Hi Thomas,
As stated in all cryptodev guides for the PMDs based on intel-ipsec-mb,
the minimum dependent intel-ipsec-mb lib version is bumped to 1.0.
Could you bump the library version and try again?
Regards,
Fan
> -----Original Message-----
> From: Akhil Goyal <gak...@marvell.com>
> Sent: Wednesday, October 20, 2021 5:24 AM
> To: Thomas Monjalon <tho...@monjalon.net>; Power, Ciara
> <ciara.po...@intel.com>; De Lara Guarch, Pablo
> <pablo.de.lara.gua...@intel.com>; Mcnamara, John
> <john.mcnam...@intel.com>
> Cc: dev@dpdk.org; Zhang, Roy Fan <roy.fan.zh...@intel.com>; Bronowski,
> PiotrX <piotrx.bronow...@intel.com>; m...@ashroe.eu;
> david.march...@redhat.com
> Subject: RE: [dpdk-dev] [EXT] [PATCH v4 00/14] drivers/crypto: introduce
> ipsec_mb framework
>
> > 18/10/2021 17:21, Akhil Goyal:
> > > > This set of patches introduces a new framework, making all common
> > code of
> > > > SW crypto PMD implementations built on top of intel-ipsec-mb library
> > > > sharable. This helps to reduce future effort on the code maintenance
> and
> > > > future updates. It also moves all SW PMD implementation specific
> details
> > > > into single files located in the crypto/ipsec_mb folder.
> > > > A CHACHA20_POLY1305 SW PMD is added based on this framework.
> > > >
> > > > Multi-process support for the PMDs is added for intel-ipsec-mb v1.1.
> > > > The minimum intel-ipsec-mb version required is bumped to 1.0.
> > > >
> > > > ZUC-256 support is added for the aesni_mb PMD, with relevant tests.
> > > >
> > > > v4:
> > >
> > > Acked-by: Akhil Goyal <gak...@marvell.com>
> > > Patches are rebased over TOT of next-crypto
> > > Release notes are updated
> > > Applied to dpdk-next-crypto
> >
> > I think compilation has not been tested.
>
> I am not sure why this is failing at your end,
> On my machine, it is getting compiled with intel-ipsec-mb v1.0.
> I am double checking compilation for all the individual patches as well.
> Will inform you once it is completed.
> On TOT of next crypto, I did a quick test touched the pmd_zuc.c and it is
> getting compiled.
> cavium@cavium-DT13:~/up/dpdk-next-crypto$ touch
> drivers/crypto/ipsec_mb/pmd_zuc.c
> cavium@cavium-DT13:~/up/dpdk-next-crypto$ ./devtools/test-meson-
> builds.sh
> ninja: Entering directory `./build-gcc-static'
> [24/24] Linking target app/test/dpdk-test.
> ninja: Entering directory `./build-gcc-shared'
> [8/8] Linking target drivers/librte_crypto_ipsec_mb.so.22.0.
> ninja: Entering directory `./build-clang-static'
> [24/24] Linking target app/test/dpdk-test.
> ninja: Entering directory `./build-clang-shared'
> [8/8] Linking target drivers/librte_crypto_ipsec_mb.so.22.0.
> ninja: Entering directory `./build-x86-generic'
> [9/9] Linking target buildtools/chkincs/chkincs.
> ninja: Entering directory `./build-x86-mingw'
> ninja: no work to do.
>
> > You need to update intel-ipsec-mb to v1.0.
> > Result:
> >
> > drivers/crypto/ipsec_mb/pmd_zuc.c:176:33: error: ‘hash_keys’ may be
> used
> > uninitialized [-Werror=maybe-uninitialized]/aesni/intel-ipsec-mb/lib/intel-
> > ipsec-mb.h:1444:11: note: in definition of macro
> ‘IMB_ZUC_EIA3_N_BUFFER’
> > 1444 | ((_mgr)->eia3_n_buffer((_key), (_iv), (_in), (_len), (_tag),
> > (_num)))
> > | ^~~~
> > drivers/crypto/ipsec_mb/pmd_zuc.c:176:33: note: by argument 1 of type
> > ‘const void * const*’ to ‘void(const void * const*, const void * const*,
> const
> > void * const*, const uint32_t *, uint32_t **, const uint32_t)’ {aka
> ‘void(const
> > void * const*, const void * const*, const void * const*, const unsigned int
> > *,
> > unsigned int **, const unsigned int)’}
> > /aesni/intel-ipsec-mb/lib/intel-ipsec-mb.h:1444:11: note: in definition of
> > macro ‘IMB_ZUC_EIA3_N_BUFFER’
> > 1444 | ((_mgr)->eia3_n_buffer((_key), (_iv), (_in), (_len), (_tag),
> > (_num)))
> > | ^~~~
> > drivers/crypto/ipsec_mb/pmd_zuc.c:145:21: note: ‘hash_keys’ declared
> here
> > 145 | const void *hash_keys[ZUC_MAX_BURST];
> > | ^~~~~~~~~
> > In file included from
> > ../../dpdk/drivers/crypto/ipsec_mb/ipsec_mb_private.h:8,
> > from ../../dpdk/drivers/crypto/ipsec_mb/pmd_zuc_priv.h:8,
> > from ../../dpdk/drivers/crypto/ipsec_mb/pmd_zuc.c:5:
> > drivers/crypto/ipsec_mb/pmd_zuc.c:176:33: error: ‘iv’ may be used
> > uninitialized [-Werror=maybe-uninitialized]
> > 176 | IMB_ZUC_EIA3_N_BUFFER(qp->mb_mgr, (const void
> > **)hash_keys,
> > /aesni/intel-ipsec-mb/lib/intel-ipsec-mb.h:1444:11: note: in definition of
> > macro ‘IMB_ZUC_EIA3_N_BUFFER’
> > 1444 | ((_mgr)->eia3_n_buffer((_key), (_iv), (_in), (_len), (_tag),
> > (_num)))
> > | ^~~~
> > drivers/crypto/ipsec_mb/pmd_zuc.c:176:33: note: by argument 2 of type
> > ‘const void * const*’ to ‘void(const void * const*, const void * const*,
> const
> > void * const*, const uint32_t *, uint32_t **, const uint32_t)’ {aka
> ‘void(const
> > void * const*, const void * const*, const void * const*, const unsigned int
> > *,
> > unsigned int **, const unsigned int)’}
> > /aesni/intel-ipsec-mb/lib/intel-ipsec-mb.h:1444:11: note: in definition of
> > macro ‘IMB_ZUC_EIA3_N_BUFFER’
> > 1444 | ((_mgr)->eia3_n_buffer((_key), (_iv), (_in), (_len), (_tag),
> > (_num)))
> > | ^~~~
> > drivers/crypto/ipsec_mb/pmd_zuc.c:144:18: note: ‘iv’ declared here
> > 144 | uint8_t *iv[ZUC_MAX_BURST];
> > | ^~
> > In file included from
> > ../../dpdk/drivers/crypto/ipsec_mb/ipsec_mb_private.h:8,
> > from ../../dpdk/drivers/crypto/ipsec_mb/pmd_zuc_priv.h:8,
> > from ../../dpdk/drivers/crypto/ipsec_mb/pmd_zuc.c:5:
> > drivers/crypto/ipsec_mb/pmd_zuc.c:176:33: error: ‘src’ may be used
> > uninitialized [-Werror=maybe-uninitialized]
> > 176 | IMB_ZUC_EIA3_N_BUFFER(qp->mb_mgr, (const void
> > **)hash_keys,
> > /aesni/intel-ipsec-mb/lib/intel-ipsec-mb.h:1444:11: note: in definition of
> > macro ‘IMB_ZUC_EIA3_N_BUFFER’
> > 1444 | ((_mgr)->eia3_n_buffer((_key), (_iv), (_in), (_len), (_tag),
> > (_num)))
> > | ^~~~
> > ../../dpdk/d(const void * const*, const void * const*, const void * const*,
> > const uint32_t *, uint32_t **, const uint32_t)’ {aka ‘void(const void *
> const*,
> > const void * const*, const void * const*, const unsigned int *, unsigned int
> > **, const unsigned int)’}
> > 176 | IMB_ZUC_EIA3_N_BUFFER(qp->mb_mgr, (const void
> > **)hash_keys,
> > /aesni/intel-ipsec-mb/lib/intel-ipsec-mb.h:1444:11: note: in definition of
> > macro ‘IMB_ZUC_EIA3_N_BUFFER’
> > 1444 | ((_mgr)->eia3_n_buffer((_key), (_iv), (_in), (_len), (_tag),
> > (_num)))
> > | ^~~~
> > drivers/crypto/ipsec_mb/pmd_zuc.c:141:18: note: ‘src’ declared here
> > 141 | uint8_t *src[ZUC_MAX_BURST];
> > | ^~~
> > In file included from
> > ../../dpdk/drivers/crypto/ipsec_mb/ipsec_mb_private.h:8,
> > from ../../dpdk/drivers/crypto/ipsec_mb/pmd_zuc_priv.h:8,
> > from ../../dpdk/drivers/crypto/ipsec_mb/pmd_zuc.c:5:
> > drivers/crypto/ipsec_mb/pmd_zuc.c:176:33: error: ‘length_in_bits’ may be
> > used uninitialized [-Werror=maybe-uninitialized]
> > 176 | IMB_ZUC_EIA3_N_BUFFER(qp->mb_mgr, (const void
> > **)hash_keys,
> > /aesni/intel-ipsec-mb/lib/intel-ipsec-mb.h:1444:11: note: in definition of
> > macro ‘IMB_ZUC_EIA3_N_BUFFER’
> > 1444 | ((_mgr)->eia3_n_buffer((_key), (_iv), (_in), (_len), (_tag),
> > (_num)))
> > | ^~~~
> > drivers/crypto/ipsec_mb/pmd_zuc.c:176:33: note: by argument 4 of type
> > ‘const uint32_t *’ {aka ‘const unsigned int *’} to ‘void(const void *
> > const*,
> > const void * const*, const void * const*, const uint32_t *, uint32_t **,
> const
> > uint32_t)’ {aka ‘void(const void * const*, const void * const*, const void *
> > const*, const unsigned int *, unsigned int **, const unsigned int)’}
> > 176 | IMB_ZUC_EIA3_N_BUFFER(qp->mb_mgr, (const void
> > **)hash_keys,
> > /aesni/intel-ipsec-mb/lib/intel-ipsec-mb.h:1444:11: note: in definition of
> > macro ‘IMB_ZUC_EIA3_N_BUFFER’
> > 1444 | ((_mgr)->eia3_n_buffer((_key), (_iv), (_in), (_len), (_tag),
> > (_num)))
> > | ^~~~
> > drivers/crypto/ipsec_mb/pmd_zuc.c:143:18: note: ‘length_in_bits’
> declared
> > here
> > 143 | uint32_t length_in_bits[ZUC_MAX_BURST];
> > | ^~~~~~~~~~~~~~
> > cc1: all warnings being treated as errors
> >
> > I think I will just drop all these patches from 21.11-rc1.
> >
> > Not sure I will accept any new revision, given how I am annoyed recently
> > by such bad work in Intel.
> >
