All of lore.kernel.org
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Martin George" <martinus.gpy@gmail.com>, linux-nvme@lists.infradead.org
Subject: Re: [PATCH 1/2] nvme-sysfs: use kfree_sensitive() for DH-HMAC-CHAP secrets
Date: Mon, 16 Nov 2026 10:14:02 +0000	[thread overview]
Message-ID: <sashiko-outbox-216698@kernel.org> (raw)
In-Reply-To: <20261116100244.7824-2-marting@netapp.com>

Thank you for your contribution! Sashiko AI review found no new issues
in this patch, but noticed 1 related pre-existing issue it does not
introduce:
- [Critical] nvme: double-free and use-after-free in
  nvme_ctrl_dhchap_secret_store() [1]
--

commit 86ec9e603dccdb522db4b07402c30b2c18b2a11b
Author: Martin George <martinus.gpy@gmail.com>

nvme-sysfs: use kfree_sensitive() for DH-HMAC-CHAP secrets

This commit replaces kfree() with kfree_sensitive() when freeing DH-HMAC-CHAP
secret strings in dhchap_secret_store() and dhchap_ctrl_secret_store().

> diff --git a/drivers/nvme/host/sysfs.c b/drivers/nvme/host/sysfs.c
> index 10018c6356b62..5a195b8971f21 100644
> --- a/drivers/nvme/host/sysfs.c
> +++ b/drivers/nvme/host/sysfs.c
> @@ -849,10 +849,10 @@ static ssize_t dhchap_secret_store(struct device *dev,
>  
>  		ret = nvme_auth_parse_key(dhchap_secret, &key);
>  		if (ret) {
> -			kfree(dhchap_secret);
> +			kfree_sensitive(dhchap_secret);
>  			return ret;
>  		}
> -		kfree(opts->dhchap_secret);
> +		kfree_sensitive(opts->dhchap_secret);
>  		opts->dhchap_secret = dhchap_secret;
>  		host_key = ctrl->host_key;
>  		mutex_lock(&ctrl->dhchap_auth_mutex);

[Severity: Critical]
This problem wasn't introduced by this patch, but can concurrent sysfs accesses
[1] trigger a double free or use-after-free on opts->dhchap_secret and
ctrl->host_key?

In dhchap_secret_store(), opts->dhchap_secret is freed and replaced without
holding a lock, and ctrl->host_key is read before acquiring
ctrl->dhchap_auth_mutex:

CPU0                                  CPU1
dhchap_secret_store()                 dhchap_secret_store()
  kfree_sensitive(opts->dhchap_secret)
                                        kfree_sensitive(opts->dhchap_secret)
                                          <-- double free
  opts->dhchap_secret = dhchap_secret
  host_key = ctrl->host_key
                                        host_key = ctrl->host_key
  mutex_lock(&ctrl->dhchap_auth_mutex)
  ctrl->host_key = key
  mutex_unlock(&ctrl->dhchap_auth_mutex)
  nvme_auth_free_key(host_key)
                                        mutex_lock(&ctrl->dhchap_auth_mutex)
                                        ctrl->host_key = key
                                        mutex_unlock(&ctrl->dhchap_auth_mutex)
                                        nvme_auth_free_key(host_key)
                                          <-- double free

A concurrent dhchap_secret_show() or strcmp() in dhchap_secret_store() can
also read opts->dhchap_secret while it is being freed, causing a
use-after-free.  The same unlocked accesses exist in
dhchap_ctrl_secret_store() and dhchap_ctrl_secret_show() for
opts->dhchap_ctrl_secret and ctrl->ctrl_key.

[1] https://sashiko.dev/#/bug/linux-d9bcd9ea-2b96-4998-b954-e0e2557a158a

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261116100244.7824-1-marting@netapp.com?part=1


  reply	other threads:[~2026-11-16 10:14 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-16 10:02 [PATCH 0/2] nvme: use kfree_sensitive() for DH-HMAC-CHAP secrets Martin George
2026-11-16 10:02 ` [PATCH 1/2] nvme-sysfs: " Martin George
2026-11-16 10:14   ` sashiko-bot [this message]
2026-11-16 13:44   ` Hannes Reinecke
2026-11-16 10:02 ` [PATCH 2/2] nvme-fabrics: use kfree_sensitive() for the old DH-HMAC-CHAP secret Martin George
2026-11-16 13:45   ` Hannes Reinecke
2026-11-16 16:18 ` [PATCH 0/2] nvme: use kfree_sensitive() for DH-HMAC-CHAP secrets Keith Busch

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-216698@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=linux-nvme@lists.infradead.org \
    --cc=martinus.gpy@gmail.com \
    --cc=sashiko-reviews@lists.linux.dev \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.