From: Olivier MATZ <olivier.matz@6wind.com>
To: Pablo de Lara <pablo.de.lara.guarch@intel.com>, dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] ring: add function to free a ring
Date: Mon, 07 Sep 2015 10:45:11 +0200 [thread overview]
Message-ID: <55ED4E97.3070505@6wind.com> (raw)
In-Reply-To: <1439906454-12860-1-git-send-email-pablo.de.lara.guarch@intel.com>
Hi Pablo,
Please find some comments below.
On 08/18/2015 04:00 PM, Pablo de Lara wrote:
> When creating a ring, a memzone is created to allocate it in memory,
> but the ring could not be freed, as memzones could not be.
>
> Since memzones can be freed now, then rings can be as well,
> taking into account if they were initialized using pre-allocated memory
> (in which case, memory should be freed externally) or using rte_memzone_reserve
> (with rte_ring_create), freeing the memory with rte_memzone_free.
>
> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> ---
> lib/librte_ring/rte_ring.c | 43 ++++++++++++++++++++++++++++++++++++
> lib/librte_ring/rte_ring.h | 7 ++++++
> lib/librte_ring/rte_ring_version.map | 7 ++++++
> 3 files changed, 57 insertions(+)
>
> diff --git a/lib/librte_ring/rte_ring.c b/lib/librte_ring/rte_ring.c
> index c9e59d4..83ce6d3 100644
> --- a/lib/librte_ring/rte_ring.c
> +++ b/lib/librte_ring/rte_ring.c
> @@ -208,6 +208,49 @@ rte_ring_create(const char *name, unsigned count, int socket_id,
> return r;
> }
>
> +/* free the ring */
> +void
> +rte_ring_free(struct rte_ring *r)
> +{
> + struct rte_ring_list *ring_list = NULL;
> + char mz_name[RTE_MEMZONE_NAMESIZE];
> + struct rte_tailq_entry *te;
> + const struct rte_memzone *mz;
> +
> + if (r == NULL)
> + return;
> +
> + snprintf(mz_name, sizeof(mz_name), "%s%s", RTE_RING_MZ_PREFIX, r->name);
> + mz = rte_memzone_lookup(mz_name);
> +
> + /*
> + * Free ring memory if it was allocated with rte_memzone_reserve,
> + * otherwise it should be freed externally
> + */
> + if (rte_memzone_free(mz) != 0)
> + return;
Should we have a log here? I think it may hide important
bugs if we just return silently here.
> +
> + ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
> + rte_rwlock_write_lock(RTE_EAL_TAILQ_RWLOCK);
> +
> + /* find out tailq entry */
> + TAILQ_FOREACH(te, ring_list, next) {
> + if (te->data == (void *) r)
> + break;
> + }
> +
> + if (te == NULL) {
> + rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
> + return;
> + }
If I understand well, a ring is in the tailq only if it was
created with rte_ring_create(). A ring that is statically created
in memory is not in the tailq. But we already returned in that
case. So (te == NULL) should not happen here, right? We could
also add an error log then.
I'm not sure we should handle the case where the ring is not allocated
with rte_ring_create() in this function. If the ring is allocated with
another mean (either in a global variable, or with another dynamic
memory allocator), this function should not be called.
What do you think?
Regards,
Olivier
> +
> + TAILQ_REMOVE(ring_list, te, next);
> +
> + rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
> +
> + rte_free(te);
> +}
> +
> /*
> * change the high water mark. If *count* is 0, water marking is
> * disabled
> diff --git a/lib/librte_ring/rte_ring.h b/lib/librte_ring/rte_ring.h
> index af68888..e75566f 100644
> --- a/lib/librte_ring/rte_ring.h
> +++ b/lib/librte_ring/rte_ring.h
> @@ -300,6 +300,13 @@ int rte_ring_init(struct rte_ring *r, const char *name, unsigned count,
> */
> struct rte_ring *rte_ring_create(const char *name, unsigned count,
> int socket_id, unsigned flags);
> +/**
> + * De-allocate all memory used by the ring.
> + *
> + * @param r
> + * Ring to free
> + */
> +void rte_ring_free(struct rte_ring *r);
>
> /**
> * Change the high water mark.
> diff --git a/lib/librte_ring/rte_ring_version.map b/lib/librte_ring/rte_ring_version.map
> index 982fdd1..5474b98 100644
> --- a/lib/librte_ring/rte_ring_version.map
> +++ b/lib/librte_ring/rte_ring_version.map
> @@ -11,3 +11,10 @@ DPDK_2.0 {
>
> local: *;
> };
> +
> +DPDK_2.2 {
> + global:
> +
> + rte_ring_free;
> +
> +} DPDK_2.0;
>
next prev parent reply other threads:[~2015-09-07 8:45 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-08-18 14:00 Pablo de Lara
2015-09-07 8:45 ` Olivier MATZ [this message]
2015-09-09 7:48 ` De Lara Guarch, Pablo
2015-09-22 13:55 ` [dpdk-dev] [PATCH v2] " Pablo de Lara
2015-10-02 14:01 ` [dpdk-dev] [PATCH v3] " Pablo de Lara
2015-10-02 14:09 ` Bruce Richardson
2015-10-02 14:27 ` De Lara Guarch, Pablo
2015-10-02 15:53 ` [dpdk-dev] [PATCH v4] " Pablo de Lara
2015-10-13 14:01 ` Olivier MATZ
2015-11-03 23:50 ` Thomas Monjalon
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=55ED4E97.3070505@6wind.com \
--to=olivier.matz@6wind.com \
--cc=dev@dpdk.org \
--cc=pablo.de.lara.guarch@intel.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).