Re: [dpdk-dev] [PATCH 2/3] examples/ipsec_secgw: fix possible NULL dereference

[Akhil Goyal <akhil.goyal@nxp.com>]

On 3/27/2019 6:49 PM, Ananyev, Konstantin wrote:
> Hi Akhil,
>
>>> Fixes: 3e5f4625dc17 ("examples/ipsec-secgw: make data-path to use IPsec library")
>>> Coverity issue: 336844
>>>
>>> Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
>>> ---
>>>    examples/ipsec-secgw/ipsec_process.c | 11 ++++-------
>>>    1 file changed, 4 insertions(+), 7 deletions(-)
>>>
>>> diff --git a/examples/ipsec-secgw/ipsec_process.c b/examples/ipsec-secgw/ipsec_process.c
>>> index e403c461a..3f9cacb8f 100644
>>> --- a/examples/ipsec-secgw/ipsec_process.c
>>> +++ b/examples/ipsec-secgw/ipsec_process.c
>>> @@ -217,16 +217,11 @@ ipsec_process(struct ipsec_ctx *ctx, struct ipsec_traffic *trf)
>>>    		pg = grp + i;
>>>    		sa = pg->id.ptr;
>>>
>>> -		/* no valid SA found */
>>> -		if (sa == NULL)
>>> -			k = 0;
>>> -
>>>    		ips = &sa->ips;
>> I think this is not fixing the NULL dereference properly. This line
>> would give fault if sa is null.
> I don't think it would - here we just get an address of ips,
> we don't try to access it.
> Konstantin
ok got it.. I missed the '&'..
>
>>> -		satp = rte_ipsec_sa_type(ips->sa);
>>>
>>>    		/* no valid HW session for that SA, try to create one */
>>> -		if (ips->crypto.ses == NULL &&
>>> -				fill_ipsec_session(ips, ctx, sa) != 0)
>>> +		if (sa == NULL || (ips->crypto.ses == NULL &&
>>> +				fill_ipsec_session(ips, ctx, sa) != 0))
>>>    			k = 0;
>>>
>>>    		/* process packets inline */
>>> @@ -234,6 +229,8 @@ ipsec_process(struct ipsec_ctx *ctx, struct ipsec_traffic *trf)
>>>    				sa->type ==
>>>    				RTE_SECURITY_ACTION_TYPE_INLINE_PROTOCOL) {
>>>
>>> +			satp = rte_ipsec_sa_type(ips->sa);
>>> +
>>>    			/*
>>>    			 * This is just to satisfy inbound_sa_check()
>>>    			 * and get_hop_for_offload_pkt().