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

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

[Shuah Khan]

While debugging an unrelated problem that got me down the path of
reviewing at all the fdget() and fdput() paths in the kernel.

While doing the review, I noticed these two places where fdput()
is called after failed fdget(). Fixing them in these two patches.
  
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 | 5 +++--
 2 files changed, 4 insertions(+), 3 deletions(-)

-- 
2.20.1

[PATCH 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. Fix it anyway since failed fdget() doesn't require
a fdput().

This was introdcued in a commit to add sync_file_range() helper.

Fixes: 22f96b3808c1 ("fs: add sync_file_range() helper")
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
---
 fs/sync.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/fs/sync.c b/fs/sync.c
index 4d1ff010bc5a..faaff835ef12 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -369,10 +369,11 @@ int ksys_sync_file_range(int fd, loff_t offset, loff_t nbytes,
 
 	ret = -EBADF;
 	f = fdget(fd);
-	if (f.file)
+	if (f.file) {
 		ret = sync_file_range(f.file, offset, nbytes, flags);
+		fdput(f);
+	}
 
-	fdput(f);
 	return ret;
 }
 
-- 
2.20.1

[PATCH 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);
-- 
2.20.1

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

[Al Viro]

On Thu, May 07, 2020 at 05:57:09PM -0600, Shuah Khan wrote:
> 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. Fix it anyway since failed fdget() doesn't require
> a fdput().
> 
> This was introdcued in a commit to add sync_file_range() helper.

Er...  What's the point microoptimizing the slow path here?

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

[Al Viro]

On Fri, May 08, 2020 at 01:05:09AM +0100, Al Viro wrote:
> On Thu, May 07, 2020 at 05:57:09PM -0600, Shuah Khan wrote:
> > 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. Fix it anyway since failed fdget() doesn't require
> > a fdput().
> > 
> > This was introdcued in a commit to add sync_file_range() helper.
> 
> Er...  What's the point microoptimizing the slow path here?

PS: I'm not saying the patch is incorrect, but Fixes: is IMO over the
top.  And looking at that thing,
{
        struct fd f = fdget(fd);
        int ret;

	if (unlikely(!f.file))
		return -EBADF;

	ret = sync_file_range(f.file, offset, nbytes, flags);
        fdput(f);
        return ret;
}

might be cleaner, but that's a matter of taste...

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

[Luis Chamberlain]

On Fri, May 08, 2020 at 01:24:22AM +0100, Al Viro wrote:
> On Fri, May 08, 2020 at 01:05:09AM +0100, Al Viro wrote:
> > On Thu, May 07, 2020 at 05:57:09PM -0600, Shuah Khan wrote:
> > > 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. Fix it anyway since failed fdget() doesn't require
> > > a fdput().
> > > 
> > > This was introdcued in a commit to add sync_file_range() helper.
> > 
> > Er...  What's the point microoptimizing the slow path here?
> 
> PS: I'm not saying the patch is incorrect, but Fixes: is IMO over the
> top.  And looking at that thing,
> {
>         struct fd f = fdget(fd);
>         int ret;
> 
> 	if (unlikely(!f.file))
> 		return -EBADF;
> 
> 	ret = sync_file_range(f.file, offset, nbytes, flags);
>         fdput(f);
>         return ret;
> }
> 
> might be cleaner, but that's a matter of taste...

This makes it easier to read.

  Luis

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

[Shuah Khan]

On 5/7/20 8:21 PM, Luis Chamberlain wrote:
> On Fri, May 08, 2020 at 01:24:22AM +0100, Al Viro wrote:
>> On Fri, May 08, 2020 at 01:05:09AM +0100, Al Viro wrote:
>>> On Thu, May 07, 2020 at 05:57:09PM -0600, Shuah Khan wrote:
>>>> 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. Fix it anyway since failed fdget() doesn't require
>>>> a fdput().
>>>>
>>>> This was introdcued in a commit to add sync_file_range() helper.
>>>
>>> Er...  What's the point microoptimizing the slow path here?
>>
>> PS: I'm not saying the patch is incorrect, but Fixes: is IMO over the
>> top.  And looking at that thing,
>> {
>>          struct fd f = fdget(fd);
>>          int ret;
>>
>> 	if (unlikely(!f.file))
>> 		return -EBADF;
>>
>> 	ret = sync_file_range(f.file, offset, nbytes, flags);
>>          fdput(f);
>>          return ret;
>> }
>>
>> might be cleaner, but that's a matter of taste...
> 
> This makes it easier to read.
> 

Yes it does. I will make the changes and send v2.

thanks,
-- Shuah