[RFC, PATCH 4/15] knfsd: map null socket to SOCK_STREAM in svc_max_payload

[RFC, PATCH 4/15] knfsd: map null socket to SOCK_STREAM in svc_max_payload

[Tom Tucker]

The svc_max_payload function looks into the sk_sock structure to determine
what the reported max payload should be. For RDMA, there is no socket. This 
change effectively maps a NULL sk_sock to SOCK_STREAM.

Signed-off-by: Tom Tucker <tom@opengridcomputing.com>
---

 net/sunrpc/svc.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
index 9fec17d..b343db8 100644
--- a/net/sunrpc/svc.c
+++ b/net/sunrpc/svc.c
@@ -1024,7 +1024,7 @@ u32 svc_max_payload(const struct svc_rqs
 {
 	int max = RPCSVC_MAXPAYLOAD_TCP;
 
-	if (rqstp->rq_sock->sk_sock->type == SOCK_DGRAM)
+	if (rqstp->rq_sock->sk_sock && rqstp->rq_sock->sk_sock->type == SOCK_DGRAM)
 		max = RPCSVC_MAXPAYLOAD_UDP;
 	if (rqstp->rq_server->sv_max_payload < max)
 		max = rqstp->rq_server->sv_max_payload;


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: [RFC, PATCH 4/15] knfsd: map null socket to SOCK_STREAM in svc_max_payload

[J. Bruce Fields]

On Fri, May 18, 2007 at 12:45:07PM -0500, Tom Tucker wrote:
> 
> The svc_max_payload function looks into the sk_sock structure to determine
> what the reported max payload should be. For RDMA, there is no socket. This 
> change effectively maps a NULL sk_sock to SOCK_STREAM.
> 
> Signed-off-by: Tom Tucker <tom@opengridcomputing.com>
> ---
> 
>  net/sunrpc/svc.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
> index 9fec17d..b343db8 100644
> --- a/net/sunrpc/svc.c
> +++ b/net/sunrpc/svc.c
> @@ -1024,7 +1024,7 @@ u32 svc_max_payload(const struct svc_rqs
>  {
>  	int max = RPCSVC_MAXPAYLOAD_TCP;
>  
> -	if (rqstp->rq_sock->sk_sock->type == SOCK_DGRAM)
> +	if (rqstp->rq_sock->sk_sock && rqstp->rq_sock->sk_sock->type == SOCK_DGRAM)
>  		max = RPCSVC_MAXPAYLOAD_UDP;
>  	if (rqstp->rq_server->sv_max_payload < max)
>  		max = rqstp->rq_server->sv_max_payload;

Don't you want to use Greg's sko_max_payload() (or whatever it ends up
being) for this?

--b.

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: [RFC, PATCH 4/15] knfsd: map null socket to SOCK_STREAM in svc_max_payload

[Tom Tucker]

On Fri, 2007-05-18 at 14:28 -0400, J. Bruce Fields wrote:
> On Fri, May 18, 2007 at 12:45:07PM -0500, Tom Tucker wrote:
> > 
> > The svc_max_payload function looks into the sk_sock structure to determine
> > what the reported max payload should be. For RDMA, there is no socket. This 
> > change effectively maps a NULL sk_sock to SOCK_STREAM.
> > 
> > Signed-off-by: Tom Tucker <tom@opengridcomputing.com>
> > ---
> > 
> >  net/sunrpc/svc.c |    2 +-
> >  1 files changed, 1 insertions(+), 1 deletions(-)
> > 
> > diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
> > index 9fec17d..b343db8 100644
> > --- a/net/sunrpc/svc.c
> > +++ b/net/sunrpc/svc.c
> > @@ -1024,7 +1024,7 @@ u32 svc_max_payload(const struct svc_rqs
> >  {
> >  	int max = RPCSVC_MAXPAYLOAD_TCP;
> >  
> > -	if (rqstp->rq_sock->sk_sock->type == SOCK_DGRAM)
> > +	if (rqstp->rq_sock->sk_sock && rqstp->rq_sock->sk_sock->type == SOCK_DGRAM)
> >  		max = RPCSVC_MAXPAYLOAD_UDP;
> >  	if (rqstp->rq_server->sv_max_payload < max)
> >  		max = rqstp->rq_server->sv_max_payload;
> 
> Don't you want to use Greg's sko_max_payload() (or whatever it ends up
> being) for this?

Yes, however, we don't know what it's going to end up being yet and I'd
like to get this code reviewed, so I can work the issues in parallel.
BTW, the current transport switch (before the mods) is up and running in
several locations. 

> 
> --b.


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: [RFC, PATCH 4/15] knfsd: map null socket to SOCK_STREAM in svc_max_payload

[J. Bruce Fields]

On Fri, May 18, 2007 at 01:40:40PM -0500, Tom Tucker wrote:
> Yes, however, we don't know what it's going to end up being yet and I'd
> like to get this code reviewed, so I can work the issues in parallel.

Sure.

But I guess some questions will be easiest to answer when the two are
completely integrated.

> BTW, the current transport switch (before the mods) is up and running in
> several locations. 

OK!

--b.

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: [RFC, PATCH 4/15] knfsd: map null socket to SOCK_STREAM in svc_max_payload

[Tom Tucker]

On Fri, 2007-05-18 at 14:48 -0400, J. Bruce Fields wrote:
> On Fri, May 18, 2007 at 01:40:40PM -0500, Tom Tucker wrote:
> > Yes, however, we don't know what it's going to end up being yet and I'd
> > like to get this code reviewed, so I can work the issues in parallel.
> 
> Sure.
> 
> But I guess some questions will be easiest to answer when the two are
> completely integrated.

I also think by considering the two together, you can get a better
understanding of what's actually needed.

> 
> > BTW, the current transport switch (before the mods) is up and running in
> > several locations. 
> 
> OK!
> 
> --b.


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs