mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] proc: potential ERR_PTR dereference on proc_fill_cache()
@ 2014-08-11  2:47 Daeseok Youn
  2014-08-11  3:25 ` Eric W. Biederman
  0 siblings, 1 reply; 3+ messages in thread
From: Daeseok Youn @ 2014-08-11  2:47 UTC (permalink / raw)
  To: akpm
  Cc: oleg, ebiederm, eparis, rgb, tixxdz, adobriyan, bill.c.roberts,
	linux-kernel

The d_hash_and_lookup() returns NULL or ERR_PTR on lookup
failure.

Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
---
 fs/proc/base.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/fs/proc/base.c b/fs/proc/base.c
index f50d4be..49bc09c 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -1673,7 +1673,7 @@ bool proc_fill_cache(struct file *file, struct dir_context *ctx,
 	ino_t ino;
 
 	child = d_hash_and_lookup(dir, &qname);
-	if (!child) {
+	if (IS_ERR_OR_NULL(child)) {
 		child = d_alloc(dir, &qname);
 		if (!child)
 			goto end_instantiate;
-- 
1.7.1


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] proc: potential ERR_PTR dereference on proc_fill_cache()
  2014-08-11  2:47 [PATCH] proc: potential ERR_PTR dereference on proc_fill_cache() Daeseok Youn
@ 2014-08-11  3:25 ` Eric W. Biederman
  2014-08-11  5:43   ` DaeSeok Youn
  0 siblings, 1 reply; 3+ messages in thread
From: Eric W. Biederman @ 2014-08-11  3:25 UTC (permalink / raw)
  To: Daeseok Youn
  Cc: akpm, oleg, eparis, rgb, tixxdz, adobriyan, bill.c.roberts, linux-kernel

Daeseok Youn <daeseok.youn@gmail.com> writes:

> The d_hash_and_lookup() returns NULL or ERR_PTR on lookup
> failure.

Given that the only error condition can come from .d_hash and that proc
does not implement .d_hash if we want to do something with error returns
let's call BUG_ON. aka:

diff --git a/fs/proc/base.c b/fs/proc/base.c
index f50d4be..49bc09c 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -1673,7 +1673,7 @@ bool proc_fill_cache(struct file *file, struct dir_context *ctx,
 	ino_t ino;
 
 	child = d_hash_and_lookup(dir, &qname);
+	BUG_ON(IS_ERR(child))
	if (!child) {
 		child = d_alloc(dir, &qname);
 		if (!child)
 			goto end_instantiate;

Or do you see something in d_hash_and_lookup that I am overlooking?

Eric


> Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
> ---
>  fs/proc/base.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/fs/proc/base.c b/fs/proc/base.c
> index f50d4be..49bc09c 100644
> --- a/fs/proc/base.c
> +++ b/fs/proc/base.c
> @@ -1673,7 +1673,7 @@ bool proc_fill_cache(struct file *file, struct dir_context *ctx,
>  	ino_t ino;
>  
>  	child = d_hash_and_lookup(dir, &qname);
> -	if (!child) {
> +	if (IS_ERR_OR_NULL(child)) {
>  		child = d_alloc(dir, &qname);
>  		if (!child)
>  			goto end_instantiate;

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] proc: potential ERR_PTR dereference on proc_fill_cache()
  2014-08-11  3:25 ` Eric W. Biederman
@ 2014-08-11  5:43   ` DaeSeok Youn
  0 siblings, 0 replies; 3+ messages in thread
From: DaeSeok Youn @ 2014-08-11  5:43 UTC (permalink / raw)
  To: Eric W. Biederman
  Cc: Andrew Morton, oleg, eparis, rgb, tixxdz, adobriyan,
	bill.c.roberts, linux-kernel

2014-08-11 12:25 GMT+09:00 Eric W. Biederman <ebiederm@xmission.com>:
> Daeseok Youn <daeseok.youn@gmail.com> writes:
>
>> The d_hash_and_lookup() returns NULL or ERR_PTR on lookup
>> failure.
>
> Given that the only error condition can come from .d_hash and that proc
> does not implement .d_hash if we want to do something with error returns
> let's call BUG_ON. aka:
Yes, you're right. proc doesn't implement .d_hash, so my patch is not needed.
It couldn't return ERR_PTR(-error) without implementing .d_hash
This patch should be dropped.

>
> diff --git a/fs/proc/base.c b/fs/proc/base.c
> index f50d4be..49bc09c 100644
> --- a/fs/proc/base.c
> +++ b/fs/proc/base.c
> @@ -1673,7 +1673,7 @@ bool proc_fill_cache(struct file *file, struct dir_context *ctx,
>         ino_t ino;
>
>         child = d_hash_and_lookup(dir, &qname);
> +       BUG_ON(IS_ERR(child))
>         if (!child) {
>                 child = d_alloc(dir, &qname);
>                 if (!child)
>                         goto end_instantiate;
>
> Or do you see something in d_hash_and_lookup that I am overlooking?
No.. I should check more carefully. sorry for that.

Thanks for review.

regards,
Daeseok Youn.

>
> Eric
>
>
>> Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
>> ---
>>  fs/proc/base.c |    2 +-
>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/fs/proc/base.c b/fs/proc/base.c
>> index f50d4be..49bc09c 100644
>> --- a/fs/proc/base.c
>> +++ b/fs/proc/base.c
>> @@ -1673,7 +1673,7 @@ bool proc_fill_cache(struct file *file, struct dir_context *ctx,
>>       ino_t ino;
>>
>>       child = d_hash_and_lookup(dir, &qname);
>> -     if (!child) {
>> +     if (IS_ERR_OR_NULL(child)) {
>>               child = d_alloc(dir, &qname);
>>               if (!child)
>>                       goto end_instantiate;

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2014-08-11  5:49 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-08-11  2:47 [PATCH] proc: potential ERR_PTR dereference on proc_fill_cache() Daeseok Youn
2014-08-11  3:25 ` Eric W. Biederman
2014-08-11  5:43   ` DaeSeok Youn

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®