[PATCH] vfs: require i_size <= SIZE_MAX in kernel_read_file()

[PATCH] vfs: require i_size <= SIZE_MAX in kernel_read_file()

[Eric Biggers]

From: Eric Biggers <ebiggers@google.com>

On 32-bit systems, the buffer allocated by kernel_read_file() is too
small if the file size is > SIZE_MAX, due to truncation to size_t.

Fortunately, since the 'count' argument to kernel_read() is also
truncated to size_t, only the allocated space is filled; then, -EIO is
returned since 'pos != i_size' after the read loop.

But this is not obvious and seems incidental.  We should be more
explicit about this case.  So, fail early if i_size > SIZE_MAX.

Signed-off-by: Eric Biggers <ebiggers@google.com>
---
 fs/exec.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/fs/exec.c b/fs/exec.c
index 1ebf6e5a521d..fc281b738a98 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -908,14 +908,14 @@ int kernel_read_file(struct file *file, void **buf, loff_t *size,
 		goto out;
 
 	i_size = i_size_read(file_inode(file));
-	if (max_size > 0 && i_size > max_size) {
-		ret = -EFBIG;
-		goto out;
-	}
 	if (i_size <= 0) {
 		ret = -EINVAL;
 		goto out;
 	}
+	if (i_size > SIZE_MAX || (max_size > 0 && i_size > max_size)) {
+		ret = -EFBIG;
+		goto out;
+	}
 
 	if (id != READING_FIRMWARE_PREALLOC_BUFFER)
 		*buf = vmalloc(i_size);
-- 
2.19.0.rc2.392.g5ba43deb5a-goog

Re: [PATCH] vfs: require i_size <= SIZE_MAX in kernel_read_file()

[Mimi Zohar]

On Fri, 2018-09-07 at 12:16 -0700, Eric Biggers wrote:
> From: Eric Biggers <ebiggers@google.com>
> 
> On 32-bit systems, the buffer allocated by kernel_read_file() is too
> small if the file size is > SIZE_MAX, due to truncation to size_t.
> 
> Fortunately, since the 'count' argument to kernel_read() is also
> truncated to size_t, only the allocated space is filled; then, -EIO is
> returned since 'pos != i_size' after the read loop.
> 
> But this is not obvious and seems incidental.  We should be more
> explicit about this case.  So, fail early if i_size > SIZE_MAX.
> 
> Signed-off-by: Eric Biggers <ebiggers@google.com>

Thanks!

> ---
>  fs/exec.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/exec.c b/fs/exec.c
> index 1ebf6e5a521d..fc281b738a98 100644
> --- a/fs/exec.c
> +++ b/fs/exec.c
> @@ -908,14 +908,14 @@ int kernel_read_file(struct file *file, void **buf, loff_t *size,
>  		goto out;
> 
>  	i_size = i_size_read(file_inode(file));
> -	if (max_size > 0 && i_size > max_size) {
> -		ret = -EFBIG;
> -		goto out;
> -	}
>  	if (i_size <= 0) {
>  		ret = -EINVAL;
>  		goto out;
>  	}
> +	if (i_size > SIZE_MAX || (max_size > 0 && i_size > max_size)) {
> +		ret = -EFBIG;
> +		goto out;
> +	}
> 
>  	if (id != READING_FIRMWARE_PREALLOC_BUFFER)
>  		*buf = vmalloc(i_size);

Re: [PATCH] vfs: require i_size <= SIZE_MAX in kernel_read_file()

[Eric Biggers]

On Fri, Sep 07, 2018 at 12:16:24PM -0700, Eric Biggers wrote:
> From: Eric Biggers <ebiggers@google.com>
> 
> On 32-bit systems, the buffer allocated by kernel_read_file() is too
> small if the file size is > SIZE_MAX, due to truncation to size_t.
> 
> Fortunately, since the 'count' argument to kernel_read() is also
> truncated to size_t, only the allocated space is filled; then, -EIO is
> returned since 'pos != i_size' after the read loop.
> 
> But this is not obvious and seems incidental.  We should be more
> explicit about this case.  So, fail early if i_size > SIZE_MAX.
> 
> Signed-off-by: Eric Biggers <ebiggers@google.com>
> ---
>  fs/exec.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/exec.c b/fs/exec.c
> index 1ebf6e5a521d..fc281b738a98 100644
> --- a/fs/exec.c
> +++ b/fs/exec.c
> @@ -908,14 +908,14 @@ int kernel_read_file(struct file *file, void **buf, loff_t *size,
>  		goto out;
>  
>  	i_size = i_size_read(file_inode(file));
> -	if (max_size > 0 && i_size > max_size) {
> -		ret = -EFBIG;
> -		goto out;
> -	}
>  	if (i_size <= 0) {
>  		ret = -EINVAL;
>  		goto out;
>  	}
> +	if (i_size > SIZE_MAX || (max_size > 0 && i_size > max_size)) {
> +		ret = -EFBIG;
> +		goto out;
> +	}
>  
>  	if (id != READING_FIRMWARE_PREALLOC_BUFFER)
>  		*buf = vmalloc(i_size);
> -- 

Al, are you planning to apply this?

- Eric