Skip to contentSkip to conversation

nfsd: Remove write permission from file content

PATCH In review [email protected] 2 participants 3 messages Jan 4, 2013 – Jan 8, 2013 1 message not in the archive

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 nfsd: Remove write permission from file content

From: Yanchuan Nian <[email protected]>

The write function doesn't be implemented in file content, and it's meaningless to write data into this file directly. Remove write permission from it.

---
 net/sunrpc/cache.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
net/sunrpc/cache.c +1 -1
index 9afa439..9f84703 100644
@@ -1614,7 +1614,7 @@ static int create_cache_proc_entries(struct cache_detail *cd, struct net *net)
 			goto out_nomem;
 	}
 	if (cd->cache_show) {
-		p = proc_create_data("content", S_IFREG|S_IRUSR|S_IWUSR,
+		p = proc_create_data("content", S_IFREG|S_IRUSR,
 				cd->u.procfs.proc_ent,
 				&content_file_operations_procfs, cd);
 		cd->u.procfs.content_ent = p;
Signature · 1 line
1.7.4.4
#1 lore raw
Quoted 4 lines · On Fri, Jan 04, 2013 at 07:45:35PM +0800, [email protected] wrote:

From: Yanchuan Nian <[email protected]>

The write function doesn't be implemented in file content, and it's meaningless to write data into this file directly. Remove write permission from it.

So does it really matter either way?

OK, applying, but I wonder.

--b.

Quoted 22 lines
---
 net/sunrpc/cache.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
net/sunrpc/cache.c +1 -1
index 9afa439..9f84703 100644
@@ -1614,7 +1614,7 @@ static int create_cache_proc_entries(struct cache_detail *cd, struct net *net)
 			goto out_nomem;
 	}
 	if (cd->cache_show) {
-		p = proc_create_data("content", S_IFREG|S_IRUSR|S_IWUSR,
+		p = proc_create_data("content", S_IFREG|S_IRUSR,
 				cd->u.procfs.proc_ent,
 				&content_file_operations_procfs, cd);
 		cd->u.procfs.content_ent = p;
Signature · 1 line
1.7.4.4

A message that is not in the archive (sent off-list or not yet received): CAC5OsZPR88SktqjTKsqJVaxKZ1LkPnvPcT2gPSFq_oPDZ1FZpw@mail.gmail.com

#3 lore raw
Quoted 11 lines · Yanchuan Nian

2013/1/5 J. Bruce Fields <[email protected]>

Quoted 9 lines · On Fri, Jan 04, 2013 at 07:45:35PM +0800, [email protected] wrote:
Quoted 3 lines

From: Yanchuan Nian <[email protected]>

The write function doesn't be implemented in file content, and it's

meaningless

Quoted 1 line

to write data into this file directly. Remove write permission from it.

So does it really matter either way?

Quoted 7 lines

No, it doesn't matter. I saw that nfs-utils communicates with nfsd through proc fs, so I tried to update the cache by writing data to proc fs directly. I found the format of "channel" and "flush" in nfs-utils, but I couldn't find the format of "content", so I read the sunrpc source code, and found that the write function doesn't be implemented, but the write permission is set when registering into proc fs. I also found that "content" in pipe fs doesn't have write permission.

OK, so if it had been read-only it might have saved you a little confusion. Fair enough.

--b.

Quoted 30 lines
Quoted 30 lines

OK, applying, but I wonder.

--b.

net/sunrpc/cache.c static int create_cache_proc_entries(struct

Quoted 11 lines
---
 net/sunrpc/cache.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
net/sunrpc/cache.c +0 -0
index 9afa439..9f84703 100644
@@ -1614,7 +1614,7 @@ static int create_cache_proc_entries(struct

cache_detail *cd, struct net *net)

Quoted 11 lines
net/sunrpc/cache.c +1 -2
                      goto out_nomem;
      }
      if (cd->cache_show) {
-             p = proc_create_data("content", S_IFREG|S_IRUSR|S_IWUSR,
+             p = proc_create_data("content", S_IFREG|S_IRUSR,
                              cd->u.procfs.proc_ent,
                              &content_file_operations_procfs, cd);
              cd->u.procfs.content_ent = p;
--

1.7.4.4