Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Subject: Re: [PATCH v2 2/2] ceph: wait the first reply of inflight
 unlink/rmdir
To:     Jeff Layton <jlayton@kernel.org>, viro@zeniv.linux.org.uk
Cc:     idryomov@gmail.com, vshankar@redhat.com,
        ceph-devel@vger.kernel.org, arnd@arndb.de, mcgrof@kernel.org,
        akpm@linux-foundation.org, linux-fsdevel@vger.kernel.org,
        linux-kernel@vger.kernel.org, kernel test robot <lkp@intel.com>
References: <20220517010316.81483-1-xiubli@redhat.com>
 <20220517010316.81483-3-xiubli@redhat.com>
 <a2d05d80e30831e915e707a48520139500befc2b.camel@kernel.org>
 <bce4ef40-277f-8bc0-6cdb-3435eddddf44@redhat.com>
 <bd2ea8d6467ff8ea98c7bd048fd417aced86e20d.camel@kernel.org>
From:   Xiubo Li <xiubli@redhat.com>
Message-ID: <718ffcfa-d211-23ec-947c-0c9dec33781c@redhat.com>
Date:   Tue, 17 May 2022 19:59:34 +0800
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.10.1
MIME-Version: 1.0
In-Reply-To: <bd2ea8d6467ff8ea98c7bd048fd417aced86e20d.camel@kernel.org>
Content-Type: text/plain; charset=iso-8859-15; format=flowed
Content-Transfer-Encoding: 7bit
Content-Language: en-US
Precedence: bulk


On 5/17/22 7:54 PM, Jeff Layton wrote:
> On Tue, 2022-05-17 at 19:49 +0800, Xiubo Li wrote:
>> On 5/17/22 7:35 PM, Jeff Layton wrote:
>>> On Tue, 2022-05-17 at 09:03 +0800, Xiubo Li wrote:
>>>> In async unlink case the kclient won't wait for the first reply
>>>> from MDS and just drop all the links and unhash the dentry and then
>>>> succeeds immediately.
>>>>
>>>> For any new create/link/rename,etc requests followed by using the
>>>> same file names we must wait for the first reply of the inflight
...
>>> I doubt you need this large a hashtable, particularly given that this is
>>> per-superblock. In most cases, we'll just have a few of these in flight
>>> at a time.
>> A global hashtable ? And set the order to 8 ?
> Per-sb is fine, IMO. 6-8 bits sounds reasonable.

Sure, let's use 8. From my snaptest I can see there had a lot of 
dentries in the hashtable at the same time some times.

-- Xiubo