Buffer overflow (-Wstringop-overflow) in matrixRow[] (FragDecoder.c) detected by GCC 10.2.1

[tinic]

Changing
uint8_t matrixRow[(FRAG_MAX_NB >> 3 ) + 1]; to uint8_t matrixRow[(FRAG_MAX_NB >> 3 ) + 1 + 16]; suppresses the warning.

[43/79] Building C object CMakeFiles/LoRaMac-node.dir/LoRa...src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c.o ../../LoRaMac-node/src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c: In function 'FragDecoderProcess': ../../LoRaMac-node/src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c:594:22: warning: writing 1 byte into a region of size 0 [-Wstringop-overflow=] 594 | matrixRow[i] = 0; | ~~~~~~~~~~~~~^~~ ../../LoRaMac-node/src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c:307:13: note: at offset 0 to object 'matrixRow' with size 3 declared here 307 | uint8_t matrixRow[(FRAG_MAX_NB >> 3 ) + 1]; | ^~~~~~~~~ ../../LoRaMac-node/src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c:594:22: warning: writing 1 byte into a region of size 0 [-Wstringop-overflow=] 594 | matrixRow[i] = 0; | ~~~~~~~~~~~~~^~~ ../../LoRaMac-node/src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c:307:13: note: at offset 0 to object 'matrixRow' with size 3 declared here 307 | uint8_t matrixRow[(FRAG_MAX_NB >> 3 ) + 1]; | ^~~~~~~~~ ../../LoRaMac-node/src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c:594:22: warning: writing 1 byte into a region of size 0 [-Wstringop-overflow=] 594 | matrixRow[i] = 0; | ~~~~~~~~~~~~~^~~ ../../LoRaMac-node/src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c:307:13: note: at offset 0 to object 'matrixRow' with size 3 declared here 307 | uint8_t matrixRow[(FRAG_MAX_NB >> 3 ) + 1]; | ^~~~~~~~~

[tinic]

Looks like the optimizer gets confused about an unsigned <-> signed compare. This fixes the issue and is probably cleaner also:

--- a/src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c
+++ b/src/apps/LoRaMac/common/LmHandler/packages/FragDecoder.c
@@ -589,7 +589,7 @@ static void FragGetParityMatrixRow( int32_t n, int32_t m, uint8_t *matrixRow )
     }

     x = 1 + ( 1001 * n );
-    for( uint16_t i = 0; i < ( ( m >> 3 ) + 1 ); i++ )
+    for( int32_t i = 0; i < ( ( m >> 3 ) + 1 ); i++ )^M
     {
         matrixRow[i] = 0;
     }

[mluis1]

Thanks for the fix. May I ask you to propose a PR containing this fix?

Another question, which compiler options have you used to generate these warnings? So far I was not able to get the same output as you. I am using this project default compiler options.