Hi Muneendra, > -----Original Message----- > From: Muneendra Kumar M [mailto:muneendra.kumar@xxxxxxxxxxxx] > Sent: Friday, February 23, 2018 3:04 AM > To: Parav Pandit <parav@xxxxxxxxxxxx>; linux-rdma@xxxxxxxxxxxxxxx > Subject: RE: [PATCH] Bug[RDMA/core]:Null pointer check is missing in > addr_resolve > > Hi Parav, > Thanks for your feedback. > As you mentioned below since addr4_resolve() or addr6_resolve() doesn't fail > ,ndev is not NULL in mycase. > And the below simpler patch which was provided by you holds good. > o.k. good. > I will update the patch and will resend the same. > Will it be ok? It looks good to me for fixing this particular crash. Few things to take care in v1 version that you will send. 1. Please send patch using git send-mail option, so that there is only one email and that also avoids patch attachment. 2. Please add fixes line below before Sign-off so that it can be back ported to older kernels. Fixes: 200298326b27 ("IB/core: Validate route when we init ah") On a side note, I am sending similar fix in ib_init_ah_attr_from_path() which surfaces from same patch. It is already under review internally and Leon will push it soon. 3. Rewrite the commit description without the code in it to explain the cause and this fix. 4. If possible, please describe the test (IPv4/6, application) how this happens so that when others see the error, they can relate to this issue and cherry pick this fix. 5. Send v1 patch with prefix of [PATCH for-rc] instead of [PATCH] ��.n��������+%������w��{.n�����{���fk��ܨ}���Ơz�j:+v�����w����ޙ��&�)ߡ�a����z�ޗ���ݢj��w�f
