[Mono-bugs] [Bug 80498][Wis] Changed - alignment of address returned by localloc

[bugzilla-daemon at bugzilla.ximian.com]

Please do not reply to this email- if you want to comment on the bug, go to the
URL shown below and enter your comments there.

Changed by massi at ximian.com.

http://bugzilla.ximian.com/show_bug.cgi?id=80498

--- shadow/80498	2007-01-15 14:44:39.000000000 -0500
+++ shadow/80498.tmp.8630	2007-01-16 14:03:16.000000000 -0500
@@ -1,12 +1,12 @@
 Bug#: 80498
 Product: Mono: Runtime
 Version: 1.2
 OS: unknown
 OS Details: 
-Status: NEW   
+Status: ASSIGNED   
 Resolution: 
 Severity: Unknown
 Priority: Wishlist
 Component: JIT
 AssignedTo: massi at ximian.com                            
 ReportedBy: erven.rohou at st.com               
@@ -53,6 +53,29 @@
 By "implement this" I mean only if localloc asks for more than 8
 bytes, align to 8, otherwise do like we do now.
 
 
 ------- Additional Comments From lupus at ximian.com  2007-01-12 11:12 -------
 It should always align to 8 bytes, it's pointless to add a check.
+
+------- Additional Comments From massi at ximian.com  2007-01-16 14:03 -------
+There's one thing I don't understand in "always align to 8 bytes".
+Currently localloc aligns following MONO_ARCH_FRAME_ALIGNMENT, which
+is 4 only on ppc and on non-Windows x86.
+I really do not think I should give MONO_ARCH_FRAME_ALIGNMENT a new
+value, but that's OK, I could just define a new macro (like
+MONO_ARCH_LOCALLOC_ALIGNMENT) defined as MONO_ARCH_FRAME_ALIGNMENT
+in almost all cases, and 8 in the others.
+Something like:
+
+#if (MONO_ARCH_FRAME_ALIGNMENT == 4)
+#define MONO_ARCH_LOCALLOC_ALIGNMEENT 8
+#else
+#define MONO_ARCH_LOCALLOC_ALIGNMENT MONO_ARCH_FRAME_ALIGNMENT
+#endif
+
+maybe at the end of "mini-arch.h".
+
+But do you really mean that asking for 4 bytes (an int32) on x86
+and ppc should align it to 8 bytes, just to keep the code general
+and simple?
+