Update CameraStreamBuffering test case

CameraStreamBuffering test case intentionally calls
setMaxFramesInFight() with very large integer value to verify that API
handles unreasonaly large value properly.  A type of its parameter is
changed from unsigned 32-bit integer to signed 32-bit integer in the
latest EVS interface definition written in the stable AIDL.  To adapt a
test case, this CL updates it to use std::numeric_limits<int32_t>::max()
instead of 0xFFFFFFFF, which is translated as -1 in signed 32-bit integer
format.

Bug: 226628778
Test: atest VtsHalEvsV1_1TargetTest
      atest VtsHalEvsTargetTest
Change-Id: Ic35c6c962617825bc271cec8544770340db7728a
diff --git a/automotive/evs/aidl/vts/VtsHalEvsTargetTest.cpp b/automotive/evs/aidl/vts/VtsHalEvsTargetTest.cpp
index a442368..09e1088 100644
--- a/automotive/evs/aidl/vts/VtsHalEvsTargetTest.cpp
+++ b/automotive/evs/aidl/vts/VtsHalEvsTargetTest.cpp
@@ -526,7 +526,7 @@
         mActiveCameras.push_back(pCam);
 
         // Ask for a very large number of buffers in flight to ensure it errors correctly
-        auto badResult = pCam->setMaxFramesInFlight(0xFFFFFFFF);
+        auto badResult = pCam->setMaxFramesInFlight(std::numeric_limits<int32_t>::max());
         EXPECT_TRUE(!badResult.isOk() && badResult.getServiceSpecificError() ==
                                                  static_cast<int>(EvsResult::INVALID_ARG));