Hi Colin,
I have submitted the patch fixing the reported issue to net-next branch.
Thank you,
Geetha.
________________________________________
From: Colin Ian King <colin.k...@canonical.com>
Sent: Monday, February 15, 2021 10:22 PM
To: Hariprasad Kelam
Cc: Sunil Kovvuri Goutham; Linu Cherian; Geethasowjanya Akula; Jerin Jacob
Kollanukkaran; Hariprasad Kelam; Subbaraya Sundeep Bhatta;
netdev@vger.kernel.org; linux-ker...@vger.kernel.org
Subject: [EXT] re: octeontx2-af: cn10k: MAC internal loopback support
External Email
----------------------------------------------------------------------
Hi,
Static analysis on linux-next today using Coverity found an issue in the
following commit:
commit 3ad3f8f93c81f81d6e28b2e286b03669cc1fb3b0
Author: Hariprasad Kelam <hke...@marvell.com>
Date: Thu Feb 11 21:28:34 2021 +0530
octeontx2-af: cn10k: MAC internal loopback support
The analysis is as follows:
723 static int rvu_cgx_config_intlbk(struct rvu *rvu, u16 pcifunc, bool en)
724 {
725 struct mac_ops *mac_ops;
1. var_decl: Declaring variable lmac_id without initializer.
726 u8 cgx_id, lmac_id;
727
2. Condition !is_cgx_config_permitted(rvu, pcifunc), taking false branch.
728 if (!is_cgx_config_permitted(rvu, pcifunc))
729 return -EPERM;
730
Uninitialized scalar variable (UNINIT)
731 mac_ops = get_mac_ops(rvu_cgx_pdata(cgx_id, rvu));
732
Uninitialized scalar variable (UNINIT)
3. uninit_use_in_call: Using uninitialized value lmac_id when calling
*mac_ops->mac_lmac_intl_lbk.
733 return mac_ops->mac_lmac_intl_lbk(rvu_cgx_pdata(cgx_id, rvu),
734 lmac_id, en);
735 }
Variables cgx_id and lmac_id are no longer being initialized and garbage
values are being passed into function calls. Originally, these
variables were being initialized with a call to rvu_get_cgx_lmac_id()
but that has now been removed.
Colin
