Patchwork fs/seq_file: Fix warning of passing zero to 'PTR_ERR'

login
register
mail settings
Submitter Gomonovych, Vasyl
Date Dec. 7, 2017, 11:03 p.m.
Message ID <1512687788-27172-1-git-send-email-gomonovych@gmail.com>
Download mbox | patch
Permalink /patch/10101233/
State New
Headers show

Comments

Gomonovych, Vasyl - Dec. 7, 2017, 11:03 p.m.
p could be NULL and passing into PTR_ERR

Signed-off-by: Vasyl Gomonovych <gomonovych@gmail.com>
---
 fs/seq_file.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
willy@infradead.org - Dec. 7, 2017, 11:20 p.m.
On Fri, Dec 08, 2017 at 12:03:07AM +0100, Vasyl Gomonovych wrote:
> p could be NULL and passing into PTR_ERR

What makes you think this is correct?

To quote the documentation:

The next function to implement is called, amazingly, next(); its job is to
move the iterator forward to the next position in the sequence.  The
example module can simply increment the position by one; more useful
modules will do what is needed to step through some data structure. The
next() function returns a new iterator, or NULL if the sequence is
complete. Here's the example version:

So if it returns NULL, we want to set err to 0 and break.  Which is, um,
exactly what the code does.

Did you test this at all?

> Signed-off-by: Vasyl Gomonovych <gomonovych@gmail.com>
> ---
>  fs/seq_file.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/seq_file.c b/fs/seq_file.c
> index 4be761c..8b700b9 100644
> --- a/fs/seq_file.c
> +++ b/fs/seq_file.c
> @@ -262,8 +262,8 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
>  		size_t offs = m->count;
>  		loff_t next = pos;
>  		p = m->op->next(m, p, &next);
> -		if (!p || IS_ERR(p)) {
> -			err = PTR_ERR(p);
> +		if (IS_ERR(p)) {
> +			err = (!p ? -EFAULT : PTR_ERR(p));
>  			break;
>  		}
>  		err = m->op->show(m, p);
> -- 
> 1.9.1
>
Al Viro - Dec. 7, 2017, 11:23 p.m.
On Fri, Dec 08, 2017 at 12:03:07AM +0100, Vasyl Gomonovych wrote:
> p could be NULL and passing into PTR_ERR
> 
> Signed-off-by: Vasyl Gomonovych <gomonovych@gmail.com>
> ---
>  fs/seq_file.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/seq_file.c b/fs/seq_file.c
> index 4be761c..8b700b9 100644
> --- a/fs/seq_file.c
> +++ b/fs/seq_file.c
> @@ -262,8 +262,8 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
>  		size_t offs = m->count;
>  		loff_t next = pos;
>  		p = m->op->next(m, p, &next);
> -		if (!p || IS_ERR(p)) {
> -			err = PTR_ERR(p);
> +		if (IS_ERR(p)) {
> +			err = (!p ? -EFAULT : PTR_ERR(p));

What does it fix, if I might ask?  And while we are at it, would
you mind explaining the reasoning behind that change?  Or, say,
testing done to it...
Al Viro - Dec. 7, 2017, 11:26 p.m.
On Thu, Dec 07, 2017 at 11:23:26PM +0000, Al Viro wrote:
> On Fri, Dec 08, 2017 at 12:03:07AM +0100, Vasyl Gomonovych wrote:
> > p could be NULL and passing into PTR_ERR
> > 
> > Signed-off-by: Vasyl Gomonovych <gomonovych@gmail.com>
> > ---
> >  fs/seq_file.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/fs/seq_file.c b/fs/seq_file.c
> > index 4be761c..8b700b9 100644
> > --- a/fs/seq_file.c
> > +++ b/fs/seq_file.c
> > @@ -262,8 +262,8 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
> >  		size_t offs = m->count;
> >  		loff_t next = pos;
> >  		p = m->op->next(m, p, &next);
> > -		if (!p || IS_ERR(p)) {
> > -			err = PTR_ERR(p);
> > +		if (IS_ERR(p)) {
> > +			err = (!p ? -EFAULT : PTR_ERR(p));
> 
> What does it fix, if I might ask?  And while we are at it, would
> you mind explaining the reasoning behind that change?  Or, say,
> testing done to it...

While we are at it, where has that -EFAULT come from?  And how
would it be ever reached, seeing that IS_ERR(NULL) is false?
Gomonovych, Vasyl - Dec. 8, 2017, 7:54 a.m.
Hi,

Guys sorry for this idiotic piece of code.
Yesterday after doc seq_file.txt read I did not catch real way of work there.
And made this shit.
Sorry.

Regards Vasyl

On Fri, Dec 8, 2017 at 12:26 AM, Al Viro <viro@zeniv.linux.org.uk> wrote:
> On Thu, Dec 07, 2017 at 11:23:26PM +0000, Al Viro wrote:
>> On Fri, Dec 08, 2017 at 12:03:07AM +0100, Vasyl Gomonovych wrote:
>> > p could be NULL and passing into PTR_ERR
>> >
>> > Signed-off-by: Vasyl Gomonovych <gomonovych@gmail.com>
>> > ---
>> >  fs/seq_file.c | 4 ++--
>> >  1 file changed, 2 insertions(+), 2 deletions(-)
>> >
>> > diff --git a/fs/seq_file.c b/fs/seq_file.c
>> > index 4be761c..8b700b9 100644
>> > --- a/fs/seq_file.c
>> > +++ b/fs/seq_file.c
>> > @@ -262,8 +262,8 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
>> >             size_t offs = m->count;
>> >             loff_t next = pos;
>> >             p = m->op->next(m, p, &next);
>> > -           if (!p || IS_ERR(p)) {
>> > -                   err = PTR_ERR(p);
>> > +           if (IS_ERR(p)) {
>> > +                   err = (!p ? -EFAULT : PTR_ERR(p));
>>
>> What does it fix, if I might ask?  And while we are at it, would
>> you mind explaining the reasoning behind that change?  Or, say,
>> testing done to it...
>
> While we are at it, where has that -EFAULT come from?  And how
> would it be ever reached, seeing that IS_ERR(NULL) is false?

Patch

diff --git a/fs/seq_file.c b/fs/seq_file.c
index 4be761c..8b700b9 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -262,8 +262,8 @@  ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
 		size_t offs = m->count;
 		loff_t next = pos;
 		p = m->op->next(m, p, &next);
-		if (!p || IS_ERR(p)) {
-			err = PTR_ERR(p);
+		if (IS_ERR(p)) {
+			err = (!p ? -EFAULT : PTR_ERR(p));
 			break;
 		}
 		err = m->op->show(m, p);