dev.dpdk.org archive mirror
 help / color / mirror / Atom feed
From: Ferruh Yigit <ferruh.yigit@intel.com>
To: John McNamara <john.mcnamara@intel.com>,
	Marko Kovacevic <marko.kovacevic@intel.com>
Cc: dev@dpdk.org, David Marchand <david.marchand@redhat.com>,
	Vamsi Attunuru <vattunuru@marvell.com>,
	Kiran Kumar K <kirankumark@marvell.com>,
	Jerin Jacob <jerinj@marvell.com>,
	Igor Ryzhov <iryzhov@nfware.com>
Subject: Re: [dpdk-dev] [PATCH v2] kni: increase kernel version requirement for VA
Date: Thu, 28 Nov 2019 12:13:16 +0000	[thread overview]
Message-ID: <c07ceb9b-b9e5-a240-a141-4077e79f055d@intel.com> (raw)
In-Reply-To: <20191120172227.42798-1-ferruh.yigit@intel.com>

On 11/20/2019 5:22 PM, Ferruh Yigit wrote:
> A build error reported related to the selected
> 'get_user_pages_remote()' kernel API:
> 
> .../kernel/linux/kni/kni_dev.h:113:8:
>   error: too few arguments to function ‘get_user_pages_remote’
>   ret = get_user_pages_remote(tsk, tsk->mm, iova, 1
>         ^~~~~~~~~~~~~~~~~~~~~
> 
> Currently there are three version of the 'get_user_pages_remote()'
> supported, based on kernel version
> < 4.9, = 4.9, > 4.9
> 
> These version based checks are not working fine with the distro kernels
> which is the cause of reported build error. The error reported by the
> kernel version 4.8, but it is using API defined in > 4.9.
> 
> To be able to take control of this, and possible more, related build
> error, increasing the minimum supported kernel version for iova=va with
> KNI to kernel version 4.9.
> 
> This leaves us with single version of the kernel API and more
> manageable.
> 
> Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
> ---
> Cc: David Marchand <david.marchand@redhat.com>
> Cc: Vamsi Attunuru <vattunuru@marvell.com>
> Cc: Kiran Kumar K <kirankumark@marvell.com>
> Cc: Jerin Jacob <jerinj@marvell.com>
> 
> v2:
> * fixed typo in comment
> Cc: Igor Ryzhov <iryzhov@nfware.com>
> ---
>  doc/guides/prog_guide/kernel_nic_interface.rst |  2 +-
>  doc/guides/rel_notes/release_19_11.rst         |  2 +-
>  kernel/linux/kni/compat.h                      | 17 ++++++-----------
>  kernel/linux/kni/kni_dev.h                     | 10 ----------
>  lib/librte_eal/linux/eal/eal.c                 |  6 +++---
>  lib/librte_kni/rte_kni.c                       |  2 +-
>  6 files changed, 12 insertions(+), 27 deletions(-)
> 
> diff --git a/doc/guides/prog_guide/kernel_nic_interface.rst b/doc/guides/prog_guide/kernel_nic_interface.rst
> index c4479ffbf..848b00253 100644
> --- a/doc/guides/prog_guide/kernel_nic_interface.rst
> +++ b/doc/guides/prog_guide/kernel_nic_interface.rst
> @@ -305,7 +305,7 @@ IOVA = VA: Support
>  
>  KNI operates in IOVA_VA scheme when
>  
> -- LINUX_VERSION_CODE >= KERNEL_VERSION(4, 6, 0) and
> +- LINUX_VERSION_CODE > KERNEL_VERSION(4, 9, 0) and
>  - EAL option `iova-mode=va` is passed or bus IOVA scheme in the DPDK is selected
>    as RTE_IOVA_VA.
>  
> diff --git a/doc/guides/rel_notes/release_19_11.rst b/doc/guides/rel_notes/release_19_11.rst
> index 21be600ab..45b58190c 100644
> --- a/doc/guides/rel_notes/release_19_11.rst
> +++ b/doc/guides/rel_notes/release_19_11.rst
> @@ -299,7 +299,7 @@ New Features
>    * Added IOVA = VA support for KNI, KNI can operate in IOVA = VA mode when
>      `iova-mode=va` EAL option is passed to the application or when bus IOVA
>      scheme is selected as RTE_IOVA_VA. This mode only works on Linux Kernel
> -    versions 4.6.0 and above.
> +    versions above 4.9.0.
>  
>    * Due to IOVA to KVA address translations, based on the KNI use case there
>      can be a performance impact. For mitigation, forcing IOVA to PA via EAL
> diff --git a/kernel/linux/kni/compat.h b/kernel/linux/kni/compat.h
> index 062b170ef..83ecac2d8 100644
> --- a/kernel/linux/kni/compat.h
> +++ b/kernel/linux/kni/compat.h
> @@ -122,16 +122,11 @@
>  #define HAVE_SIGNAL_FUNCTIONS_OWN_HEADER
>  #endif
>  
> -#if KERNEL_VERSION(4, 6, 0) <= LINUX_VERSION_CODE
> -
> +/*
> + * iova to kva mapping support can be provided since 4.6.0, but required
> + * kernel version increased to > 4.9.0 because of the updates in
> + * get_user_pages_remote() kernel API
> + */
> +#if KERNEL_VERSION(4, 9, 0) < LINUX_VERSION_CODE
>  #define HAVE_IOVA_TO_KVA_MAPPING_SUPPORT
> -
> -#if KERNEL_VERSION(4, 9, 0) > LINUX_VERSION_CODE
> -#define GET_USER_PAGES_REMOTE_API_V1
> -#elif KERNEL_VERSION(4, 9, 0) == LINUX_VERSION_CODE
> -#define GET_USER_PAGES_REMOTE_API_V2
> -#else
> -#define GET_USER_PAGES_REMOTE_API_V3
> -#endif
> -
>  #endif
> diff --git a/kernel/linux/kni/kni_dev.h b/kernel/linux/kni/kni_dev.h
> index fb641b696..5e75c6371 100644
> --- a/kernel/linux/kni/kni_dev.h
> +++ b/kernel/linux/kni/kni_dev.h
> @@ -101,18 +101,8 @@ static inline phys_addr_t iova_to_phys(struct task_struct *tsk,
>  	offset = iova & (PAGE_SIZE - 1);
>  
>  	/* Read one page struct info */
> -#ifdef GET_USER_PAGES_REMOTE_API_V3
>  	ret = get_user_pages_remote(tsk, tsk->mm, iova, 1,
>  				    FOLL_TOUCH, &page, NULL, NULL);
> -#endif
> -#ifdef GET_USER_PAGES_REMOTE_API_V2
> -	ret = get_user_pages_remote(tsk, tsk->mm, iova, 1,
> -				    FOLL_TOUCH, &page, NULL);
> -#endif
> -#ifdef GET_USER_PAGES_REMOTE_API_V1
> -	ret = get_user_pages_remote(tsk, tsk->mm, iova, 1
> -				    0, 0, &page, NULL);
> -#endif
>  	if (ret < 0)
>  		return 0;
>  
> diff --git a/lib/librte_eal/linux/eal/eal.c b/lib/librte_eal/linux/eal/eal.c
> index b5b71500c..5879e33e5 100644
> --- a/lib/librte_eal/linux/eal/eal.c
> +++ b/lib/librte_eal/linux/eal/eal.c
> @@ -1073,7 +1073,7 @@ rte_eal_init(int argc, char **argv)
>  				 */
>  				iova_mode = RTE_IOVA_VA;
>  				RTE_LOG(DEBUG, EAL, "Physical addresses are unavailable, selecting IOVA as VA mode.\n");
> -#if defined(RTE_LIBRTE_KNI) && LINUX_VERSION_CODE >= KERNEL_VERSION(4, 6, 0)
> +#if defined(RTE_LIBRTE_KNI) && LINUX_VERSION_CODE > KERNEL_VERSION(4, 9, 0)
>  			} else if (rte_eal_check_module("rte_kni") == 1) {
>  				iova_mode = RTE_IOVA_PA;
>  				RTE_LOG(DEBUG, EAL, "KNI is loaded, selecting IOVA as PA mode for better KNI perfomance.\n");

Btw, since issue is revisited, I have a question about this check,

This check is to favor PA when bus mode is DC for the KNI performance concerns,
For this selection, why we need the kernel version check at all?
KNI can work in the PA mode without any specific kernel version dependency, so
why not if bus mode is DC and kni kernel module is inserted, select iova mode as
PA independent from kernel version?

  parent reply	other threads:[~2019-11-28 12:13 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-20 16:00 [dpdk-dev] [PATCH] kni: increase kernel version requirement for VA Ferruh Yigit
2019-11-20 16:24 ` Igor Ryzhov
2019-11-20 16:26   ` Ferruh Yigit
2019-11-20 17:22 ` [dpdk-dev] [PATCH v2] " Ferruh Yigit
2019-11-20 17:28   ` David Marchand
2019-11-20 22:30     ` David Marchand
2019-11-28 10:44       ` Andrew Rybchenko
2019-11-28 10:50         ` Thomas Monjalon
2019-11-28 11:01         ` Jerin Jacob
2019-11-28 11:02         ` David Marchand
2019-11-28 11:05           ` Ferruh Yigit
2019-11-21  1:58     ` Jerin Jacob
2019-11-21  2:18       ` [dpdk-dev] [EXT] " Vamsi Krishna Attunuru
2019-11-28 12:13   ` Ferruh Yigit [this message]
2019-11-28 14:58     ` Vamsi Krishna Attunuru

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=c07ceb9b-b9e5-a240-a141-4077e79f055d@intel.com \
    --to=ferruh.yigit@intel.com \
    --cc=david.marchand@redhat.com \
    --cc=dev@dpdk.org \
    --cc=iryzhov@nfware.com \
    --cc=jerinj@marvell.com \
    --cc=john.mcnamara@intel.com \
    --cc=kirankumark@marvell.com \
    --cc=marko.kovacevic@intel.com \
    --cc=vattunuru@marvell.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).