Re: [PATCH RFC v5 0/2] nfsd: Initial implementation of NFSv4 Courteous Server

[dai.ngo@oracle.com]

On 11/17/21 6:14 AM, J. Bruce Fields wrote:
> On Tue, Nov 16, 2021 at 03:06:32PM -0800, dai.ngo@oracle.com wrote:
>> Just a reminder that this patch is still waiting for your review.
> Yeah, I was procrastinating and hoping yo'ud figure out the pynfs
> failure for me....

Last time I ran 4.0 OPEN18 test by itself and it passed. I will run
all OPEN tests together with 5.15-rc7 to see if the problem you've
seen still there.

-Dai

>   I'll see if I can get some time today.--b.
>
>> Thanks,
>> -Dai
>>
>> On 10/1/21 2:41 PM, dai.ngo@oracle.com wrote:
>>> On 10/1/21 1:53 PM, J. Bruce Fields wrote:
>>>> On Tue, Sep 28, 2021 at 08:56:39PM -0400, Dai Ngo wrote:
>>>>> Hi Bruce,
>>>>>
>>>>> This series of patches implement the NFSv4 Courteous Server.
>>>> Apologies, I keep meaning to get back to this and haven't yet.
>>>>
>>>> I do notice I'm seeing a timeout on pynfs 4.0 test OPEN18.
>>> It's weird, this test passes on my system:
>>>
>>>
>>> [root@nfsvmf25 nfs4.0]# ./testserver.py $server --rundeps -v OPEN18
>>> INIT     st_setclientid.testValid : RUNNING
>>> INIT     st_setclientid.testValid : PASS
>>> MKFILE   st_open.testOpen : RUNNING
>>> MKFILE   st_open.testOpen : PASS
>>> OPEN18   st_open.testShareConflict1 : RUNNING
>>> OPEN18   st_open.testShareConflict1 : PASS
>>> **************************************************
>>> INIT     st_setclientid.testValid : PASS
>>> OPEN18   st_open.testShareConflict1 : PASS
>>> MKFILE   st_open.testOpen : PASS
>>> **************************************************
>>> Command line asked for 3 of 673 tests
>>> Of those: 0 Skipped, 0 Failed, 0 Warned, 3 Passed
>>> [root@nfsvmf25 nfs4.0]#
>>>
>>> Do you have a network trace?
>>>
>>> -Dai
>>>
>>>> --b.
>>>>
>>>>> A server which does not immediately expunge the state on lease
>>>>> expiration
>>>>> is known as a Courteous Server.  A Courteous Server continues
>>>>> to recognize
>>>>> previously generated state tokens as valid until conflict
>>>>> arises between
>>>>> the expired state and the requests from another client, or the server
>>>>> reboots.
>>>>>
>>>>> The v2 patch includes the following:
>>>>>
>>>>> . add new callback, lm_expire_lock, to lock_manager_operations to
>>>>>     allow the lock manager to take appropriate action with
>>>>> conflict lock.
>>>>>
>>>>> . handle conflicts of NFSv4 locks with NFSv3/NLM and local locks.
>>>>>
>>>>> . expire courtesy client after 24hr if client has not reconnected.
>>>>>
>>>>> . do not allow expired client to become courtesy client if there are
>>>>>     waiters for client's locks.
>>>>>
>>>>> . modify client_info_show to show courtesy client and seconds from
>>>>>     last renew.
>>>>>
>>>>> . fix a problem with NFSv4.1 server where the it keeps returning
>>>>>     SEQ4_STATUS_CB_PATH_DOWN in the successful SEQUENCE reply, after
>>>>>     the courtesy client re-connects, causing the client to keep sending
>>>>>     BCTS requests to server.
>>>>>
>>>>> The v3 patch includes the following:
>>>>>
>>>>> . modified posix_test_lock to check and resolve conflict locks
>>>>>     to handle NLM TEST and NFSv4 LOCKT requests.
>>>>>
>>>>> . separate out fix for back channel stuck in SEQ4_STATUS_CB_PATH_DOWN.
>>>>>
>>>>> The v4 patch includes:
>>>>>
>>>>> . rework nfsd_check_courtesy to avoid dead lock of fl_lock and
>>>>> client_lock
>>>>>     by asking the laudromat thread to destroy the courtesy client.
>>>>>
>>>>> . handle NFSv4 share reservation conflicts with courtesy client. This
>>>>>     includes conflicts between access mode and deny mode and vice versa.
>>>>>
>>>>> . drop the patch for back channel stuck in SEQ4_STATUS_CB_PATH_DOWN.
>>>>>
>>>>> The v5 patch includes:
>>>>>
>>>>> . fix recursive locking of file_rwsem from posix_lock_file.
>>>>>
>>>>> . retest with LOCKDEP enabled.
>>>>>
>>>>> NOTE: I will submit pynfs tests for courteous server including tests
>>>>> for share reservation conflicts in a separate patch.
>>>>>