Re: [PATCH rdma-next 12/13] RDMA/rxe: Implement disassociate_ucontext callback

Zhu Yanjun <[email protected]>
Newsgroups org.kernel.vger.cgroups,org.kernel.vger.linux-kselftest,org.kernel.vger.linux-rdma,org.kernel.vger.linux-s390,org.kernel.vger.netdev
Message-ID <[email protected]>
在 2026/7/9 2:55, Jiri Pirko 写道:
> From: Jiri Pirko <[email protected]>
>
> Implement an empty disassociate_ucontext() callback so the RDMA core
> can move rxe devices between net namespaces. The core requires this
> callback to reset user contexts without waiting for userspace.
>
> rxe needs no teardown here: its user-mapped queues live in
> reference-counted vmalloc memory (see rxe_mmap.c) that stays valid
> while userspace holds the mappings.

The logic here is correct for enabling netns migration. Since RXE's 
user-mapped

queues rely on reference-counted vmalloc memory, this empty stub is safe 
enough

to prevent kernel panics or Use-After-Free (UAF) during device movement.


However, as a note for future improvement: keeping this callback 
entirely empty

means userspace won't be immediately notified (via SIGBUS or page table 
zapping)

when the underlying device is moved out of its netns. For now, this 
serves well as

a functional enabler, but we might want to introduce proper PTE 
tearing/zap_vma_ptes()

down the road if applications require strict, immediate disconnection 
semantics.


Thanks Jiri.

Reviewed-by: Yanjun Zhu <[email protected]>

Zhu Yanjun

> Signed-off-by: Jiri Pirko <[email protected]>
> ---
>   drivers/infiniband/sw/rxe/rxe_verbs.c | 5 +++++
>   1 file changed, 5 insertions(+)
>
> diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.c b/drivers/infiniband/sw/rxe/rxe_verbs.c
> index 1ec130fee8ea..6eb10d2f0653 100644
> --- a/drivers/infiniband/sw/rxe/rxe_verbs.c
> +++ b/drivers/infiniband/sw/rxe/rxe_verbs.c
> @@ -240,6 +240,10 @@ static void rxe_dealloc_ucontext(struct ib_ucontext *ibuc)
>   		rxe_err_uc(uc, "cleanup failed, err = %d\n", err);
>   }
>   
> +static void rxe_disassociate_ucontext(struct ib_ucontext *ibuc)
> +{
> +}
> +
>   /* pd */
>   static int rxe_alloc_pd(struct ib_pd *ibpd, struct ib_udata *udata)
>   {
> @@ -1478,6 +1482,7 @@ static const struct ib_device_ops rxe_dev_ops = {
>   	.destroy_srq = rxe_destroy_srq,
>   	.detach_mcast = rxe_detach_mcast,
>   	.device_group = &rxe_attr_group,
> +	.disassociate_ucontext = rxe_disassociate_ucontext,
>   	.enable_driver = rxe_enable_driver,
>   	.get_dma_mr = rxe_get_dma_mr,
>   	.get_hw_stats = rxe_ib_get_hw_stats,

-- 
Best Regards,
Yanjun.Zhu
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.