[PATCH] octeontx2-af: Fix possible buffer overflow

[PATCH] octeontx2-af: Fix possible buffer overflow

[Elena Salomatkina]

A loop in rvu_mbox_handler_nix_bandprof_free() contains
a break if (idx == MAX_BANDPROF_PER_PFFUNC),
but if idx may reach MAX_BANDPROF_PER_PFFUNC
buffer '(*req->prof_idx)[layer]' overflow happens before that check.

The patch moves the break to the
beginning of the loop.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: e8e095b3b370 ("octeontx2-af: cn10k: Bandwidth profiles config support").
Signed-off-by: Elena Salomatkina <elena.salomatkina.cmc@gmail.com>
---
 drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
index 23c2f2ed2fb8..c112c71ff576 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
@@ -5505,6 +5505,8 @@ int rvu_mbox_handler_nix_bandprof_free(struct rvu *rvu,
 
 		ipolicer = &nix_hw->ipolicer[layer];
 		for (idx = 0; idx < req->prof_count[layer]; idx++) {
+			if (idx == MAX_BANDPROF_PER_PFFUNC)
+				break;
 			prof_idx = req->prof_idx[layer][idx];
 			if (prof_idx >= ipolicer->band_prof.max ||
 			    ipolicer->pfvf_map[prof_idx] != pcifunc)
@@ -5518,8 +5520,6 @@ int rvu_mbox_handler_nix_bandprof_free(struct rvu *rvu,
 			ipolicer->pfvf_map[prof_idx] = 0x00;
 			ipolicer->match_id[prof_idx] = 0;
 			rvu_free_rsrc(&ipolicer->band_prof, prof_idx);
-			if (idx == MAX_BANDPROF_PER_PFFUNC)
-				break;
 		}
 	}
 	mutex_unlock(&rvu->rsrc_lock);
-- 
2.34.1

Re: [PATCH] octeontx2-af: Fix possible buffer overflow

[Simon Horman]

On Thu, Nov 23, 2023 at 08:36:30PM +0300, Elena Salomatkina wrote:
> A loop in rvu_mbox_handler_nix_bandprof_free() contains
> a break if (idx == MAX_BANDPROF_PER_PFFUNC),
> but if idx may reach MAX_BANDPROF_PER_PFFUNC
> buffer '(*req->prof_idx)[layer]' overflow happens before that check.
> 
> The patch moves the break to the
> beginning of the loop.
> 
> Found by Linux Verification Center (linuxtesting.org) with SVACE.
> 
> Fixes: e8e095b3b370 ("octeontx2-af: cn10k: Bandwidth profiles config support").
> Signed-off-by: Elena Salomatkina <elena.salomatkina.cmc@gmail.com>

Thanks Elena,

I agree with your analysis and that this seems to be
an appropriate fix for the problem.

As this is a fix, it should be targeted at the net, as opposed to net-next,
tree.  Please keep this in mind for future patch submissions.

	Subject: [PATCH net] ...

Link https://docs.kernel.org/process/maintainer-netdev.html

The above nit notwithstanding,

Reviewed-by: Simon Horman <horms@kernel.org>