qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Daniel P. Berrangé" <berrange@redhat.com>
To: Alejandro Zeise <alejandro.zeise@seagate.com>
Cc: qemu-arm@nongnu.org, kris.conklin@seagate.com,
	jonathan.henze@seagate.com, evan.burgess@seagate.com,
	clg@kaod.org, peter.maydell@linaro.org, qemu-devel@nongnu.org
Subject: Re: [PATCH v4 07/15] crypto/hash: Implement and use new hash API
Date: Thu, 8 Aug 2024 17:21:58 +0100	[thread overview]
Message-ID: <ZrTwphQDG20aGtsP@redhat.com> (raw)
In-Reply-To: <20240807195122.2827364-8-alejandro.zeise@seagate.com>

On Wed, Aug 07, 2024 at 07:51:14PM +0000, Alejandro Zeise wrote:
> Changes the public hash API implementation to support accumulative hashing.
> 
> Implementations for the public functions are added to call the new
> driver functions that implement context creation, updating,
> finalization, and destruction.
> 
> Additionally changes the "shortcut" functions to use these 4 new core
> functions.
> 
> Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
> ---
>  crypto/hash.c | 163 +++++++++++++++++++++++++++++++++++++++-----------
>  1 file changed, 128 insertions(+), 35 deletions(-)
> 
> diff --git a/crypto/hash.c b/crypto/hash.c
> index b0f8228bdc..f3c18cdd74 100644
> --- a/crypto/hash.c
> +++ b/crypto/hash.c
> @@ -1,6 +1,7 @@
>  /*
>   * QEMU Crypto hash algorithms
>   *
> + * Copyright (c) 2024 Seagate Technology LLC and/or its Affiliates
>   * Copyright (c) 2015 Red Hat, Inc.
>   *
>   * This library is free software; you can redistribute it and/or
> @@ -45,23 +46,20 @@ int qcrypto_hash_bytesv(QCryptoHashAlgorithm alg,
>                          size_t *resultlen,
>                          Error **errp)
>  {
> -#ifdef CONFIG_AF_ALG
> -    int ret;
> -    /*
> -     * TODO:
> -     * Maybe we should treat some afalg errors as fatal
> -     */
> -    ret = qcrypto_hash_afalg_driver.hash_bytesv(alg, iov, niov,
> -                                                result, resultlen,
> -                                                NULL);
> -    if (ret == 0) {
> -        return ret;
> +    int fail;
> +    QCryptoHash *ctx = qcrypto_hash_new(alg, errp);
> +
> +    if (ctx) {
> +        fail = qcrypto_hash_updatev(ctx, iov, niov, errp) ||
> +               qcrypto_hash_finalize_bytes(ctx, result, resultlen, errp);
> +
> +        /* Ensure context is always freed regardless of error */
> +        qcrypto_hash_free(ctx);
> +    } else {
> +        fail = -1;
>      }
> -#endif

Generally I'd prefer immediate return on error, and you can
use g_autoptr to free the object too. With that you'll end
up without the extra 'fail' variable:

    g_autoptr(QCryptoHash) ctx = qcrypto_hash_new(alg, errp);

    if (!ctx) {
        return -1;
    }

    if (qcrypto_hash_updatev(ctx, iov, niov, errp) < 0 ||
        qcrypto_hash_finalize_bytes(ctx, result, resultlen, errp) < 0) {
	return -1;
    }

    return 0;

> @@ -77,30 +75,121 @@ int qcrypto_hash_bytes(QCryptoHashAlgorithm alg,
>      return qcrypto_hash_bytesv(alg, &iov, 1, result, resultlen, errp);
>  }
>  
> -static const char hex[] = "0123456789abcdef";
> -
> -int qcrypto_hash_digestv(QCryptoHashAlgorithm alg,
> +int qcrypto_hash_updatev(QCryptoHash *hash,
>                           const struct iovec *iov,
>                           size_t niov,
> -                         char **digest,
>                           Error **errp)
>  {
> +#ifdef CONFIG_AF_ALG
> +    return qcrypto_hash_afalg_driver.hash_update(hash, iov, niov, errp);
> +#else
> +    return qcrypto_hash_lib_driver.hash_update(hash, iov, niov, errp);
> +#endif /* CONFIG_AF_ALG */
> +}

This isn't quite the same as the old code. That would try afalg, and then
dynamically fallback to the userspace driver.

I think we need to deal with this all in the qcrypto_hash_new() method.
Attempt to create an AF_ALG driver instance via its 'hash_new' method,
and if that fails create the normal driver 'hash_new' method. We need
to record which we created in QCryptoHash struct, and then in the
update, finalize & free methods we need to call the corresponding
driver.

> +
> +int qcrypto_hash_update(QCryptoHash *hash,
> +                        const char *buf,
> +                        size_t len,
> +                        Error **errp)
> +{
> +    struct iovec iov = { .iov_base = (char *)buf, .iov_len = len };
> +
> +    return qcrypto_hash_updatev(hash, &iov, 1, errp);
> +}
> +
> +QCryptoHash *qcrypto_hash_new(QCryptoHashAlgorithm alg, Error **errp)
> +{

Here we should call

   if (!qcrypto_hash_supports(alg)) {
      error_setg(errp, "Unsupported hash algorithm %s",
                 QCryptoHashAlgorithm_lookup[alg]);
      return NULL;
   }

That way, we avoid repeating this check in every single
'hash_new' method impl

> +#ifdef CONFIG_AF_ALG
> +    return qcrypto_hash_afalg_driver.hash_new(alg, errp);
> +#else
> +    return qcrypto_hash_lib_driver.hash_new(alg, errp);
> +#endif /* CONFIG_AF_ALG */
> +}
> +
> +void qcrypto_hash_free(QCryptoHash *hash)
> +{
> +#ifdef CONFIG_AF_ALG
> +    qcrypto_hash_afalg_driver.hash_free(hash);
> +#else
> +    qcrypto_hash_lib_driver.hash_free(hash);
> +#endif /* CONFIG_AF_ALG */
> +}
> +
> +int qcrypto_hash_finalize_bytes(QCryptoHash *hash,
> +                                uint8_t **result,
> +                                size_t *result_len,
> +                                Error **errp)
> +{
> +#ifdef CONFIG_AF_ALG
> +    return qcrypto_hash_afalg_driver.hash_finalize(hash, result, result_len,
> +                                                   errp);
> +#else
> +    return qcrypto_hash_lib_driver.hash_finalize(hash, result, result_len, errp);
> +#endif /* CONFIG_AF_ALG */
> +}
> +
> +static const char hex[] = "0123456789abcdef";
> +
> +int qcrypto_hash_finalize_digest(QCryptoHash *hash,
> +                                 char **digest,
> +                                 Error **errp)
> +{
> +    int ret;
>      uint8_t *result = NULL;

declare this

  g_autofree uint8_t *result = NULL;

then you can drop the later g_free(result)

>      size_t resultlen = 0;
>      size_t i;
>  
> -    if (qcrypto_hash_bytesv(alg, iov, niov, &result, &resultlen, errp) < 0) {
> -        return -1;
> +    ret = qcrypto_hash_finalize_bytes(hash, &result, &resultlen, errp);
> +    if (ret == 0) {
> +        *digest = g_new0(char, (resultlen * 2) + 1);
> +        for (i = 0 ; i < resultlen ; i++) {
> +            (*digest)[(i * 2)] = hex[(result[i] >> 4) & 0xf];
> +            (*digest)[(i * 2) + 1] = hex[result[i] & 0xf];
> +        }
> +        (*digest)[resultlen * 2] = '\0';
> +        g_free(result);
> +    }
> +
> +    return ret;
> +}
> +
> +int qcrypto_hash_finalize_base64(QCryptoHash *hash,
> +                                 char **base64,
> +                                 Error **errp)
> +{
> +    int ret;
> +    uint8_t *result = NULL;

g_autofree for this too

> +    size_t resultlen = 0;
> +
> +    ret = qcrypto_hash_finalize_bytes(hash, &result, &resultlen, errp);
> +    if (ret == 0) {
> +        *base64 = g_base64_encode(result, resultlen);
> +        g_free(result);
>      }
>  
> -    *digest = g_new0(char, (resultlen * 2) + 1);
> -    for (i = 0 ; i < resultlen ; i++) {
> -        (*digest)[(i * 2)] = hex[(result[i] >> 4) & 0xf];
> -        (*digest)[(i * 2) + 1] = hex[result[i] & 0xf];
> +    return ret;
> +}
> +
> +int qcrypto_hash_digestv(QCryptoHashAlgorithm alg,
> +                         const struct iovec *iov,
> +                         size_t niov,
> +                         char **digest,
> +                         Error **errp)
> +{
> +    bool fail;
> +    QCryptoHash *ctx = qcrypto_hash_new(alg, errp);
> +
> +    if (ctx) {
> +        fail = qcrypto_hash_updatev(ctx, iov, niov, errp) ||
> +               qcrypto_hash_finalize_digest(ctx, digest, errp);
> +
> +        /* Ensure context is always freed regardless of error */
> +        qcrypto_hash_free(ctx);
> +    } else {
> +        fail = false;
>      }
> -    (*digest)[resultlen * 2] = '\0';
> -    g_free(result);
> -    return 0;
> +
> +    return fail;
>  }

Same comment about re-arranging the code that I mentioned higher
up against qcrypto_hash_bytesv

>  
>  int qcrypto_hash_digest(QCryptoHashAlgorithm alg,
> @@ -120,16 +209,20 @@ int qcrypto_hash_base64v(QCryptoHashAlgorithm alg,
>                           char **base64,
>                           Error **errp)
>  {
> -    uint8_t *result = NULL;
> -    size_t resultlen = 0;
> +    bool fail;
> +    QCryptoHash *ctx = qcrypto_hash_new(alg, errp);
> +
> +    if (ctx) {
> +        fail = qcrypto_hash_updatev(ctx, iov, niov, errp) ||
> +               qcrypto_hash_finalize_base64(ctx, base64, errp);
>  
> -    if (qcrypto_hash_bytesv(alg, iov, niov, &result, &resultlen, errp) < 0) {
> -        return -1;
> +        /* Ensure context is always freed regardless of error */
> +        qcrypto_hash_free(ctx);
> +    } else {
> +        fail = 1;
>      }
>  
> -    *base64 = g_base64_encode(result, resultlen);
> -    g_free(result);
> -    return 0;
> +    return fail;
>  }

And same comment again.

>  
>  int qcrypto_hash_base64(QCryptoHashAlgorithm alg,
> -- 
> 2.34.1
> 

With regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|



  reply	other threads:[~2024-08-08 16:23 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-07 19:51 [PATCH v4 00/15] hw/misc/aspeed_hace: Fix SG Accumulative Hash Calculations Alejandro Zeise
2024-08-07 19:51 ` [PATCH v4 01/15] crypto: accumulative hashing API Alejandro Zeise
2024-08-08 16:04   ` Daniel P. Berrangé
2024-08-08 17:00     ` Markus Armbruster
2024-08-07 19:51 ` [PATCH v4 02/15] crypto/hash-glib: Implement new hash API Alejandro Zeise
2024-08-08 16:58   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 03/15] crypto/hash-gcrypt: " Alejandro Zeise
2024-08-08 17:00   ` Daniel P. Berrangé
2024-08-08 17:05   ` Daniel P. Berrangé
2024-08-08 18:24     ` Alejandro Zeise
2024-08-09  8:37       ` Daniel P. Berrangé
2024-08-08 17:10   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 04/15] crypto/hash-gnutls: " Alejandro Zeise
2024-08-08 17:10   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 05/15] crypto/hash-nettle: " Alejandro Zeise
2024-08-07 19:51 ` [PATCH v4 06/15] crypto/hash-afalg: " Alejandro Zeise
2024-08-08 17:16   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 07/15] crypto/hash: Implement and use " Alejandro Zeise
2024-08-08 16:21   ` Daniel P. Berrangé [this message]
2024-08-07 19:51 ` [PATCH v4 08/15] tests/unit/test-crypto-hash: accumulative hashing Alejandro Zeise
2024-08-08 17:18   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 09/15] crypto/hash-glib: Remove old hash API functions Alejandro Zeise
2024-08-08 17:19   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 10/15] crypto/hash-gcrypt: " Alejandro Zeise
2024-08-08 17:19   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 11/15] crypto/hash-gnutls: " Alejandro Zeise
2024-08-08 17:20   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 12/15] crypto/hash-nettle: " Alejandro Zeise
2024-08-08 17:22   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 13/15] crypto/hash-afalg: " Alejandro Zeise
2024-08-08 17:23   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 14/15] crypto/hashpriv: Remove old hash API function Alejandro Zeise
2024-08-08 17:24   ` Daniel P. Berrangé
2024-08-07 19:51 ` [PATCH v4 15/15] hw/misc/aspeed_hace: Fix SG Accumulative hashing Alejandro Zeise
2024-08-27 13:53   ` Cédric Le Goater
2024-08-07 20:01 ` [PATCH v4 00/15] hw/misc/aspeed_hace: Fix SG Accumulative Hash Calculations Philippe Mathieu-Daudé
2024-08-07 20:30   ` Alejandro Zeise
2024-08-08  8:46   ` Daniel P. Berrangé

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=ZrTwphQDG20aGtsP@redhat.com \
    --to=berrange@redhat.com \
    --cc=alejandro.zeise@seagate.com \
    --cc=clg@kaod.org \
    --cc=evan.burgess@seagate.com \
    --cc=jonathan.henze@seagate.com \
    --cc=kris.conklin@seagate.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-arm@nongnu.org \
    --cc=qemu-devel@nongnu.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;
as well as URLs for NNTP newsgroup(s).