[PATCH v2 0/2] fs: avoid fdput() after failed fdget()

[PATCH v2 0/2] fs: avoid fdput() after failed fdget()

[Shuah Khan]

While debugging an unrelated problem, I noticed these two cases fdput()
is called after failed fdget() while reviewing at all the fdget() and
fdput() paths in the kernel.

Changes since v1:
Patch 1:
  Changed to address review comments to refine the code for improved
  readability in addition to the change to avoid fdput() on failed
  fdget()
Patch 2:
  No change to v1. Including it in the series to keep the patches
  together.

Shuah Khan (2):
  fs: avoid fdput() after failed fdget() in ksys_sync_file_range()
  fs: avoid fdput() after failed fdget() in kernel_read_file_from_fd()

 fs/exec.c |  2 +-
 fs/sync.c | 12 ++++++------
 2 files changed, 7 insertions(+), 7 deletions(-)

-- 
2.25.1

[PATCH v2 1/2] fs: avoid fdput() after failed fdget() in ksys_sync_file_range()

[Shuah Khan]

Fix ksys_sync_file_range() to avoid fdput() after a failed fdget().
fdput() doesn't do fput() on this file since FDPUT_FPUT isn't set
in fd.flags.

Change it anyway since failed fdget() doesn't require a fdput(). Refine
the code path a bit for it to read more clearly.
Reference: commit 22f96b3808c1 ("fs: add sync_file_range() helper")

Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
---
 fs/sync.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/fs/sync.c b/fs/sync.c
index 4d1ff010bc5a..300ca73ec87c 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -364,15 +364,15 @@ int sync_file_range(struct file *file, loff_t offset, loff_t nbytes,
 int ksys_sync_file_range(int fd, loff_t offset, loff_t nbytes,
 			 unsigned int flags)
 {
-	int ret;
-	struct fd f;
+	int ret = -EBADF;
+	struct fd f = fdget(fd);
 
-	ret = -EBADF;
-	f = fdget(fd);
-	if (f.file)
-		ret = sync_file_range(f.file, offset, nbytes, flags);
+	if (!f.file)
+		goto out;
 
+	ret = sync_file_range(f.file, offset, nbytes, flags);
 	fdput(f);
+out:
 	return ret;
 }
 
-- 
2.25.1

[PATCH v2 2/2] fs: avoid fdput() after failed fdget() in kernel_read_file_from_fd()

[Shuah Khan]

Fix kernel_read_file_from_fd() to avoid fdput() after a failed fdget().
fdput() doesn't do fput() on this file since FDPUT_FPUT isn't set
in fd.flags. Fix it anyway since failed fdget() doesn't require
a fdput().

This was introduced in a commit that added kernel_read_file_from_fd() as
a wrapper for the VFS common kernel_read_file().

Fixes: b844f0ecbc56 ("vfs: define kernel_copy_file_from_fd()")
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
---
 fs/exec.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/exec.c b/fs/exec.c
index 06b4c550af5d..ea24bdce939d 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -1021,8 +1021,8 @@ int kernel_read_file_from_fd(int fd, void **buf, loff_t *size, loff_t max_size,
 		goto out;
 
 	ret = kernel_read_file(f.file, buf, size, max_size, id);
-out:
 	fdput(f);
+out:
 	return ret;
 }
 EXPORT_SYMBOL_GPL(kernel_read_file_from_fd);
-- 
2.25.1

Re: [PATCH v2 1/2] fs: avoid fdput() after failed fdget() in ksys_sync_file_range()

[Al Viro]

On Tue, May 12, 2020 at 01:43:04PM -0600, Shuah Khan wrote:

> @@ -364,15 +364,15 @@ int sync_file_range(struct file *file, loff_t offset, loff_t nbytes,
>  int ksys_sync_file_range(int fd, loff_t offset, loff_t nbytes,
>  			 unsigned int flags)
>  {
> -	int ret;
> -	struct fd f;
> +	int ret = -EBADF;
> +	struct fd f = fdget(fd);
>  
> -	ret = -EBADF;
> -	f = fdget(fd);
> -	if (f.file)
> -		ret = sync_file_range(f.file, offset, nbytes, flags);
> +	if (!f.file)
> +		goto out;
>  
> +	ret = sync_file_range(f.file, offset, nbytes, flags);
>  	fdput(f);
> +out:
>  	return ret;

IDGI...  What's the point of that goto out, when it leads straight to return?

Re: [PATCH v2 2/2] fs: avoid fdput() after failed fdget() in kernel_read_file_from_fd()

[Al Viro]

On Tue, May 12, 2020 at 01:43:05PM -0600, Shuah Khan wrote:
> Fix kernel_read_file_from_fd() to avoid fdput() after a failed fdget().
> fdput() doesn't do fput() on this file since FDPUT_FPUT isn't set
> in fd.flags. Fix it anyway since failed fdget() doesn't require
> a fdput().
> 
> This was introduced in a commit that added kernel_read_file_from_fd() as
> a wrapper for the VFS common kernel_read_file().
> 
> Fixes: b844f0ecbc56 ("vfs: define kernel_copy_file_from_fd()")
> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
> ---
>  fs/exec.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/exec.c b/fs/exec.c
> index 06b4c550af5d..ea24bdce939d 100644
> --- a/fs/exec.c
> +++ b/fs/exec.c
> @@ -1021,8 +1021,8 @@ int kernel_read_file_from_fd(int fd, void **buf, loff_t *size, loff_t max_size,
>  		goto out;
>  
>  	ret = kernel_read_file(f.file, buf, size, max_size, id);
> -out:
>  	fdput(f);
> +out:
>  	return ret;

Again, that goto is a pointless obfuscation; just return -EBADF
and be done with that.

Incidentally, why is that thing exported?

Re: [PATCH v2 2/2] fs: avoid fdput() after failed fdget() in kernel_read_file_from_fd()

[Luis Chamberlain]

On Wed, May 13, 2020 at 06:49:50AM +0100, Al Viro wrote:
> On Tue, May 12, 2020 at 01:43:05PM -0600, Shuah Khan wrote:
> > diff --git a/fs/exec.c b/fs/exec.c
> > index 06b4c550af5d..ea24bdce939d 100644
> > --- a/fs/exec.c
> > +++ b/fs/exec.c
> > @@ -1021,8 +1021,8 @@ int kernel_read_file_from_fd(int fd, void **buf, loff_t *size, loff_t max_size,
> >  		goto out;
> >  
> >  	ret = kernel_read_file(f.file, buf, size, max_size, id);
> > -out:
> >  	fdput(f);
> > +out:
> >  	return ret;
> 
> Incidentally, why is that thing exported?

Both kernel_read_file_from_fd() and kernel_read_file() are exported
because they have users, however kernel_read_file() only has security
stuff as a user. Do we want to get rid of the lsm hook for it?

I also have some non-posted patches which tucks away these kernel_read*()
exports under a symbol namespace, to avoid wide-spread use / abuse on
areas in the kernel, so I'd be happy to take this on if we want to
remove it export / lsm hook as part of my series. I did this as there
is another series of patches for a new driver which extend these family
of functions with a now pread() variant....

  Luis

Re: [PATCH v2 2/2] fs: avoid fdput() after failed fdget() in kernel_read_file_from_fd()

[Luis Chamberlain]

On Wed, May 13, 2020 at 7:13 AM Luis Chamberlain <mcgrof@kernel.org> wrote:
>
> On Wed, May 13, 2020 at 06:49:50AM +0100, Al Viro wrote:
> > On Tue, May 12, 2020 at 01:43:05PM -0600, Shuah Khan wrote:
> > > diff --git a/fs/exec.c b/fs/exec.c
> > > index 06b4c550af5d..ea24bdce939d 100644
> > > --- a/fs/exec.c
> > > +++ b/fs/exec.c
> > > @@ -1021,8 +1021,8 @@ int kernel_read_file_from_fd(int fd, void **buf, loff_t *size, loff_t max_size,
> > >             goto out;
> > >
> > >     ret = kernel_read_file(f.file, buf, size, max_size, id);
> > > -out:
> > >     fdput(f);
> > > +out:
> > >     return ret;
> >
> > Incidentally, why is that thing exported?
>
> Both kernel_read_file_from_fd() and kernel_read_file() are exported
> because they have users, however kernel_read_file() only has security
> stuff as a user. Do we want to get rid of the lsm hook for it?

Alright, yeah just the export needs to be removed. I have a patch
series dealing with these callers so will add it to my queue.

  Luis

Re: [PATCH v2 2/2] fs: avoid fdput() after failed fdget() in kernel_read_file_from_fd()

[Shuah Khan]

On 5/12/20 11:49 PM, Al Viro wrote:
> On Tue, May 12, 2020 at 01:43:05PM -0600, Shuah Khan wrote:
>> Fix kernel_read_file_from_fd() to avoid fdput() after a failed fdget().
>> fdput() doesn't do fput() on this file since FDPUT_FPUT isn't set
>> in fd.flags. Fix it anyway since failed fdget() doesn't require
>> a fdput().
>>
>> This was introduced in a commit that added kernel_read_file_from_fd() as
>> a wrapper for the VFS common kernel_read_file().
>>
>> Fixes: b844f0ecbc56 ("vfs: define kernel_copy_file_from_fd()")
>> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
>> ---
>>   fs/exec.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/fs/exec.c b/fs/exec.c
>> index 06b4c550af5d..ea24bdce939d 100644
>> --- a/fs/exec.c
>> +++ b/fs/exec.c
>> @@ -1021,8 +1021,8 @@ int kernel_read_file_from_fd(int fd, void **buf, loff_t *size, loff_t max_size,
>>   		goto out;
>>   
>>   	ret = kernel_read_file(f.file, buf, size, max_size, id);
>> -out:
>>   	fdput(f);
>> +out:
>>   	return ret;
> 
> Again, that goto is a pointless obfuscation; just return -EBADF
> and be done with that.
> 

My reasoning is if this routine ends up growing it might be useful
to handle the return this way.

In any case, I will send v3 for both of these patches.

thanks,
-- Shuah

Re: [PATCH v2 2/2] fs: avoid fdput() after failed fdget() in kernel_read_file_from_fd()

[Scott Branden]

Hi Luis,

On 2020-05-13 7:19 a.m., Luis Chamberlain wrote:
> On Wed, May 13, 2020 at 7:13 AM Luis Chamberlain <mcgrof@kernel.org> wrote:
>> On Wed, May 13, 2020 at 06:49:50AM +0100, Al Viro wrote:
>>> On Tue, May 12, 2020 at 01:43:05PM -0600, Shuah Khan wrote:
>>>> diff --git a/fs/exec.c b/fs/exec.c
>>>> index 06b4c550af5d..ea24bdce939d 100644
>>>> --- a/fs/exec.c
>>>> +++ b/fs/exec.c
>>>> @@ -1021,8 +1021,8 @@ int kernel_read_file_from_fd(int fd, void **buf, loff_t *size, loff_t max_size,
>>>>              goto out;
>>>>
>>>>      ret = kernel_read_file(f.file, buf, size, max_size, id);
>>>> -out:
>>>>      fdput(f);
>>>> +out:
>>>>      return ret;
>>> Incidentally, why is that thing exported?
>> Both kernel_read_file_from_fd() and kernel_read_file() are exported
>> because they have users, however kernel_read_file() only has security
>> stuff as a user. Do we want to get rid of the lsm hook for it?
> Alright, yeah just the export needs to be removed. I have a patch
> series dealing with these callers so will add it to my queue.
When will these changes make it into linux-next?
It is difficult for me to complete my patch series without these other 
misc. changes in place.
>
>    Luis
Regards,
  Scott

Re: [PATCH v2 2/2] fs: avoid fdput() after failed fdget() in kernel_read_file_from_fd()

[Scott Branden]

On 2020-05-22 2:59 p.m., Scott Branden wrote:
> Hi Luis,
>
> On 2020-05-13 7:19 a.m., Luis Chamberlain wrote:
>> On Wed, May 13, 2020 at 7:13 AM Luis Chamberlain <mcgrof@kernel.org> 
>> wrote:
>>> On Wed, May 13, 2020 at 06:49:50AM +0100, Al Viro wrote:
>>>> On Tue, May 12, 2020 at 01:43:05PM -0600, Shuah Khan wrote:
>>>>> diff --git a/fs/exec.c b/fs/exec.c
>>>>> index 06b4c550af5d..ea24bdce939d 100644
>>>>> --- a/fs/exec.c
>>>>> +++ b/fs/exec.c
>>>>> @@ -1021,8 +1021,8 @@ int kernel_read_file_from_fd(int fd, void 
>>>>> **buf, loff_t *size, loff_t max_size,
>>>>>              goto out;
>>>>>
>>>>>      ret = kernel_read_file(f.file, buf, size, max_size, id);
>>>>> -out:
>>>>>      fdput(f);
>>>>> +out:
>>>>>      return ret;
>>>> Incidentally, why is that thing exported?
>>> Both kernel_read_file_from_fd() and kernel_read_file() are exported
>>> because they have users, however kernel_read_file() only has security
>>> stuff as a user. Do we want to get rid of the lsm hook for it?
>> Alright, yeah just the export needs to be removed. I have a patch
>> series dealing with these callers so will add it to my queue.
> When will these changes make it into linux-next?
> It is difficult for me to complete my patch series without these other 
> misc. changes in place.
Sorry, I see the patch series is still being worked on (missing 
changelog, comments, etc).
Hopefully the patches stabilize so I can apply my changes on top fairly 
soon.
>>
>>    Luis
> Regards,
>  Scott

Re: [PATCH v2 2/2] fs: avoid fdput() after failed fdget() in kernel_read_file_from_fd()

[Luis Chamberlain]

On Fri, May 22, 2020 at 03:14:59PM -0700, Scott Branden wrote:
> 
> 
> On 2020-05-22 2:59 p.m., Scott Branden wrote:
> > Hi Luis,
> > 
> > On 2020-05-13 7:19 a.m., Luis Chamberlain wrote:
> > > On Wed, May 13, 2020 at 7:13 AM Luis Chamberlain <mcgrof@kernel.org>
> > > wrote:
> > > > On Wed, May 13, 2020 at 06:49:50AM +0100, Al Viro wrote:
> > > > > On Tue, May 12, 2020 at 01:43:05PM -0600, Shuah Khan wrote:
> > > > > > diff --git a/fs/exec.c b/fs/exec.c
> > > > > > index 06b4c550af5d..ea24bdce939d 100644
> > > > > > --- a/fs/exec.c
> > > > > > +++ b/fs/exec.c
> > > > > > @@ -1021,8 +1021,8 @@ int kernel_read_file_from_fd(int
> > > > > > fd, void **buf, loff_t *size, loff_t max_size,
> > > > > >              goto out;
> > > > > > 
> > > > > >      ret = kernel_read_file(f.file, buf, size, max_size, id);
> > > > > > -out:
> > > > > >      fdput(f);
> > > > > > +out:
> > > > > >      return ret;
> > > > > Incidentally, why is that thing exported?
> > > > Both kernel_read_file_from_fd() and kernel_read_file() are exported
> > > > because they have users, however kernel_read_file() only has security
> > > > stuff as a user. Do we want to get rid of the lsm hook for it?
> > > Alright, yeah just the export needs to be removed. I have a patch
> > > series dealing with these callers so will add it to my queue.
> > When will these changes make it into linux-next?
> > It is difficult for me to complete my patch series without these other
> > misc. changes in place.
> Sorry, I see the patch series is still being worked on (missing changelog,
> comments, etc).
> Hopefully the patches stabilize so I can apply my changes on top fairly
> soon.

Yeah I have to redo that series to take into account feedback. I'll be
sure cc you on that. I have a few other series to attend to before that,
so I think this will take a week.

  Luis