X-Git-Url: http://repo.macrolet.net/gitweb/?a=blobdiff_plain;ds=sidebyside;f=src%2Fcode%2Ftarget-hash-table.lisp;h=79a9cbd0078606a53e34340108bd9c23a4cfd5d8;hb=ea3a2433c72ee97c5691c29d882a63e4d86f0a32;hp=c28e640a305ed8220cf8665f39da84c1875babe1;hpb=084168e1524a6493bc0f9d1697753d31239b158d;p=sbcl.git diff --git a/src/code/target-hash-table.lisp b/src/code/target-hash-table.lisp index c28e640..79a9cbd 100644 --- a/src/code/target-hash-table.lisp +++ b/src/code/target-hash-table.lisp @@ -14,6 +14,23 @@ ;;;; utilities +;;; Without the locking the next vector can get cyclic causing +;;; looping in a WITHOUT-GCING form, SHRINK-VECTOR can corrupt memory +;;; and who knows what else. +;;; +;;; WITHOUT-GCING implies WITHOUT-INTERRUPTS. +(defmacro with-spinlock-and-without-gcing ((spinlock) &body body) + #!-sb-thread + (declare (ignore spinlock)) + `(without-gcing + (unwind-protect + (progn + #!+sb-thread + (sb!thread::get-spinlock ,spinlock) + ,@body) + #!+sb-thread + (sb!thread::release-spinlock ,spinlock)))) + (eval-when (:compile-toplevel :load-toplevel :execute) (defconstant max-hash sb!xc:most-positive-fixnum)) @@ -34,12 +51,21 @@ (defun eq-hash (key) (declare (values hash (member t nil))) (values (pointer-hash key) - (oddp (get-lisp-obj-address key)))) + (oddp (get-lisp-obj-address key)))) #!-sb-fluid (declaim (inline equal-hash)) (defun equal-hash (key) (declare (values hash (member t nil))) - (values (sxhash key) nil)) + (typecase key + ;; For some types the definition of EQUAL implies a special hash + ((or string cons number bit-vector pathname) + (values (sxhash key) nil)) + ;; Otherwise use an EQ hash, rather than SXHASH, since the values + ;; of SXHASH will be extremely badly distributed due to the + ;; requirements of the spec fitting badly with our implementation + ;; strategy. + (t + (eq-hash key)))) #!-sb-fluid (declaim (inline eql-hash)) (defun eql-hash (key) @@ -50,7 +76,13 @@ (defun equalp-hash (key) (declare (values hash (member t nil))) - (values (psxhash key) nil)) + (typecase key + ;; Types requiring special treatment. Note that PATHNAME and + ;; HASH-TABLE are caught by the STRUCTURE-OBJECT test. + ((or array cons number character structure-object) + (values (psxhash key) nil)) + (t + (eq-hash key)))) (defun almost-primify (num) (declare (type index num)) @@ -72,26 +104,22 @@ #!+sb-doc "Define a new kind of hash table test." (declare (type symbol name) - (type function test-fun hash-fun)) + (type function test-fun hash-fun)) (setf *hash-table-tests* - (cons (list name test-fun hash-fun) - (remove name *hash-table-tests* :test #'eq :key #'car))) + (cons (list name test-fun hash-fun) + (remove name *hash-table-tests* :test #'eq :key #'car))) name) ;;;; construction and simple accessors (defconstant +min-hash-table-size+ 16) (defconstant +min-hash-table-rehash-threshold+ (float 1/16 1.0)) -;; as explained by pmai on openprojects #lisp IRC 2002-07-30: #x80000000 -;; is bigger than any possible nonEQ hash value, and thus indicates an -;; empty slot; and EQ hash tables don't use HASH-TABLE-HASH-VECTOR -(defconstant +magic-hash-vector-value+ #x80000000) (defun make-hash-table (&key (test 'eql) - (size +min-hash-table-size+) - (rehash-size 1.5) - (rehash-threshold 1) - (weak-p nil)) + (size +min-hash-table-size+) + (rehash-size 1.5) + (rehash-threshold 1) + (weakness nil)) #!+sb-doc "Create and return a new hash table. The keywords are as follows: :TEST -- Indicates what kind of test to use. @@ -105,47 +133,54 @@ forcing a rehash. Can be any positive number <=1, with density approaching zero as the threshold approaches 0. Density 1 means an average of one entry per bucket. - :WEAK-P -- (This is an extension from CMU CL, not currently supported - in SBCL 0.6.6, but perhaps supported in a future version.) If T, - don't keep entries if the key would otherwise be garbage." + :WEAKNESS -- IF NIL (the default) it is a normal non-weak hash table. + If one of :KEY, :VALUE, :KEY-AND-VALUE, :KEY-OR-VALUE it is a weak + hash table. + Depending on the type of weakness the lack of references to the + key and the value may allow for removal of the entry. If WEAKNESS + is :KEY and the key would otherwise be garbage the entry is eligible + for removal from the hash table. Similarly, if WEAKNESS is :VALUE + the life of an entry depends on its value's references. If WEAKNESS + is :KEY-AND-VALUE and either the key or the value would otherwise be + garbage the entry can be removed. If WEAKNESS is :KEY-OR-VALUE and + both the key and the value would otherwise be garbage the entry can + be removed." (declare (type (or function symbol) test)) (declare (type unsigned-byte size)) - (when weak-p - (error "stub: unsupported WEAK-P option")) (multiple-value-bind (test test-fun hash-fun) (cond ((or (eq test #'eq) (eq test 'eq)) - (values 'eq #'eq #'eq-hash)) - ((or (eq test #'eql) (eq test 'eql)) - (values 'eql #'eql #'eql-hash)) - ((or (eq test #'equal) (eq test 'equal)) - (values 'equal #'equal #'equal-hash)) - ((or (eq test #'equalp) (eq test 'equalp)) - (values 'equalp #'equalp #'equalp-hash)) - (t - ;; FIXME: I'd like to remove *HASH-TABLE-TESTS* stuff. - ;; Failing that, I'd like to rename it to - ;; *USER-HASH-TABLE-TESTS*. - (dolist (info *hash-table-tests* - (error "unknown :TEST for MAKE-HASH-TABLE: ~S" - test)) - (destructuring-bind (test-name test-fun hash-fun) info - (when (or (eq test test-name) (eq test test-fun)) - (return (values test-name test-fun hash-fun))))))) + (values 'eq #'eq #'eq-hash)) + ((or (eq test #'eql) (eq test 'eql)) + (values 'eql #'eql #'eql-hash)) + ((or (eq test #'equal) (eq test 'equal)) + (values 'equal #'equal #'equal-hash)) + ((or (eq test #'equalp) (eq test 'equalp)) + (values 'equalp #'equalp #'equalp-hash)) + (t + ;; FIXME: I'd like to remove *HASH-TABLE-TESTS* stuff. + ;; Failing that, I'd like to rename it to + ;; *USER-HASH-TABLE-TESTS*. + (dolist (info *hash-table-tests* + (error "unknown :TEST for MAKE-HASH-TABLE: ~S" + test)) + (destructuring-bind (test-name test-fun hash-fun) info + (when (or (eq test test-name) (eq test test-fun)) + (return (values test-name test-fun hash-fun))))))) (let* ((size (max +min-hash-table-size+ - (min size - ;; SIZE is just a hint, so if the user asks - ;; for a SIZE which'd be too big for us to - ;; easily implement, we bump it down. - (floor array-dimension-limit 1024)))) - (rehash-size (if (integerp rehash-size) - rehash-size - (float rehash-size 1.0))) - ;; FIXME: Original REHASH-THRESHOLD default should be 1.0, - ;; not 1, to make it easier for the compiler to avoid - ;; boxing. - (rehash-threshold (max +min-hash-table-rehash-threshold+ - (float rehash-threshold 1.0))) - (size+1 (1+ size)) ; The first element is not usable. + (min size + ;; SIZE is just a hint, so if the user asks + ;; for a SIZE which'd be too big for us to + ;; easily implement, we bump it down. + (floor array-dimension-limit 1024)))) + (rehash-size (if (integerp rehash-size) + rehash-size + (float rehash-size 1.0))) + ;; FIXME: Original REHASH-THRESHOLD default should be 1.0, + ;; not 1, to make it easier for the compiler to avoid + ;; boxing. + (rehash-threshold (max +min-hash-table-rehash-threshold+ + (float rehash-threshold 1.0))) + (size+1 (1+ size)) ; The first element is not usable. ;; KLUDGE: The most natural way of expressing the below is ;; (round (/ (float size+1) rehash-threshold)), and indeed ;; it was expressed like that until 0.7.0. However, @@ -154,41 +189,47 @@ ;; but only for truncating; therefore, we fudge this issue ;; a little. The other uses of truncate, below, similarly ;; used to be round. -- CSR, 2002-10-01 - ;; - ;; Note that this has not yet been audited for - ;; correctness. It just seems to work. -- CSR, 2002-11-02 - (scaled-size (truncate (/ (float size+1) rehash-threshold))) - (length (almost-primify (max scaled-size - (1+ +min-hash-table-size+)))) - (index-vector (make-array length - :element-type '(unsigned-byte 32) - :initial-element 0)) - ;; needs to be the same length as the KV vector - ;; (FIXME: really? why doesn't the code agree?) - (next-vector (make-array size+1 - :element-type '(unsigned-byte 32))) - (kv-vector (make-array (* 2 size+1) - :initial-element +empty-ht-slot+)) - (table (%make-hash-table - :test test - :test-fun test-fun - :hash-fun hash-fun - :rehash-size rehash-size - :rehash-threshold rehash-threshold - :rehash-trigger size - :table kv-vector - :weak-p weak-p - :index-vector index-vector - :next-vector next-vector - :hash-vector (unless (eq test 'eq) - (make-array size+1 - :element-type '(unsigned-byte 32) - :initial-element +magic-hash-vector-value+))))) + ;; + ;; Note that this has not yet been audited for + ;; correctness. It just seems to work. -- CSR, 2002-11-02 + (scaled-size (truncate (/ (float size+1) rehash-threshold))) + (length (almost-primify (max scaled-size + (1+ +min-hash-table-size+)))) + (index-vector (make-array length + :element-type + '(unsigned-byte #.sb!vm:n-word-bits) + :initial-element 0)) + ;; Needs to be the half the length of the KV vector to link + ;; KV entries - mapped to indeces at 2i and 2i+1 - + ;; together. + (next-vector (make-array size+1 + :element-type + '(unsigned-byte #.sb!vm:n-word-bits))) + (kv-vector (make-array (* 2 size+1) + :initial-element +empty-ht-slot+)) + (table (%make-hash-table + :test test + :test-fun test-fun + :hash-fun hash-fun + :rehash-size rehash-size + :rehash-threshold rehash-threshold + :rehash-trigger size + :table kv-vector + :weakness weakness + :index-vector index-vector + :next-vector next-vector + :hash-vector + (unless (eq test 'eq) + (make-array size+1 + :element-type '(unsigned-byte + #.sb!vm:n-word-bits) + :initial-element +magic-hash-vector-value+)) + :spinlock (sb!thread::make-spinlock)))) (declare (type index size+1 scaled-size length)) ;; Set up the free list, all free. These lists are 0 terminated. (do ((i 1 (1+ i))) - ((>= i size)) - (setf (aref next-vector i) (1+ i))) + ((>= i size)) + (setf (aref next-vector i) (1+ i))) (setf (aref next-vector size) 0) (setf (hash-table-next-free-kv table) 1) (setf (hash-table-needing-rehash table) 0) @@ -199,7 +240,7 @@ #!+sb-doc "Return the number of entries in the given HASH-TABLE." (declare (type hash-table hash-table) - (values index)) + (values index)) (hash-table-number-entries hash-table)) #!+sb-doc @@ -222,9 +263,9 @@ "Return the test HASH-TABLE was created with.") #!+sb-doc -(setf (fdocumentation 'hash-table-weak-p 'function) - "Return T if HASH-TABLE will not keep entries for keys that would - otherwise be garbage, and NIL if it will.") +(setf (fdocumentation 'hash-table-weakness 'function) + "Return the WEAKNESS of HASH-TABLE which is one of NIL, :KEY, +:VALUE, :KEY-AND-VALUE, :KEY-OR-VALUE.") ;;;; accessing functions @@ -233,37 +274,44 @@ (defun rehash (table) (declare (type hash-table table)) (let* ((old-kv-vector (hash-table-table table)) - (old-next-vector (hash-table-next-vector table)) - (old-hash-vector (hash-table-hash-vector table)) - (old-size (length old-next-vector)) - (new-size - (let ((rehash-size (hash-table-rehash-size table))) - (etypecase rehash-size - (fixnum - (+ rehash-size old-size)) - (float - (the index (truncate (* rehash-size old-size))))))) - (new-kv-vector (make-array (* 2 new-size) - :initial-element +empty-ht-slot+)) - (new-next-vector (make-array new-size - :element-type '(unsigned-byte 32) - :initial-element 0)) - (new-hash-vector (when old-hash-vector - (make-array new-size - :element-type '(unsigned-byte 32) - :initial-element +magic-hash-vector-value+))) - (old-index-vector (hash-table-index-vector table)) - (new-length (almost-primify - (truncate (/ (float new-size) - (hash-table-rehash-threshold table))))) - (new-index-vector (make-array new-length - :element-type '(unsigned-byte 32) - :initial-element 0))) + (old-next-vector (hash-table-next-vector table)) + (old-hash-vector (hash-table-hash-vector table)) + (old-size (length old-next-vector)) + (new-size + (let ((rehash-size (hash-table-rehash-size table))) + (etypecase rehash-size + (fixnum + (+ rehash-size old-size)) + (float + (the index (truncate (* rehash-size old-size))))))) + (new-kv-vector (make-array (* 2 new-size) + :initial-element +empty-ht-slot+)) + (new-next-vector + (make-array new-size + :element-type '(unsigned-byte #.sb!vm:n-word-bits) + :initial-element 0)) + (new-hash-vector + (when old-hash-vector + (make-array new-size + :element-type '(unsigned-byte #.sb!vm:n-word-bits) + :initial-element +magic-hash-vector-value+))) + (old-index-vector (hash-table-index-vector table)) + (new-length (almost-primify + (truncate (/ (float new-size) + (hash-table-rehash-threshold table))))) + (new-index-vector + (make-array new-length + :element-type '(unsigned-byte #.sb!vm:n-word-bits) + :initial-element 0))) (declare (type index new-size new-length old-size)) ;; Disable GC tricks on the OLD-KV-VECTOR. (set-header-data old-kv-vector sb!vm:vector-normal-subtype) + ;; Non-empty weak hash tables always need GC support. + (when (and (hash-table-weakness table) (plusp (hash-table-count table))) + (set-header-data new-kv-vector sb!vm:vector-valid-hashing-subtype)) + ;; FIXME: here and in several other places in the hash table code, ;; loops like this one are used when FILL or REPLACE would be ;; appropriate. why are standard CL functions not used? @@ -278,56 +326,58 @@ ;; Copy over the hash-vector. (when old-hash-vector (dotimes (i old-size) - (setf (aref new-hash-vector i) (aref old-hash-vector i)))) + (setf (aref new-hash-vector i) (aref old-hash-vector i)))) (setf (hash-table-next-free-kv table) 0) (setf (hash-table-needing-rehash table) 0) ;; Rehash all the entries; last to first so that after the pushes ;; the chains are first to last. (do ((i (1- new-size) (1- i))) - ((zerop i)) + ((zerop i)) + (declare (type index/2 i)) (let ((key (aref new-kv-vector (* 2 i))) - (value (aref new-kv-vector (1+ (* 2 i))))) - (cond ((and (eq key +empty-ht-slot+) - (eq value +empty-ht-slot+)) - ;; Slot is empty, push it onto the free list. - (setf (aref new-next-vector i) - (hash-table-next-free-kv table)) - (setf (hash-table-next-free-kv table) i)) - ((and new-hash-vector - (not (= (aref new-hash-vector i) +magic-hash-vector-value+))) - ;; Can use the existing hash value (not EQ based) - (let* ((hashing (aref new-hash-vector i)) - (index (rem hashing new-length)) - (next (aref new-index-vector index))) - (declare (type index index) - (type hash hashing)) - ;; Push this slot into the next chain. - (setf (aref new-next-vector i) next) - (setf (aref new-index-vector index) i))) - (t - ;; EQ base hash. - ;; Enable GC tricks. - (set-header-data new-kv-vector - sb!vm:vector-valid-hashing-subtype) - (let* ((hashing (pointer-hash key)) - (index (rem hashing new-length)) - (next (aref new-index-vector index))) - (declare (type index index) - (type hash hashing)) - ;; Push this slot onto the next chain. - (setf (aref new-next-vector i) next) - (setf (aref new-index-vector index) i)))))) + (value (aref new-kv-vector (1+ (* 2 i))))) + (cond ((and (eq key +empty-ht-slot+) + (eq value +empty-ht-slot+)) + ;; Slot is empty, push it onto the free list. + (setf (aref new-next-vector i) + (hash-table-next-free-kv table)) + (setf (hash-table-next-free-kv table) i)) + ((and new-hash-vector + (not (= (aref new-hash-vector i) + +magic-hash-vector-value+))) + ;; Can use the existing hash value (not EQ based) + (let* ((hashing (aref new-hash-vector i)) + (index (rem hashing new-length)) + (next (aref new-index-vector index))) + (declare (type index index) + (type hash hashing)) + ;; Push this slot into the next chain. + (setf (aref new-next-vector i) next) + (setf (aref new-index-vector index) i))) + (t + ;; EQ base hash. + ;; Enable GC tricks. + (set-header-data new-kv-vector + sb!vm:vector-valid-hashing-subtype) + (let* ((hashing (pointer-hash key)) + (index (rem hashing new-length)) + (next (aref new-index-vector index))) + (declare (type index index) + (type hash hashing)) + ;; Push this slot onto the next chain. + (setf (aref new-next-vector i) next) + (setf (aref new-index-vector index) i)))))) (setf (hash-table-table table) new-kv-vector) (setf (hash-table-index-vector table) new-index-vector) (setf (hash-table-next-vector table) new-next-vector) (setf (hash-table-hash-vector table) new-hash-vector) ;; Shrink the old vectors to 0 size to help the conservative GC. - (shrink-vector old-kv-vector 0) - (shrink-vector old-index-vector 0) - (shrink-vector old-next-vector 0) + (%shrink-vector old-kv-vector 0) + (%shrink-vector old-index-vector 0) + (%shrink-vector old-next-vector 0) (when old-hash-vector - (shrink-vector old-hash-vector 0)) + (%shrink-vector old-hash-vector 0)) (setf (hash-table-rehash-trigger table) new-size)) (values)) @@ -335,16 +385,18 @@ (defun rehash-without-growing (table) (declare (type hash-table table)) (let* ((kv-vector (hash-table-table table)) - (next-vector (hash-table-next-vector table)) - (hash-vector (hash-table-hash-vector table)) - (size (length next-vector)) - (index-vector (hash-table-index-vector table)) - (length (length index-vector))) + (next-vector (hash-table-next-vector table)) + (hash-vector (hash-table-hash-vector table)) + (size (length next-vector)) + (index-vector (hash-table-index-vector table)) + (length (length index-vector))) (declare (type index size length)) - ;; Disable GC tricks, they will be re-enabled during the re-hash - ;; if necesary. - (set-header-data kv-vector sb!vm:vector-normal-subtype) + ;; Non-empty weak hash tables always need GC support. + (unless (and (hash-table-weakness table) (plusp (hash-table-count table))) + ;; Disable GC tricks, they will be re-enabled during the re-hash + ;; if necessary. + (set-header-data kv-vector sb!vm:vector-normal-subtype)) ;; Rehash all the entries. (setf (hash-table-next-free-kv table) 0) @@ -354,52 +406,54 @@ (dotimes (i length) (setf (aref index-vector i) 0)) (do ((i (1- size) (1- i))) - ((zerop i)) + ((zerop i)) + (declare (type index/2 i)) (let ((key (aref kv-vector (* 2 i))) - (value (aref kv-vector (1+ (* 2 i))))) - (cond ((and (eq key +empty-ht-slot+) - (eq value +empty-ht-slot+)) - ;; Slot is empty, push it onto free list. - (setf (aref next-vector i) (hash-table-next-free-kv table)) - (setf (hash-table-next-free-kv table) i)) - ((and hash-vector (not (= (aref hash-vector i) +magic-hash-vector-value+))) - ;; Can use the existing hash value (not EQ based) - (let* ((hashing (aref hash-vector i)) - (index (rem hashing length)) - (next (aref index-vector index))) - (declare (type index index)) - ;; Push this slot into the next chain. - (setf (aref next-vector i) next) - (setf (aref index-vector index) i))) - (t - ;; EQ base hash. - ;; Enable GC tricks. - (set-header-data kv-vector sb!vm:vector-valid-hashing-subtype) - (let* ((hashing (pointer-hash key)) - (index (rem hashing length)) - (next (aref index-vector index))) - (declare (type index index) - (type hash hashing)) - ;; Push this slot into the next chain. - (setf (aref next-vector i) next) - (setf (aref index-vector index) i))))))) + (value (aref kv-vector (1+ (* 2 i))))) + (cond ((and (eq key +empty-ht-slot+) + (eq value +empty-ht-slot+)) + ;; Slot is empty, push it onto free list. + (setf (aref next-vector i) (hash-table-next-free-kv table)) + (setf (hash-table-next-free-kv table) i)) + ((and hash-vector (not (= (aref hash-vector i) + +magic-hash-vector-value+))) + ;; Can use the existing hash value (not EQ based) + (let* ((hashing (aref hash-vector i)) + (index (rem hashing length)) + (next (aref index-vector index))) + (declare (type index index)) + ;; Push this slot into the next chain. + (setf (aref next-vector i) next) + (setf (aref index-vector index) i))) + (t + ;; EQ base hash. + ;; Enable GC tricks. + (set-header-data kv-vector sb!vm:vector-valid-hashing-subtype) + (let* ((hashing (pointer-hash key)) + (index (rem hashing length)) + (next (aref index-vector index))) + (declare (type index index) + (type hash hashing)) + ;; Push this slot into the next chain. + (setf (aref next-vector i) next) + (setf (aref index-vector index) i))))))) (values)) (defun flush-needing-rehash (table) (let* ((kv-vector (hash-table-table table)) - (index-vector (hash-table-index-vector table)) - (next-vector (hash-table-next-vector table)) - (length (length index-vector))) + (index-vector (hash-table-index-vector table)) + (next-vector (hash-table-next-vector table)) + (length (length index-vector))) (do ((next (hash-table-needing-rehash table))) - ((zerop next)) - (declare (type index next)) + ((zerop next)) + (declare (type index/2 next)) (let* ((key (aref kv-vector (* 2 next))) - (hashing (pointer-hash key)) - (index (rem hashing length)) - (temp (aref next-vector next))) - (setf (aref next-vector next) (aref index-vector index)) - (setf (aref index-vector index) next) - (setf next temp)))) + (hashing (pointer-hash key)) + (index (rem hashing length)) + (temp (aref next-vector next))) + (setf (aref next-vector next) (aref index-vector index)) + (setf (aref index-vector index) next) + (setf next temp)))) (setf (hash-table-needing-rehash table) 0) (values)) @@ -409,40 +463,63 @@ value and T as multiple values, or returns DEFAULT and NIL if there is no such entry. Entries can be added using SETF." (declare (type hash-table hash-table) - (values t (member t nil))) - (without-gcing + (values t (member t nil))) + (gethash3 key hash-table default)) + +(defun gethash2 (key hash-table) + #!+sb-doc + "Two argument version of GETHASH" + (declare (type hash-table hash-table) + (values t (member t nil))) + (gethash3 key hash-table nil)) + +(defun gethash3 (key hash-table default) + #!+sb-doc + "Three argument version of GETHASH" + (declare (type hash-table hash-table) + (values t (member t nil))) + (with-spinlock-and-without-gcing ((hash-table-spinlock hash-table)) (cond ((= (get-header-data (hash-table-table hash-table)) - sb!vm:vector-must-rehash-subtype) - (rehash-without-growing hash-table)) - ((not (zerop (hash-table-needing-rehash hash-table))) - (flush-needing-rehash hash-table))) - ;; Search for key in the hash table. - (multiple-value-bind (hashing eq-based) - (funcall (hash-table-hash-fun hash-table) key) - (declare (type hash hashing)) - (let* ((index-vector (hash-table-index-vector hash-table)) - (length (length index-vector)) - (index (rem hashing length)) - (next (aref index-vector index)) - (table (hash-table-table hash-table)) - (next-vector (hash-table-next-vector hash-table)) - (hash-vector (hash-table-hash-vector hash-table)) - (test-fun (hash-table-test-fun hash-table))) - (declare (type index index)) - ;; Search next-vector chain for a matching key. - (if (or eq-based (not hash-vector)) - (do ((next next (aref next-vector next))) - ((zerop next) (values default nil)) - (declare (type index next)) - (when (eq key (aref table (* 2 next))) - (return (values (aref table (1+ (* 2 next))) t)))) - (do ((next next (aref next-vector next))) - ((zerop next) (values default nil)) - (declare (type index next)) - (when (and (= hashing (aref hash-vector next)) - (funcall test-fun key (aref table (* 2 next)))) - ;; Found. - (return (values (aref table (1+ (* 2 next))) t))))))))) + sb!vm:vector-must-rehash-subtype) + (rehash-without-growing hash-table)) + ((not (zerop (hash-table-needing-rehash hash-table))) + (flush-needing-rehash hash-table))) + + ;; First check the cache. Use EQ here for speed. + (let ((cache (hash-table-cache hash-table)) + (table (hash-table-table hash-table))) + + (if (and cache (< cache (length table)) (eq (aref table cache) key)) + (values (aref table (1+ cache)) t) + + ;; Search for key in the hash table. + (multiple-value-bind (hashing eq-based) + (funcall (hash-table-hash-fun hash-table) key) + (declare (type hash hashing)) + (let* ((index-vector (hash-table-index-vector hash-table)) + (length (length index-vector)) + (index (rem hashing length)) + (next (aref index-vector index)) + (next-vector (hash-table-next-vector hash-table)) + (hash-vector (hash-table-hash-vector hash-table)) + (test-fun (hash-table-test-fun hash-table))) + (declare (type index index)) + ;; Search next-vector chain for a matching key. + (if (or eq-based (not hash-vector)) + (do ((next next (aref next-vector next))) + ((zerop next) (values default nil)) + (declare (type index/2 next)) + (when (eq key (aref table (* 2 next))) + (setf (hash-table-cache hash-table) (* 2 next)) + (return (values (aref table (1+ (* 2 next))) t)))) + (do ((next next (aref next-vector next))) + ((zerop next) (values default nil)) + (declare (type index/2 next)) + (when (and (= hashing (aref hash-vector next)) + (funcall test-fun key (aref table (* 2 next)))) + ;; Found. + (setf (hash-table-cache hash-table) (* 2 next)) + (return (values (aref table (1+ (* 2 next))) t))))))))))) ;;; so people can call #'(SETF GETHASH) (defun (setf gethash) (new-value key table &optional default) @@ -452,76 +529,92 @@ (defun %puthash (key hash-table value) (declare (type hash-table hash-table)) (aver (hash-table-index-vector hash-table)) - (without-gcing + (with-spinlock-and-without-gcing ((hash-table-spinlock hash-table)) ;; We need to rehash here so that a current key can be found if it ;; exists. Check that there is room for one more entry. May not be ;; needed if the key is already present. (cond ((zerop (hash-table-next-free-kv hash-table)) - (rehash hash-table)) - ((= (get-header-data (hash-table-table hash-table)) - sb!vm:vector-must-rehash-subtype) - (rehash-without-growing hash-table)) - ((not (zerop (hash-table-needing-rehash hash-table))) - (flush-needing-rehash hash-table))) - - ;; Search for key in the hash table. - (multiple-value-bind (hashing eq-based) - (funcall (hash-table-hash-fun hash-table) key) - (declare (type hash hashing)) - (let* ((index-vector (hash-table-index-vector hash-table)) - (length (length index-vector)) - (index (rem hashing length)) - (next (aref index-vector index)) - (kv-vector (hash-table-table hash-table)) - (next-vector (hash-table-next-vector hash-table)) - (hash-vector (hash-table-hash-vector hash-table)) - (test-fun (hash-table-test-fun hash-table))) - (declare (type index index)) - - (cond ((or eq-based (not hash-vector)) - (when eq-based - (set-header-data kv-vector sb!vm:vector-valid-hashing-subtype)) - - ;; Search next-vector chain for a matching key. - (do ((next next (aref next-vector next))) - ((zerop next)) - (declare (type index next)) - (when (eq key (aref kv-vector (* 2 next))) - ;; Found, just replace the value. - (setf (aref kv-vector (1+ (* 2 next))) value) - (return-from %puthash value)))) - (t - ;; Search next-vector chain for a matching key. - (do ((next next (aref next-vector next))) - ((zerop next)) - (declare (type index next)) - (when (and (= hashing (aref hash-vector next)) - (funcall test-fun key - (aref kv-vector (* 2 next)))) - ;; Found, just replace the value. - (setf (aref kv-vector (1+ (* 2 next))) value) - (return-from %puthash value))))) - - ;; Pop a KV slot off the free list - (let ((free-kv-slot (hash-table-next-free-kv hash-table))) - ;; Double-check for overflow. - (aver (not (zerop free-kv-slot))) - (setf (hash-table-next-free-kv hash-table) - (aref next-vector free-kv-slot)) - (incf (hash-table-number-entries hash-table)) - - (setf (aref kv-vector (* 2 free-kv-slot)) key) - (setf (aref kv-vector (1+ (* 2 free-kv-slot))) value) - - ;; Setup the hash-vector if necessary. - (when hash-vector - (if (not eq-based) - (setf (aref hash-vector free-kv-slot) hashing) - (aver (= (aref hash-vector free-kv-slot) +magic-hash-vector-value+)))) - - ;; Push this slot into the next chain. - (setf (aref next-vector free-kv-slot) next) - (setf (aref index-vector index) free-kv-slot))))) + (rehash hash-table)) + ((= (get-header-data (hash-table-table hash-table)) + sb!vm:vector-must-rehash-subtype) + (rehash-without-growing hash-table)) + ((not (zerop (hash-table-needing-rehash hash-table))) + (flush-needing-rehash hash-table))) + + (let ((cache (hash-table-cache hash-table)) + (kv-vector (hash-table-table hash-table))) + + ;; Check the cache + (if (and cache (< cache (length kv-vector)) + (eq (aref kv-vector cache) key)) + ;; If cached, just store here + (setf (aref kv-vector (1+ cache)) value) + + ;; Search for key in the hash table. + (multiple-value-bind (hashing eq-based) + (funcall (hash-table-hash-fun hash-table) key) + (declare (type hash hashing)) + (let* ((index-vector (hash-table-index-vector hash-table)) + (length (length index-vector)) + (index (rem hashing length)) + (next (aref index-vector index)) + (kv-vector (hash-table-table hash-table)) + (next-vector (hash-table-next-vector hash-table)) + (hash-vector (hash-table-hash-vector hash-table)) + (test-fun (hash-table-test-fun hash-table))) + (declare (type index index next)) + (when (hash-table-weakness hash-table) + (set-header-data kv-vector sb!vm:vector-valid-hashing-subtype)) + (cond ((or eq-based (not hash-vector)) + (when eq-based + (set-header-data kv-vector + sb!vm:vector-valid-hashing-subtype)) + + ;; Search next-vector chain for a matching key. + (do ((next next (aref next-vector next))) + ((zerop next)) + (declare (type index/2 next)) + (when (eq key (aref kv-vector (* 2 next))) + ;; Found, just replace the value. + (setf (hash-table-cache hash-table) (* 2 next)) + (setf (aref kv-vector (1+ (* 2 next))) value) + (return-from %puthash value)))) + (t + ;; Search next-vector chain for a matching key. + (do ((next next (aref next-vector next))) + ((zerop next)) + (declare (type index/2 next)) + (when (and (= hashing (aref hash-vector next)) + (funcall test-fun key + (aref kv-vector (* 2 next)))) + ;; Found, just replace the value. + (setf (hash-table-cache hash-table) (* 2 next)) + (setf (aref kv-vector (1+ (* 2 next))) value) + (return-from %puthash value))))) + + ;; Pop a KV slot off the free list + (let ((free-kv-slot (hash-table-next-free-kv hash-table))) + (declare (type index/2 free-kv-slot)) + ;; Double-check for overflow. + (aver (not (zerop free-kv-slot))) + (setf (hash-table-next-free-kv hash-table) + (aref next-vector free-kv-slot)) + (incf (hash-table-number-entries hash-table)) + + (setf (hash-table-cache hash-table) (* 2 free-kv-slot)) + (setf (aref kv-vector (* 2 free-kv-slot)) key) + (setf (aref kv-vector (1+ (* 2 free-kv-slot))) value) + + ;; Setup the hash-vector if necessary. + (when hash-vector + (if (not eq-based) + (setf (aref hash-vector free-kv-slot) hashing) + (aver (= (aref hash-vector free-kv-slot) + +magic-hash-vector-value+)))) + + ;; Push this slot into the next chain. + (setf (aref next-vector free-kv-slot) next) + (setf (aref index-vector index) free-kv-slot))))))) value) (defun remhash (key hash-table) @@ -529,30 +622,35 @@ "Remove the entry in HASH-TABLE associated with KEY. Return T if there was such an entry, or NIL if not." (declare (type hash-table hash-table) - (values (member t nil))) - (without-gcing + (values (member t nil))) + (with-spinlock-and-without-gcing ((hash-table-spinlock hash-table)) ;; We need to rehash here so that a current key can be found if it ;; exists. (cond ((= (get-header-data (hash-table-table hash-table)) - sb!vm:vector-must-rehash-subtype) - (rehash-without-growing hash-table)) - ((not (zerop (hash-table-needing-rehash hash-table))) - (flush-needing-rehash hash-table))) + sb!vm:vector-must-rehash-subtype) + (rehash-without-growing hash-table)) + ((not (zerop (hash-table-needing-rehash hash-table))) + (flush-needing-rehash hash-table))) + + ;; For now, just clear the cache + (setf (hash-table-cache hash-table) nil) ;; Search for key in the hash table. (multiple-value-bind (hashing eq-based) (funcall (hash-table-hash-fun hash-table) key) (declare (type hash hashing)) (let* ((index-vector (hash-table-index-vector hash-table)) - (length (length index-vector)) - (index (rem hashing length)) - (next (aref index-vector index)) - (table (hash-table-table hash-table)) - (next-vector (hash-table-next-vector hash-table)) - (hash-vector (hash-table-hash-vector hash-table)) - (test-fun (hash-table-test-fun hash-table))) - (declare (type index index next)) + (length (length index-vector)) + (index (rem hashing length)) + (next (aref index-vector index)) + (table (hash-table-table hash-table)) + (next-vector (hash-table-next-vector hash-table)) + (hash-vector (hash-table-hash-vector hash-table)) + (test-fun (hash-table-test-fun hash-table))) + (declare (type index index) + (type index/2 next)) (flet ((clear-slot (chain-vector prior-slot-location slot-location) + (declare (type index/2 slot-location)) ;; Mark slot as empty. (setf (aref table (* 2 slot-location)) +empty-ht-slot+ (aref table (1+ (* 2 slot-location))) +empty-ht-slot+) @@ -564,7 +662,8 @@ (hash-table-next-free-kv hash-table)) (setf (hash-table-next-free-kv hash-table) slot-location) (when hash-vector - (setf (aref hash-vector slot-location) +magic-hash-vector-value+)) + (setf (aref hash-vector slot-location) + +magic-hash-vector-value+)) (decf (hash-table-number-entries hash-table)) t)) (cond ((zerop next) @@ -588,45 +687,43 @@ (do ((prior next next) (next (aref next-vector next) (aref next-vector next))) ((zerop next) nil) - (declare (type index next)) + (declare (type index/2 next)) (when (and (= hashing (aref hash-vector next)) (funcall test-fun key (aref table (* 2 next)))) - (return-from remhash (clear-slot next-vector prior next))))))))))) + (return-from remhash + (clear-slot next-vector prior next))))))))))) (defun clrhash (hash-table) #!+sb-doc "This removes all the entries from HASH-TABLE and returns the hash table itself." - (let* ((kv-vector (hash-table-table hash-table)) - (kv-length (length kv-vector)) - (next-vector (hash-table-next-vector hash-table)) - (hash-vector (hash-table-hash-vector hash-table)) - (size (length next-vector)) - (index-vector (hash-table-index-vector hash-table)) - (length (length index-vector))) - ;; Disable GC tricks. - (set-header-data kv-vector sb!vm:vector-normal-subtype) - ;; Mark all slots as empty by setting all keys and values to magic - ;; tag. - (do ((i 2 (1+ i))) - ((>= i kv-length)) - (setf (aref kv-vector i) +empty-ht-slot+)) - (aver (eq (aref kv-vector 0) hash-table)) - ;; Set up the free list, all free. - (do ((i 1 (1+ i))) - ((>= i (1- size))) - (setf (aref next-vector i) (1+ i))) - (setf (aref next-vector (1- size)) 0) - (setf (hash-table-next-free-kv hash-table) 1) - (setf (hash-table-needing-rehash hash-table) 0) - ;; Clear the index-vector. - (dotimes (i length) - (setf (aref index-vector i) 0)) - ;; Clear the hash-vector. - (when hash-vector - (dotimes (i size) - (setf (aref hash-vector i) +magic-hash-vector-value+)))) - (setf (hash-table-number-entries hash-table) 0) + (declare (optimize speed)) + (with-spinlock-and-without-gcing ((hash-table-spinlock hash-table)) + (let* ((kv-vector (hash-table-table hash-table)) + (next-vector (hash-table-next-vector hash-table)) + (hash-vector (hash-table-hash-vector hash-table)) + (size (length next-vector)) + (index-vector (hash-table-index-vector hash-table))) + ;; Disable GC tricks. + (set-header-data kv-vector sb!vm:vector-normal-subtype) + ;; Mark all slots as empty by setting all keys and values to magic + ;; tag. + (aver (eq (aref kv-vector 0) hash-table)) + (fill kv-vector +empty-ht-slot+ :start 2) + ;; Set up the free list, all free. + (do ((i 1 (1+ i))) + ((>= i (1- size))) + (setf (aref next-vector i) (1+ i))) + (setf (aref next-vector (1- size)) 0) + (setf (hash-table-next-free-kv hash-table) 1) + (setf (hash-table-needing-rehash hash-table) 0) + ;; Clear the index-vector. + (fill index-vector 0) + ;; Clear the hash-vector. + (when hash-vector + (fill hash-vector +magic-hash-vector-value+))) + (setf (hash-table-cache hash-table) nil) + (setf (hash-table-number-entries hash-table) 0)) hash-table) ;;;; MAPHASH @@ -640,43 +737,44 @@ (declaim (inline maphash)) (defun maphash (function-designator hash-table) #!+sb-doc - "For each entry in HASH-TABLE, call the designated two-argument function - on the key and value of the entry. Return NIL." + "For each entry in HASH-TABLE, call the designated two-argument function on +the key and value of the entry. Return NIL." + ;; This essentially duplicates WITH-HASH-TABLE-ITERATOR, so + ;; any changes here should be reflected there as well. (let ((fun (%coerce-callable-to-fun function-designator)) - (size (length (hash-table-next-vector hash-table)))) + (size (length (hash-table-next-vector hash-table)))) (declare (type function fun)) (do ((i 1 (1+ i))) - ((>= i size)) - (declare (type index i)) + ((>= i size)) + (declare (type index/2 i)) (let* ((kv-vector (hash-table-table hash-table)) - (key (aref kv-vector (* 2 i))) - (value (aref kv-vector (1+ (* 2 i))))) - (unless (and (eq key +empty-ht-slot+) - (eq value +empty-ht-slot+)) - (funcall fun key value)))))) + (key (aref kv-vector (* 2 i))) + (value (aref kv-vector (1+ (* 2 i))))) + ;; We are running without locking or WITHOUT-GCING. For a weak + ;; :VALUE hash table it's possible that the GC hit after KEY + ;; was read and now the entry is gone. So check if either the + ;; key or the value is empty. + (unless (or (eq key +empty-ht-slot+) + (eq value +empty-ht-slot+)) + (funcall fun key value)))))) ;;;; methods on HASH-TABLE ;;; Return a list of keyword args and values to use for MAKE-HASH-TABLE ;;; when reconstructing HASH-TABLE. (defun %hash-table-ctor-args (hash-table) - (when (hash-table-weak-p hash-table) - ;; FIXME: This might actually work with no trouble, but as of - ;; sbcl-0.6.12.10 when this code was written, weak hash tables - ;; weren't working yet, so I couldn't test it. When weak hash - ;; tables are supported again, this should be fixed. - (error "can't dump weak hash tables readably")) ; defensive programming.. `(:test ',(hash-table-test hash-table) :size ',(hash-table-size hash-table) :rehash-size ',(hash-table-rehash-size hash-table) - :rehash-threshold ',(hash-table-rehash-threshold hash-table))) + :rehash-threshold ',(hash-table-rehash-threshold hash-table) + :weakness ',(hash-table-weakness hash-table))) ;;; Return an association list representing the same data as HASH-TABLE. (defun %hash-table-alist (hash-table) (let ((result nil)) (maphash (lambda (key value) - (push (cons key value) result)) - hash-table) + (push (cons key value) result)) + hash-table) result)) ;;; Stuff an association list into HASH-TABLE. Return the hash table, @@ -690,23 +788,21 @@ (def!method print-object ((hash-table hash-table) stream) (declare (type stream stream)) - (cond ((not *print-readably*) - (print-unreadable-object (hash-table stream :type t :identity t) - (format stream - ":TEST ~S :COUNT ~S" - (hash-table-test hash-table) - (hash-table-count hash-table)))) - ((not *read-eval*) - (error "can't print hash tables readably without *READ-EVAL*")) - (t - (with-standard-io-syntax - (format stream - "#.~W" - `(%stuff-hash-table (make-hash-table ,@(%hash-table-ctor-args - hash-table)) - ',(%hash-table-alist hash-table))))))) + (cond ((or (not *print-readably*) (not *read-eval*)) + (print-unreadable-object (hash-table stream :type t :identity t) + (format stream + ":TEST ~S :COUNT ~S" + (hash-table-test hash-table) + (hash-table-count hash-table)))) + (t + (with-standard-io-syntax + (format stream + "#.~W" + `(%stuff-hash-table (make-hash-table ,@(%hash-table-ctor-args + hash-table)) + ',(%hash-table-alist hash-table))))))) (def!method make-load-form ((hash-table hash-table) &optional environment) (declare (ignore environment)) (values `(make-hash-table ,@(%hash-table-ctor-args hash-table)) - `(%stuff-hash-table ,hash-table ',(%hash-table-alist hash-table)))) + `(%stuff-hash-table ,hash-table ',(%hash-table-alist hash-table))))