DPDK patches and discussions
 help / color / mirror / Atom feed
From: Ferruh Yigit <ferruh.yigit@intel.com>
To: "Lipiec, Herakliusz" <herakliusz.lipiec@intel.com>,
	"Wiles, Keith" <keith.wiles@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>,
	"rasland@mellanox.com" <rasland@mellanox.com>,
	"stable@dpdk.org" <stable@dpdk.org>
Subject: Re: [dpdk-dev] [dpdk-stable] [PATCH] net/tap: ipc add check for number of messages received
Date: Fri, 19 Apr 2019 18:17:40 +0100	[thread overview]
Message-ID: <4239c177-f0be-2881-b801-0877f475d85c@intel.com> (raw)
Message-ID: <20190419171740.I5dH6ZST09b149qsX9K9jTh7cQxQ8vH9NU5DFxRlN1M@z> (raw)
In-Reply-To: <EB47986807B11C41AD0C5C13A40DE8835BF161@IRSMSX104.ger.corp.intel.com>

On 4/19/2019 5:39 PM, Lipiec, Herakliusz wrote:
> On 4/18/2019 7:13, Ferruh Yigit worte:
>> On 4/18/2019 6:19 PM, Herakliusz Lipiec wrote:
>>> A sucessfull call to rte_mp_request_sync does not guarantee that there
>>> are valid messages in the buffer, and this should be checked for
>>> before accessing data in the message.
>>>
>>> Fixes: c9aa56edec8e ("net/tap: access primary process queues from
>>> secondary")
>>> Cc: rasland@mellanox.com
>>> Cc: stable@dpdk.org
>>> Signed-off-by: Herakliusz Lipiec <herakliusz.lipiec@intel.com>
>>> ---
>>>  drivers/net/tap/rte_eth_tap.c | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/net/tap/rte_eth_tap.c
>>> b/drivers/net/tap/rte_eth_tap.c index e9fda8cf6..a619a8850 100644
>>> --- a/drivers/net/tap/rte_eth_tap.c
>>> +++ b/drivers/net/tap/rte_eth_tap.c
>>> @@ -2101,7 +2101,7 @@ tap_mp_attach_queues(const char *port_name,
>> struct rte_eth_dev *dev)
>>>  	request.len_param = sizeof(*request_param);
>>>  	/* Send request and receive reply */
>>>  	ret = rte_mp_request_sync(&request, &replies, &timeout);
>>> -	if (ret < 0) {
>>> +	if (ret < 0 || replies.n_receieved != 1) {
>>
>> The API documentation says:
>>
>> ||   * @return
>>
>>
>>
>> ||   *  - On success, return 0.
>>
>>
>>
>> ||   *  - On failure, return -1, and the reason will be stored in rte_errno.
>>
>> So if the API returns 0, why the reply is not valid, also if reply is not valid how
>> can you rely on a value in 'replies'
>>
>> What do you think updating the 'rte_mp_request_sync()' API to return error
>> whenever the reply is not valid?
> The reply is not valid, because there is no valid msg pointer in replies.msg (should be null)
> replies.nb_received should be either 0 (if replies carries no message) or 1 (if there is a message).
> 
> There are two other code paths that can return a success, but have no (valid) message.
> In rte_mp_request_sync there is a call to mp_request_sync which may return 0 with no message in case of:
> - failure to send the message on behalf of remote
> - the caller not caring about reply message.
> I propose to add a check for nb_received to net/tap since this seems to be done in everywhere 
> else when rte_mp_request_sync is called (this will do no harm), and also I think that return codes should be fixed,
> but that can be done irrelevant of this. 
> 

I see, "replies.nb_received" can be relied on since it is set in the begging of
the 'rte_mp_request_sync()'

And I can see you have created a defect for the API fix [1], it is OK to get tap
fix for rc2, but for next release it would be appreciated if you own the defect
you have submitted.

Thanks,
ferruh

[1]
https://bugs.dpdk.org/show_bug.cgi?id=257

  parent reply	other threads:[~2019-04-19 17:17 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-18 17:19 [dpdk-dev] " Herakliusz Lipiec
2019-04-18 17:19 ` Herakliusz Lipiec
2019-04-18 18:13 ` [dpdk-dev] [dpdk-stable] " Ferruh Yigit
2019-04-18 18:13   ` Ferruh Yigit
2019-04-19 16:39   ` Lipiec, Herakliusz
2019-04-19 16:39     ` Lipiec, Herakliusz
2019-04-19 17:17     ` Ferruh Yigit [this message]
2019-04-19 17:17       ` Ferruh Yigit
2019-04-19 10:28 ` [dpdk-dev] [PATCH v2] " Herakliusz Lipiec
2019-04-19 10:28   ` Herakliusz Lipiec
2019-04-19 17:37   ` Ferruh Yigit
2019-04-19 17:37     ` Ferruh Yigit

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=4239c177-f0be-2881-b801-0877f475d85c@intel.com \
    --to=ferruh.yigit@intel.com \
    --cc=dev@dpdk.org \
    --cc=herakliusz.lipiec@intel.com \
    --cc=keith.wiles@intel.com \
    --cc=rasland@mellanox.com \
    --cc=stable@dpdk.org \
    /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).