Re: [PATCH v9 15/17] fuse: {io-uring} Prevent mount point hang on fuse-server termination

[Bernd Schubert <bernd@bsbernd.com>]

On 1/7/25 17:14, Luis Henriques wrote:
> On Tue, Jan 07 2025, Bernd Schubert wrote:
> 
>> When the fuse-server terminates while the fuse-client or kernel
>> still has queued URING_CMDs, these commands retain references
>> to the struct file used by the fuse connection. This prevents
>> fuse_dev_release() from being invoked, resulting in a hung mount
>> point.
>>
>> This patch addresses the issue by making queued URING_CMDs
>> cancelable, allowing fuse_dev_release() to proceed as expected
>> and preventing the mount point from hanging.
>>
>> Signed-off-by: Bernd Schubert <bschubert@ddn.com>
>> ---
>>   fs/fuse/dev.c         |  2 ++
>>   fs/fuse/dev_uring.c   | 71 ++++++++++++++++++++++++++++++++++++++++++++++++---
>>   fs/fuse/dev_uring_i.h |  9 +++++++
>>   3 files changed, 79 insertions(+), 3 deletions(-)
>>
>> diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
>> index afafa960d4725d9b64b22f17bf09c846219396d6..1b593b23f7b8c319ec38c7e726dabf516965500e 100644
>> --- a/fs/fuse/dev.c
>> +++ b/fs/fuse/dev.c
>> @@ -599,8 +599,10 @@ static int fuse_request_queue_background(struct fuse_req *req)
>>   	}
>>   	__set_bit(FR_ISREPLY, &req->flags);
>>   
>> +#ifdef CONFIG_FUSE_IO_URING
>>   	if (fuse_uring_ready(fc))
>>   		return fuse_request_queue_background_uring(fc, req);
>> +#endif
> 

Oh sorry, I had tried to remove the ifdef in v9, but didn't succeed
and added back in the wrong patch.


Thanks,
Bernd