[Lldb-commits] [PATCH] D14985: Add 64/128 bit arm neon register definitions on linux

[Tamas Berghammer via lldb-commits]

tberghammer added a comment.

I made some of the tests in TestRegisters.py arm compatible (see http://reviews.llvm.org/D15010) and we also test register read/write in some extent with the gdb remote protocol tests (they caught several issues while I was doing this patch)


================
Comment at: source/Plugins/Process/Utility/RegisterInfos_arm.h:41
@@ -40,3 +42,1 @@
-#ifndef DBG_OFFSET
-#error DBG_OFFSET_NAME must be defined before including this header file
 #endif
----------------
labath wrote:
> Was removing DBG_OFFSET intentional?
Yes because we don't use it in this file

================
Comment at: source/Plugins/Process/Utility/RegisterInfos_arm.h:316
@@ +315,3 @@
+static uint32_t g_q5_contains[] =  { fpu_d10, fpu_d11, fpu_s20, fpu_s21, fpu_s22, fpu_s23, LLDB_INVALID_REGNUM };
+static uint32_t g_q6_contains[] =  { fpu_d12, fpu_d13, fpu_s24, fpu_s25, fpu_s24, fpu_s27, LLDB_INVALID_REGNUM };
+static uint32_t g_q7_contains[] =  { fpu_d14, fpu_d15, fpu_s28, fpu_s29, fpu_s28, fpu_s31, LLDB_INVALID_REGNUM };
----------------
omjavaid wrote:
> here fpu_s24 is being repeated twice instead of fpu_s26.
Nice catch

================
Comment at: source/Plugins/Process/Utility/RegisterInfos_arm.h:317
@@ +316,3 @@
+static uint32_t g_q6_contains[] =  { fpu_d12, fpu_d13, fpu_s24, fpu_s25, fpu_s24, fpu_s27, LLDB_INVALID_REGNUM };
+static uint32_t g_q7_contains[] =  { fpu_d14, fpu_d15, fpu_s28, fpu_s29, fpu_s28, fpu_s31, LLDB_INVALID_REGNUM };
+static uint32_t g_q8_contains[] =  { fpu_d16, fpu_d17, LLDB_INVALID_REGNUM };
----------------
omjavaid wrote:
> here fpu_s28 is being repeated twice instead of fpu_s30.
Nice catch


http://reviews.llvm.org/D14985