DPDK patches and discussions
 help / color / mirror / Atom feed
From: Tyler Retzlaff <roretzla@linux.microsoft.com>
To: Robin Jarry <rjarry@redhat.com>
Cc: dev@dpdk.org, Jerin Jacob <jerinj@marvell.com>,
	Kiran Kumar K <kirankumark@marvell.com>,
	Nithin Dabilpuram <ndabilpuram@marvell.com>,
	Zhirun Yan <yanzhirun_163@163.com>
Subject: Re: [PATCH v2] graph: expose node context as pointers
Date: Fri, 22 Mar 2024 09:56:15 -0700	[thread overview]
Message-ID: <20240322165615.GA31848@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net> (raw)
In-Reply-To: <20240322163130.671185-2-rjarry@redhat.com>

On Fri, Mar 22, 2024 at 05:31:31PM +0100, Robin Jarry wrote:
> In some cases, the node context data is used to store two pointers
> because the data is larger than the reserved 16 bytes. Having to define
> intermediate structures just to be able to cast is tedious. Add two
> pointers that take the same space than ctx.
> 
> Signed-off-by: Robin Jarry <rjarry@redhat.com>
> ---
> 
> Notes:
>     v2:
>     
>     * Added __extension__ (not sure where it is needed, I don't have access to windows).

i can answer this!

windows toolchains will only require __extension__ qualification on use
of statement expressions, so msvc won't require any use of __extension__
in this patch.

as a general rule of thumb __extension__ is something you may choose to
use for any gcc compiled code that is an extension to standard C and you
intend to use the -pedantic flag (i.e. -std=c11 && -pedantic used together)

i've commented inline below.

>     * It still fails the header check for C++. It seems not possible to align an unnamed union...
>       Tyler, do you have an idea about how to fix that?

yes, see below what i suspect you want.

>     * Added static_assert to ensure the anonymous union is not larger than RTE_NODE_CTX_SZ.
> 
>  lib/graph/rte_graph_worker_common.h | 12 +++++++++++-
>  1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/lib/graph/rte_graph_worker_common.h b/lib/graph/rte_graph_worker_common.h
> index 36d864e2c14e..a60c2bc3f0c3 100644
> --- a/lib/graph/rte_graph_worker_common.h
> +++ b/lib/graph/rte_graph_worker_common.h
> @@ -112,7 +112,14 @@ struct __rte_cache_aligned rte_node {
>  	};
>  	/* Fast path area  */
>  #define RTE_NODE_CTX_SZ 16
> -	alignas(RTE_CACHE_LINE_SIZE) uint8_t ctx[RTE_NODE_CTX_SZ]; /**< Node Context. */
> +	__extension__ alignas(RTE_CACHE_LINE_SIZE) union {

__extension__ should not be on the anonymous union (since they are standard C11).

anonymous union declaration is actually a type with no name and then a data
field of that type so __rte_aligned is most likely what you want, since
you're using RTE_CACHE_LINE_SIZE we can use __rte_cache_aligned.

union __rte_cache_aligned {
   ... your union fields ...
};

and i think checkpatches still gives a warning unrelated to alignment
for this but it can be safely ignored. it's the warning about alignment
that we care about and should be fixed.

> +		uint8_t ctx[RTE_NODE_CTX_SZ];
> +		/* Convenience aliases to store pointers without complex casting. */
> +		__extension__ struct {

this is correct/recommended since anonymous structs aren't standard,
with the __extension__ -pedantic won't emit a warning (our intention).

> +			void *ctx_ptr;
> +			void *ctx_ptr2;
> +		};
> +	}; /**< Node Context. */
>  	uint16_t size;		/**< Total number of objects available. */
>  	uint16_t idx;		/**< Number of objects used. */
>  	rte_graph_off_t off;	/**< Offset of node in the graph reel. */
> @@ -130,6 +137,9 @@ struct __rte_cache_aligned rte_node {
>  	alignas(RTE_CACHE_LINE_MIN_SIZE) struct rte_node *nodes[]; /**< Next nodes. */
>  };
>  
> +static_assert(offsetof(struct rte_node, size) - offsetof(struct rte_node, ctx) == RTE_NODE_CTX_SZ,
> +	"The node context anonymous union cannot be larger than RTE_NODE_CTX_SZ");
> +

you should include directly include <stddef.h> in this file for use of offsetof.
you should include directly include <assert.h> in this file for use of the static_assert.

hope this helps!

ty

>  /**
>   * @internal
>   *
> -- 
> 2.44.0

  reply	other threads:[~2024-03-22 16:56 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-22 16:31 Robin Jarry
2024-03-22 16:56 ` Tyler Retzlaff [this message]
2024-03-22 23:41   ` Robin Jarry

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=20240322165615.GA31848@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net \
    --to=roretzla@linux.microsoft.com \
    --cc=dev@dpdk.org \
    --cc=jerinj@marvell.com \
    --cc=kirankumark@marvell.com \
    --cc=ndabilpuram@marvell.com \
    --cc=rjarry@redhat.com \
    --cc=yanzhirun_163@163.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).