diff mbox series

[v2] ima: Fix NULL pointer dereference in ima_file_hash

Message ID 20200916124931.1254990-1-kpsingh@chromium.org (mailing list archive)
State New, archived
Headers show
Series [v2] ima: Fix NULL pointer dereference in ima_file_hash | expand

Commit Message

KP Singh Sept. 16, 2020, 12:49 p.m. UTC
From: KP Singh <kpsingh@google.com>

ima_file_hash can be called when there is no iint->ima_hash available
even though the inode exists in the integrity cache.

An example where this can happen (suggested by Jann Horn):

Process A does:

	while(1) {
		unlink("/tmp/imafoo");
		fd = open("/tmp/imafoo", O_RDWR|O_CREAT|O_TRUNC, 0700);
		if (fd == -1) {
			perror("open");
			continue;
		}
		write(fd, "A", 1);
		close(fd);
	}

and Process B does:

	while (1) {
		int fd = open("/tmp/imafoo", O_RDONLY);
		if (fd == -1)
			continue;
    		char *mapping = mmap(NULL, 0x1000, PROT_READ|PROT_EXEC,
			 	     MAP_PRIVATE, fd, 0);
		if (mapping != MAP_FAILED)
			munmap(mapping, 0x1000);
		close(fd);
  	}

Due to the race to get the iint->mutex between ima_file_hash and
process_measurement iint->ima_hash could still be NULL.

Fixes: 6beea7afcc72 ("ima: add the ability to query the cached hash of a given file")
Signed-off-by: KP Singh <kpsingh@google.com>
Reviewed-by: Florent Revest <revest@chromium.org>
---
 security/integrity/ima/ima_main.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

Comments

KP Singh Sept. 16, 2020, 1:36 p.m. UTC | #1
On Wed, Sep 16, 2020 at 2:49 PM KP Singh <kpsingh@chromium.org> wrote:
>
> From: KP Singh <kpsingh@google.com>

[...]

Another attempt to get this on the lists.

> diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c
> index 8a91711ca79b..4c86cd4eece0 100644
> --- a/security/integrity/ima/ima_main.c
> +++ b/security/integrity/ima/ima_main.c
> @@ -531,6 +531,16 @@ int ima_file_hash(struct file *file, char *buf, size_t buf_size)
>                 return -EOPNOTSUPP;
>
>         mutex_lock(&iint->mutex);
> +
> +       /*
> +        * ima_file_hash can be called when ima_collect_measurement has still
> +        * not been called, we might not always have a hash.
> +        */
> +       if (!iint->ima_hash) {
> +               mutex_unlock(&iint->mutex);
> +               return -EOPNOTSUPP;
> +       }
> +
>         if (buf) {
>                 size_t copied_size;
>
> --
> 2.28.0.526.ge36021eeef-goog
>
Mimi Zohar Sept. 16, 2020, 4 p.m. UTC | #2
On Wed, 2020-09-16 at 14:49 +0200, KP Singh wrote:
> From: KP Singh <kpsingh@google.com>
> 
> ima_file_hash can be called when there is no iint->ima_hash available
> even though the inode exists in the integrity cache.
> 
> An example where this can happen (suggested by Jann Horn):
> 
> Process A does:
> 
> 	while(1) {
> 		unlink("/tmp/imafoo");
> 		fd = open("/tmp/imafoo", O_RDWR|O_CREAT|O_TRUNC, 0700);
> 		if (fd == -1) {
> 			perror("open");
> 			continue;
> 		}
> 		write(fd, "A", 1);
> 		close(fd);
> 	}
> 
> and Process B does:
> 
> 	while (1) {
> 		int fd = open("/tmp/imafoo", O_RDONLY);
> 		if (fd == -1)
> 			continue;
>     		char *mapping = mmap(NULL, 0x1000, PROT_READ|PROT_EXEC,
> 			 	     MAP_PRIVATE, fd, 0);
> 		if (mapping != MAP_FAILED)
> 			munmap(mapping, 0x1000);
> 		close(fd);
>   	}
> 
> Due to the race to get the iint->mutex between ima_file_hash and
> process_measurement iint->ima_hash could still be NULL.
> 
> Fixes: 6beea7afcc72 ("ima: add the ability to query the cached hash of a given file")
> Signed-off-by: KP Singh <kpsingh@google.com>
> Reviewed-by: Florent Revest <revest@chromium.org>
> ---
>  security/integrity/ima/ima_main.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
> diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c
> index 8a91711ca79b..4c86cd4eece0 100644
> --- a/security/integrity/ima/ima_main.c
> +++ b/security/integrity/ima/ima_main.c
> @@ -531,6 +531,16 @@ int ima_file_hash(struct file *file, char *buf, size_t buf_size)
>  		return -EOPNOTSUPP;
>  
>  	mutex_lock(&iint->mutex);
> +
> +	/*
> +	 * ima_file_hash can be called when ima_collect_measurement has still
> +	 * not been called, we might not always have a hash.
> +	 */
> +	if (!iint->ima_hash) {
> +		mutex_unlock(&iint->mutex);
> +		return -EOPNOTSUPP;
> +	}
> +

Not having a file hash is rather common (e.g. mknodat, prior to the
file being closed).  Before appraising the integrity of a file, it
checks whether it is a new file (eg. IMA_NEW_FILE), but, unfortunately,
the flag is only set for those files in the appraise policy.

The patch looks fine, but you might want to reflect not having a file
hash is common in the patch description.

Mimi

>  	if (buf) {
>  		size_t copied_size;
>
KP Singh Sept. 16, 2020, 4:59 p.m. UTC | #3
On Wed, Sep 16, 2020 at 6:00 PM Mimi Zohar <zohar@linux.ibm.com> wrote:
>
> On Wed, 2020-09-16 at 14:49 +0200, KP Singh wrote:
> > From: KP Singh <kpsingh@google.com>
> >
> > ima_file_hash can be called when there is no iint->ima_hash available
> > even though the inode exists in the integrity cache.
> >
> > An example where this can happen (suggested by Jann Horn):
> >
> > Process A does:
> >
> >       while(1) {
> >               unlink("/tmp/imafoo");
> >               fd = open("/tmp/imafoo", O_RDWR|O_CREAT|O_TRUNC, 0700);
> >               if (fd == -1) {
> >                       perror("open");
> >                       continue;
> >               }
> >               write(fd, "A", 1);
> >               close(fd);
> >       }
> >
> > and Process B does:
> >
> >       while (1) {
> >               int fd = open("/tmp/imafoo", O_RDONLY);
> >               if (fd == -1)
> >                       continue;
> >               char *mapping = mmap(NULL, 0x1000, PROT_READ|PROT_EXEC,
> >                                    MAP_PRIVATE, fd, 0);
> >               if (mapping != MAP_FAILED)
> >                       munmap(mapping, 0x1000);
> >               close(fd);
> >       }
> >
> > Due to the race to get the iint->mutex between ima_file_hash and
> > process_measurement iint->ima_hash could still be NULL.
> >
> > Fixes: 6beea7afcc72 ("ima: add the ability to query the cached hash of a given file")
> > Signed-off-by: KP Singh <kpsingh@google.com>
> > Reviewed-by: Florent Revest <revest@chromium.org>
> > ---
> >  security/integrity/ima/ima_main.c | 10 ++++++++++
> >  1 file changed, 10 insertions(+)
> >
> > diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c
> > index 8a91711ca79b..4c86cd4eece0 100644
> > --- a/security/integrity/ima/ima_main.c
> > +++ b/security/integrity/ima/ima_main.c
> > @@ -531,6 +531,16 @@ int ima_file_hash(struct file *file, char *buf, size_t buf_size)
> >               return -EOPNOTSUPP;
> >
> >       mutex_lock(&iint->mutex);
> > +
> > +     /*
> > +      * ima_file_hash can be called when ima_collect_measurement has still
> > +      * not been called, we might not always have a hash.
> > +      */
> > +     if (!iint->ima_hash) {
> > +             mutex_unlock(&iint->mutex);
> > +             return -EOPNOTSUPP;
> > +     }
> > +
>
> Not having a file hash is rather common (e.g. mknodat, prior to the
> file being closed).  Before appraising the integrity of a file, it
> checks whether it is a new file (eg. IMA_NEW_FILE), but, unfortunately,
> the flag is only set for those files in the appraise policy.

Makes sense.

>
> The patch looks fine, but you might want to reflect not having a file
> hash is common in the patch description.
>

Thanks! Will send another revision with an updated description.

- KP

> Mimi
>
> >       if (buf) {
> >               size_t copied_size;
> >
>
>
diff mbox series

Patch

diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c
index 8a91711ca79b..4c86cd4eece0 100644
--- a/security/integrity/ima/ima_main.c
+++ b/security/integrity/ima/ima_main.c
@@ -531,6 +531,16 @@  int ima_file_hash(struct file *file, char *buf, size_t buf_size)
 		return -EOPNOTSUPP;
 
 	mutex_lock(&iint->mutex);
+
+	/*
+	 * ima_file_hash can be called when ima_collect_measurement has still
+	 * not been called, we might not always have a hash.
+	 */
+	if (!iint->ima_hash) {
+		mutex_unlock(&iint->mutex);
+		return -EOPNOTSUPP;
+	}
+
 	if (buf) {
 		size_t copied_size;