public inbox for linux-security-module@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] ima: abort file hash computation on fatal signal
@ 2026-03-22 11:10 Shigeru Yoshida
  2026-03-22 14:17 ` Eric Biggers
  0 siblings, 1 reply; 2+ messages in thread
From: Shigeru Yoshida @ 2026-03-22 11:10 UTC (permalink / raw)
  To: Mimi Zohar, Roberto Sassu, Dmitry Kasatkin, Eric Snowberg,
	Paul Moore, James Morris, Serge E. Hallyn
  Cc: Shigeru Yoshida, syzbot+6ed94e81a1492fe1d512, linux-integrity,
	linux-security-module, linux-kernel

ima_calc_file_hash_atfm() and ima_calc_file_hash_tfm() compute a hash
over the entire file contents without checking for pending fatal
signals. When a very large file is being hashed during mmap (via
ima_file_mmap), the computation can take an extended period. If a
coredump is initiated by another thread in the same thread group during
this time, the dumper thread waits in coredump_wait() for all other
threads to exit. However, the hashing thread cannot exit until the hash
loop completes, resulting in a hung task.

Add fatal_signal_pending() checks to both the ahash and shash file
hashing loops so that the computation is aborted promptly when SIGKILL
is received.

Fixes: 3323eec921ef ("integrity: IMA as an integrity service provider")
Reported-by: syzbot+6ed94e81a1492fe1d512@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=6ed94e81a1492fe1d512
Tested-by: syzbot+6ed94e81a1492fe1d512@syzkaller.appspotmail.com
Signed-off-by: Shigeru Yoshida <syoshida@redhat.com>
---
 security/integrity/ima/ima_crypto.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/security/integrity/ima/ima_crypto.c b/security/integrity/ima/ima_crypto.c
index aff61643415d..7b721b9c944f 100644
--- a/security/integrity/ima/ima_crypto.c
+++ b/security/integrity/ima/ima_crypto.c
@@ -17,6 +17,7 @@
 #include <linux/crypto.h>
 #include <linux/scatterlist.h>
 #include <linux/err.h>
+#include <linux/sched/signal.h>
 #include <linux/slab.h>
 #include <crypto/hash.h>
 
@@ -416,6 +417,12 @@ static int ima_calc_file_hash_atfm(struct file *file,
 
 		if (rbuf[1])
 			active = !active; /* swap buffers, if we use two */
+
+		if (fatal_signal_pending(current)) {
+			ahash_wait(ahash_rc, &wait);
+			rc = -EINTR;
+			goto out3;
+		}
 	}
 	/* wait for the last update request to complete */
 	rc = ahash_wait(ahash_rc, &wait);
@@ -491,6 +498,10 @@ static int ima_calc_file_hash_tfm(struct file *file,
 		rc = crypto_shash_update(shash, rbuf, rbuf_len);
 		if (rc)
 			break;
+		if (fatal_signal_pending(current)) {
+			rc = -EINTR;
+			break;
+		}
 	}
 	kfree(rbuf);
 out:
-- 
2.52.0


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH] ima: abort file hash computation on fatal signal
  2026-03-22 11:10 [PATCH] ima: abort file hash computation on fatal signal Shigeru Yoshida
@ 2026-03-22 14:17 ` Eric Biggers
  0 siblings, 0 replies; 2+ messages in thread
From: Eric Biggers @ 2026-03-22 14:17 UTC (permalink / raw)
  To: Shigeru Yoshida
  Cc: Mimi Zohar, Roberto Sassu, Dmitry Kasatkin, Eric Snowberg,
	Paul Moore, James Morris, Serge E. Hallyn,
	syzbot+6ed94e81a1492fe1d512, linux-integrity,
	linux-security-module, linux-kernel

On Sun, Mar 22, 2026 at 08:10:19PM +0900, Shigeru Yoshida wrote:
> ima_calc_file_hash_atfm() and ima_calc_file_hash_tfm() compute a hash
> over the entire file contents without checking for pending fatal
> signals. When a very large file is being hashed during mmap (via
> ima_file_mmap), the computation can take an extended period. If a
> coredump is initiated by another thread in the same thread group during
> this time, the dumper thread waits in coredump_wait() for all other
> threads to exit. However, the hashing thread cannot exit until the hash
> loop completes, resulting in a hung task.
> 
> Add fatal_signal_pending() checks to both the ahash and shash file
> hashing loops so that the computation is aborted promptly when SIGKILL
> is received.
> 
> Fixes: 3323eec921ef ("integrity: IMA as an integrity service provider")
> Reported-by: syzbot+6ed94e81a1492fe1d512@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=6ed94e81a1492fe1d512
> Tested-by: syzbot+6ed94e81a1492fe1d512@syzkaller.appspotmail.com
> Signed-off-by: Shigeru Yoshida <syoshida@redhat.com>
> ---
>  security/integrity/ima/ima_crypto.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/security/integrity/ima/ima_crypto.c b/security/integrity/ima/ima_crypto.c
> index aff61643415d..7b721b9c944f 100644
> --- a/security/integrity/ima/ima_crypto.c
> +++ b/security/integrity/ima/ima_crypto.c
> @@ -17,6 +17,7 @@
>  #include <linux/crypto.h>
>  #include <linux/scatterlist.h>
>  #include <linux/err.h>
> +#include <linux/sched/signal.h>
>  #include <linux/slab.h>
>  #include <crypto/hash.h>
>  
> @@ -416,6 +417,12 @@ static int ima_calc_file_hash_atfm(struct file *file,
>  
>  		if (rbuf[1])
>  			active = !active; /* swap buffers, if we use two */
> +
> +		if (fatal_signal_pending(current)) {
> +			ahash_wait(ahash_rc, &wait);
> +			rc = -EINTR;
> +			goto out3;
> +		}

I think you'll need to rebase onto
https://git.kernel.org/pub/scm/linux/kernel/git/zohar/linux-integrity.git/log/?h=next-integrity
since there is a patch queued up that removes ima_calc_file_hash_atfm().
So only ima_calc_file_hash_tfm() will need to be updated.

- Eric

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2026-03-22 14:17 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-03-22 11:10 [PATCH] ima: abort file hash computation on fatal signal Shigeru Yoshida
2026-03-22 14:17 ` Eric Biggers

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox