DPDK patches and discussions
 help / color / mirror / Atom feed
From: Santosh Shukla <santosh.shukla@caviumnetworks.com>
To: Shreyansh Jain <shreyansh.jain@nxp.com>
Cc: olivier.matz@6wind.com, dev@dpdk.org, hemant.agrawal@nxp.com
Subject: Re: [dpdk-dev] [RFC PATCH] test/test: support default mempool autotest
Date: Fri, 31 Mar 2017 20:41:16 +0530	[thread overview]
Message-ID: <20170331151100.GA11069@santosh-Latitude-E5530-non-vPro> (raw)
In-Reply-To: <1490955469-16216-1-git-send-email-shreyansh.jain@nxp.com>

On Fri, Mar 31, 2017 at 03:47:49PM +0530, Shreyansh Jain wrote:
> Mempool test currently supports:
>  * ring_mp_mc
>  * stack
> 
> In case a new mempool handler is added, there are multiple options
> for supporting that in the mempool autotest:
> 1. Like the patch below, adding a new default pool options
>    So, ring* + stack + default (which can be 'stack' or 'ring')
>  * This way, whatever the value of RTE_MBUF_DEFAULT_MEMPOOL_OPS is set,
>    it would be verified.
>  * even if that means duplicating some test (for example when "stack" is
>    set as default and it already part of standard test)
> 
> 2. Removing stack handler as standard, and moving only to one specified
>    by RTE_MBUF_DEFAULT_MEMPOOL_OPS (+ existing ring*)
>  * It still leaves space for duplication of ring_mp_mc in case that is
>    set to default (as in case of master tree)
> 
> 3. Iterating over the list of mempool handlers and performing a set
>    or predefined tests
>  * reqiures quite a lot of rewrite of mempool autotest
>  * specially, allowing some special tests (cache/no-cache) cases when
>    a set of variables in loop are being used, would be tricky
> 
> 4. only checking the default pool set by RTE_* macro
>  * In case a user has build DPDK using a configured value, probably it
>    expected that application (or custom applications) would use that
>    default handler.
>  * would also mean that non-default (non RTE_* value) would not be tested
>    even though they are being used.
> 
> The decision above would impact how new mempool handlers are added and
> how their testing (API verification) can be done without impacting the
> mempool_autotest file everytime.
> 
> Signed-off-by: Shreyansh Jain <shreyansh.jain@nxp.com>
> 
> ---
>  test/test/test_mempool.c | 35 +++++++++++++++++++++++++++++++++--
>  1 file changed, 33 insertions(+), 2 deletions(-)
>

I am too using option-2) for my ext-mempool driver and Yes, there is a chance
of 'ring_mp_mc' duplication if left as default.

BTW: we need similar changes at mempool_perf test too, right? 
Are you planning to send, if not then I will propose one. Let me know.

Thanks.

> diff --git a/test/test/test_mempool.c b/test/test/test_mempool.c
> index b9880b3..aefbf80 100644
> --- a/test/test/test_mempool.c
> +++ b/test/test/test_mempool.c
> @@ -509,9 +509,11 @@ walk_cb(struct rte_mempool *mp, void *userdata __rte_unused)
>  static int
>  test_mempool(void)
>  {
> +	int ret = -1;
>  	struct rte_mempool *mp_cache = NULL;
>  	struct rte_mempool *mp_nocache = NULL;
>  	struct rte_mempool *mp_stack = NULL;
> +	struct rte_mempool *default_pool = NULL;
>  
>  	rte_atomic32_init(&synchro);
>  
> @@ -561,6 +563,30 @@ test_mempool(void)
>  	}
>  	rte_mempool_obj_iter(mp_stack, my_obj_init, NULL);
>  
> +	/* Create a mempool based on Default handler, if not "stack" */
> +	printf("Testing %s mempool handler\n",
> +	       RTE_MBUF_DEFAULT_MEMPOOL_OPS);
> +	default_pool = rte_mempool_create_empty("default_pool",
> +					MEMPOOL_SIZE,
> +					MEMPOOL_ELT_SIZE,
> +					RTE_MEMPOOL_CACHE_MAX_SIZE, 0,
> +					SOCKET_ID_ANY, 0);
> +
> +	if (default_pool == NULL) {
> +		printf("cannot allocate default mempool\n");
> +		goto err;
> +	}
> +	if (rte_mempool_set_ops_byname(default_pool,
> +			RTE_MBUF_DEFAULT_MEMPOOL_OPS, NULL) < 0) {
> +		printf("cannot set default handler\n");
> +		goto err;
> +	}
> +	if (rte_mempool_populate_default(default_pool) < 0) {
> +		printf("cannot populate default mempool\n");
> +		goto err;
> +	}
> +	rte_mempool_obj_iter(default_pool, my_obj_init, NULL);
> +
>  	/* retrieve the mempool from its name */
>  	if (rte_mempool_lookup("test_nocache") != mp_nocache) {
>  		printf("Cannot lookup mempool from its name\n");
> @@ -605,15 +631,20 @@ test_mempool(void)
>  	if (test_mempool_basic(mp_stack, 1) < 0)
>  		goto err;
>  
> +	if (test_mempool_basic(default_pool, 1) < 0)
> +		goto err;
> +
>  	rte_mempool_list_dump(stdout);
>  
> -	return 0;
> +	ret = 0;
>  
>  err:
>  	rte_mempool_free(mp_nocache);
>  	rte_mempool_free(mp_cache);
>  	rte_mempool_free(mp_stack);
> -	return -1;
> +	rte_mempool_free(default_pool);
> +
> +	return ret;
>  }
>  
>  REGISTER_TEST_COMMAND(mempool_autotest, test_mempool);
> -- 
> 2.7.4
> 

  parent reply	other threads:[~2017-03-31 15:11 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-03-31 10:17 Shreyansh Jain
2017-03-31 13:59 ` Olivier Matz
2017-03-31 14:11   ` Hemant Agrawal
2017-03-31 15:11 ` Santosh Shukla [this message]
2017-04-03  9:27   ` Hemant Agrawal
2017-04-03  9:00 ` [dpdk-dev] [PATCH v2 1/2] test/test: free mempool on exit Hemant Agrawal
2017-04-03  9:00   ` [dpdk-dev] [PATCH v2 2/2] test/test: support default mempool autotest Hemant Agrawal
2017-04-04  5:57   ` [dpdk-dev] [PATCH v3 1/2] test/test: free mempool on exit Hemant Agrawal
2017-04-04  5:57     ` [dpdk-dev] [PATCH v3 2/2] test/test: support default mempool autotest Hemant Agrawal
2017-04-05  7:02       ` santosh
2017-04-05  7:21         ` Shreyansh Jain
2017-04-04 15:28     ` [dpdk-dev] [PATCH v3 1/2] test/test: free mempool on exit santosh
2017-04-05  9:35     ` [dpdk-dev] [PATCH v4 " Shreyansh Jain
2017-04-05  9:35       ` [dpdk-dev] [PATCH v4 2/2] test/test: support default mempool autotest Shreyansh Jain
2017-04-07 15:42         ` Olivier Matz
2017-04-19 13:10           ` Thomas Monjalon
2017-04-07 15:42       ` [dpdk-dev] [PATCH v4 1/2] test/test: free mempool on exit Olivier Matz

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=20170331151100.GA11069@santosh-Latitude-E5530-non-vPro \
    --to=santosh.shukla@caviumnetworks.com \
    --cc=dev@dpdk.org \
    --cc=hemant.agrawal@nxp.com \
    --cc=olivier.matz@6wind.com \
    --cc=shreyansh.jain@nxp.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).