By using a regular non-overflow-checking add, the MediaTek icc-emi
driver will happy wrap at U32_MAX + 1 to 0. As it's common for the
interconnect core to fill in INT_MAX values, this is not a hypothetical
situation, but something that actually happens in regular use. This
would be pretty disasterous if anything used this driver.

Replace the addition with an overflow-checked addition from overflow.h,
and saturate to U32_MAX if an overflow is detected.

Fixes: b45293799f75 ("interconnect: mediatek: Add MediaTek MT8183/8195 EMI Interconnect driver")
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
---
 drivers/interconnect/mediatek/icc-emi.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/interconnect/mediatek/icc-emi.c b/drivers/interconnect/mediatek/icc-emi.c
index 182aa2b0623a..dfa3a9cd9399 100644
--- a/drivers/interconnect/mediatek/icc-emi.c
+++ b/drivers/interconnect/mediatek/icc-emi.c
@@ -12,6 +12,7 @@
 #include <linux/module.h>
 #include <linux/of.h>
 #include <linux/of_platform.h>
+#include <linux/overflow.h>
 #include <linux/platform_device.h>
 #include <linux/soc/mediatek/dvfsrc.h>
 
@@ -22,7 +23,9 @@ static int mtk_emi_icc_aggregate(struct icc_node *node, u32 tag, u32 avg_bw,
 {
 	struct mtk_icc_node *in = node->data;
 
-	*agg_avg += avg_bw;
+	if (check_add_overflow(*agg_avg, avg_bw, agg_avg))
+		*agg_avg = U32_MAX;
+
 	*agg_peak = max_t(u32, *agg_peak, peak_bw);
 
 	in->sum_avg = *agg_avg;

-- 
2.52.0