[PATCH AUTOSEL 5.13 184/189] flow_offload: action should not be NULL when it is referenced

From: gushengxian <redacted>

[ Upstream commit 9ea3e52c5bc8bb4a084938dc1e3160643438927a ]

"action" should not be NULL when it is referenced.

Signed-off-by: gushengxian <redacted>
Signed-off-by: gushengxian <redacted>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 include/net/flow_offload.h | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
index dc5c1e69cd9f..69c9eabf8325 100644
--- a/include/net/flow_offload.h
+++ b/include/net/flow_offload.h

@@ -319,12 +319,14 @@ flow_action_mixed_hw_stats_check(const struct flow_action *action,
 	if (flow_offload_has_one_action(action))
 		return true;
 
-	flow_action_for_each(i, action_entry, action) {
-		if (i && action_entry->hw_stats != last_hw_stats) {
-			NL_SET_ERR_MSG_MOD(extack, "Mixing HW stats types for actions is not supported");
-			return false;
+	if (action) {
+		flow_action_for_each(i, action_entry, action) {
+			if (i && action_entry->hw_stats != last_hw_stats) {
+				NL_SET_ERR_MSG_MOD(extack, "Mixing HW stats types for actions is not supported");
+				return false;
+			}
+			last_hw_stats = action_entry->hw_stats;
 		}
-		last_hw_stats = action_entry->hw_stats;
 	}
 	return true;
 }

-- 
2.30.2

On 7/6/21 1:14 PM, Sasha Levin wrote:

This patch makes no sense really.

If action is NULL, a crash would happen earlier anyway in

if (flow_offload_has_one_action(action))
    return true;

Also, I wonder why it has been backported to stable version,
there was no Fixes: tag in the submission.

On Fri, Aug 06, 2021 at 11:30:04AM +0200, Eric Dumazet wrote:

On 7/6/21 1:14 PM, Sasha Levin wrote:

quoted

From: gushengxian <redacted>

[ Upstream commit 9ea3e52c5bc8bb4a084938dc1e3160643438927a ]

"action" should not be NULL when it is referenced.

Signed-off-by: gushengxian <redacted>
Signed-off-by: gushengxian <redacted>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 include/net/flow_offload.h | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
index dc5c1e69cd9f..69c9eabf8325 100644
--- a/include/net/flow_offload.h
+++ b/include/net/flow_offload.h

@@ -319,12 +319,14 @@ flow_action_mixed_hw_stats_check(const struct flow_action *action,
 	if (flow_offload_has_one_action(action))
 		return true;

-	flow_action_for_each(i, action_entry, action) {
-		if (i && action_entry->hw_stats != last_hw_stats) {
-			NL_SET_ERR_MSG_MOD(extack, "Mixing HW stats types for actions is not supported");
-			return false;
+	if (action) {
+		flow_action_for_each(i, action_entry, action) {
+			if (i && action_entry->hw_stats != last_hw_stats) {
+				NL_SET_ERR_MSG_MOD(extack, "Mixing HW stats types for actions is not supported");
+				return false;
+			}
+			last_hw_stats = action_entry->hw_stats;
 		}
-		last_hw_stats = action_entry->hw_stats;
 	}
 	return true;
 }

This patch makes no sense really.

If action is NULL, a crash would happen earlier anyway in

if (flow_offload_has_one_action(action))
   return true;

Also, I wonder why it has been backported to stable version,
there was no Fixes: tag in the submission.

That's the AUTOSEL logic: it will attempt to find fixes in commits that
don't have a fixes: tag. Those get longer review times and aren't
backported as quickly as commits with a stable tag or a fixes tag.

-- 
Thanks,
Sasha