On Tue, 2017-05-09 at 11:19 +0200, Christoph Hellwig wrote:
> p_count is the only writeable memeber of struct rpc_procinfo, which
> is
> a good candidate to be const-ified as it contains function pointers.
>
> This patch just removes it and returns zero in the /proc file as the
> count of calls per procedure doesn't seem all that useful. But it's
> just a dumb hack and we might need a proper fix instead.
>
> Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> ---
> include/linux/sunrpc/clnt.h | 1 -
> net/sunrpc/clnt.c | 1 -
> net/sunrpc/stats.c | 3 +--
> 3 files changed, 1 insertion(+), 4 deletions(-)
>
> diff --git a/include/linux/sunrpc/clnt.h
> b/include/linux/sunrpc/clnt.h
> index 6095ecba0dde..d1f6778c7da7 100644
> --- a/include/linux/sunrpc/clnt.h
> +++ b/include/linux/sunrpc/clnt.h
> @@ -99,7 +99,6 @@ struct rpc_procinfo {
> kxdrdproc_t p_decode; /* XDR decode
> function */
> unsigned int p_arglen; /* argument hdr
> length (u32) */
> unsigned int p_replen; /* reply hdr
> length (u32) */
> - unsigned int p_count; /* call count */
> unsigned int p_timer; /* Which RTT
> timer to use */
> u32 p_statidx; /* Which
> procedure to account */
> const char * p_name; /* name
> of procedure */
> diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
> index bdb75f7488e9..e6d2ae49d791 100644
> --- a/net/sunrpc/clnt.c
> +++ b/net/sunrpc/clnt.c
> @@ -1532,7 +1532,6 @@ call_start(struct rpc_task *task)
> (RPC_IS_ASYNC(task) ? "async" : "sync"));
>
> /* Increment call count */
> - task->tk_msg.rpc_proc->p_count++;
> clnt->cl_stats->rpccnt++;
> task->tk_action = call_reserve;
> }
> diff --git a/net/sunrpc/stats.c b/net/sunrpc/stats.c
> index caeb01ad2b5a..88b1e18c0dd6 100644
> --- a/net/sunrpc/stats.c
> +++ b/net/sunrpc/stats.c
> @@ -55,8 +55,7 @@ static int rpc_proc_show(struct seq_file *seq, void
> *v) {
> seq_printf(seq, "proc%u %u",
> vers->number, vers-
> >nrprocs);
> for (j = 0; j < vers->nrprocs; j++)
> - seq_printf(seq, " %u",
> - vers->procs[j].p_count);
> + seq_printf(seq, " %u", 0);
> seq_putc(seq, '\n');
> }
> return 0;
Again, we really do not want to break nfsstat or mountstats. A separate
statistics structure seems the way to go.
--
Trond Myklebust
Linux NFS client maintainer, PrimaryData
trond.myklebust@xxxxxxxxxxxxxxx
��.n��������+%������w��{.n�����{��w��� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
