: Use proc_create() to setup ->proc_fops first

PATCH Applied Wang Chen 3 participants 5 messages Feb 28, 2008 – Feb 29, 2008

Subsystems (4)

Networking Network drivers NFS client NFS server

lore

Share & tools

lore

Wang Chen

author

2008-02-28 10:58 UTC #0 lore raw

PATCH 3/15 : Use proc_create() to setup ->proc_fops first

Use proc_create() to make sure that ->proc_fops be setup before gluing PDE to main tree.

Signed-off-by Wang Chen <[email protected]>

---
 net/sunrpc/cache.c |   14 ++++++--------
 net/sunrpc/stats.c |    3 +--
 2 files changed, 7 insertions(+), 10 deletions(-)

net/sunrpc/cache.c +6 -8

index 636c8e0..b5f2786 100644
@@ -316,31 +316,29 @@ static int create_cache_proc_entries(struct cache_detail *cd)
 	cd->proc_ent->owner = cd->owner;
 	cd->channel_ent = cd->content_ent = NULL;
 
-	p = create_proc_entry("flush", S_IFREG|S_IRUSR|S_IWUSR, cd->proc_ent);
+	p = proc_create("flush", S_IFREG|S_IRUSR|S_IWUSR,
+			cd->proc_ent, &cache_flush_operations);
 	cd->flush_ent = p;
 	if (p == NULL)
 		goto out_nomem;
-	p->proc_fops = &cache_flush_operations;
 	p->owner = cd->owner;
 	p->data = cd;
 
 	if (cd->cache_request || cd->cache_parse) {
-		p = create_proc_entry("channel", S_IFREG|S_IRUSR|S_IWUSR,
-				      cd->proc_ent);
+		p = proc_create("channel", S_IFREG|S_IRUSR|S_IWUSR,
+				cd->proc_ent, &cache_file_operations);
 		cd->channel_ent = p;
 		if (p == NULL)
 			goto out_nomem;
-		p->proc_fops = &cache_file_operations;
 		p->owner = cd->owner;
 		p->data = cd;
 	}
 	if (cd->cache_show) {
-		p = create_proc_entry("content", S_IFREG|S_IRUSR|S_IWUSR,
-				      cd->proc_ent);
+		p = proc_create("content", S_IFREG|S_IRUSR|S_IWUSR,
+				cd->proc_ent, &content_file_operations);
 		cd->content_ent = p;
 		if (p == NULL)
 			goto out_nomem;
-		p->proc_fops = &content_file_operations;
 		p->owner = cd->owner;
 		p->data = cd;
 	}

net/sunrpc/stats.c +1 -2

index 5a16875..c6061a4 100644
@@ -229,9 +229,8 @@ do_register(const char *name, void *data, const struct file_operations *fops)
 	rpc_proc_init();
 	dprintk("RPC:       registering /proc/net/rpc/%s\n", name);
 
-	ent = create_proc_entry(name, 0, proc_net_rpc);
+	ent = proc_create(name, 0, proc_net_rpc, fops);
 	if (ent) {
-		ent->proc_fops = fops;
 		ent->data = data;
 	}
 	return ent;

Signature · 1 line

WCN

David Miller

maintainer Applied

2008-02-28 22:02 UTC #1 lore raw

From: Wang Chen <[email protected]>
Date: Thu, 28 Feb 2008 18:55:40 +0800

Quoted 4 lines

Use proc_create() to make sure that ->proc_fops be setup before gluing PDE to main tree.
Signed-off-by Wang Chen <[email protected]>

Applied.

net/sunrpc/stats.c do_register(const char *name, void *data, const struct file_operations *fops) · line 236

Quoted 11 lines

net/sunrpc/stats.c +1 -2

@@ -229,9 +229,8 @@ do_register(const char *name, void *data, const struct file_operations *fops)
 	rpc_proc_init();
 	dprintk("RPC:       registering /proc/net/rpc/%s\n", name);
 
-	ent = create_proc_entry(name, 0, proc_net_rpc);
+	ent = proc_create(name, 0, proc_net_rpc, fops);
 	if (ent) {
-		ent->proc_fops = fops;
 		ent->data = data;
 	}
 	return ent;

For this case it appears that ent->data has the same kind of visibility problem that ent->proc_fops does.

Wang Chen

author

2008-02-29 05:29 UTC #2 lore raw

David Miller said the following on 2008-2-29 6:02:

Quoted 14 lines

From: Wang Chen <[email protected]>
Date: Thu, 28 Feb 2008 18:55:40 +0800
Quoted 4 lines
Use proc_create() to make sure that ->proc_fops be setup before gluing PDE to main tree.
Signed-off-by Wang Chen <[email protected]>
Applied.
Quoted 4 lines
net/sunrpc/stats.c +0 -0
@@ -229,9 +229,8 @@ do_register(const char *name, void *data, const struct file_operations *fops)
 	rpc_proc_init();
 	dprintk("RPC:       registering /proc/net/rpc/%s\n", name);
 

Quoted 11 lines

net/sunrpc/stats.c do_register(const char *name, void *data, const struct file_operations *fops) · line 236
Quoted 7 lines
net/sunrpc/stats.c +1 -2
-	ent = create_proc_entry(name, 0, proc_net_rpc);
+	ent = proc_create(name, 0, proc_net_rpc, fops);
 	if (ent) {
-		ent->proc_fops = fops;
 		ent->data = data;
 	}
 	return ent;
For this case it appears that ent->data has the same kind of visibility problem that ent->proc_fops does.

Thanks Dave, I will check whether ->data also has the visibility problem.

Wang Chen

author

2008-02-29 08:37 UTC #3 lore raw

Wang Chen said the following on 2008-2-29 13:26:

Quoted 25 lines

David Miller said the following on 2008-2-29 6:02:
Quoted 13 lines
From: Wang Chen <[email protected]>
Date: Thu, 28 Feb 2008 18:55:40 +0800
Quoted 4 lines
Use proc_create() to make sure that ->proc_fops be setup before gluing PDE to main tree.
Signed-off-by Wang Chen <[email protected]>
Applied.
Quoted 4 lines
net/sunrpc/stats.c +0 -0
@@ -229,9 +229,8 @@ do_register(const char *name, void *data, const struct file_operations *fops)
 	rpc_proc_init();
 	dprintk("RPC:       registering /proc/net/rpc/%s\n", name);
 
Quoted 10 lines
net/sunrpc/stats.c do_register(const char *name, void *data, const struct file_operations *fops) · line 236
Quoted 7 lines
net/sunrpc/stats.c +1 -2
-	ent = create_proc_entry(name, 0, proc_net_rpc);
+	ent = proc_create(name, 0, proc_net_rpc, fops);
 	if (ent) {
-		ent->proc_fops = fops;
 		ent->data = data;
 	}
 	return ent;
For this case it appears that ent->data has the same kind of visibility problem that ent->proc_fops does.

Quoted 2 lines

Thanks Dave, I will check whether ->data also has the visibility problem.

I have looked at the proc_create().
The reason for why we need to setup pde->proc_fops in proc_create() before the pde be visible, is that proc_fops will be setuped in proc_register() and and NULL of proc_fops will make proc_register() give pde an improper fops.

But, ->data and ->owner will not be affected by this instance. So, it's safe to setup ->data and ->owner after visibility of pde.

Alexey Dobriyan

2008-02-29 09:32 UTC #4 lore raw

Quoted 29 lines · Wang Chen

Wang Chen said the following on 2008-2-29 13:26:
Quoted 25 lines
David Miller said the following on 2008-2-29 6:02:
Quoted 13 lines
From: Wang Chen <[email protected]>
Date: Thu, 28 Feb 2008 18:55:40 +0800
Quoted 4 lines
Use proc_create() to make sure that ->proc_fops be setup before gluing PDE to main tree.
Signed-off-by Wang Chen <[email protected]>
Applied.
Quoted 4 lines
net/sunrpc/stats.c +0 -0
@@ -229,9 +229,8 @@ do_register(const char *name, void *data, const struct file_operations *fops)
 	rpc_proc_init();
 	dprintk("RPC:       registering /proc/net/rpc/%s\n", name);
 
Quoted 10 lines
net/sunrpc/stats.c do_register(const char *name, void *data, const struct file_operations *fops) · line 236
Quoted 7 lines
net/sunrpc/stats.c +1 -2
-	ent = create_proc_entry(name, 0, proc_net_rpc);
+	ent = proc_create(name, 0, proc_net_rpc, fops);
 	if (ent) {
-		ent->proc_fops = fops;
 		ent->data = data;
 	}
 	return ent;
For this case it appears that ent->data has the same kind of visibility problem that ent->proc_fops does.
Quoted 2 lines
Thanks Dave, I will check whether ->data also has the visibility problem.

Quoted 7 lines

I have looked at the proc_create().
The reason for why we need to setup pde->proc_fops in proc_create() before the pde be visible, is that proc_fops will be setuped in proc_register() and and NULL of proc_fops will make proc_register() give pde an improper fops.
But, ->data and ->owner will not be affected by this instance. So, it's safe to setup ->data and ->owner after visibility of pde.

->owner is buggy, believed to be unnecessary and will thus die so don't bother with it. ;-)

->data looks safe to setup separately while module is loading:
before ->data will be used by code which is unprepared to it being NULL, VFS will pin module, which can't be done during module load.

I think the existence of proc_create_data() depends entirely on cases which do all the above from code that is not module_init hook. Does anyone know some?