diff mbox series

[v3,15/19] IMA: Add support for file reads without contents

Message ID 20200724213640.389191-16-keescook@chromium.org (mailing list archive)
State Superseded
Headers show
Series Introduce partial kernel_read_file() support | expand

Commit Message

Kees Cook July 24, 2020, 9:36 p.m. UTC 
From: Scott Branden <scott.branden@broadcom.com>

When the kernel_read_file LSM hook is called with contents=false, IMA
can appraise the file directly, without requiring a filled buffer. When
such a buffer is available, though, IMA can continue to use it instead
of forcing a double read here.

Signed-off-by: Scott Branden <scott.branden@broadcom.com>
Link: https://lore.kernel.org/lkml/20200706232309.12010-10-scott.branden@broadcom.com/
Signed-off-by: Kees Cook <keescook@chromium.org>
---
 security/integrity/ima/ima_main.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)

Comments

Mimi Zohar July 27, 2020, 1:23 p.m. UTC | #1 
On Fri, 2020-07-24 at 14:36 -0700, Kees Cook wrote:
> From: Scott Branden <scott.branden@broadcom.com>
> 
> When the kernel_read_file LSM hook is called with contents=false, IMA
> can appraise the file directly, without requiring a filled buffer. When
> such a buffer is available, though, IMA can continue to use it instead
> of forcing a double read here.
> 
> Signed-off-by: Scott Branden <scott.branden@broadcom.com>
> Link: https://lore.kernel.org/lkml/20200706232309.12010-10-scott.branden@broadcom.com/
> Signed-off-by: Kees Cook <keescook@chromium.org>

After adjusting the comment below.

Reviewed-by: Mimi Zohar <zohar@linux.ibm.com>

> ---
>  security/integrity/ima/ima_main.c | 22 ++++++++++++++++------
>  1 file changed, 16 insertions(+), 6 deletions(-)
> 
> diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c
> index dc4f90660aa6..459e50526a12 100644
> --- a/security/integrity/ima/ima_main.c
> +++ b/security/integrity/ima/ima_main.c
> @@ -613,11 +613,8 @@ void ima_post_path_mknod(struct dentry *dentry)
>  int ima_read_file(struct file *file, enum kernel_read_file_id read_id,
>  		  bool contents)
>  {
> -	/* Reject all partial reads during appraisal. */
> -	if (!contents) {
> -		if (ima_appraise & IMA_APPRAISE_ENFORCE)
> -			return -EACCES;
> -	}
> +	enum ima_hooks func;
> +	u32 secid;
>  
>  	/*
>  	 * Do devices using pre-allocated memory run the risk of the
> @@ -626,7 +623,20 @@ int ima_read_file(struct file *file, enum kernel_read_file_id read_id,
>  	 * buffers? It may be desirable to include the buffer address
>  	 * in this API and walk all the dma_map_single() mappings to check.
>  	 */
> -	return 0;
> +
> +	/*
> +	 * There will be a call made to ima_post_read_file() with
> +	 * a filled buffer, so we don't need to perform an extra
> +	 * read early here.
> +	 */
> +	if (contents)
> +		return 0;
> +
> +	/* Read entire file for all partial reads during appraisal. */

In addition to verifying the file signature, the file might be
included in the IMA measurement list or the file hash may be used to
augment the audit record.  Please remove "during appraisal" from the
comment.

> +	func = read_idmap[read_id] ?: FILE_CHECK;
> +	security_task_getsecid(current, &secid);
> +	return process_measurement(file, current_cred(), secid, NULL,
> +				   0, MAY_READ, func);
>  }
>  
>  const int read_idmap[READING_MAX_ID] = {
Kees Cook July 28, 2020, 7:44 p.m. UTC | #2 
On Mon, Jul 27, 2020 at 09:23:34AM -0400, Mimi Zohar wrote:
> On Fri, 2020-07-24 at 14:36 -0700, Kees Cook wrote:
> > From: Scott Branden <scott.branden@broadcom.com>
> > 
> > When the kernel_read_file LSM hook is called with contents=false, IMA
> > can appraise the file directly, without requiring a filled buffer. When
> > such a buffer is available, though, IMA can continue to use it instead
> > of forcing a double read here.
> > 
> > Signed-off-by: Scott Branden <scott.branden@broadcom.com>
> > Link: https://lore.kernel.org/lkml/20200706232309.12010-10-scott.branden@broadcom.com/
> > Signed-off-by: Kees Cook <keescook@chromium.org>
> 
> After adjusting the comment below.
> 
> Reviewed-by: Mimi Zohar <zohar@linux.ibm.com>

Sure!

Greg, shall I send a v4 with added Reviews and the comment change or is
that minor enough that you're able to do it?

Thanks for the reviews Mimi!

-Kees

> 
> > ---
> >  security/integrity/ima/ima_main.c | 22 ++++++++++++++++------
> >  1 file changed, 16 insertions(+), 6 deletions(-)
> > 
> > diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c
> > index dc4f90660aa6..459e50526a12 100644
> > --- a/security/integrity/ima/ima_main.c
> > +++ b/security/integrity/ima/ima_main.c
> > @@ -613,11 +613,8 @@ void ima_post_path_mknod(struct dentry *dentry)
> >  int ima_read_file(struct file *file, enum kernel_read_file_id read_id,
> >  		  bool contents)
> >  {
> > -	/* Reject all partial reads during appraisal. */
> > -	if (!contents) {
> > -		if (ima_appraise & IMA_APPRAISE_ENFORCE)
> > -			return -EACCES;
> > -	}
> > +	enum ima_hooks func;
> > +	u32 secid;
> >  
> >  	/*
> >  	 * Do devices using pre-allocated memory run the risk of the
> > @@ -626,7 +623,20 @@ int ima_read_file(struct file *file, enum kernel_read_file_id read_id,
> >  	 * buffers? It may be desirable to include the buffer address
> >  	 * in this API and walk all the dma_map_single() mappings to check.
> >  	 */
> > -	return 0;
> > +
> > +	/*
> > +	 * There will be a call made to ima_post_read_file() with
> > +	 * a filled buffer, so we don't need to perform an extra
> > +	 * read early here.
> > +	 */
> > +	if (contents)
> > +		return 0;
> > +
> > +	/* Read entire file for all partial reads during appraisal. */
> 
> In addition to verifying the file signature, the file might be
> included in the IMA measurement list or the file hash may be used to
> augment the audit record.  Please remove "during appraisal" from the
> comment.
> 
> > +	func = read_idmap[read_id] ?: FILE_CHECK;
> > +	security_task_getsecid(current, &secid);
> > +	return process_measurement(file, current_cred(), secid, NULL,
> > +				   0, MAY_READ, func);
> >  }
> >  
> >  const int read_idmap[READING_MAX_ID] = {
>
Greg KH July 28, 2020, 7:56 p.m. UTC | #3 
On Tue, Jul 28, 2020 at 12:44:50PM -0700, Kees Cook wrote:
> On Mon, Jul 27, 2020 at 09:23:34AM -0400, Mimi Zohar wrote:
> > On Fri, 2020-07-24 at 14:36 -0700, Kees Cook wrote:
> > > From: Scott Branden <scott.branden@broadcom.com>
> > > 
> > > When the kernel_read_file LSM hook is called with contents=false, IMA
> > > can appraise the file directly, without requiring a filled buffer. When
> > > such a buffer is available, though, IMA can continue to use it instead
> > > of forcing a double read here.
> > > 
> > > Signed-off-by: Scott Branden <scott.branden@broadcom.com>
> > > Link: https://lore.kernel.org/lkml/20200706232309.12010-10-scott.branden@broadcom.com/
> > > Signed-off-by: Kees Cook <keescook@chromium.org>
> > 
> > After adjusting the comment below.
> > 
> > Reviewed-by: Mimi Zohar <zohar@linux.ibm.com>
> 
> Sure!
> 
> Greg, shall I send a v4 with added Reviews and the comment change or is
> that minor enough that you're able to do it?

v4 is needed, as this series is a mess of reviewes and you will have to
redo at least one patch and drop some others, right?

thanks,

greg k-h
Kees Cook July 28, 2020, 8:12 p.m. UTC | #4 
On Tue, Jul 28, 2020 at 09:56:40PM +0200, Greg Kroah-Hartman wrote:
> On Tue, Jul 28, 2020 at 12:44:50PM -0700, Kees Cook wrote:
> > On Mon, Jul 27, 2020 at 09:23:34AM -0400, Mimi Zohar wrote:
> > > On Fri, 2020-07-24 at 14:36 -0700, Kees Cook wrote:
> > > > From: Scott Branden <scott.branden@broadcom.com>
> > > > 
> > > > When the kernel_read_file LSM hook is called with contents=false, IMA
> > > > can appraise the file directly, without requiring a filled buffer. When
> > > > such a buffer is available, though, IMA can continue to use it instead
> > > > of forcing a double read here.
> > > > 
> > > > Signed-off-by: Scott Branden <scott.branden@broadcom.com>
> > > > Link: https://lore.kernel.org/lkml/20200706232309.12010-10-scott.branden@broadcom.com/
> > > > Signed-off-by: Kees Cook <keescook@chromium.org>
> > > 
> > > After adjusting the comment below.
> > > 
> > > Reviewed-by: Mimi Zohar <zohar@linux.ibm.com>
> > 
> > Sure!
> > 
> > Greg, shall I send a v4 with added Reviews and the comment change or is
> > that minor enough that you're able to do it?
> 
> v4 is needed, as this series is a mess of reviewes and you will have to
> redo at least one patch and drop some others, right?

Well, I wasn't sure what your desire was, given the weirdness of taking
some and reverting others. I will do a v4 based on driver-core-next.

Thanks!
diff mbox series

Patch

diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c
index dc4f90660aa6..459e50526a12 100644
--- a/security/integrity/ima/ima_main.c
+++ b/security/integrity/ima/ima_main.c
@@ -613,11 +613,8 @@  void ima_post_path_mknod(struct dentry *dentry)
 int ima_read_file(struct file *file, enum kernel_read_file_id read_id,
 		  bool contents)
 {
-	/* Reject all partial reads during appraisal. */
-	if (!contents) {
-		if (ima_appraise & IMA_APPRAISE_ENFORCE)
-			return -EACCES;
-	}
+	enum ima_hooks func;
+	u32 secid;
 
 	/*
 	 * Do devices using pre-allocated memory run the risk of the
@@ -626,7 +623,20 @@  int ima_read_file(struct file *file, enum kernel_read_file_id read_id,
 	 * buffers? It may be desirable to include the buffer address
 	 * in this API and walk all the dma_map_single() mappings to check.
 	 */
-	return 0;
+
+	/*
+	 * There will be a call made to ima_post_read_file() with
+	 * a filled buffer, so we don't need to perform an extra
+	 * read early here.
+	 */
+	if (contents)
+		return 0;
+
+	/* Read entire file for all partial reads during appraisal. */
+	func = read_idmap[read_id] ?: FILE_CHECK;
+	security_task_getsecid(current, &secid);
+	return process_measurement(file, current_cred(), secid, NULL,
+				   0, MAY_READ, func);
 }
 
 const int read_idmap[READING_MAX_ID] = {