[PATCH] SUNRPC: skip dead but not buried clients on PipeFS events

[PATCH] SUNRPC: skip dead but not buried clients on PipeFS events

[Stanislav Kinsbursky]

These clients can't be safely dereferenced if their counter in 0.

Signee-off-by: Stanislav Kinsbursky <skinsbursky@parallels.com>

Signed-off-by: Stanislav Kinsbursky <skinsbursky@parallels.com>

---
 net/sunrpc/clnt.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 6797246..591994d 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -218,7 +218,8 @@ static struct rpc_clnt *rpc_get_client_for_event(struct net *net, int event)
 		if (((event == RPC_PIPEFS_MOUNT) && clnt->cl_dentry) ||
 		    ((event == RPC_PIPEFS_UMOUNT) && !clnt->cl_dentry))
 			continue;
-		atomic_inc(&clnt->cl_count);
+		if (atomic_inc_return(&clnt->cl_count) == 1)
+			continue;
 		spin_unlock(&sn->rpc_client_lock);
 		return clnt;
 	}

Re: [PATCH] SUNRPC: skip dead but not buried clients on PipeFS events

[Stanislav Kinsbursky]

Sorry, but ignore this patch too.
It can't be that simple because of these cl_count tricks in rpc_release_client...

19.04.2012 14:57, Stanislav Kinsbursky пишет:
> These clients can't be safely dereferenced if their counter in 0.
>
> Signee-off-by: Stanislav Kinsbursky<skinsbursky-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
>
> Signed-off-by: Stanislav Kinsbursky<skinsbursky-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
>
> ---
>   net/sunrpc/clnt.c |    3 ++-
>   1 files changed, 2 insertions(+), 1 deletions(-)
>
> diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
> index 6797246..591994d 100644
> --- a/net/sunrpc/clnt.c
> +++ b/net/sunrpc/clnt.c
> @@ -218,7 +218,8 @@ static struct rpc_clnt *rpc_get_client_for_event(struct net *net, int event)
>   		if (((event == RPC_PIPEFS_MOUNT)&&  clnt->cl_dentry) ||
>   		    ((event == RPC_PIPEFS_UMOUNT)&&  !clnt->cl_dentry))
>   			continue;
> -		atomic_inc(&clnt->cl_count);
> +		if (atomic_inc_return(&clnt->cl_count) == 1)
> +			continue;
>   		spin_unlock(&sn->rpc_client_lock);
>   		return clnt;
>   	}
>


-- 
Best regards,
Stanislav Kinsbursky
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] SUNRPC: skip dead but not buried clients on PipeFS events

[bfields]

On Thu, Apr 19, 2012 at 03:36:57PM +0400, Stanislav Kinsbursky wrote:
> Sorry, but ignore this patch too.
> It can't be that simple because of these cl_count tricks in rpc_release_client...

OK.  Very minor whine:

> 
> 19.04.2012 14:57, Stanislav Kinsbursky пишет:
> >These clients can't be safely dereferenced if their counter in 0.
> >
> >Signee-off-by: Stanislav Kinsbursky<skinsbursky@parallels.com>
> >
> >Signed-off-by: Stanislav Kinsbursky<skinsbursky@parallels.com>

I don't mind fixing up trivial slips every now and then, but that double
signed-off-by seems to happen on a lot of your posts; could you figure
out what's up with your scripts?

--b.

> >
> >---
> >  net/sunrpc/clnt.c |    3 ++-
> >  1 files changed, 2 insertions(+), 1 deletions(-)
> >
> >diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
> >index 6797246..591994d 100644
> >--- a/net/sunrpc/clnt.c
> >+++ b/net/sunrpc/clnt.c
> >@@ -218,7 +218,8 @@ static struct rpc_clnt *rpc_get_client_for_event(struct net *net, int event)
> >  		if (((event == RPC_PIPEFS_MOUNT)&&  clnt->cl_dentry) ||
> >  		    ((event == RPC_PIPEFS_UMOUNT)&&  !clnt->cl_dentry))
> >  			continue;
> >-		atomic_inc(&clnt->cl_count);
> >+		if (atomic_inc_return(&clnt->cl_count) == 1)
> >+			continue;
> >  		spin_unlock(&sn->rpc_client_lock);
> >  		return clnt;
> >  	}
> >
> 
> 
> -- 
> Best regards,
> Stanislav Kinsbursky
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH] SUNRPC: skip dead but not buried clients on PipeFS events

[Stanislav Kinsbursky]

20.04.2012 01:40, bfields-uC3wQj2KruNg9hUCZPvPmw@public.gmane.org написал:
> On Thu, Apr 19, 2012 at 03:36:57PM +0400, Stanislav Kinsbursky wrote:
>> Sorry, but ignore this patch too.
>> It can't be that simple because of these cl_count tricks in rpc_release_client...
> OK.  Very minor whine:
>
>> 19.04.2012 14:57, Stanislav Kinsbursky пишет:
>>> These clients can't be safely dereferenced if their counter in 0.
>>>
>>> Signee-off-by: Stanislav Kinsbursky<skinsbursky-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
>>>
>>> Signed-off-by: Stanislav Kinsbursky<skinsbursky-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>
> I don't mind fixing up trivial slips every now and then, but that double
> signed-off-by seems to happen on a lot of your posts; could you figure
> out what's up with your scripts?

Yes, sure. Sorry.
I've noticed this problem just yesterday.
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html