Re: [PATCH] iommu/arm-smmu-v3: Fix undefined behavior in GBPA_UPDATE

From: xenia
Date: Fri Jul 01 2022 - 12:13:55 EST


On 7/1/22 17:34, Will Deacon wrote:
On Thu, Jun 30, 2022 at 09:39:59AM +0300, Xenia Ragiadakou wrote:
The expression 1 << 31 results in undefined behaviour because the type of
integer constant 1 is (signed) int and the result of shifting 1 by 31 bits
is not representable in the (signed) int type.

Change the type of 1 to unsigned int by adding the U suffix.

Signed-off-by: Xenia Ragiadakou <burzalodowa@xxxxxxxxx>
---
drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h
index cd48590ada30..44fbd499edea 100644
--- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h
+++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h
@@ -96,7 +96,7 @@
#define CR2_E2H (1 << 0)
#define ARM_SMMU_GBPA 0x44
-#define GBPA_UPDATE (1 << 31)
+#define GBPA_UPDATE (1U << 31)
There are loads of these kicking around in the kernel sources and we compile
with -fno-strict-overflow.

If you really want to change these, then let's use the BIT() macro instead,
but I think it's really just churn.

Will
Hi Will,

I thought that since in commit 587e6c10a7ce89a5924fdbeff2ec524fbd6a124b there was a similar fix to Q_OVERFLOW_FLAG (see below)

--- a/drivers/iommu/arm-smmu-v3.c
+++ b/drivers/iommu/arm-smmu-v3.c
@@ -183,7 +183,7 @@

 #define Q_IDX(llq, p)                  ((p) & ((1 << (llq)->max_n_shift) - 1))
 #define Q_WRP(llq, p)                  ((p) & (1 << (llq)->max_n_shift))
-#define Q_OVERFLOW_FLAG                        (1 << 31)
+#define Q_OVERFLOW_FLAG                        (1U << 31)
 #define Q_OVF(p)                       ((p) & Q_OVERFLOW_FLAG)
 #define Q_ENT(q, p)                    ((q)->base +                    \
                                         Q_IDX(&((q)->llq), p) *        \

then it would make sense to fix GBPA_UPDATE in the same way.

Xenia