From: "Burakov, Anatoly" <anatoly.burakov@intel.com>
To: "Carrillo, Erik G" <erik.g.carrillo@intel.com>,
"rsanford@akamai.com" <rsanford@akamai.com>,
"thomas@monjalon.net" <thomas@monjalon.net>
Cc: "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] [PATCH] timer: fix resource leak in finalize
Date: Thu, 2 May 2019 14:03:42 +0100 [thread overview]
Message-ID: <7b2dfc9d-4638-f8ab-1f4b-ad268635b386@intel.com> (raw)
Message-ID: <20190502130342.PrPrAPR_ygEUyqZNYO7KFRIKNo9cjZzHkY5R1WCMa8o@z> (raw)
In-Reply-To: <BE54F058557D9A4FAC1D84E2FC6D875723402CB9@fmsmsx115.amr.corp.intel.com>
On 02-May-19 1:19 PM, Carrillo, Erik G wrote:
>
>
>> -----Original Message-----
>> From: Burakov, Anatoly
>> Sent: Thursday, May 2, 2019 4:18 AM
>> To: Carrillo, Erik G <erik.g.carrillo@intel.com>; rsanford@akamai.com;
>> thomas@monjalon.net
>> Cc: dev@dpdk.org
>> Subject: Re: [dpdk-dev] [PATCH] timer: fix resource leak in finalize
>>
>> On 01-May-19 8:00 PM, Erik Gabriel Carrillo wrote:
>>> The finalize function should free the memzone created in the init
>>> function, rather than freeing the allocation the memzone references,
>>> otherwise a memzone descriptor can be leaked.
>>>
>>> Fixes: c0749f7096c7 ("timer: allow management in shared memory")
>>>
>>> Signed-off-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com>
>>> ---
>>> lib/librte_timer/rte_timer.c | 12 ++++++++++--
>>> 1 file changed, 10 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/lib/librte_timer/rte_timer.c
>>> b/lib/librte_timer/rte_timer.c index eb46009..fb7a87e 100644
>>> --- a/lib/librte_timer/rte_timer.c
>>> +++ b/lib/librte_timer/rte_timer.c
>>> @@ -60,6 +60,7 @@ struct rte_timer_data {
>>> };
>>>
>>> #define RTE_MAX_DATA_ELS 64
>>> +static const struct rte_memzone *rte_timer_data_mz;
>>> static struct rte_timer_data *rte_timer_data_arr;
>>> static const uint32_t default_data_id;
>>> static uint32_t rte_timer_subsystem_initialized; @@ -164,6 +165,7 @@
>>> rte_timer_subsystem_init_v1905(void)
>>> if (mz == NULL)
>>> return -EEXIST;
>>>
>>> + rte_timer_data_mz = mz;
>>> rte_timer_data_arr = mz->addr;
>>>
>>> rte_timer_data_arr[default_data_id].internal_flags |= @@ -
>> 180,6
>>> +182,7 @@ rte_timer_subsystem_init_v1905(void)
>>> if (mz == NULL)
>>> return -ENOMEM;
>>>
>>> + rte_timer_data_mz = mz;
>>> rte_timer_data_arr = mz->addr;
>>>
>>> for (i = 0; i < RTE_MAX_DATA_ELS; i++) { @@ -205,8 +208,13 @@
>>> BIND_DEFAULT_SYMBOL(rte_timer_subsystem_init, _v1905, 19.05);
>>> void __rte_experimental
>>> rte_timer_subsystem_finalize(void)
>>> {
>>> - if (rte_timer_data_arr)
>>> - rte_free(rte_timer_data_arr);
>>> + if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>>> + return;
>>> +
>>> + if (!rte_timer_subsystem_initialized)
>>> + return;
>>> +
>>> + rte_memzone_free(rte_timer_data_mz);
>>
>> The patch is a correct fix, but the whole idea of this looks dangerous to me.
>>
>> If we exit the primary while secondaries are still running, wouldn't it basically
>> pull out timer data from under secondaries' feet?
>>
>
> Ah yes - that’s right. Perhaps it would be better to maintain a reference count of some sort such that the last process to exit could cause the memzone_free.
>
It feels like a hack, to be honest. A process can crash or exit without
calling rte_eal_cleanup(), which will lead to a memory leak due to
refcount being stuck at a value that's not representing reality. It will
be saf-er than current approach, but still not ideal.
However, i guess it's a good compromise, if i were to choose between a
memory leak and a segfault :D I wonder if there is a better approach.
> Thanks,
> Erik
>
>>>
>>> rte_timer_subsystem_initialized = 0;
>>> }
>>>
>>
>>
>> --
>> Thanks,
>> Anatoly
--
Thanks,
Anatoly
next prev parent reply other threads:[~2019-05-02 13:03 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-05-01 19:00 Erik Gabriel Carrillo
2019-05-01 19:00 ` Erik Gabriel Carrillo
2019-05-02 9:18 ` Burakov, Anatoly
2019-05-02 9:18 ` Burakov, Anatoly
2019-05-02 12:19 ` Carrillo, Erik G
2019-05-02 12:19 ` Carrillo, Erik G
2019-05-02 13:03 ` Burakov, Anatoly [this message]
2019-05-02 13:03 ` Burakov, Anatoly
2019-05-02 13:48 ` Carrillo, Erik G
2019-05-02 13:48 ` Carrillo, Erik G
2019-05-03 22:54 ` [dpdk-dev] [PATCH v2] " Erik Gabriel Carrillo
2019-05-03 22:54 ` Erik Gabriel Carrillo
2019-05-07 11:03 ` Burakov, Anatoly
2019-05-07 11:03 ` Burakov, Anatoly
2019-05-07 22:04 ` Carrillo, Erik G
2019-05-07 22:04 ` Carrillo, Erik G
2019-05-08 8:49 ` Burakov, Anatoly
2019-05-08 8:49 ` Burakov, Anatoly
2019-05-08 23:01 ` Carrillo, Erik G
2019-05-08 23:01 ` Carrillo, Erik G
2019-05-09 7:44 ` Thomas Monjalon
2019-05-09 7:44 ` Thomas Monjalon
2019-05-08 22:35 ` [dpdk-dev] [PATCH v3] " Erik Gabriel Carrillo
2019-05-08 22:35 ` Erik Gabriel Carrillo
2019-05-09 8:29 ` Burakov, Anatoly
2019-05-09 8:29 ` Burakov, Anatoly
2019-06-05 9:33 ` Thomas Monjalon
2019-06-05 9:47 ` Burakov, Anatoly
2019-06-25 16:11 ` [dpdk-dev] [PATCH 0/2] Fix timer resource leak Anatoly Burakov
2019-07-05 13:20 ` [dpdk-dev] [PATCH v2 0/1] " Anatoly Burakov
2019-07-05 17:22 ` [dpdk-dev] [PATCH v3 " Anatoly Burakov
2019-07-05 17:22 ` [dpdk-dev] [PATCH v3 1/1] timer: fix resource leak in finalize Anatoly Burakov
2019-07-05 22:06 ` Thomas Monjalon
2019-07-05 13:20 ` [dpdk-dev] [PATCH v2 " Anatoly Burakov
2019-06-25 16:11 ` [dpdk-dev] [PATCH 1/2] eal: add internal locks for timer lib into EAL Anatoly Burakov
2019-06-27 18:41 ` Carrillo, Erik G
2019-07-04 9:09 ` David Marchand
2019-07-04 10:44 ` Burakov, Anatoly
2019-06-25 16:11 ` [dpdk-dev] [PATCH 2/2] timer: fix resource leak in finalize Anatoly Burakov
2019-06-27 18:48 ` Carrillo, Erik G
2019-07-04 9:10 ` David Marchand
2019-07-04 10:45 ` Burakov, Anatoly
2019-07-04 10:50 ` David Marchand
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=7b2dfc9d-4638-f8ab-1f4b-ad268635b386@intel.com \
--to=anatoly.burakov@intel.com \
--cc=dev@dpdk.org \
--cc=erik.g.carrillo@intel.com \
--cc=rsanford@akamai.com \
--cc=thomas@monjalon.net \
/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).