Skip to contentSkip to conversation

staging: vchiq_arm: Clear VLA warning

PATCH Superseded Tobin C. Harding 4 participants 4 messages Mar 9, 2018

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 staging: vchiq_arm: Clear VLA warning

The kernel would like to have all stack VLA usage removed[1]. The array here is fixed (declared with a const variable) but it appears like VLA to the compiler. We can use a pre-processor define to quiet the compiler.

[1]: https://lkml.org/lkml/2018/3/7/621

---

The name of this constant may need changing, there is already a pre-processor constant VCHIQ_MAX_SERVICES

 .../staging/vc04_services/interface/vchiq_arm/vchiq_arm.c  | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c +8 -6
index f5cefda49b22..c972869a0333 100644
@@ -3434,13 +3434,15 @@ vchiq_release_service(VCHIQ_SERVICE_HANDLE_T handle)
 	return ret;
 }
 
+/* Only dump 64 services */
+#define VCHIQ_LOCAL_MAX_SERVICES 64
+
 void
 vchiq_dump_service_use_state(VCHIQ_STATE_T *state)
 {
 	VCHIQ_ARM_STATE_T *arm_state = vchiq_platform_get_arm_state(state);
 	int i, j = 0;
-	/* Only dump 64 services */
-	static const int local_max_services = 64;
+
 	/* If there's more than 64 services, only dump ones with
 	 * non-zero counts */
 	int only_nonzero = 0;
@@ -3455,7 +3457,7 @@ vchiq_dump_service_use_state(VCHIQ_STATE_T *state)
 		int fourcc;
 		int clientid;
 		int use_count;
-	} service_data[local_max_services];
+	} service_data[VCHIQ_LOCAL_MAX_SERVICES];
 
 	if (!arm_state)
 		return;
@@ -3466,10 +3468,10 @@ vchiq_dump_service_use_state(VCHIQ_STATE_T *state)
 	peer_count = arm_state->peer_use_count;
 	vc_use_count = arm_state->videocore_use_count;
 	active_services = state->unused_service;
-	if (active_services > local_max_services)
+	if (active_services > VCHIQ_LOCAL_MAX_SERVICES)
 		only_nonzero = 1;
 
-	for (i = 0; (i < active_services) && (j < local_max_services); i++) {
+	for (i = 0; (i < active_services) && (j < VCHIQ_LOCAL_MAX_SERVICES); i++) {
 		VCHIQ_SERVICE_T *service_ptr = state->services[i];
 
 		if (!service_ptr)
@@ -3499,7 +3501,7 @@ vchiq_dump_service_use_state(VCHIQ_STATE_T *state)
 		vchiq_log_warning(vchiq_susp_log_level, "Too many active "
 			"services (%d).  Only dumping up to first %d services "
 			"with non-zero use-count", active_services,
-			local_max_services);
+			VCHIQ_LOCAL_MAX_SERVICES);
 
 	for (i = 0; i < j; i++) {
 		vchiq_log_warning(vchiq_susp_log_level,
Signature · 1 line
2.7.4
#1 lore raw
Quoted 5 lines · Tobin C. Harding

The kernel would like to have all stack VLA usage removed[1]. The array here is fixed (declared with a const variable) but it appears like VLA to the compiler. We can use a pre-processor define to quiet the compiler.

Quoted 7 lines

[1]: https://lkml.org/lkml/2018/3/7/621

---

The name of this constant may need changing, there is already a pre-processor constant VCHIQ_MAX_SERVICES

Maybe just use ARRAY_SIZE(local_max_services) and not have the #define?

#2 lore raw
Quoted 9 lines · Eric Anholt
Quoted 7 lines · Tobin C. Harding

The kernel would like to have all stack VLA usage removed[1]. The array here is fixed (declared with a const variable) but it appears like VLA to the compiler. We can use a pre-processor define to quiet the compiler.

[1]: https://lkml.org/lkml/2018/3/7/621

Quoted 7 lines
Quoted 5 lines

---

The name of this constant may need changing, there is already a pre-processor constant VCHIQ_MAX_SERVICES

Maybe just use ARRAY_SIZE(local_max_services) and not have the #define?

I think you mean ARRAY_SIZE(service_data) ? In that case, yeah, it seems like a raw "64" for the array size can be used instead.

-Kees

Signature · 2 lines
Kees Cook
Pixel Security
#3 lore raw
Quoted 3 lines · Kees Cook

I think you mean ARRAY_SIZE(service_data) ? In that case, yeah, it seems like a raw "64" for the array size can be used instead.

I think 64 is much too big anyway. That's 768 bytes of stack data that is used to check stuff deep in some transfer call chain.

So that code is broken from a stack usage model anyway. If it's just a "random big number", then cut it down in size.

I don't know. The code was imported in 2013, and has seen very little change since. I'm not sure how many users it has. But while changing this, just do the stack size limitation at the same time.

Linus