diff mbox series

[RFC,06/21] NFSD: Remove spurious cb_setup_err tracepoint

Message ID 162066193457.94415.10829735588517134118.stgit@klimt.1015granger.net (mailing list archive)
State New
Headers show
Series NFSD callback and lease management observability | expand

Commit Message

Chuck Lever III May 10, 2021, 3:52 p.m. UTC
This path is not really an error path, so the tracepoint I added
there is just noise.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---
 fs/nfsd/nfs4callback.c |    4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

Comments

J. Bruce Fields May 10, 2021, 8:27 p.m. UTC | #1
On Mon, May 10, 2021 at 11:52:14AM -0400, Chuck Lever wrote:
> This path is not really an error path,

What's the non-error case for this path?

On a quick look it seems like that'd mean a 4.1 client doesn't have a
connection available for the backchannel, which sounds bad.

But I'm probably overlooking something....

--b.

> so the tracepoint I added
> there is just noise.
> 
> Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
> ---
>  fs/nfsd/nfs4callback.c |    4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
> index ab1836381e22..15ba16c54793 100644
> --- a/fs/nfsd/nfs4callback.c
> +++ b/fs/nfsd/nfs4callback.c
> @@ -915,10 +915,8 @@ static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *c
>  		args.authflavor = clp->cl_cred.cr_flavor;
>  		clp->cl_cb_ident = conn->cb_ident;
>  	} else {
> -		if (!conn->cb_xprt) {
> -			trace_nfsd_cb_setup_err(clp, -EINVAL);
> +		if (!conn->cb_xprt)
>  			return -EINVAL;
> -		}
>  		clp->cl_cb_conn.cb_xprt = conn->cb_xprt;
>  		clp->cl_cb_session = ses;
>  		args.bc_xprt = conn->cb_xprt;
>
Chuck Lever III May 10, 2021, 8:29 p.m. UTC | #2
> On May 10, 2021, at 4:27 PM, J. Bruce Fields <bfields@fieldses.org> wrote:
> 
> On Mon, May 10, 2021 at 11:52:14AM -0400, Chuck Lever wrote:
>> This path is not really an error path,
> 
> What's the non-error case for this path?

From what I can tell, it appears to be the default exit for when
there is a session and backchannel. Feel free to straighten me
out, but it just seemed to always fire for NFSv4.1 mounts.


> On a quick look it seems like that'd mean a 4.1 client doesn't have a
> connection available for the backchannel, which sounds bad.
> 
> But I'm probably overlooking something....
> 
> --b.
> 
>> so the tracepoint I added
>> there is just noise.
>> 
>> Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
>> ---
>> fs/nfsd/nfs4callback.c |    4 +---
>> 1 file changed, 1 insertion(+), 3 deletions(-)
>> 
>> diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
>> index ab1836381e22..15ba16c54793 100644
>> --- a/fs/nfsd/nfs4callback.c
>> +++ b/fs/nfsd/nfs4callback.c
>> @@ -915,10 +915,8 @@ static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *c
>> 		args.authflavor = clp->cl_cred.cr_flavor;
>> 		clp->cl_cb_ident = conn->cb_ident;
>> 	} else {
>> -		if (!conn->cb_xprt) {
>> -			trace_nfsd_cb_setup_err(clp, -EINVAL);
>> +		if (!conn->cb_xprt)
>> 			return -EINVAL;
>> -		}
>> 		clp->cl_cb_conn.cb_xprt = conn->cb_xprt;
>> 		clp->cl_cb_session = ses;
>> 		args.bc_xprt = conn->cb_xprt;
>> 

--
Chuck Lever
J. Bruce Fields May 11, 2021, 5:44 p.m. UTC | #3
On Mon, May 10, 2021 at 08:29:32PM +0000, Chuck Lever III wrote:
> 
> 
> > On May 10, 2021, at 4:27 PM, J. Bruce Fields <bfields@fieldses.org> wrote:
> > 
> > On Mon, May 10, 2021 at 11:52:14AM -0400, Chuck Lever wrote:
> >> This path is not really an error path,
> > 
> > What's the non-error case for this path?
> 
> >From what I can tell, it appears to be the default exit for when
> there is a session and backchannel. Feel free to straighten me
> out, but it just seemed to always fire for NFSv4.1 mounts.

I'd be curious to know why.  I'll see if I can find some time to
experiment.

--b.

> > On a quick look it seems like that'd mean a 4.1 client doesn't have a
> > connection available for the backchannel, which sounds bad.
> > 
> > But I'm probably overlooking something....
> > 
> > --b.
> > 
> >> so the tracepoint I added
> >> there is just noise.
> >> 
> >> Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
> >> ---
> >> fs/nfsd/nfs4callback.c |    4 +---
> >> 1 file changed, 1 insertion(+), 3 deletions(-)
> >> 
> >> diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
> >> index ab1836381e22..15ba16c54793 100644
> >> --- a/fs/nfsd/nfs4callback.c
> >> +++ b/fs/nfsd/nfs4callback.c
> >> @@ -915,10 +915,8 @@ static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *c
> >> 		args.authflavor = clp->cl_cred.cr_flavor;
> >> 		clp->cl_cb_ident = conn->cb_ident;
> >> 	} else {
> >> -		if (!conn->cb_xprt) {
> >> -			trace_nfsd_cb_setup_err(clp, -EINVAL);
> >> +		if (!conn->cb_xprt)
> >> 			return -EINVAL;
> >> -		}
> >> 		clp->cl_cb_conn.cb_xprt = conn->cb_xprt;
> >> 		clp->cl_cb_session = ses;
> >> 		args.bc_xprt = conn->cb_xprt;
> >> 
> 
> --
> Chuck Lever
> 
>
diff mbox series

Patch

diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index ab1836381e22..15ba16c54793 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -915,10 +915,8 @@  static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *c
 		args.authflavor = clp->cl_cred.cr_flavor;
 		clp->cl_cb_ident = conn->cb_ident;
 	} else {
-		if (!conn->cb_xprt) {
-			trace_nfsd_cb_setup_err(clp, -EINVAL);
+		if (!conn->cb_xprt)
 			return -EINVAL;
-		}
 		clp->cl_cb_conn.cb_xprt = conn->cb_xprt;
 		clp->cl_cb_session = ses;
 		args.bc_xprt = conn->cb_xprt;