Re: [PATCH rdma-next v3 03/14] RDMA/core: Support renaming a device when changing its net namespace

[Jiri Pirko <jiri@resnulli.us>]

Fri, Jul 17, 2026 at 03:24:28PM +0200, sashiko-bot@kernel.org wrote:
>Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
>- [Medium] The netlink handler fails to pass the requested device name to `ib_device_set_netns_put` during a network namespace change, and ignores namespace moves entirely if a rename is also requested.
>--
>
>--- Patch [3]: [PATCH rdma-next v3 03/14] RDMA/core: Support renaming a device when changing its net namespace ---
>commit 65e19be4bcb60d41b06e567081d99cee8fd2722c
>Author: Jiri Pirko <jiri@resnulli.us>
>
>RDMA/core: Support renaming a device when changing its net namespace
>
>This commit adds support for renaming a device at the same time it is moved
>to a new network namespace. It allows namespace moves to request a specific
>destination device name, falling back to a trusted kernel-controlled pattern
>if none is provided.
>
>> diff --git a/drivers/infiniband/core/nldev.c b/drivers/infiniband/core/nldev.c
>> index 32b6c4d68ca0a..77a7580801486 100644
>> --- a/drivers/infiniband/core/nldev.c
>> +++ b/drivers/infiniband/core/nldev.c
>> @@ -1195,7 +1195,7 @@ static int nldev_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
>>  		u32 ns_fd;
>>  
>>  		ns_fd = nla_get_u32(tb[RDMA_NLDEV_NET_NS_FD]);
>> -		err = ib_device_set_netns_put(skb, device, ns_fd);
>> +		err = ib_device_set_netns_put(skb, device, ns_fd, NULL);
>
>[Severity: Medium]
>Does this hardcoded NULL prevent the core feature from being reached?
>
>It appears the newly added name parameter in ib_device_set_netns_put() is 
>never populated with the requested device name from the netlink attributes 
>in nldev_set_doit(). 
>
>If a user provides both RDMA_NLDEV_ATTR_DEV_NAME and RDMA_NLDEV_NET_NS_FD,
>the rename-during-move functionality cannot be used. Additionally, if the
>name attribute is parsed in a separate block, the function appears to take
>an early exit path, bypassing the namespace move entirely.
>
>Could this be updated to properly process both attributes and pass the
>requested device name during the move?

False possitive. Next patch changes this.


>
>>  		if (err == -EEXIST)
>>  			NL_SET_ERR_MSG(extack,
>>  				       "Device name already exists in the target net namespace");
>
>-- 
>Sashiko AI review · https://sashiko.dev/#/patchset/20260716132316.1495242-1-jiri@resnulli.us?part=3