Changeset 13748

Timestamp:

May 30, 2010, 1:57:18 PM (10 years ago)

Author:

gb

Message:

Complain about the lack of "trapped FP exceptions" on VFPv3.

File:

• branches/arm/lisp-kernel/ARM-notes.txt (modified) (1 diff)

branches/arm/lisp-kernel/ARM-notes.txt

@@ -19 +19 @@
 don't see any reason to try to optimize for older variants.)
 
+More: recent ARM variants (ARMv7, at least) support vfp v3 with NEON
+SIMD extenstions.  If vfp v3, the exception-enable bits in the FPSCR
+always read as 0; that means that FP exceptions can't cause SIGFPE.
+
+On a jailbroken iPod Touch, an enabled FP exception seems to force
+a reboot instead of (or as well as) SIGFPE.
+
+Checking to see if an FP operation generated an interesting FP
+exception involves something like:
+
+ (fp-operation fp1 fp2 fp3)
+ (mov temp1 fpscr) ; the "exception occurred" bits get set and can be read
+ (load temp2 (tcr.lisp-exception-bits))
+ (ands temp3 temp1 temp2) ; set flags
+ (uuo-if (:? ne) ....)
+
+That just sucks; it makes FP exception checking a time/space/safety tradeoff.
+
+ 
+
+
 
 - instruction set