X-Git-Url: http://repo.macrolet.net/gitweb/?a=blobdiff_plain;f=src%2Fcompiler%2Fx86-64%2Fsystem.lisp;h=1edf735b59671384a49035d25354a3a97ae099dd;hb=a7ec2f1af6881c93585d97fb2fdcdfd98006f217;hp=e411567985d7331fe1ec6cab0e2879608ecf44f1;hpb=4d5a8689d1d303f65c2fa933bb8757063641a8f9;p=sbcl.git diff --git a/src/compiler/x86-64/system.lisp b/src/compiler/x86-64/system.lisp index e411567..1edf735 100644 --- a/src/compiler/x86-64/system.lisp +++ b/src/compiler/x86-64/system.lisp @@ -44,7 +44,7 @@ (inst jmp :ne DONE) ;; Pick off fixnums. - (inst and al-tn 7) + (inst and al-tn fixnum-tag-mask) (inst jmp :e DONE) ;; must be an other immediate @@ -126,14 +126,16 @@ (storew eax x 0 other-pointer-lowtag) (move res x))) -(define-vop (make-fixnum) +(define-vop (pointer-hash) + (:translate pointer-hash) (:args (ptr :scs (any-reg descriptor-reg) :target res)) (:results (res :scs (any-reg descriptor-reg))) + (:policy :fast-safe) (:generator 1 - ;; Some code (the hash table code) depends on this returning a - ;; positive number so make sure it does. (move res ptr) - (inst shl res 4) + ;; Mask the lowtag, and shift the whole address into a positive + ;; fixnum. + (inst and res (lognot lowtag-mask)) (inst shr res 1))) (define-vop (make-other-immediate-type) @@ -234,9 +236,6 @@ (define-source-transform %closure-fun (closure) `(%simple-fun-self ,closure)) -(define-source-transform %funcallable-instance-fun (fin) - `(%simple-fun-self ,fin)) - (define-vop (%set-fun-self) (:policy :fast-safe) (:translate (setf %simple-fun-self)) @@ -251,20 +250,6 @@ fun-pointer-lowtag))) (storew temp function simple-fun-self-slot fun-pointer-lowtag) (move result new-self))) - -;;; KLUDGE: This seems to be some kind of weird override of the way -;;; that the objdef.lisp code would ordinarily set up the slot -;;; accessor. It's inherited from CMU CL, and it works, and naively -;;; deleting it seemed to cause problems, but it's not obvious why -;;; it's done this way. Any ideas? -- WHN 2001-08-02 -(defknown ((setf %funcallable-instance-fun)) (function function) function - (unsafe)) -;;; CMU CL comment: -;;; We would have really liked to use a source-transform for this, but -;;; they don't work with SETF functions. -;;; FIXME: Can't we just use DEFSETF or something? -(deftransform (setf %funcallable-instance-fun) ((value fin)) - '(setf (%simple-fun-self fin) value)) ;;;; other miscellaneous VOPs @@ -305,7 +290,68 @@ (note-next-instruction vop :internal-error) (inst wait))) -;;;; dynamic vop count collection support +;;;; Miscellany + +;;; the RDTSC instruction (present on Pentium processors and +;;; successors) allows you to access the time-stamp counter, a 64-bit +;;; model-specific register that counts executed cycles. The +;;; instruction returns the low cycle count in EAX and high cycle +;;; count in EDX. +;;; +;;; In order to obtain more significant results on out-of-order +;;; processors (such as the Pentium II and later), we issue a +;;; serializing CPUID instruction before and after reading the cycle +;;; counter. This instruction is used for its side effect of emptying +;;; the processor pipeline, to ensure that the RDTSC instruction is +;;; executed once all pending instructions have been completed and +;;; before any others. CPUID writes to EBX and ECX in addition to EAX +;;; and EDX, so they need to be added as temporaries. +;;; +;;; Note that cache effects mean that the cycle count can vary for +;;; different executions of the same code (it counts cycles, not +;;; retired instructions). Furthermore, the results are per-processor +;;; and not per-process, so are unreliable on multiprocessor machines +;;; where processes can migrate between processors. +;;; +;;; This method of obtaining a cycle count has the advantage of being +;;; very fast (around 20 cycles), and of not requiring a system call. +;;; However, you need to know your processor's clock speed to translate +;;; this into real execution time. +;;; +;;; FIXME: This about the WITH-CYCLE-COUNTER interface a bit, and then +;;; perhaps export it from SB-SYS. + +(defknown %read-cycle-counter () (values (unsigned-byte 32) (unsigned-byte 32)) ()) + +(define-vop (%read-cycle-counter) + (:policy :fast-safe) + (:translate %read-cycle-counter) + (:temporary (:sc unsigned-reg :offset eax-offset :target lo) eax) + (:temporary (:sc unsigned-reg :offset edx-offset :target hi) edx) + (:temporary (:sc unsigned-reg :offset ebx-offset) ebx) + (:temporary (:sc unsigned-reg :offset ecx-offset) ecx) + (:ignore ebx ecx) + (:results (hi :scs (unsigned-reg)) + (lo :scs (unsigned-reg))) + (:result-types unsigned-num unsigned-num) + (:generator 5 + (zeroize eax) + ;; Intel docs seem quite consistent on only using CPUID before RDTSC, + ;; not both before and after. Go figure. + (inst cpuid) + (inst rdtsc) + (move lo eax) + (move hi edx))) + +(defmacro with-cycle-counter (&body body) + "Returns the primary value of BODY as the primary value, and the +number of CPU cycles elapsed as secondary value. EXPERIMENTAL." + (with-unique-names (hi0 hi1 lo0 lo1) + `(multiple-value-bind (,hi0 ,lo0) (%read-cycle-counter) + (values (locally ,@body) + (multiple-value-bind (,hi1 ,lo1) (%read-cycle-counter) + (+ (ash (- ,hi1 ,hi0) 32) + (- ,lo1 ,lo0))))))) #!+sb-dyncount (define-vop (count-me) @@ -315,3 +361,36 @@ (inst inc (make-ea :qword :base count-vector :disp (- (* (+ vector-data-offset index) n-word-bytes) other-pointer-lowtag))))) + +;;;; Memory barrier support + +#!+memory-barrier-vops +(define-vop (%compiler-barrier) + (:policy :fast-safe) + (:translate %compiler-barrier) + (:generator 3)) + +#!+memory-barrier-vops +(define-vop (%memory-barrier) + (:policy :fast-safe) + (:translate %memory-barrier) + (:generator 3 + (inst mfence))) + +#!+memory-barrier-vops +(define-vop (%read-barrier) + (:policy :fast-safe) + (:translate %read-barrier) + (:generator 3)) + +#!+memory-barrier-vops +(define-vop (%write-barrier) + (:policy :fast-safe) + (:translate %write-barrier) + (:generator 3)) + +#!+memory-barrier-vops +(define-vop (%data-dependency-barrier) + (:policy :fast-safe) + (:translate %data-dependency-barrier) + (:generator 3))