Linux Security Modules development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Yuansheng Yang" <0xbeac@gmail.com>,
	"Jarkko Sakkinen" <jarkko@kernel.org>
Cc: "David Howells" <dhowells@redhat.com>,
	"Paul Moore" <paul@paul-moore.com>,
	"James Morris" <jmorris@namei.org>,
	"Serge E. Hallyn" <serge@hallyn.com>,
	keyrings@vger.kernel.org, linux-security-module@vger.kernel.org
Subject: Re: [PATCH v5] KEYS: Protect pkey method checks against type removal
Date: Sun, 08 Nov 2026 13:32:35 +0000	[thread overview]
Message-ID: <sashiko-outbox-203780@kernel.org> (raw)
In-Reply-To: <20261108131417.77287-1-0xbeac@gmail.com>

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Medium] KEYS: unlocked key->type dereferences in keyctl_pkey method
  checks
--

commit 0ed416869e71a7b6cc7a35c0985435f3666b373b
Author: Yuansheng Yang <0xbeac@gmail.com>

KEYS: Protect pkey method checks against type removal

This commit takes key->sem for reading around the four asymmetric key
method calls in security/keys/keyctl_pkey.c and re-checks the method
pointer under the lock to prevent races with key type unregistration.

> diff --git a/security/keys/keyctl_pkey.c b/security/keys/keyctl_pkey.c
> index 15b6bf6399b59..0ea6f2790e9de 100644
> --- a/security/keys/keyctl_pkey.c
> +++ b/security/keys/keyctl_pkey.c

[ ... ]

> @@ -305,31 +332,41 @@ long keyctl_pkey_verify(const struct keyctl_pkey_params __user *_params,
>  	ret = keyctl_pkey_params_get_2(_params, _info, KEYCTL_PKEY_VERIFY,
>  				       &params);
>  	if (ret < 0)
> -		goto error_params;
> +		goto out_free_params;
>  
>  	ret = -EOPNOTSUPP;
>  	if (!params.key->type->asym_verify_signature)
> -		goto error_params;
> +		goto out_free_params;

[Severity: Medium]
Can the remaining unlocked checks of key->type still fault on freed module
memory when a key type module is unloaded?

Checking key->type->asym_verify_signature here (and the corresponding
unlocked checks of key->type->asym_eds_op in keyctl_pkey_e_d_s() and
key->type->asym_query in keyctl_pkey_params_get()) reads key->type and
then dereferences the struct key_type in module memory without holding
key->sem:

CPU0                                  CPU1
keyctl_pkey_verify()
  reg = params.key->type
                                      unregister_key_type()
                                        ...
                                        down_write(&key->sem)
                                        key->type = &key_type_dead
                                        up_write(&key->sem)
                                      free_module()
  if (!reg->asym_verify_signature)  <-- use-after-free

security/keys/keyctl_pkey.c:keyctl_pkey_params_get() {
	...
	key_ref = lookup_user_key(id, 0, KEY_NEED_SEARCH);
	if (IS_ERR(key_ref))
		return PTR_ERR(key_ref);
	params->key = key_ref_to_ptr(key_ref);

	if (!params->key->type->asym_query)
		return -EOPNOTSUPP;

	return 0;
}

Both callers of keyctl_pkey_params_get() immediately take key->sem and
re-check asym_query before doing anything else, so the unlocked check in
keyctl_pkey_params_get() can be dropped, and the early asym_eds_op and
asym_verify_signature checks can be performed under key->sem inside
keyctl_pkey_params_get_2() while still returning -EOPNOTSUPP before
memdup_user().

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261108131417.77287-1-0xbeac@gmail.com?part=1

  reply	other threads:[~2026-11-08 13:32 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-03  3:29 [PATCH] KEYS: Protect pkey method checks against type removal BeaCox
2026-11-03  3:39 ` sashiko-bot
2026-11-03 15:17 ` Jarkko Sakkinen
2026-11-04  2:14   ` [PATCH v2] " BeaCox
2026-11-04  2:25     ` sashiko-bot
2026-11-04 14:49     ` Jarkko Sakkinen
2026-11-04 23:59     ` Jarkko Sakkinen
2026-11-05  2:24       ` Yuansheng Yang
2026-11-05  3:28         ` Jarkko Sakkinen
2026-11-05  4:38           ` [PATCH v3] " Yuansheng Yang
2026-11-05  4:48             ` sashiko-bot
2026-11-05 16:55             ` Jarkko Sakkinen
2026-11-05 19:36               ` [PATCH v4] " Yuansheng Yang
2026-11-05 19:48                 ` sashiko-bot
2026-11-06 14:08                 ` Jarkko Sakkinen
2026-11-08 12:59                   ` Yuansheng Yang
2026-11-08 13:14                   ` [PATCH v5] " Yuansheng Yang
2026-11-08 13:32                     ` sashiko-bot [this message]
2026-11-08 15:45                     ` Jarkko Sakkinen
2026-11-09  5:52                       ` [PATCH v6] " Yuansheng Yang
2026-11-09  6:02                         ` sashiko-bot
2026-11-09 15:04                         ` Jarkko Sakkinen
2026-11-09 15:06                           ` Jarkko Sakkinen
2026-11-10  1:49                         ` Jarkko Sakkinen
2026-11-10  3:02                           ` Yuansheng Yang
2026-11-10  3:02                           ` [PATCH v7] " Yuansheng Yang
2026-11-10  3:06                             ` Jarkko Sakkinen
2026-11-10  3:13                             ` sashiko-bot

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=sashiko-outbox-203780@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=0xbeac@gmail.com \
    --cc=dhowells@redhat.com \
    --cc=jarkko@kernel.org \
    --cc=jmorris@namei.org \
    --cc=keyrings@vger.kernel.org \
    --cc=linux-security-module@vger.kernel.org \
    --cc=paul@paul-moore.com \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=serge@hallyn.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox