drm/amdgpu: enable -msse2 for GCC 7.1+ users
authorNick Desaulniers <ndesaulniers@google.com>
Wed, 16 Oct 2019 23:02:09 +0000 (16:02 -0700)
committerAlex Deucher <alexander.deucher@amd.com>
Wed, 30 Oct 2019 15:56:20 +0000 (11:56 -0400)
A final attempt at enabling sse2 for GCC users.

Orininally attempted in:
commit 10117450735c ("drm/amd/display: add -msse2 to prevent Clang from emitting libcalls to undefined SW FP routines")

Reverted due to "reported instability" in:
commit 193392ed9f69 ("Revert "drm/amd/display: add -msse2 to prevent Clang from emitting libcalls to undefined SW FP routines"")

Re-added just for Clang in:
commit 0f0727d971f6 ("drm/amd/display: readd -msse2 to prevent Clang from emitting libcalls to undefined SW FP routines")

The original report didn't have enough information to know if the GPF
was due to misalignment, but I suspect that it was. (The missing
information was the disassembly of the function at the bottom of the
trace, to see if the instruction pointer pointed to an instruction with
16B alignment memory operand requirements.  The stack trace does show
the stack was only 8B but not 16B aligned though, which makes this a
strong possibility).

Now that the stack misalignment issue has been fixed for users of GCC
7.1+, reattempt adding -msse2. This matches Clang.

It will likely never be safe to enable this for pre-GCC 7.1 AND use a
16B aligned stack in these translation units.

This is only a functional change for GCC 7.1+ users, and should be boot
tested.

Link: https://bugs.freedesktop.org/show_bug.cgi?id=109487
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/display/dc/calcs/Makefile
drivers/gpu/drm/amd/display/dc/dcn20/Makefile
drivers/gpu/drm/amd/display/dc/dcn21/Makefile
drivers/gpu/drm/amd/display/dc/dml/Makefile
drivers/gpu/drm/amd/display/dc/dsc/Makefile

index a1af55a..26c6d73 100644 (file)
@@ -37,9 +37,7 @@ ifdef IS_OLD_GCC
 # GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
 # (8B stack alignment).
 calcs_ccflags += -mpreferred-stack-boundary=4
-endif
-
-ifdef CONFIG_CC_IS_CLANG
+else
 calcs_ccflags += -msse2
 endif
 
index cb0ac13..63f3bdd 100644 (file)
@@ -23,9 +23,7 @@ ifdef IS_OLD_GCC
 # GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
 # (8B stack alignment).
 CFLAGS_$(AMDDALPATH)/dc/dcn20/dcn20_resource.o += -mpreferred-stack-boundary=4
-endif
-
-ifdef CONFIG_CC_IS_CLANG
+else
 CFLAGS_$(AMDDALPATH)/dc/dcn20/dcn20_resource.o += -msse2
 endif
 
index f92320d..ff50ae7 100644 (file)
@@ -16,9 +16,7 @@ ifdef IS_OLD_GCC
 # GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
 # (8B stack alignment).
 CFLAGS_$(AMDDALPATH)/dc/dcn21/dcn21_resource.o += -mpreferred-stack-boundary=4
-endif
-
-ifdef CONFIG_CC_IS_CLANG
+else
 CFLAGS_$(AMDDALPATH)/dc/dcn21/dcn21_resource.o += -msse2
 endif
 
index ef1bdd2..8df2516 100644 (file)
@@ -37,9 +37,7 @@ ifdef IS_OLD_GCC
 # GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
 # (8B stack alignment).
 dml_ccflags += -mpreferred-stack-boundary=4
-endif
-
-ifdef CONFIG_CC_IS_CLANG
+else
 dml_ccflags += -msse2
 endif
 
index 3f78408..9707372 100644 (file)
@@ -14,9 +14,7 @@ ifdef IS_OLD_GCC
 # GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
 # (8B stack alignment).
 dsc_ccflags += -mpreferred-stack-boundary=4
-endif
-
-ifdef CONFIG_CC_IS_CLANG
+else
 dsc_ccflags += -msse2
 endif