Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: linux-5.4.80: "refcount_t: underflow; use-after-free" in
 rpc_async_schedule->rpc_free_task->nfs4_put_copy
From:   Timo Rothenpieler <timo@rothenpieler.org>
To:     Linux NFS Mailing List <linux-nfs@vger.kernel.org>
References: <f310cd4d-c36e-ee14-2296-aa13b12cb438@rothenpieler.org>
 <81fa8f5a-5ebf-07cb-f9a9-8b3da63f330f@rothenpieler.org>
Message-ID: <d078f7e9-9e42-9783-a0c4-7b05a925d375@rothenpieler.org>
Date:   Mon, 30 Nov 2020 15:54:34 +0100
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0) Gecko/20100101
 Thunderbird/78.5.0
MIME-Version: 1.0
In-Reply-To: <81fa8f5a-5ebf-07cb-f9a9-8b3da63f330f@rothenpieler.org>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 30.11.2020 14:12, Timo Rothenpieler wrote:
> Quick update: I applied that commit on top of the 5.4.80 Kernel, and 
> have not been able to reproduce the issue anymore in a quick test.
> Will leave it running like this for a bit and see if it's fixed.
> 
> I don't know if this is in any way correct though.

Partial success, I guess.

The use after free did not occur again, but users still get their 
singularity pulls stuck, waiting on copy_file_range() to return.
Must have just been a coincidence that those two always happened at the 
same time.

It works fine on a fresh mount, until at some random point all attempts 
lead to it getting stuck.
Re-Mounting, rebooting client or rebooting server all seem to resolve it 
temporarily, until it happens again.