[v2,05/10] nfs: track whether server sets MAY_NOTIFY_LOCK flag
diff mbox

Message ID 1473446870-1831-6-git-send-email-jlayton@redhat.com
State New
Headers show

Commit Message

Jeff Layton Sept. 9, 2016, 6:47 p.m. UTC
If it does, then always have the client sleep for the max time before
repolling for the lock. If it doesn't then we can skip all of the
waitqueue handling as well.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
 fs/nfs/nfs4_fs.h  | 1 +
 fs/nfs/nfs4proc.c | 2 ++
 2 files changed, 3 insertions(+)

Comments

Schumaker, Anna Sept. 12, 2016, 8:19 p.m. UTC | #1
Hi Jeff,

On 09/09/2016 02:47 PM, Jeff Layton wrote:
> If it does, then always have the client sleep for the max time before
> repolling for the lock. If it doesn't then we can skip all of the
> waitqueue handling as well.
> 
> Signed-off-by: Jeff Layton <jlayton@redhat.com>
> ---
>  fs/nfs/nfs4_fs.h  | 1 +
>  fs/nfs/nfs4proc.c | 2 ++
>  2 files changed, 3 insertions(+)
> 
> diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
> index 9bf64eacba5b..91e4f135a5f2 100644
> --- a/fs/nfs/nfs4_fs.h
> +++ b/fs/nfs/nfs4_fs.h
> @@ -156,6 +156,7 @@ enum {
>  	NFS_STATE_RECLAIM_NOGRACE,	/* OPEN stateid needs to recover state */
>  	NFS_STATE_POSIX_LOCKS,		/* Posix locks are supported */
>  	NFS_STATE_RECOVERY_FAILED,	/* OPEN stateid state recovery failed */
> +	NFS_STATE_MAY_NOTIFY_LOCK,	/* server may CB_NOTIFY_LOCK */

Looks like the reason I don't have this flag is because it's added as part of the server patches.  I'll coordinate with the nfsd merge to make sure everything is added in the right order!

Thanks,
Anna
>  };
>  
>  struct nfs4_state {
> diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> index 90e8ded0ef82..627a9185822f 100644
> --- a/fs/nfs/nfs4proc.c
> +++ b/fs/nfs/nfs4proc.c
> @@ -2537,6 +2537,8 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
>  		goto out;
>  	if (server->caps & NFS_CAP_POSIX_LOCK)
>  		set_bit(NFS_STATE_POSIX_LOCKS, &state->flags);
> +	if (opendata->o_res.rflags & NFS4_OPEN_RESULT_MAY_NOTIFY_LOCK)
> +		set_bit(NFS_STATE_MAY_NOTIFY_LOCK, &state->flags);
>  
>  	dentry = opendata->dentry;
>  	if (d_really_is_negative(dentry)) {
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Jeff Layton Sept. 12, 2016, 8:30 p.m. UTC | #2
On Mon, 2016-09-12 at 16:19 -0400, Anna Schumaker wrote:
> Hi Jeff,
> 
> On 09/09/2016 02:47 PM, Jeff Layton wrote:
> > 
> > If it does, then always have the client sleep for the max time before
> > repolling for the lock. If it doesn't then we can skip all of the
> > waitqueue handling as well.
> > 
> > > > Signed-off-by: Jeff Layton <jlayton@redhat.com>
> > ---
> >  fs/nfs/nfs4_fs.h  | 1 +
> >  fs/nfs/nfs4proc.c | 2 ++
> >  2 files changed, 3 insertions(+)
> > 
> > diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
> > index 9bf64eacba5b..91e4f135a5f2 100644
> > --- a/fs/nfs/nfs4_fs.h
> > +++ b/fs/nfs/nfs4_fs.h
> > @@ -156,6 +156,7 @@ enum {
> > > > > >  	NFS_STATE_RECLAIM_NOGRACE,	/* OPEN stateid needs to recover state */
> > > > > >  	NFS_STATE_POSIX_LOCKS,		/* Posix locks are supported */
> > > > > >  	NFS_STATE_RECOVERY_FAILED,	/* OPEN stateid state recovery failed */
> > > > > > +	NFS_STATE_MAY_NOTIFY_LOCK,	/* server may CB_NOTIFY_LOCK */
> 
> Looks like the reason I don't have this flag is because it's added as part of the server patches.  I'll coordinate with the nfsd merge to make sure everything is added in the right order!
> 
> Thanks,
> Anna

Oh! Yes, that would explain it -- sorry...

Would it help at all to break out the addition of that field into a
separate patch? Just wondering what the right approach is for future
reference.

Thanks,
Jeff

> > 
> >  };
> >  
> >  struct nfs4_state {
> > diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> > index 90e8ded0ef82..627a9185822f 100644
> > --- a/fs/nfs/nfs4proc.c
> > +++ b/fs/nfs/nfs4proc.c
> > @@ -2537,6 +2537,8 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
> > > >  		goto out;
> > > >  	if (server->caps & NFS_CAP_POSIX_LOCK)
> > > >  		set_bit(NFS_STATE_POSIX_LOCKS, &state->flags);
> > > > +	if (opendata->o_res.rflags & NFS4_OPEN_RESULT_MAY_NOTIFY_LOCK)
> > > > +		set_bit(NFS_STATE_MAY_NOTIFY_LOCK, &state->flags);
> >  
> > > >  	dentry = opendata->dentry;
> > > >  	if (d_really_is_negative(dentry)) {
> > 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
Schumaker, Anna Sept. 12, 2016, 8:38 p.m. UTC | #3
On 09/12/2016 04:30 PM, Jeff Layton wrote:
> On Mon, 2016-09-12 at 16:19 -0400, Anna Schumaker wrote:
>> Hi Jeff,
>>
>> On 09/09/2016 02:47 PM, Jeff Layton wrote:
>>>
>>> If it does, then always have the client sleep for the max time before
>>> repolling for the lock. If it doesn't then we can skip all of the
>>> waitqueue handling as well.
>>>
>>>>> Signed-off-by: Jeff Layton <jlayton@redhat.com>
>>> ---
>>>  fs/nfs/nfs4_fs.h  | 1 +
>>>  fs/nfs/nfs4proc.c | 2 ++
>>>  2 files changed, 3 insertions(+)
>>>
>>> diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
>>> index 9bf64eacba5b..91e4f135a5f2 100644
>>> --- a/fs/nfs/nfs4_fs.h
>>> +++ b/fs/nfs/nfs4_fs.h
>>> @@ -156,6 +156,7 @@ enum {
>>>>>>>  	NFS_STATE_RECLAIM_NOGRACE,	/* OPEN stateid needs to recover state */
>>>>>>>  	NFS_STATE_POSIX_LOCKS,		/* Posix locks are supported */
>>>>>>>  	NFS_STATE_RECOVERY_FAILED,	/* OPEN stateid state recovery failed */
>>>>>>> +	NFS_STATE_MAY_NOTIFY_LOCK,	/* server may CB_NOTIFY_LOCK */
>>
>> Looks like the reason I don't have this flag is because it's added as part of the server patches.  I'll coordinate with the nfsd merge to make sure everything is added in the right order!
>>
>> Thanks,
>> Anna
> 
> Oh! Yes, that would explain it -- sorry...
> 
> Would it help at all to break out the addition of that field into a
> separate patch? Just wondering what the right approach is for future
> reference.

Hmm ... it might, actually.  I think we've done that in the past with changes to other common files, and then merge order doesn't matter as much.

Thanks,
Anna

> 
> Thanks,
> Jeff
> 
>>>
>>>  };
>>>  
>>>  struct nfs4_state {
>>> diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
>>> index 90e8ded0ef82..627a9185822f 100644
>>> --- a/fs/nfs/nfs4proc.c
>>> +++ b/fs/nfs/nfs4proc.c
>>> @@ -2537,6 +2537,8 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
>>>>>  		goto out;
>>>>>  	if (server->caps & NFS_CAP_POSIX_LOCK)
>>>>>  		set_bit(NFS_STATE_POSIX_LOCKS, &state->flags);
>>>>> +	if (opendata->o_res.rflags & NFS4_OPEN_RESULT_MAY_NOTIFY_LOCK)
>>>>> +		set_bit(NFS_STATE_MAY_NOTIFY_LOCK, &state->flags);
>>>  
>>>>>  	dentry = opendata->dentry;
>>>>>  	if (d_really_is_negative(dentry)) {
>>>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch
diff mbox

diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
index 9bf64eacba5b..91e4f135a5f2 100644
--- a/fs/nfs/nfs4_fs.h
+++ b/fs/nfs/nfs4_fs.h
@@ -156,6 +156,7 @@  enum {
 	NFS_STATE_RECLAIM_NOGRACE,	/* OPEN stateid needs to recover state */
 	NFS_STATE_POSIX_LOCKS,		/* Posix locks are supported */
 	NFS_STATE_RECOVERY_FAILED,	/* OPEN stateid state recovery failed */
+	NFS_STATE_MAY_NOTIFY_LOCK,	/* server may CB_NOTIFY_LOCK */
 };
 
 struct nfs4_state {
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 90e8ded0ef82..627a9185822f 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -2537,6 +2537,8 @@  static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
 		goto out;
 	if (server->caps & NFS_CAP_POSIX_LOCK)
 		set_bit(NFS_STATE_POSIX_LOCKS, &state->flags);
+	if (opendata->o_res.rflags & NFS4_OPEN_RESULT_MAY_NOTIFY_LOCK)
+		set_bit(NFS_STATE_MAY_NOTIFY_LOCK, &state->flags);
 
 	dentry = opendata->dentry;
 	if (d_really_is_negative(dentry)) {