linux-kernel - Re: [PATCH v2 5/6] crypto: stm32/hash: Support Ux500 hash

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 20 Jan 2023 18:21:35 +0800
From:   Herbert Xu <herbert@...dor.apana.org.au>
To:     Linus Walleij <linus.walleij@...aro.org>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Rob Herring <robh+dt@...nel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        Maxime Coquelin <mcoquelin.stm32@...il.com>,
        Alexandre Torgue <alexandre.torgue@...s.st.com>,
        Lionel Debieve <lionel.debieve@...s.st.com>,
        linux-crypto@...r.kernel.org, devicetree@...r.kernel.org,
        linux-stm32@...md-mailman.stormreply.com,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 5/6] crypto: stm32/hash: Support Ux500 hash

On Tue, Jan 10, 2023 at 08:19:16PM +0100, Linus Walleij wrote:
>  
> +static void stm32_hash_emptymsg_fallback(struct ahash_request *req)
> +{
> +	struct crypto_ahash *ahash = crypto_ahash_reqtfm(req);
> +	struct stm32_hash_ctx *ctx = crypto_ahash_ctx(ahash);
> +	struct stm32_hash_request_ctx *rctx = ahash_request_ctx(req);
> +	struct stm32_hash_dev *hdev = rctx->hdev;
> +	struct crypto_shash *xtfm;
> +	struct shash_desc *sdesc;
> +	size_t len;
> +	int ret;
> +
> +	dev_dbg(hdev->dev, "use fallback message size 0 key size %d\n",
> +		ctx->keylen);
> +	xtfm = crypto_alloc_shash(crypto_ahash_alg_name(ahash),
> +				  0, CRYPTO_ALG_NEED_FALLBACK);
> +	if (IS_ERR(xtfm)) {
> +		dev_err(hdev->dev, "failed to allocate synchronous fallback\n");
> +		return;
> +	}
> +
> +	len = sizeof(*sdesc) + crypto_shash_descsize(xtfm);
> +	sdesc = kmalloc(len, GFP_KERNEL);
> +	if (!sdesc)
> +		goto err_hashkey_sdesc;
> +	sdesc->tfm = xtfm;
> +
> +	if (ctx->keylen) {
> +		ret = crypto_shash_setkey(xtfm, ctx->key, ctx->keylen);
> +		if (ret) {
> +			dev_err(hdev->dev, "failed to set key ret=%d\n", ret);
> +			goto err_hashkey;
> +		}
> +	}
> +
> +	ret = crypto_shash_init(sdesc);
> +	if (ret) {
> +		dev_err(hdev->dev, "shash init error ret=%d\n", ret);
> +		goto err_hashkey;
> +	}
> +
> +	ret = crypto_shash_finup(sdesc, NULL, 0, rctx->digest);
> +	if (ret)
> +		dev_err(hdev->dev, "shash finup error\n");
> +err_hashkey:
> +	kfree(sdesc);
> +err_hashkey_sdesc:
> +	crypto_free_shash(xtfm);
> +}

Calling crypto_alloc_shash is not allowed in this context.  For
example, we might have been called down from the block layer due to
swapping.  Even if you intermediate this with kernel threads, it
still doesn't change the nature of the dead-lock.

So if you need a fallback for zero-length messages, just allocate
it unconditionally in the init_tfm function.

Cheers,
-- 
Email: Herbert Xu <herbert@...dor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt