[EXT] [PATCH v2] cryptodev: avoid algorithm strings null pointers

[Akhil Goyal]

> -----Original Message-----
> From: liu xixin <liuxixin2020 at outlook.com>
> Sent: Friday, June 9, 2023 12:31 PM
> To: Akhil Goyal <gakhil at marvell.com>; dev at dpdk.org
> Subject: 答复: [EXT] [PATCH v2] cryptodev: avoid algorithm strings null pointers
> 
> > Subject: [EXT] [PATCH v2] cryptodev: avoid algorithm strings null
> > pointers
> >
> > The crypto algorithm strings identifiers that are Continuous may be
> > null, so there is needed to add null judgment.
> > When testing with dpdk-test-crypto-perf and passing in the parameter
> > --auth-algo sm3-hmac, The program caused a segfault due to a null
> > pointer passed in by strcmp.
> > Adding this patch can solve the segfault problem.
> 
> I believe this is a fix and you should add fixes tag for this and need to be
> backported.
> 
> >
> > Signed-off-by: xixin.liu <liuxixin2020 at outlook.com>
> 
> Signoff format is not correct.
> Please follow https://urldefense.proofpoint.com/v2/url?u=https-
> 3A__doc.dpdk.org_guides_contributing_patches.html&d=DwIGoQ&c=nKjWec2b
> 6R0mOyPaz7xtfQ&r=DnL7Si2wl_PRwpZ9TWey3eu68gBzn7DkPwuqhd6WNyo&m
> =6mzvm07Go5-
> S_2jVk9KRF1mOl0Juay2Sa2WI2XiNgTbg6ZhMcm75GNceSVhe0Doj&s=tvqhxvElc
> m_8h3e7YIBym6IAHk6BxUAFx2RKmjJ6Ibw&e=
> 
> 
> > ---
> >  lib/cryptodev/rte_cryptodev.c | 8 ++++++++
> >  1 file changed, 8 insertions(+)
> >
> > diff --git a/lib/cryptodev/rte_cryptodev.c
> > b/lib/cryptodev/rte_cryptodev.c index a96114b2da..41c23fc596 100644
> > --- a/lib/cryptodev/rte_cryptodev.c
> > +++ b/lib/cryptodev/rte_cryptodev.c
> > @@ -346,6 +346,8 @@ rte_cryptodev_get_cipher_algo_enum(enum
> > rte_crypto_cipher_algorithm *algo_enum,
> >  	int ret = -1;	/* Invalid string */
> >
> >  	for (i = 1; i < RTE_DIM(crypto_cipher_algorithm_strings); i++) {
> > +		if (crypto_cipher_algorithm_strings[i] == NULL)
> > +			continue;
> 
> crypto_cipher_algorithm_strings is a fixed size array with all non-NULL known
> values and the for loop is iterating over it. So, this check does not make sense to
> me.
> ----> Not every element of the array is defined, eg. it is NULL that the first
> element [1],  if not check "strcmp(algo_string,
> crypto_cipher_algorithm_strings[i]" will fail

The loop starts from 1 so 0th element = NULL will not matter.
And for all other values it is not null. If something is missing,
then it can be added.
But this check is not needed.