[PATCH] ipv4 -- proc_net_fops_create

[PATCH] ipv4 -- proc_net_fops_create

[Stephen Hemminger]

Use proc_net_fops_create it is cleaner.

diff -Nru a/net/ipv4/proc.c b/net/ipv4/proc.c
--- a/net/ipv4/proc.c	Thu Sep 11 15:36:06 2003
+++ b/net/ipv4/proc.c	Thu Sep 11 15:36:06 2003
@@ -240,20 +240,12 @@
 	int rc = 0;
 	struct proc_dir_entry *p;
 
-	p = create_proc_entry("netstat", S_IRUGO, proc_net);
-	if (!p)
+	if (!proc_net_fops_create("netstat", S_IRUGO, &netstat_seq_fops))
 		goto out_netstat;
-	p->proc_fops = &netstat_seq_fops;
-
-	p = create_proc_entry("snmp", S_IRUGO, proc_net);
-	if (!p)
+	if (!proc_net_fops_create("snmp", S_IRUGO, &snmp_seq_fops))
 		goto out_snmp;
-	p->proc_fops = &snmp_seq_fops;
-
-	p = create_proc_entry("sockstat", S_IRUGO, proc_net);
-	if (!p)
+	if (!proc_net_fops_create("sockstat", S_IRUGO, &sockstat_seq_fops))
 		goto out_sockstat;
-	p->proc_fops = &sockstat_seq_fops;
 out:
 	return rc;
 out_sockstat:

Re: [PATCH] ipv4 -- proc_net_fops_create

[YOSHIFUJI Hideaki / 吉藤英明]

In article <20030911154357.6b31891c.shemminger@osdl.org> (at Thu, 11 Sep 2003 15:43:57 -0700), Stephen Hemminger <shemminger@osdl.org> says:

> Use proc_net_fops_create it is cleaner.

I sent this kind of patch for net/{core,ipv4,ipv6} to David / netdev on 
8th Sep. David?

-- 
Hideaki YOSHIFUJI @ USAGI Project <yoshfuji@linux-ipv6.org>
GPG FP: 9022 65EB 1ECF 3AD1 0BDF  80D8 4807 F894 E062 0EEA

Re: [PATCH] ipv4 -- proc_net_fops_create

[David S. Miller]

On Fri, 12 Sep 2003 07:50:03 +0900 (JST)
YOSHIFUJI Hideaki / _$B5HF#1QL@ <yoshfuji@linux-ipv6.org> wrote:

> I sent this kind of patch for net/{core,ipv4,ipv6} to David / netdev on 
> 8th Sep. David?

I'm heavily backlogged and haven't been able to apply any patches.
Sorry.  I'll try to do some today.

[PATCH] proc_net macro's for non-proc-fs case

[Stephen Hemminger]

The following should work better. They have the desired properties that:
	- return value makes sense for those that might check
	- arguments fops and info don't have to be defined (ie can be #ifdef'd out)
	- generate no warnings for null effect.

diff -Nru a/include/linux/proc_fs.h b/include/linux/proc_fs.h
--- a/include/linux/proc_fs.h	Fri Sep 12 11:19:07 2003
+++ b/include/linux/proc_fs.h	Fri Sep 12 11:19:07 2003
@@ -185,9 +185,10 @@
 #else
 
 #define proc_root_driver NULL
-#define proc_net_fops_create(name,mode,fops) do {} while(0)
-static inline struct proc_dir_entry *proc_net_create(const char *name, mode_t mode, 
-	get_info_t *get_info) {return NULL;}
+#define proc_net NULL
+
+#define proc_net_fops_create(name, mode, fops)  ({ (void)(mode), NULL; })
+#define proc_net_create(name, mode, info)	({ (void)(mode), NULL; })
 static inline void proc_net_remove(const char *name) {}
 
 static inline struct dentry *proc_pid_unhash(struct task_struct *p) { return NULL; }

Re: [PATCH] proc_net macro's for non-proc-fs case

[David S. Miller]

On Fri, 12 Sep 2003 11:21:20 -0700
Stephen Hemminger <shemminger@osdl.org> wrote:

> The following should work better. They have the desired properties that:
> 	- return value makes sense for those that might check
> 	- arguments fops and info don't have to be defined (ie can be #ifdef'd out)
> 	- generate no warnings for null effect.

Looks good, applied thanks.