Discussion:
[libav-devel] [PATCH] configure: Don't assume a 16 byte aligned stack on BSDs on i386
Martin Storsjö
2018-03-16 19:37:10 UTC
Permalink
With GCC, request it to maintain 16 byte alignment, and the existing
entry points already align it via attribute_align_arg.

With clang, do the same as for mingw; disable the aligned stack
and let the assembly functions that require it do the alignment
instead.
---
configure | 22 ++++++++++++++++++++--
1 file changed, 20 insertions(+), 2 deletions(-)

diff --git a/configure b/configure
index 95e6006440..78a2065208 100755
--- a/configure
+++ b/configure
@@ -4957,16 +4957,34 @@ elif enabled gcc; then
check_cflags -Werror=format-security
check_cflags -fdiagnostics-color=auto
enabled extra_warnings || check_disable_warning -Wno-maybe-uninitialized
+ if enabled x86_32; then
+ case $target_os in
+ *bsd*)
+ # BSDs don't guarantee a 16 byte aligned stack, but we can
+ # request GCC to try to maintain 16 byte alignment throughout
+ # function calls. Library entry points that might call assembly
+ # functions align the stack. (The parameter means 2^4 bytes.)
+ check_cflags -mpreferred-stack-boundary=4
+ ;;
+ esac
+ fi
elif enabled llvm_gcc; then
check_cflags -mllvm -stack-alignment=16
elif enabled clang; then
- if [ "$target_os" = "mingw32" -o "$target_os" = "win32" ] && enabled x86_32; then
+ if enabled x86_32; then
# Clang doesn't support maintaining alignment without assuming the
# same alignment in every function. If 16 byte alignment would be
# enabled, one would also have to either add attribute_align_arg on
# every single entry point into the libraries or enable -mstackrealign
# (doing stack realignment in every single function).
- disable aligned_stack
+ case $target_os in
+ mingw32|win32|*bsd*)
+ disable aligned_stack
+ ;;
+ *)
+ check_cflags -mllvm -stack-alignment=16
+ ;;
+ esac
else
check_cflags -mllvm -stack-alignment=16
fi
--
2.14.3 (Apple Git-98)
Luca Barbato
2018-03-17 07:26:51 UTC
Permalink
Post by Martin Storsjö
With GCC, request it to maintain 16 byte alignment, and the existing
entry points already align it via attribute_align_arg.
With clang, do the same as for mingw; disable the aligned stack
and let the assembly functions that require it do the alignment
instead.
---
configure | 22 ++++++++++++++++++++--
1 file changed, 20 insertions(+), 2 deletions(-)
diff --git a/configure b/configure
index 95e6006440..78a2065208 100755
--- a/configure
+++ b/configure
@@ -4957,16 +4957,34 @@ elif enabled gcc; then
check_cflags -Werror=format-security
check_cflags -fdiagnostics-color=auto
enabled extra_warnings || check_disable_warning -Wno-maybe-uninitialized
+ if enabled x86_32; then
+ case $target_os in
+ *bsd*)
+ # BSDs don't guarantee a 16 byte aligned stack, but we can
+ # request GCC to try to maintain 16 byte alignment throughout
+ # function calls. Library entry points that might call assembly
+ # functions align the stack. (The parameter means 2^4 bytes.)
+ check_cflags -mpreferred-stack-boundary=4
+ ;;
+ esac
+ fi
elif enabled llvm_gcc; then
check_cflags -mllvm -stack-alignment=16
elif enabled clang; then
- if [ "$target_os" = "mingw32" -o "$target_os" = "win32" ] && enabled x86_32; then
+ if enabled x86_32; then
# Clang doesn't support maintaining alignment without assuming the
# same alignment in every function. If 16 byte alignment would be
# enabled, one would also have to either add attribute_align_arg on
# every single entry point into the libraries or enable -mstackrealign
# (doing stack realignment in every single function).
- disable aligned_stack
+ case $target_os in
+ mingw32|win32|*bsd*)
+ disable aligned_stack
+ ;;
+ *)
+ check_cflags -mllvm -stack-alignment=16
+ ;;
+ esac
else
check_cflags -mllvm -stack-alignment=16
fi
Ok.

Loading...