Skip to contentSkip to conversation

NFSv4: open context can be NULL in nfs_put_open_context

PATCH In review [email protected] 3 participants 4 messages Oct 13, 2014 1 message not in the archive

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 NFSv4: open context can be NULL in nfs_put_open_context

From: Andy Adamson <[email protected]>

If a pgio error is caught before the nfs_pgio_header args->context is set with get_nfs_open_context, the context can be NULL.

---
 fs/nfs/inode.c | 3 +++
 1 file changed, 3 insertions(+)
fs/nfs/inode.c +3 -0
index 141c9f4..a2b148e 100644
@@ -798,6 +798,9 @@ static void __put_nfs_open_context(struct nfs_open_context *ctx, int is_sync)
 	struct inode *inode = ctx->dentry->d_inode;
 	struct super_block *sb = ctx->dentry->d_sb;
 
+	if (ctx == NULL)
+		return;
+
 	if (!list_empty(&ctx->list)) {
 		if (!atomic_dec_and_lock(&ctx->lock_context.count, &inode->i_lock))
 			return;
Signature · 1 line
1.8.3.1
#1 lore raw
Quoted 4 lines · On Oct 13, 2014, at 10:08 AM, [email protected] wrote:

From: Andy Adamson <[email protected]>

If a pgio error is caught before the nfs_pgio_header args->context is set with get_nfs_open_context, the context can be NULL.

It looks like all three calls to nfs_pgio_error would hit this!

The fix looks good to me.

-dros

Quoted 27 lines

---
fs/nfs/inode.c | 3 +++

1 file changed, 3 insertions(+)
fs/nfs/inode.c +3 -0
index 141c9f4..a2b148e 100644
@@ -798,6 +798,9 @@ static void __put_nfs_open_context(struct nfs_open_context *ctx, int is_sync)
 	struct inode *inode = ctx->dentry->d_inode;
 	struct super_block *sb = ctx->dentry->d_sb;
 
+	if (ctx == NULL)
+		return;
+
 	if (!list_empty(&ctx->list)) {
 		if (!atomic_dec_and_lock(&ctx->lock_context.count, &inode->i_lock))
 			return;
Signature · 6 lines
1.8.3.1

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

A message that is not in the archive (sent off-list or not yet received): CAHQdGtT12WC896sZwHQkgXgiPx0ODckX2dhNZb0O+XHBUN2E-A@mail.gmail.com

#3 lore raw

Resending to list... Apologies for the duplicate...

Quoted 49 lines · Trond Myklebust

Hi Andy,

Quoted 26 lines · On Mon, Oct 13, 2014 at 10:08 AM, <[email protected]> wrote:

From: Andy Adamson <[email protected]>

If a pgio error is caught before the nfs_pgio_header args->context is set with get_nfs_open_context, the context can be NULL.

---
 fs/nfs/inode.c | 3 +++
 1 file changed, 3 insertions(+)
fs/nfs/inode.c +3 -0
index 141c9f4..a2b148e 100644
@@ -798,6 +798,9 @@ static void __put_nfs_open_context(struct
 nfs_open_context *ctx, int is_sync)
        struct inode *inode = ctx->dentry->d_inode;
        struct super_block *sb = ctx->dentry->d_sb;
 
+       if (ctx == NULL)
+               return;
+
        if (!list_empty(&ctx->list)) {
                if (!atomic_dec_and_lock(&ctx->lock_context.count,

&inode->i_lock))

Quoted 4 lines
                        return;

--
1.8.3.1

AFAICS, there is only one callsite where we need to check the context, so we can avoid the check on all calls to put_nfs_open_context. Please see PATCH 1/2 that I just posted in response to an off-list bug report from Steve Dickson.

Cheers
Trond
--

Trond Myklebust

Linux NFS client maintainer, PrimaryData

[email protected]

Signature · 5 lines
Trond Myklebust

Linux NFS client maintainer, PrimaryData

[email protected]
#4 lore raw
Quoted 43 lines · Trond Myklebust

Resending to list... Apologies for the duplicate...

Quoted 39 lines · Trond Myklebust

Hi Andy,

Quoted 26 lines · On Mon, Oct 13, 2014 at 10:08 AM, <[email protected]> wrote:

From: Andy Adamson <[email protected]>

If a pgio error is caught before the nfs_pgio_header args->context is set with get_nfs_open_context, the context can be NULL.

---
fs/nfs/inode.c | 3 +++

1 file changed, 3 insertions(+)
fs/nfs/inode.c +3 -0
index 141c9f4..a2b148e 100644
@@ -798,6 +798,9 @@ static void __put_nfs_open_context(struct
 nfs_open_context *ctx, int is_sync)
       struct inode *inode = ctx->dentry->d_inode;
       struct super_block *sb = ctx->dentry->d_sb;
 
+       if (ctx == NULL)
+               return;
+
       if (!list_empty(&ctx->list)) {
               if (!atomic_dec_and_lock(&ctx->lock_context.count,

&inode->i_lock))

Quoted 4 lines
                       return;

--
1.8.3.1

AFAICS, there is only one callsite where we need to check the context, so we can avoid the check on all calls to put_nfs_open_context. Please see PATCH 1/2 that I just posted in response to an off-list bug report from Steve Dickson.

Thanks Trond,

Both patches should fix the OOPS, but I think your reasoning to avoid the check in all calls to put_nfs_open_context makes sense.

-dros

Quoted 23 lines
Quoted 10 lines

Cheers
Trond
--

Trond Myklebust

Linux NFS client maintainer, PrimaryData

[email protected]

Signature · 9 lines
Trond Myklebust

Linux NFS client maintainer, PrimaryData

[email protected]
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html