From: Luca Boccassi <bluca@debian.org>
To: "De Lara Guarch, Pablo" <pablo.de.lara.guarch@intel.com>
Cc: "stable@dpdk.org" <stable@dpdk.org>
Subject: Re: [PATCH 20.11 1/2] crypto/ipsec_mb: fix length and offset settings
Date: Mon, 21 Mar 2022 22:12:29 +0000 [thread overview]
Message-ID: <CAMw=ZnRV+ut6+jktLjjz8ftJ4x--USBnr6WH9qySoMw1tkbVCw@mail.gmail.com> (raw)
In-Reply-To: <DM8PR11MB559161E51BA353FCE9006CC184169@DM8PR11MB5591.namprd11.prod.outlook.com>
On Mon, 21 Mar 2022 at 19:57, De Lara Guarch, Pablo
<pablo.de.lara.guarch@intel.com> wrote:
>
> Hi Luca,
>
> > -----Original Message-----
> > From: Luca Boccassi <bluca@debian.org>
> > Sent: Monday, March 14, 2022 12:20 PM
> > To: De Lara Guarch, Pablo <pablo.de.lara.guarch@intel.com>; stable@dpdk.org
> > Subject: Re: [PATCH 20.11 1/2] crypto/ipsec_mb: fix length and offset settings
> >
> > On Mon, 2022-03-14 at 11:05 +0000, Pablo de Lara wrote:
> > > [ upstream commit a501609ea6466ed8526c0dfadedee332a4d4a451 ]
> > >
> > > KASUMI, SNOW3G and ZUC require lengths and offsets to be set in bits
> > > or bytes depending on the algorithm.
> > > There were some algorithms that were mixing these two, so this commit
> > > is fixing this issue.
> > >
> > > Fixes: ae8e085c608d ("crypto/aesni_mb: support KASUMI F8/F9")
> > > Fixes: 6c42e0cf4d12 ("crypto/aesni_mb: support SNOW3G-UEA2/UIA2")
> > > Fixes: fd8df85487c4 ("crypto/aesni_mb: support ZUC-EEA3/EIA3")
> > >
> > > Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> > > ---
> > > drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c | 121
> > > +++++++++++++++------
> > > 1 file changed, 86 insertions(+), 35 deletions(-)
> > >
> > > diff --git a/drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c
> > > b/drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c
> > > index f4ffb21e10..07f5caa76f 100644
> > > --- a/drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c
> > > +++ b/drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c
> > > @@ -1057,7 +1057,9 @@ get_session(struct aesni_mb_qp *qp, struct
> > > rte_crypto_op *op)
> > >
> > >
> > > static inline uint64_t
> > > auth_start_offset(struct rte_crypto_op *op, struct aesni_mb_session
> > *session,
> > > - uint32_t oop)
> > > + uint32_t oop, const uint32_t auth_offset,
> > > + const uint32_t cipher_offset, const uint32_t auth_length,
> > > + const uint32_t cipher_length)
> > > {
> > > struct rte_mbuf *m_src, *m_dst;
> > > uint8_t *p_src, *p_dst;
> > > @@ -1066,7 +1068,7 @@ auth_start_offset(struct rte_crypto_op *op,
> > > struct aesni_mb_session *session,
> > >
> > >
> > > /* Only cipher then hash needs special calculation. */
> > > if (!oop || session->chain_order != CIPHER_HASH)
> > > - return op->sym->auth.data.offset;
> > > + return auth_offset;
> > >
> > >
> > > m_src = op->sym->m_src;
> > > m_dst = op->sym->m_dst;
> > > @@ -1074,24 +1076,23 @@ auth_start_offset(struct rte_crypto_op *op,
> > struct aesni_mb_session *session,
> > > p_src = rte_pktmbuf_mtod(m_src, uint8_t *);
> > > p_dst = rte_pktmbuf_mtod(m_dst, uint8_t *);
> > > u_src = (uintptr_t)p_src;
> > > - u_dst = (uintptr_t)p_dst + op->sym->auth.data.offset;
> > > + u_dst = (uintptr_t)p_dst + auth_offset;
> > >
> > >
> > > /**
> > > * Copy the content between cipher offset and auth offset for
> > generating
> > > * correct digest.
> > > */
> > > - if (op->sym->cipher.data.offset > op->sym->auth.data.offset)
> > > - memcpy(p_dst + op->sym->auth.data.offset,
> > > - p_src + op->sym->auth.data.offset,
> > > - op->sym->cipher.data.offset -
> > > - op->sym->auth.data.offset);
> > > -
> > > + if (cipher_offset > auth_offset)
> > > + memcpy(p_dst + auth_offset,
> > > + p_src + auth_offset,
> > > + cipher_offset -
> > > + auth_offset);
> > > /**
> > > * Copy the content between (cipher offset + length) and (auth offset +
> > > * length) for generating correct digest
> > > */
> > > - cipher_end = op->sym->cipher.data.offset + op->sym-
> > >cipher.data.length;
> > > - auth_end = op->sym->auth.data.offset + op->sym->auth.data.length;
> > > + cipher_end = cipher_offset + cipher_length;
> > > + auth_end = auth_offset + auth_length;
> > > if (cipher_end < auth_end)
> > > memcpy(p_dst + cipher_end, p_src + cipher_end,
> > > auth_end - cipher_end);
> > > @@ -1246,6 +1247,10 @@ set_mb_job_params(JOB_AES_HMAC *job, struct
> > aesni_mb_qp *qp,
> > > struct rte_mbuf *m_src = op->sym->m_src, *m_dst;
> > > struct aesni_mb_session *session;
> > > uint32_t m_offset, oop;
> > > + uint32_t auth_off_in_bytes;
> > > + uint32_t ciph_off_in_bytes;
> > > + uint32_t auth_len_in_bytes;
> > > + uint32_t ciph_len_in_bytes;
> > >
> > >
> > > session = get_session(qp, op);
> > > if (session == NULL) {
> > > @@ -1362,6 +1367,7 @@ set_mb_job_params(JOB_AES_HMAC *job, struct
> > aesni_mb_qp *qp,
> > > if (job->cipher_mode == IMB_CIPHER_ZUC_EEA3) {
> > > job->aes_enc_key_expanded = session->cipher.zuc_cipher_key;
> > > job->aes_dec_key_expanded = session->cipher.zuc_cipher_key;
> > > + m_offset >>= 3;
> > > } else if (job->cipher_mode == IMB_CIPHER_SNOW3G_UEA2_BITLEN) {
> > > job->enc_keys = &session->cipher.pKeySched_snow3g_cipher;
> > > m_offset = 0;
> > > @@ -1418,9 +1424,6 @@ set_mb_job_params(JOB_AES_HMAC *job, struct
> > > aesni_mb_qp *qp,
> > >
> > >
> > > switch (job->hash_alg) {
> > > case AES_CCM:
> > > - job->cipher_start_src_offset_in_bytes =
> > > - op->sym->aead.data.offset;
> > > - job->msg_len_to_cipher_in_bytes = op->sym-
> > >aead.data.length;
> > > job->hash_start_src_offset_in_bytes = op->sym-
> > >aead.data.offset;
> > > job->msg_len_to_hash_in_bytes = op->sym->aead.data.length;
> > >
> > >
> > > @@ -1430,19 +1433,11 @@ set_mb_job_params(JOB_AES_HMAC *job, struct
> > > aesni_mb_qp *qp,
> > >
> > >
> > > case AES_GMAC:
> > > if (session->cipher.mode == GCM) {
> > > - job->cipher_start_src_offset_in_bytes =
> > > - op->sym->aead.data.offset;
> > > job->hash_start_src_offset_in_bytes =
> > > op->sym->aead.data.offset;
> > > - job->msg_len_to_cipher_in_bytes =
> > > - op->sym->aead.data.length;
> > > job->msg_len_to_hash_in_bytes =
> > > op->sym->aead.data.length;
> > > } else {
> > > - job->cipher_start_src_offset_in_bytes =
> > > - op->sym->auth.data.offset;
> > > - job->hash_start_src_offset_in_bytes =
> > > - op->sym->auth.data.offset;
> > > job->msg_len_to_cipher_in_bytes = 0;
> > > job->msg_len_to_hash_in_bytes = 0;
> > > }
> > > @@ -1453,10 +1448,7 @@ set_mb_job_params(JOB_AES_HMAC *job, struct
> > > aesni_mb_qp *qp,
> > >
> > >
> > > #if IMB_VERSION(0, 54, 3) <= IMB_VERSION_NUM
> > > case IMB_AUTH_CHACHA20_POLY1305:
> > > - job->cipher_start_src_offset_in_bytes = op->sym-
> > >aead.data.offset;
> > > job->hash_start_src_offset_in_bytes = op->sym-
> > >aead.data.offset;
> > > - job->msg_len_to_cipher_in_bytes =
> > > - op->sym->aead.data.length;
> > > job->msg_len_to_hash_in_bytes =
> > > op->sym->aead.data.length;
> > >
> > >
> > > @@ -1464,26 +1456,85 @@ set_mb_job_params(JOB_AES_HMAC *job, struct
> > aesni_mb_qp *qp,
> > > session->iv.offset);
> > > break;
> > > #endif
> > > - default:
> > > - /* For SNOW3G, length and offsets are already in bits */
> > > - job->cipher_start_src_offset_in_bytes =
> > > - op->sym->cipher.data.offset;
> > > - job->msg_len_to_cipher_in_bytes = op->sym-
> > >cipher.data.length;
> > > +#if IMB_VERSION(0, 53, 3) <= IMB_VERSION_NUM
> > > + /* ZUC and SNOW3G require length in bits and offset in bytes */
> > > + case IMB_AUTH_ZUC_EIA3_BITLEN:
> > > + case IMB_AUTH_SNOW3G_UIA2_BITLEN:
> > > + auth_off_in_bytes = op->sym->auth.data.offset >> 3;
> > > + ciph_off_in_bytes = op->sym->cipher.data.offset >> 3;
> > > + auth_len_in_bytes = op->sym->auth.data.length >> 3;
> > > + ciph_len_in_bytes = op->sym->cipher.data.length >> 3;
> > > +
> > > + job->hash_start_src_offset_in_bytes = auth_start_offset(op,
> > > + session, oop, auth_off_in_bytes,
> > > + ciph_off_in_bytes, auth_len_in_bytes,
> > > + ciph_len_in_bytes);
> > > + job->msg_len_to_hash_in_bits = op->sym->auth.data.length;
> > > +
> > > + job->iv = rte_crypto_op_ctod_offset(op, uint8_t *,
> > > + session->iv.offset);
> > > + break;
> > > +
> > > + /* KASUMI requires lengths and offset in bytes */
> > > + case IMB_AUTH_KASUMI_UIA1:
> > > + auth_off_in_bytes = op->sym->auth.data.offset >> 3;
> > > + ciph_off_in_bytes = op->sym->cipher.data.offset >> 3;
> > > + auth_len_in_bytes = op->sym->auth.data.length >> 3;
> > > + ciph_len_in_bytes = op->sym->cipher.data.length >> 3;
> > >
> > >
> > > job->hash_start_src_offset_in_bytes = auth_start_offset(op,
> > > - session, oop);
> > > + session, oop, auth_off_in_bytes,
> > > + ciph_off_in_bytes, auth_len_in_bytes,
> > > + ciph_len_in_bytes);
> > > + job->msg_len_to_hash_in_bytes = auth_len_in_bytes;
> > > +
> > > + job->iv = rte_crypto_op_ctod_offset(op, uint8_t *,
> > > + session->iv.offset);
> > > + break;
> > > +#endif
> > > +
> > > + default:
> > > + job->hash_start_src_offset_in_bytes = auth_start_offset(op,
> > > + session, oop, op->sym->auth.data.offset,
> > > + op->sym->cipher.data.offset,
> > > + op->sym->auth.data.length,
> > > + op->sym->cipher.data.length);
> > > job->msg_len_to_hash_in_bytes = op->sym->auth.data.length;
> > >
> > >
> > > job->iv = rte_crypto_op_ctod_offset(op, uint8_t *,
> > > session->iv.offset);
> > > }
> > >
> > >
> > > + switch (job->cipher_mode) {
> > > #if IMB_VERSION(0, 53, 3) <= IMB_VERSION_NUM
> > > - if (job->cipher_mode == IMB_CIPHER_ZUC_EEA3)
> > > - job->msg_len_to_cipher_in_bytes >>= 3;
> > > - else if (job->hash_alg == IMB_AUTH_KASUMI_UIA1)
> > > - job->msg_len_to_hash_in_bytes >>= 3;
> > > + /* ZUC requires length and offset in bytes */
> > > + case IMB_CIPHER_ZUC_EEA3:
> > > + job->cipher_start_src_offset_in_bytes =
> > > + op->sym->cipher.data.offset >> 3;
> > > + job->msg_len_to_cipher_in_bytes =
> > > + op->sym->cipher.data.length >> 3;
> > > + break;
> > > + /* ZUC and SNOW3G require length and offset in bits */
> > > + case IMB_CIPHER_SNOW3G_UEA2_BITLEN:
> > > + case IMB_CIPHER_KASUMI_UEA1_BITLEN:
> > > + job->cipher_start_src_offset_in_bits =
> > > + op->sym->cipher.data.offset;
> > > + job->msg_len_to_cipher_in_bits =
> > > + op->sym->cipher.data.length;
> > > + break;
> > > #endif
> > > + case IMB_CIPHER_CCM:
> > > + case IMB_CIPHER_GCM:
> > > + case IMB_CIPHER_CHACHA20_POLY1305:
> > > + job->cipher_start_src_offset_in_bytes =
> > > + op->sym->aead.data.offset;
> > > + job->msg_len_to_cipher_in_bytes = op->sym-
> > >aead.data.length;
> > > + break;
> > > + default:
> > > + job->cipher_start_src_offset_in_bytes =
> > > + op->sym->cipher.data.offset;
> > > + job->msg_len_to_cipher_in_bytes = op->sym-
> > >cipher.data.length;
> > > + }
> > >
> > >
> > > /* Set user data to be crypto operation data struct */
> > > job->user_data = op;
> >
> > This breaks the build on Ubuntu 20.04:
> >
> > [ 1944s] ../drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c: In function
> > ‘set_mb_job_params’:
> > [ 1944s] ../drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c:1526:7: error:
> > ‘IMB_CIPHER_GCM’ undeclared (first use in this function) [ 1944s] 1526 | case
> > IMB_CIPHER_GCM:
> > [ 1944s] | ^~~~~~~~~~~~~~
> > [ 1944s] ../drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c:1526:7: note: each
> > undeclared identifier is reported only once for each function it appears in [
> > 1944s] ../drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c:1527:31: error:
> > ‘IMB_CIPHER_NULL’ undeclared (first use in this function)
> > [ 1944s] 1527 | if (session->cipher.mode == IMB_CIPHER_NULL) {
> > [ 1944s] | ^~~~~~~~~~~~~~~
> > [ 1944s] ../drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c:1538:7: error:
> > ‘IMB_CIPHER_CCM’ undeclared (first use in this function) [ 1944s] 1538 | case
> > IMB_CIPHER_CCM:
> > [ 1944s] | ^~~~~~~~~~~~~~
> > [ 1944s] ../drivers/crypto/aesni_mb/rte_aesni_mb_pmd.c:1539:7: error:
> > ‘IMB_CIPHER_CHACHA20_POLY1305’ undeclared (first use in this function); did
> > you mean ‘RTE_CRYPTO_AEAD_CHACHA20_POLY1305’?
> > [ 1944s] 1539 | case IMB_CIPHER_CHACHA20_POLY1305:
> > [ 1944s] | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > [ 1944s] | RTE_CRYPTO_AEAD_CHACHA20_POLY1305
> >
> > https://build.opensuse.org/package/live_build_log/home:bluca:dpdk/dpdk-
> > 20.11/Ubuntu_20.04/x86_64
>
> Apologies, I did not see this till today. I don't think this patch is actually breaking the build.
> To me, it looks like an environment issue (IMB_CIPHER_GCM for instance is already present in the code).
> I wonder if that system has the right version of IPSec Multi buffer library.
> Only versions from v1.0 onwards are supported.
>
> Thanks,
> Pablo
In 20.11 ipsec-mb v0.52.0 and over are supported:
https://git.dpdk.org/dpdk-stable/tree/drivers/crypto/aesni_mb/meson.build?h=20.11#n4
If this patch requires a new version and breaks backward
compatibility, then it does not appear to be not suitable for
backporting.
next prev parent reply other threads:[~2022-03-21 22:12 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-14 11:05 Pablo de Lara
2022-03-14 11:05 ` [PATCH 20.11 2/2] crypto/ipsec_mb: fix GMAC parameters setting Pablo de Lara
2022-03-14 12:20 ` [PATCH 20.11 1/2] crypto/ipsec_mb: fix length and offset settings Luca Boccassi
2022-03-21 19:57 ` De Lara Guarch, Pablo
2022-03-21 22:12 ` Luca Boccassi [this message]
2022-03-22 10:39 ` Kevin Traynor
2022-03-22 12:25 ` De Lara Guarch, Pablo
2022-03-15 11:10 ` Kevin Traynor
2022-03-15 11:18 ` Kevin Traynor
2022-03-21 20:02 ` De Lara Guarch, Pablo
2022-03-22 10:39 ` Kevin Traynor
2022-03-22 13:39 Pablo de Lara
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='CAMw=ZnRV+ut6+jktLjjz8ftJ4x--USBnr6WH9qySoMw1tkbVCw@mail.gmail.com' \
--to=bluca@debian.org \
--cc=pablo.de.lara.guarch@intel.com \
--cc=stable@dpdk.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).