From: Chuck Lever III <chuck.lever@oracle.com>
To: Trond Myklebust <trondmy@hammerspace.com>
Cc: Linux NFS Mailing List <linux-nfs@vger.kernel.org>,
Bruce Fields <bfields@fieldses.org>,
Jeff Layton <jlayton@redhat.com>, Dai Ngo <dai.ngo@oracle.com>
Subject: Re: [PATCH RFC 1/2] NFSD: nfsd4_release_lockowner() should drop clp->cl_lock sooner
Date: Thu, 12 May 2022 14:03:48 +0000 [thread overview]
Message-ID: <1AC6F575-35C2-4FD6-B0D0-71876732A1DB@oracle.com> (raw)
In-Reply-To: <3092c3fc34621e72d39c54f03f5a6753ceeed216.camel@hammerspace.com>
> On May 11, 2022, at 6:04 PM, Trond Myklebust <trondmy@hammerspace.com> wrote:
>
> On Wed, 2022-05-11 at 17:52 -0400, Chuck Lever wrote:
>> nfsd4_release_lockowner() mustn't hold clp->cl_lock when
>> check_for_locks() invokes nfsd_file_put(), which can sleep.
>>
>> Reported-by: Dai Ngo <dai.ngo@oracle.com>
>> Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
>> Cc: <stable@vger.kernel.org>
>> ---
>> fs/nfsd/nfs4state.c | 56 +++++++++++++++++++++++------------------
>> ----------
>> 1 file changed, 25 insertions(+), 31 deletions(-)
>>
>> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
>> index 234e852fcdfa..e2eb6d031643 100644
>> --- a/fs/nfsd/nfs4state.c
>> +++ b/fs/nfsd/nfs4state.c
>> @@ -6611,8 +6611,8 @@ nfs4_set_lock_denied(struct file_lock *fl,
>> struct nfsd4_lock_denied *deny)
>> deny->ld_type = NFS4_WRITE_LT;
>> }
>>
>> -static struct nfs4_lockowner *
>> -find_lockowner_str_locked(struct nfs4_client *clp, struct xdr_netobj
>> *owner)
>> +static struct nfs4_stateowner *
>> +find_stateowner_str_locked(struct nfs4_client *clp, struct
>> xdr_netobj *owner)
>> {
>> unsigned int strhashval = ownerstr_hashval(owner);
>> struct nfs4_stateowner *so;
>> @@ -6624,11 +6624,22 @@ find_lockowner_str_locked(struct nfs4_client
>> *clp, struct xdr_netobj *owner)
>> if (so->so_is_open_owner)
>> continue;
>> if (same_owner_str(so, owner))
>> - return lockowner(nfs4_get_stateowner(so));
>> + return nfs4_get_stateowner(so);
>
> Hmm... If nfs4_get_stateowner() can fail here, don't you want to
> continue searching the list when it does?
You've lost me on this.
494 static inline struct nfs4_stateowner *
495 nfs4_get_stateowner(struct nfs4_stateowner *sop)
496 {
497 atomic_inc(&sop->so_count);
498 return sop;
499 }
500
How would nfs4_get_stateowner() fail?
>> }
>> return NULL;
>> }
>>
>> +static struct nfs4_lockowner *
>> +find_lockowner_str_locked(struct nfs4_client *clp, struct xdr_netobj
>> *owner)
>> +{
>> + struct nfs4_stateowner *so;
>> +
>> + so = find_stateowner_str_locked(clp, owner);
>> + if (!so)
>> + return NULL;
>> + return lockowner(so);
>> +}
>> +
>> static struct nfs4_lockowner *
>> find_lockowner_str(struct nfs4_client *clp, struct xdr_netobj
>> *owner)
>> {
>> @@ -7305,10 +7316,8 @@ nfsd4_release_lockowner(struct svc_rqst
>> *rqstp,
>> struct nfsd4_release_lockowner *rlockowner = &u-
>>> release_lockowner;
>> clientid_t *clid = &rlockowner->rl_clientid;
>> struct nfs4_stateowner *sop;
>> - struct nfs4_lockowner *lo = NULL;
>> + struct nfs4_lockowner *lo;
>> struct nfs4_ol_stateid *stp;
>> - struct xdr_netobj *owner = &rlockowner->rl_owner;
>> - unsigned int hashval = ownerstr_hashval(owner);
>> __be32 status;
>> struct nfsd_net *nn = net_generic(SVC_NET(rqstp),
>> nfsd_net_id);
>> struct nfs4_client *clp;
>> @@ -7322,32 +7331,18 @@ nfsd4_release_lockowner(struct svc_rqst
>> *rqstp,
>> return status;
>>
>> clp = cstate->clp;
>> - /* Find the matching lock stateowner */
>> spin_lock(&clp->cl_lock);
>> - list_for_each_entry(sop, &clp->cl_ownerstr_hashtbl[hashval],
>> - so_strhash) {
>> -
>> - if (sop->so_is_open_owner || !same_owner_str(sop,
>> owner))
>> - continue;
>> -
>> - /* see if there are still any locks associated with
>> it */
>> - lo = lockowner(sop);
>> - list_for_each_entry(stp, &sop->so_stateids,
>> st_perstateowner) {
>> - if (check_for_locks(stp->st_stid.sc_file,
>> lo)) {
>> - status = nfserr_locks_held;
>> - spin_unlock(&clp->cl_lock);
>> - return status;
>> - }
>> - }
>> + sop = find_stateowner_str_locked(clp, &rlockowner->rl_owner);
>> + spin_unlock(&clp->cl_lock);
>> + if (!sop)
>> + return nfs_ok;
>>
>> - nfs4_get_stateowner(sop);
>> - break;
>> - }
>> - if (!lo) {
>> - spin_unlock(&clp->cl_lock);
>> - return status;
>> - }
>> + lo = lockowner(sop);
>> + list_for_each_entry(stp, &sop->so_stateids, st_perstateowner)
>> + if (check_for_locks(stp->st_stid.sc_file, lo))
>> + return nfserr_locks_held;
>
> Isn't this line now leaking the reference to sop?
Indeed. Fixed.
>>
>> + spin_lock(&clp->cl_lock);
>> unhash_lockowner_locked(lo);
>> while (!list_empty(&lo->lo_owner.so_stateids)) {
>> stp = list_first_entry(&lo->lo_owner.so_stateids,
>> @@ -7360,8 +7355,7 @@ nfsd4_release_lockowner(struct svc_rqst *rqstp,
>> free_ol_stateid_reaplist(&reaplist);
>> remove_blocked_locks(lo);
>> nfs4_put_stateowner(&lo->lo_owner);
>> -
>> - return status;
>> + return nfs_ok;
>> }
>>
>> static inline struct nfs4_client_reclaim *
>>
>>
>
> --
> Trond Myklebust
> Linux NFS client maintainer, Hammerspace
> trond.myklebust@hammerspace.com
>
>
--
Chuck Lever
next prev parent reply other threads:[~2022-05-12 14:03 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-11 21:52 [PATCH RFC 0/2] Fix "sleep while locked" in RELEASE_LOCKOWNER Chuck Lever
2022-05-11 21:52 ` [PATCH RFC 1/2] NFSD: nfsd4_release_lockowner() should drop clp->cl_lock sooner Chuck Lever
2022-05-11 22:04 ` Trond Myklebust
2022-05-12 14:03 ` Chuck Lever III [this message]
2022-05-12 10:07 ` Jeff Layton
2022-05-12 14:10 ` Chuck Lever III
2022-05-11 21:52 ` [PATCH RFC 2/2] NFSD: nfsd_file_put() can sleep Chuck Lever
2022-05-12 10:08 ` Jeff Layton
[not found] ` <PH0PR14MB549335582C7A4D4F4D2269B2AAD99@PH0PR14MB5493.namprd14.prod.outlook.com>
2022-05-26 23:39 ` [PATCH RFC 0/2] Fix "sleep while locked" in RELEASE_LOCKOWNER Chuck Lever III
2022-05-27 0:21 ` Charles Hedrick
2022-05-27 15:32 ` Chuck Lever III
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1AC6F575-35C2-4FD6-B0D0-71876732A1DB@oracle.com \
--to=chuck.lever@oracle.com \
--cc=bfields@fieldses.org \
--cc=dai.ngo@oracle.com \
--cc=jlayton@redhat.com \
--cc=linux-nfs@vger.kernel.org \
--cc=trondmy@hammerspace.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.