Linux-mtd Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Zhihao Cheng <chengzhihao1@huawei.com>
To: <shao.mingyin@zte.com.cn>, <richard@nod.at>,
	<linux-mtd@lists.infradead.org>
Cc: <linux-kernel@vger.kernel.org>, <stable@vger.kernel.org>
Subject: Re: [PATCH] ubifs: propagate ubifs_shash_copy_state() errors
Date: Fri, 30 Oct 2026 12:44:28 +0800	[thread overview]
Message-ID: <4bdb09dd-bbaf-111f-adb6-4d9970ebedcc@huawei.com> (raw)
In-Reply-To: <20261030114406371EZw_orIqSRlCTxnz_L39Z@zte.com.cn>

在 2026/10/30 11:44, shao.mingyin@zte.com.cn 写道:
> From: Shao Mingyin <shao.mingyin@zte.com.cn>
> 
> ubifs_shash_copy_state() duplicates the rolling hash state with
> crypto_shash_export() and crypto_shash_import().  The export buffer is
> allocated with kmalloc(), so the function can fail with -ENOMEM, and
> crypto_shash_import() can fail as well.
> 
> Five callers ignore the return value.  The target descriptor then has no
> valid state, and the crypto_shash_final() that follows computes a digest
> from uninitialized data.  When preparing an auth node this stores a
> wrong HMAC on the media, so the affected journal head no longer
> authenticates after a remount.  During replay it makes the verification
> of a valid image fail with -EPERM, and in replay_bud() and
> add_replay_bud() it leaves the replay hash state inconsistent with the
> media, which corrupts the auth nodes written by later commits.
> 
> Check the error at all five call sites and propagate it.  In
> add_replay_bud() the descriptor belongs to the new bud, so free it on
> the added error path.
> 
> Fixes: 6a98bc4614de ("ubifs: Add authentication nodes to journal")
> Fixes: da8ef65f9573 ("ubifs: Authenticate replayed journal")
> Cc: stable@vger.kernel.org
> Signed-off-by: Shao Mingyin <shao.mingyin@zte.com.cn>
> ---

Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
> diff --git a/fs/ubifs/auth.c b/fs/ubifs/auth.c
> index 1f770795ee70..62dfacfb6d6b 100644
> --- a/fs/ubifs/auth.c
> +++ b/fs/ubifs/auth.c
> @@ -70,7 +70,10 @@ int ubifs_prepare_auth_node(struct ubifs_info *c, void *node,
>   		SHASH_DESC_ON_STACK(hash_desc, c->hash_tfm);
> 
>   		hash_desc->tfm = c->hash_tfm;
> -		ubifs_shash_copy_state(c, inhash, hash_desc);
> +
> +		err = ubifs_shash_copy_state(c, inhash, hash_desc);
> +		if (err)
> +			return err;
> 
>   		err = crypto_shash_final(hash_desc, hash);
>   		if (err)
> diff --git a/fs/ubifs/log.c b/fs/ubifs/log.c
> index c6edcbf35ebc..42cadf558a4f 100644
> --- a/fs/ubifs/log.c
> +++ b/fs/ubifs/log.c
> @@ -412,7 +412,11 @@ int ubifs_log_start_commit(struct ubifs_info *c, int *ltail_lnum)
>   					 UBIFS_REF_NODE_SZ);
>   		if (err)
>   			goto out;
> -		ubifs_shash_copy_state(c, c->log_hash, c->jheads[i].log_hash);
> +
> +		err = ubifs_shash_copy_state(c, c->log_hash,
> +					     c->jheads[i].log_hash);
> +		if (err)
> +			goto out;
>   	}
> 
>   	ubifs_pad(c, buf + len, ALIGN(len, c->min_io_size) - len);
> diff --git a/fs/ubifs/replay.c b/fs/ubifs/replay.c
> index 263045e05cf1..0b82c67a2dcc 100644
> --- a/fs/ubifs/replay.c
> +++ b/fs/ubifs/replay.c
> @@ -566,10 +566,14 @@ authenticate_sleb_hash(struct ubifs_info *c,
>   		       struct shash_desc *log_hash, u8 *hash)
>   {
>   	SHASH_DESC_ON_STACK(hash_desc, c->hash_tfm);
> +	int err;
> 
>   	hash_desc->tfm = c->hash_tfm;
> 
> -	ubifs_shash_copy_state(c, log_hash, hash_desc);
> +	err = ubifs_shash_copy_state(c, log_hash, hash_desc);
> +	if (err)
> +		return err;
> +
>   	return crypto_shash_final(hash_desc, hash);
>   }
> 
> @@ -694,8 +698,10 @@ static int replay_bud(struct ubifs_info *c, struct bud_entry *b)
>   		goto out;
>   	}
> 
> -	ubifs_shash_copy_state(c, b->bud->log_hash,
> -			       c->jheads[b->bud->jhead].log_hash);
> +	err = ubifs_shash_copy_state(c, b->bud->log_hash,
> +				     c->jheads[b->bud->jhead].log_hash);
> +	if (err)
> +		goto out;
> 
>   	/*
>   	 * The bud does not have to start from offset zero - the beginning of
> @@ -915,7 +921,9 @@ static int add_replay_bud(struct ubifs_info *c, int lnum, int offs, int jhead,
>   		goto out;
>   	}
> 
> -	ubifs_shash_copy_state(c, c->log_hash, bud->log_hash);
> +	err = ubifs_shash_copy_state(c, c->log_hash, bud->log_hash);
> +	if (err)
> +		goto out_free_log_hash;
> 
>   	ubifs_add_bud(c, bud);
> 
> @@ -924,6 +932,8 @@ static int add_replay_bud(struct ubifs_info *c, int lnum, int offs, int jhead,
>   	list_add_tail(&b->list, &c->replay_buds);
> 
>   	return 0;
> +out_free_log_hash:
> +	kfree(bud->log_hash);
>   out:
>   	kfree(bud);
>   	kfree(b);
> .
> 


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

      reply	other threads:[~2026-10-30  4:44 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-30  3:44 [PATCH] ubifs: propagate ubifs_shash_copy_state() errors shao.mingyin
2026-10-30  4:44 ` Zhihao Cheng [this message]

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=4bdb09dd-bbaf-111f-adb6-4d9970ebedcc@huawei.com \
    --to=chengzhihao1@huawei.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mtd@lists.infradead.org \
    --cc=richard@nod.at \
    --cc=shao.mingyin@zte.com.cn \
    --cc=stable@vger.kernel.org \
    /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