patches for DPDK stable branches
 help / color / mirror / Atom feed
From: "Burakov, Anatoly" <anatoly.burakov@intel.com>
To: Mingjin Ye <mingjinx.ye@intel.com>, <dev@dpdk.org>
Cc: <stable@dpdk.org>
Subject: Re: [PATCH 3/3] net/vdev: fix insert vdev core dump
Date: Thu, 11 Jul 2024 18:10:28 +0200	[thread overview]
Message-ID: <48edd8b1-715e-4363-b627-37a910d2dd70@intel.com> (raw)
In-Reply-To: <20240314093630.1066948-4-mingjinx.ye@intel.com>

On 3/14/2024 10:36 AM, Mingjin Ye wrote:
> Inserting a vdev device when the device arguments are already stored
> in devargs_list, the rte_devargs_insert function replaces the supplied
> new devargs with the found devargs and frees the new devargs. As a
> result, the use of free devargs results in a core dump.
> 
> This patch fixes the issue by using valid devargs.
> 
> Fixes: f3a1188cee4a ("devargs: make device representation generic")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Mingjin Ye <mingjinx.ye@intel.com>

I am not too familiar with how devargs works, so I have a trivial question.

I understand the point of this patch, and it is roughly as follows:

1) we enter insert_vdev and allocated new devargs structure (and copy 
the `name` into devargs)
2) we set dev->device.name to devargs->name
3) we insert devargs into the list using rte_devargs_insert
4) if devargs list already had devargs with the same name, our devargs 
is destroyed and replaced with devargs that is in the list
5) because of this, dev->device.name becomes invalid as that specific 
devargs has been freed - it points to name from the old devargs

We do need to store devargs->name in dev->device.name, and we need to do 
so after calling rte_devargs_insert to avoid keeping reference to memory 
that was freed. So, provisionally,

Reviewed-by: Anatoly Burakov <anatoly.burakov@intel.com>

My question is, under what circumstances would there be duplicate 
entries in devargs list? I assume there is an answer to that question 
because rte_devargs_insert() expects this case, so this is just for my 
understanding - I can't seem to figure out how we would arrive at a 
situation where we have duplicate devargs.

> ---
>   drivers/bus/vdev/vdev.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/bus/vdev/vdev.c b/drivers/bus/vdev/vdev.c
> index 1a6cc7d12d..53fc4a6e21 100644
> --- a/drivers/bus/vdev/vdev.c
> +++ b/drivers/bus/vdev/vdev.c
> @@ -286,7 +286,6 @@ insert_vdev(const char *name, const char *args, struct rte_vdev_device **p_dev)
>   
>   	dev->device.bus = &rte_vdev_bus;
>   	dev->device.numa_node = SOCKET_ID_ANY;
> -	dev->device.name = devargs->name;
>   
>   	if (find_vdev(name)) {
>   		/*
> @@ -300,6 +299,7 @@ insert_vdev(const char *name, const char *args, struct rte_vdev_device **p_dev)
>   
>   	rte_devargs_insert(&devargs);
>   	dev->device.devargs = devargs;
> +	dev->device.name = devargs->name;
>   	TAILQ_INSERT_TAIL(&vdev_device_list, dev, next);
>   
>   	if (p_dev)

-- 
Thanks,
Anatoly


  parent reply	other threads:[~2024-07-11 16:11 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20240314093630.1066948-1-mingjinx.ye@intel.com>
2024-03-14  9:36 ` [PATCH 1/3] bus/vdev: revert fix devargs in secondary process Mingjin Ye
2024-06-19 20:12   ` Thomas Monjalon
2024-03-14  9:36 ` [PATCH 2/3] bus/vdev: revert fix devargs after multi-process bus scan Mingjin Ye
2024-06-19 20:15   ` Thomas Monjalon
2024-06-20  6:10     ` Ye, MingjinX
2024-07-11 15:56   ` Burakov, Anatoly
2024-03-14  9:36 ` [PATCH 3/3] net/vdev: fix insert vdev core dump Mingjin Ye
2024-03-15  5:51   ` Jiang, YuX
2024-06-19 20:16   ` Thomas Monjalon
2024-06-20  6:41     ` Ye, MingjinX
2024-07-11 16:10   ` Burakov, Anatoly [this message]
2024-07-12  2:18     ` Ye, MingjinX
2024-07-12  8:38       ` Burakov, Anatoly
2024-07-16  9:53   ` [PATCH v2] " Mingjin Ye
2024-07-22 12:39     ` Burakov, Anatoly
2024-07-23 15:25       ` 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=48edd8b1-715e-4363-b627-37a910d2dd70@intel.com \
    --to=anatoly.burakov@intel.com \
    --cc=dev@dpdk.org \
    --cc=mingjinx.ye@intel.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).