3 deletions(-)
Comments
Hello Ferruh,
(Somehow it seems that patch was not delivered to dev@dpdk ML).
One comment below:
<Snip>
[]
>
> There is no update in header content and since ethdev.h included by
> ethdev_driver.h, nothing changed from driver point of view, only
> logically grouping of APIs. From applications point of view they can't
> access to driver specific APIs anymore and they shouldn't.
>
> More PMD specific data structures still remain in ethdev.h because of
> inline functions in header use them. Those will be handled separately.
+1
>
> Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
> ---
[...]
> diff --git a/drivers/bus/dpaa/include/fman.h
> b/drivers/bus/dpaa/include/fman.h
> index 9890e09ce..fec154545 100644
> --- a/drivers/bus/dpaa/include/fman.h
> +++ b/drivers/bus/dpaa/include/fman.h
> @@ -44,7 +44,7 @@
> #include <stdbool.h>
> #include <net/if.h>
>
> -#include <rte_ethdev.h>
> +#include <rte_ethdev_driver.h>
> #include <rte_ether.h>
>
> #include <compat.h>
> diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
> index 480857e57..8f4993f3a 100644
> --- a/drivers/bus/fslmc/fslmc_bus.c
> +++ b/drivers/bus/fslmc/fslmc_bus.c
> @@ -40,7 +40,7 @@
> #include <rte_malloc.h>
> #include <rte_devargs.h>
> #include <rte_memcpy.h>
> -#include <rte_ethdev.h>
> +#include <rte_ethdev_driver.h>
>
> #include <rte_fslmc.h>
> #include <fslmc_vfio.h>
> diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
> index 7831201ad..eec5cdd94 100644
> --- a/drivers/bus/fslmc/fslmc_vfio.c
> +++ b/drivers/bus/fslmc/fslmc_vfio.c
> @@ -48,7 +48,7 @@
>
> #include <eal_filesystem.h>
> #include <rte_mbuf.h>
> -#include <rte_ethdev.h>
> +#include <rte_ethdev_driver.h>
> #include <rte_malloc.h>
> #include <rte_memcpy.h>
> #include <rte_string_fns.h>
> diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
> b/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
> index 334e1f5ad..55c38fd8d 100644
> --- a/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
> +++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
> @@ -45,7 +45,7 @@
> #include <rte_cycles.h>
> #include <rte_kvargs.h>
> #include <rte_dev.h>
> -#include <rte_ethdev.h>
> +#include <rte_ethdev_driver.h>
>
> #include <fslmc_logs.h>
> #include <rte_fslmc.h>
> diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpci.c
> b/drivers/bus/fslmc/portal/dpaa2_hw_dpci.c
> index ae189c72a..7a86d047a 100644
> --- a/drivers/bus/fslmc/portal/dpaa2_hw_dpci.c
> +++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpci.c
> @@ -44,7 +44,7 @@
> #include <rte_cycles.h>
> #include <rte_kvargs.h>
> #include <rte_dev.h>
> -#include <rte_ethdev.h>
> +#include <rte_ethdev_driver.h>
>
> #include <fslmc_logs.h>
> #include <rte_fslmc.h>
> diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
> b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
> index f00070f3f..6a6c8c3b6 100644
> --- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
> +++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
> @@ -50,7 +50,7 @@
> #include<sys/eventfd.h>
>
> #include <rte_mbuf.h>
> -#include <rte_ethdev.h>
> +#include <rte_ethdev_driver.h>
> #include <rte_malloc.h>
> #include <rte_memcpy.h>
> #include <rte_string_fns.h>
> diff --git a/drivers/event/dpaa2/dpaa2_eventdev.c
> b/drivers/event/dpaa2/dpaa2_eventdev.c
> index eeeb23127..a20a556d4 100644
> --- a/drivers/event/dpaa2/dpaa2_eventdev.c
> +++ b/drivers/event/dpaa2/dpaa2_eventdev.c
> @@ -52,7 +52,7 @@
> #include <rte_memory.h>
> #include <rte_pci.h>
> #include <rte_bus_vdev.h>
> -#include <rte_ethdev.h>
> +#include <rte_ethdev_driver.h>
> #include <rte_event_eth_rx_adapter.h>
>
> #include <fslmc_vfio.h>
> diff --git a/drivers/event/dpaa2/dpaa2_hw_dpcon.c
> b/drivers/event/dpaa2/dpaa2_hw_dpcon.c
> index 005e6234d..a8da9803b 100644
> --- a/drivers/event/dpaa2/dpaa2_hw_dpcon.c
> +++ b/drivers/event/dpaa2/dpaa2_hw_dpcon.c
> @@ -44,7 +44,7 @@
> #include <rte_cycles.h>
> #include <rte_kvargs.h>
> #include <rte_dev.h>
> -#include <rte_ethdev.h>
> +#include <rte_ethdev_driver.h>
>
> #include <rte_fslmc.h>
> #include <mc/fsl_dpcon.h>
With respect to DPAA and DPAA2 and overall aim of patch:
Acked-by: Shreyansh Jain <shreyansh.jain@nxp.com>
[...]
> diff --git a/lib/librte_ether/rte_ethdev_driver.h
> b/lib/librte_ether/rte_ethdev_driver.h
> new file mode 100644
> index 000000000..3e77d1439
> --- /dev/null
> +++ b/lib/librte_ether/rte_ethdev_driver.h
> @@ -0,0 +1,163 @@
> +/*-
> + * BSD LICENSE
> + *
> + * Copyright(c) 2017 Intel Corporation. All rights reserved.
> + * All rights reserved.
> + *
> + * Redistribution and use in source and binary forms, with or without
> + * modification, are permitted provided that the following conditions
> + * are met:
> + *
> + * * Redistributions of source code must retain the above copyright
> + * notice, this list of conditions and the following disclaimer.
> + * * Redistributions in binary form must reproduce the above copyright
> + * notice, this list of conditions and the following disclaimer in
> + * the documentation and/or other materials provided with the
> + * distribution.
> + * * Neither the name of Intel Corporation nor the names of its
> + * contributors may be used to endorse or promote products derived
> + * from this software without specific prior written permission.
> + *
> + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
> + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
> + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
> + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
> + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
> + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
> + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
> + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
> + */
> +
> +#ifndef _RTE_ETHDEV_DRIVER_H_
> +#define _RTE_ETHDEV_DRIVER_H_
Now that it is almost agreed to move towards SPDX based licensing text [1][2], and this new file is being introduced - do you think it makes sense to do it for this file?
[1] http://dpdk.org/ml/archives/dev/2017-November/082382.html
[2] http://dpdk.org/ml/archives/dev/2017-November/082639.html
[...]
resending after registering with the new email domain ;)
Please, see my comments below.
On 12/01/2017 05:17 PM, Vlad Zolotarov wrote:
>
>
> On 11/30/2017 09:29 PM, Ferruh Yigit wrote:
>> remove RTE_ETHDEV_HAS_LRO_SUPPORT flag from header.
>>
>> Flag seems added with the patch that adds LRO support, and intention
>> looks like giving a pointer to application that library supports LRO.
>
> Exactly. Removing this flag may make the existing application "think"
> that LRO is not supported.
> Why do you want to remove it to begin with?
>
>> Fixes: 8eecb3295aed ("ixgbe: add LRO support")
>> Cc:vladz@cloudius-systems.com
>>
>> Signed-off-by: Ferruh Yigit<ferruh.yigit@intel.com>
>> ---
>> lib/librte_ether/rte_ethdev.h | 3 ---
>> 1 file changed, 3 deletions(-)
>>
>> diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
>> index 341c2d624..e620c3706 100644
>> --- a/lib/librte_ether/rte_ethdev.h
>> +++ b/lib/librte_ether/rte_ethdev.h
>> @@ -172,9 +172,6 @@ extern "C" {
>>
>> #include <stdint.h>
>>
>> -/* Use this macro to check if LRO API is supported */
>> -#define RTE_ETHDEV_HAS_LRO_SUPPORT
>> -
>> #include <rte_log.h>
>> #include <rte_interrupts.h>
>> #include <rte_dev.h>
>
On 12/1/2017 2:17 PM, Vlad Zolotarov wrote:
>
>
> On 11/30/2017 09:29 PM, Ferruh Yigit wrote:
>> remove RTE_ETHDEV_HAS_LRO_SUPPORT flag from header.
>>
>> Flag seems added with the patch that adds LRO support, and intention
>> looks like giving a pointer to application that library supports LRO.
>
> Exactly. Removing this flag may make the existing application "think" that LRO
> is not supported.
> Why do you want to remove it to begin with?
I agree that this is a little controversial, and I don't have a strong opinion
about it, I thought twice before sending or not sending the patch :)
This flag can be useful for the applications that use different versions of the
DPDK library (with the ones does and doesn't support LRO), so that application
can be more portable.
I would understand a dynamic approach, which would be useful for distributed
applications that you don't know and can't control what version of the DPDK
library used.
But here to benefit from this flag you need to compile your application against
DPDK library, and if you are compiling it you already know if your DPDK supports
LRO or not. And this is not something keeps changing platform to platform etc,
after a specific release LRO is always supported.
And this is the only capability support flag in the ethdev, there are many new
features introduced in each release but they are not advertised by a capability
flag. Only having LRO flag can be confusing.
_Perhaps_ DPDK should have a way to expose the supported features, and a dynamic
runtime way can be better option than compile time macros, but it should be
generic nothing specific to LRO support.
>
>>
>> Fixes: 8eecb3295aed ("ixgbe: add LRO support")
>> Cc: vladz@cloudius-systems.com
>>
>> Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
>> ---
>> lib/librte_ether/rte_ethdev.h | 3 ---
>> 1 file changed, 3 deletions(-)
>>
>> diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
>> index 341c2d624..e620c3706 100644
>> --- a/lib/librte_ether/rte_ethdev.h
>> +++ b/lib/librte_ether/rte_ethdev.h
>> @@ -172,9 +172,6 @@ extern "C" {
>>
>> #include <stdint.h>
>>
>> -/* Use this macro to check if LRO API is supported */
>> -#define RTE_ETHDEV_HAS_LRO_SUPPORT
>> -
>> #include <rte_log.h>
>> #include <rte_interrupts.h>
>> #include <rte_dev.h>
>
On 12/01/2017 06:51 PM, Ferruh Yigit wrote:
> On 12/1/2017 2:17 PM, Vlad Zolotarov wrote:
>>
>> On 11/30/2017 09:29 PM, Ferruh Yigit wrote:
>>> remove RTE_ETHDEV_HAS_LRO_SUPPORT flag from header.
>>>
>>> Flag seems added with the patch that adds LRO support, and intention
>>> looks like giving a pointer to application that library supports LRO.
>> Exactly. Removing this flag may make the existing application "think" that LRO
>> is not supported.
>> Why do you want to remove it to begin with?
> I agree that this is a little controversial, and I don't have a strong opinion
> about it, I thought twice before sending or not sending the patch :)
>
> This flag can be useful for the applications that use different versions of the
> DPDK library (with the ones does and doesn't support LRO), so that application
> can be more portable.
> I would understand a dynamic approach, which would be useful for distributed
> applications that you don't know and can't control what version of the DPDK
> library used.
> But here to benefit from this flag you need to compile your application against
> DPDK library, and if you are compiling it you already know if your DPDK supports
> LRO or not. And this is not something keeps changing platform to platform etc,
> after a specific release LRO is always supported.
True but DPDK is usually not a part of your source tree - it (should)
come as a library and currently there isn't any proper way to query
feature set.
This flag was also added after a lot of consideration as a rather ugly
compromise since there wasn't (and there isn't) a proper way to do that
at that (this) time. See more on that below.
>
> And this is the only capability support flag in the ethdev, there are many new
> features introduced in each release but they are not advertised by a capability
> flag. Only having LRO flag can be confusing.
>
> _Perhaps_ DPDK should have a way to expose the supported features, and a dynamic
> runtime way can be better option than compile time macros, but it should be
> generic nothing specific to LRO support.
IMO it's not "perhaps" it's a "must" _however_ you can't remove this
flag without giving some other tool to do the same.
Querying a DPDK version would be a wrong direction because some Linux
distributions (yes, I'm talking about you, Ubuntu) tend to have it's own
trees with their own backports and these trees may be light years ahead
in their patch level compared to vanilla trees with the same version.
>
>>> Fixes: 8eecb3295aed ("ixgbe: add LRO support")
>>> Cc: vladz@cloudius-systems.com
>>>
>>> Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
>>> ---
>>> lib/librte_ether/rte_ethdev.h | 3 ---
>>> 1 file changed, 3 deletions(-)
>>>
>>> diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
>>> index 341c2d624..e620c3706 100644
>>> --- a/lib/librte_ether/rte_ethdev.h
>>> +++ b/lib/librte_ether/rte_ethdev.h
>>> @@ -172,9 +172,6 @@ extern "C" {
>>>
>>> #include <stdint.h>
>>>
>>> -/* Use this macro to check if LRO API is supported */
>>> -#define RTE_ETHDEV_HAS_LRO_SUPPORT
>>> -
>>> #include <rte_log.h>
>>> #include <rte_interrupts.h>
>>> #include <rte_dev.h>
On 12/1/2017 6:22 PM, Vlad Zolotarov wrote:
>
>
> On 12/01/2017 06:51 PM, Ferruh Yigit wrote:
>> On 12/1/2017 2:17 PM, Vlad Zolotarov wrote:
>>>
>>> On 11/30/2017 09:29 PM, Ferruh Yigit wrote:
>>>> remove RTE_ETHDEV_HAS_LRO_SUPPORT flag from header.
>>>>
>>>> Flag seems added with the patch that adds LRO support, and intention
>>>> looks like giving a pointer to application that library supports LRO.
>>> Exactly. Removing this flag may make the existing application "think" that LRO
>>> is not supported.
>>> Why do you want to remove it to begin with?
>> I agree that this is a little controversial, and I don't have a strong opinion
>> about it, I thought twice before sending or not sending the patch :)
>>
>> This flag can be useful for the applications that use different versions of the
>> DPDK library (with the ones does and doesn't support LRO), so that application
>> can be more portable.
>> I would understand a dynamic approach, which would be useful for distributed
>> applications that you don't know and can't control what version of the DPDK
>> library used.
>> But here to benefit from this flag you need to compile your application against
>> DPDK library, and if you are compiling it you already know if your DPDK supports
>> LRO or not. And this is not something keeps changing platform to platform etc,
>> after a specific release LRO is always supported.
>
> True but DPDK is usually not a part of your source tree - it (should) come as a
> library and currently there isn't any proper way to query feature set.
> This flag was also added after a lot of consideration as a rather ugly
> compromise since there wasn't (and there isn't) a proper way to do that at that
> (this) time. See more on that below.
>
>>
>> And this is the only capability support flag in the ethdev, there are many new
>> features introduced in each release but they are not advertised by a capability
>> flag. Only having LRO flag can be confusing.
>>
>> _Perhaps_ DPDK should have a way to expose the supported features, and a dynamic
>> runtime way can be better option than compile time macros, but it should be
>> generic nothing specific to LRO support.
>
> IMO it's not "perhaps" it's a "must" _however_
This is even wider than ethdev, more like in DPDK scope, we need someone to
attack this issue.
> you can't remove this flag without giving some other tool to do the same.
Fair enough, I will drop the patch.
> Querying a DPDK version would be a wrong direction because some Linux
> distributions (yes, I'm talking about you, Ubuntu) tend to have it's own trees
> with their own backports and these trees may be light years ahead in their patch
> level compared to vanilla trees with the same version.
+1 to have better way than version check.
>
>>
>>>> Fixes: 8eecb3295aed ("ixgbe: add LRO support")
>>>> Cc: vladz@cloudius-systems.com
>>>>
>>>> Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
>>>> ---
>>>> lib/librte_ether/rte_ethdev.h | 3 ---
>>>> 1 file changed, 3 deletions(-)
>>>>
>>>> diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
>>>> index 341c2d624..e620c3706 100644
>>>> --- a/lib/librte_ether/rte_ethdev.h
>>>> +++ b/lib/librte_ether/rte_ethdev.h
>>>> @@ -172,9 +172,6 @@ extern "C" {
>>>>
>>>> #include <stdint.h>
>>>>
>>>> -/* Use this macro to check if LRO API is supported */
>>>> -#define RTE_ETHDEV_HAS_LRO_SUPPORT
>>>> -
>>>> #include <rte_log.h>
>>>> #include <rte_interrupts.h>
>>>> #include <rte_dev.h>
>
On 12/04/2017 12:54 PM, Ferruh Yigit wrote:
> On 12/1/2017 6:22 PM, Vlad Zolotarov wrote:
>>
>> On 12/01/2017 06:51 PM, Ferruh Yigit wrote:
>>> On 12/1/2017 2:17 PM, Vlad Zolotarov wrote:
>>>> On 11/30/2017 09:29 PM, Ferruh Yigit wrote:
>>>>> remove RTE_ETHDEV_HAS_LRO_SUPPORT flag from header.
>>>>>
>>>>> Flag seems added with the patch that adds LRO support, and intention
>>>>> looks like giving a pointer to application that library supports LRO.
>>>> Exactly. Removing this flag may make the existing application "think" that LRO
>>>> is not supported.
>>>> Why do you want to remove it to begin with?
>>> I agree that this is a little controversial, and I don't have a strong opinion
>>> about it, I thought twice before sending or not sending the patch :)
>>>
>>> This flag can be useful for the applications that use different versions of the
>>> DPDK library (with the ones does and doesn't support LRO), so that application
>>> can be more portable.
>>> I would understand a dynamic approach, which would be useful for distributed
>>> applications that you don't know and can't control what version of the DPDK
>>> library used.
>>> But here to benefit from this flag you need to compile your application against
>>> DPDK library, and if you are compiling it you already know if your DPDK supports
>>> LRO or not. And this is not something keeps changing platform to platform etc,
>>> after a specific release LRO is always supported.
>> True but DPDK is usually not a part of your source tree - it (should) come as a
>> library and currently there isn't any proper way to query feature set.
>> This flag was also added after a lot of consideration as a rather ugly
>> compromise since there wasn't (and there isn't) a proper way to do that at that
>> (this) time. See more on that below.
>>
>>> And this is the only capability support flag in the ethdev, there are many new
>>> features introduced in each release but they are not advertised by a capability
>>> flag. Only having LRO flag can be confusing.
>>>
>>> _Perhaps_ DPDK should have a way to expose the supported features, and a dynamic
>>> runtime way can be better option than compile time macros, but it should be
>>> generic nothing specific to LRO support.
>> IMO it's not "perhaps" it's a "must" _however_
> This is even wider than ethdev, more like in DPDK scope, we need someone to
> attack this issue.
absolutely!
>
>> you can't remove this flag without giving some other tool to do the same.
> Fair enough, I will drop the patch.
>
>> Querying a DPDK version would be a wrong direction because some Linux
>> distributions (yes, I'm talking about you, Ubuntu) tend to have it's own trees
>> with their own backports and these trees may be light years ahead in their patch
>> level compared to vanilla trees with the same version.
> +1 to have better way than version check.
>
>>>>> Fixes: 8eecb3295aed ("ixgbe: add LRO support")
>>>>> Cc: vladz@cloudius-systems.com
>>>>>
>>>>> Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
>>>>> ---
>>>>> lib/librte_ether/rte_ethdev.h | 3 ---
>>>>> 1 file changed, 3 deletions(-)
>>>>>
>>>>> diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
>>>>> index 341c2d624..e620c3706 100644
>>>>> --- a/lib/librte_ether/rte_ethdev.h
>>>>> +++ b/lib/librte_ether/rte_ethdev.h
>>>>> @@ -172,9 +172,6 @@ extern "C" {
>>>>>
>>>>> #include <stdint.h>
>>>>>
>>>>> -/* Use this macro to check if LRO API is supported */
>>>>> -#define RTE_ETHDEV_HAS_LRO_SUPPORT
>>>>> -
>>>>> #include <rte_log.h>
>>>>> #include <rte_interrupts.h>
>>>>> #include <rte_dev.h>
@@ -172,9 +172,6 @@ extern "C" {
#include <stdint.h>
-/* Use this macro to check if LRO API is supported */
-#define RTE_ETHDEV_HAS_LRO_SUPPORT
-
#include <rte_log.h>
#include <rte_interrupts.h>
#include <rte_dev.h>