racket/c/thread.c
dyb 82b2cda639 compress-level parameter, improvement in lz4 compression, and various other related improvements
- added compress-level parameter to select a compression level for
  file writing and changed the default for lz4 compression to do a
  better job compressing.  finished splitting glz input routines
  apart from glz output routines and did a bit of other restructuring.
  removed gzxfile struct-as-bytevector wrapper and moved its fd
  into glzFile.  moved DEACTIVATE to before glzdopen_input calls
  in S_new_open_input_fd and S_compress_input_fd, since glzdopen_input
  reads from the file and could block.  the compress format and now
  level are now recorded directly the thread context.  replaced
  as-gz? flag bit in compressed bytevector header word with a small
  number of bits recording the compression format at the bottom of
  the header word.  flushed a couple of bytevector compression mats
  that depended on the old representation.  (these last few changes
  should make adding new compression formats easier.)  added
  s-directory build options to choose whether to compress and, if
  so, the format and level.
    compress-io.h, compress-io.c, new-io.c, equates.h, system.h,
    scheme.c, gc.c,
    io.ss, cmacros.ss, back.ss, bytevector.ss, primdata.ss, s/Mf-base,
    io.ms, mat.ss, bytevector.ms, root-experr*,
    release_notes.stex, io.stex, system.stex, objects.stex
- improved the effectiveness of LZ4 boot-file compression to within
  15% of gzip by increasing the lz4 output-port in_buffer size to
  1<<18.  With the previous size (1<<14) LZ4-compressed boot files
  were about 50% larger.  set the lz4 input-port in_buffer and
  out_buffer sizes to 1<<12 and 1<<14.  there's no clear win at
  present for larger input-port buffer sizes.
    compress-io.c
- To reduce the memory hit for the increased output-port in_buffer
  size and the corresponding increase in computed out_buffer size,
  one output-side out_buffer is now allocated (lazily) per thread
  and stored in the thread context.  The other buffers are now
  directly a part of the lz4File_out and lz4File_in structures
  rather than allocated separately.
    compress-io.c, scheme.c, gc.c,
    cmacros.ss
- split out the buffer emit code from glzwrite_lz4 into a
  separate glzemit_lz4 helper that is now also used by gzclose
  so we can avoid dealing with a NULL buffer in glzwrite_lz4.
  glzwrite_lz4 also uses it to writing large buffers directly and
  avoid the memcpy.
    compress-io.c
- replaced lz4File_out and lz4File_in mode enumeration with the
  compress format and inputp boolean.  using switch to check and
  raising exceptions for unexpected values to further simplify
  adding new compression formats in the future.
    compress-io.c
- replaced the never-defined struct lz4File pointer in glzFile
  union with the more specific struct lz4File_in_r and Lz4File_out_r
  pointers.
    compress-io.h, compress-io.c
- added free of lz4 structures to gzclose.  also changed file-close
  logic generally so that (1) port is marked closed before anything is
  freed to avoid dangling pointers in the case of an interrupt or
  error, and (2) structures are freed even in the case of a write
  or close error, before the error is reported.  also now mallocing
  glz and lz4 structures after possibility of errors have passed where
  possible and freeing them when not.
    compress-io.c,
    io.ss
- added return-value checks to malloc calls and to a couple of other
  C-library calls.
    compress-io.c
- corrected EINTR checks to look at errno rather than return codes.
    compress-io.c
- added S_ prefixes to the glz* exports
    externs.h, compress-io.c, new-io.c, scheme.c, fasl.c
- added entries for mutex-name and mutex-thread
    threads.stex

original commit: 722ffabef4c938bc92c0fe07f789a9ba350dc6c6
2019-04-18 05:47:19 -07:00

463 lines
12 KiB
C

/* thread.c
* Copyright 1984-2017 Cisco Systems, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "system.h"
/* locally defined functions */
#ifdef PTHREADS
static s_thread_rv_t start_thread PROTO((void *tc));
static IBOOL destroy_thread PROTO((ptr tc));
#endif
void S_thread_init() {
if (S_boot_time) {
S_protect(&S_G.threadno);
S_G.threadno = FIX(0);
#ifdef PTHREADS
/* this is also reset in scheme.c after heap restoration */
s_thread_mutex_init(&S_tc_mutex.pmutex);
S_tc_mutex.owner = s_thread_self();
S_tc_mutex.count = 0;
s_thread_cond_init(&S_collect_cond);
S_tc_mutex_depth = 0;
#endif /* PTHREADS */
}
}
/* this needs to be reworked. currently, S_create_thread_object is
called from main to create the base thread, from fork_thread when
there is already an active current thread, and from S_activate_thread
when there is no current thread. we have to avoid thread-local
allocation in at least the latter case, so we call vector_in and
cons_in and arrange for S_thread to use find_room rather than
thread_find_room. scheme.c does part of the initialization of the
base thread (e.g., parameters, current input/output ports) in one
or more places. */
ptr S_create_thread_object(who, p_tc) const char *who; ptr p_tc; {
ptr thread, tc;
INT i;
tc_mutex_acquire()
if (S_threads == Snil) {
tc = (ptr)S_G.thread_context;
} else { /* clone parent */
ptr p_v = PARAMETERS(p_tc);
iptr i, n = Svector_length(p_v);
/* use S_vector_in to avoid thread-local allocation */
ptr v = S_vector_in(space_new, 0, n);
tc = (ptr)malloc(size_tc);
if (tc == (ptr)0)
S_error(who, "unable to malloc thread data structure");
memcpy((void *)tc, (void *)p_tc, size_tc);
for (i = 0; i < n; i += 1)
INITVECTIT(v, i) = Svector_ref(p_v, i);
PARAMETERS(tc) = v;
CODERANGESTOFLUSH(tc) = Snil;
}
/* override nonclonable tc fields */
THREADNO(tc) = S_G.threadno;
S_G.threadno = S_add(S_G.threadno, FIX(1));
CCHAIN(tc) = Snil;
WINDERS(tc) = Snil;
STACKLINK(tc) = SYMVAL(S_G.null_continuation_id);
STACKCACHE(tc) = Snil;
/* S_reset_scheme_stack initializes stack, size, esp, and sfp */
S_reset_scheme_stack(tc, stack_slop);
FRAME(tc,0) = (ptr)&CODEIT(S_G.dummy_code_object,size_rp_header);
/* S_reset_allocation_pointer initializes ap and eap */
S_reset_allocation_pointer(tc);
RANDOMSEED(tc) = most_positive_fixnum < 0xffffffff ? most_positive_fixnum : 0xffffffff;
X(tc) = Y(tc) = U(tc) = V(tc) = W(tc) = FIX(0);
TIMERTICKS(tc) = Sfalse;
DISABLECOUNT(tc) = Sfixnum(0);
SIGNALINTERRUPTPENDING(tc) = Sfalse;
KEYBOARDINTERRUPTPENDING(tc) = Sfalse;
TARGETMACHINE(tc) = S_intern((const unsigned char *)MACHINE_TYPE);
/* choosing not to clone virtual registers */
for (i = 0 ; i < virtual_register_count ; i += 1) {
VIRTREG(tc, i) = FIX(0);
}
/* S_thread had better not do thread-local allocation */
thread = S_thread(tc);
/* use S_cons_in to avoid thread-local allocation */
S_threads = S_cons_in(space_new, 0, thread, S_threads);
S_nthreads += 1;
SETSYMVAL(S_G.active_threads_id,
FIX(UNFIX(SYMVAL(S_G.active_threads_id)) + 1));
ACTIVE(tc) = 1;
/* collect request is only thing that can be pending for new thread.
must do this after we're on the thread list in case the cons
adding us onto the thread list set collect-request-pending */
SOMETHINGPENDING(tc) = SYMVAL(S_G.collect_request_pending_id);
GUARDIANENTRIES(tc) = Snil;
LZ4OUTBUFFER(tc) = NULL;
tc_mutex_release()
return thread;
}
#ifdef PTHREADS
IBOOL Sactivate_thread() { /* create or reactivate current thread */
ptr tc = get_thread_context();
if (tc == (ptr)0) { /* thread created by someone else */
ptr thread;
/* borrow base thread for now */
thread = S_create_thread_object("Sactivate_thread", S_G.thread_context);
s_thread_setspecific(S_tc_key, (ptr)THREADTC(thread));
return 1;
} else {
reactivate_thread(tc)
return 0;
}
}
int S_activate_thread() { /* Like Sactivate_thread(), but returns a mode to revert the effect */
ptr tc = get_thread_context();
if (tc == (ptr)0) {
Sactivate_thread();
return unactivate_mode_destroy;
} else if (!ACTIVE(tc)) {
reactivate_thread(tc);
return unactivate_mode_deactivate;
} else
return unactivate_mode_noop;
}
void S_unactivate_thread(int mode) { /* Reverts a previous S_activate_thread() effect */
switch (mode) {
case unactivate_mode_deactivate:
Sdeactivate_thread();
break;
case unactivate_mode_destroy:
Sdestroy_thread();
break;
case unactivate_mode_noop:
default:
break;
}
}
void Sdeactivate_thread() { /* deactivate current thread */
ptr tc = get_thread_context();
if (tc != (ptr)0) deactivate_thread(tc)
}
int Sdestroy_thread() { /* destroy current thread */
ptr tc = get_thread_context();
if (tc != (ptr)0 && destroy_thread(tc)) {
s_thread_setspecific(S_tc_key, 0);
return 1;
}
return 0;
}
static IBOOL destroy_thread(tc) ptr tc; {
ptr *ls; IBOOL status;
status = 0;
tc_mutex_acquire()
ls = &S_threads;
while (*ls != Snil) {
ptr thread = Scar(*ls);
if (THREADTC(thread) == (uptr)tc) {
*ls = Scdr(*ls);
S_nthreads -= 1;
/* process remembered set before dropping allocation area */
S_scan_dirty((ptr **)EAP(tc), (ptr **)REAL_EAP(tc));
/* process guardian entries */
{
ptr target, ges, obj, next; seginfo *si;
target = S_G.guardians[0];
for (ges = GUARDIANENTRIES(tc); ges != Snil; ges = next) {
obj = GUARDIANOBJ(ges);
next = GUARDIANNEXT(ges);
if (!IMMEDIATE(obj) && (si = MaybeSegInfo(ptr_get_segment(obj))) != NULL && si->generation != static_generation) {
INITGUARDIANNEXT(ges) = target;
target = ges;
}
}
S_G.guardians[0] = target;
}
/* deactivate thread */
if (ACTIVE(tc)) {
SETSYMVAL(S_G.active_threads_id,
FIX(UNFIX(SYMVAL(S_G.active_threads_id)) - 1));
if (Sboolean_value(SYMVAL(S_G.collect_request_pending_id))
&& SYMVAL(S_G.active_threads_id) == FIX(0)) {
s_thread_cond_signal(&S_collect_cond);
}
}
if (LZ4OUTBUFFER(tc) != NULL) free(LZ4OUTBUFFER(tc));
free((void *)tc);
THREADTC(thread) = 0; /* mark it dead */
status = 1;
break;
}
ls = &Scdr(*ls);
}
tc_mutex_release()
return status;
}
ptr S_fork_thread(thunk) ptr thunk; {
ptr thread;
int status;
/* pass the current thread's context as the parent thread */
thread = S_create_thread_object("fork-thread", get_thread_context());
CP(THREADTC(thread)) = thunk;
if ((status = s_thread_create(start_thread, (void *)THREADTC(thread))) != 0) {
destroy_thread((ptr)THREADTC(thread));
S_error1("fork-thread", "failed: ~a", S_strerror(status));
}
return thread;
}
static s_thread_rv_t start_thread(p) void *p; {
ptr tc = (ptr)p; ptr cp;
s_thread_setspecific(S_tc_key, tc);
cp = CP(tc);
CP(tc) = Svoid; /* should hold calling code object, which we don't have */
TRAP(tc) = (ptr)default_timer_ticks;
Scall0(cp);
/* caution: calling into Scheme may result into a collection, so we
can't access any Scheme objects, e.g., cp, after this point. But tc
is static, so we can access it. */
/* find and destroy our thread */
destroy_thread(tc);
s_thread_setspecific(S_tc_key, (ptr)0);
s_thread_return;
}
scheme_mutex_t *S_make_mutex() {
scheme_mutex_t *m;
m = (scheme_mutex_t *)malloc(sizeof(scheme_mutex_t));
if (m == (scheme_mutex_t *)0)
S_error("make-mutex", "unable to malloc mutex");
s_thread_mutex_init(&m->pmutex);
m->owner = s_thread_self();
m->count = 0;
return m;
}
void S_mutex_free(m) scheme_mutex_t *m; {
s_thread_mutex_destroy(&m->pmutex);
free(m);
}
void S_mutex_acquire(m) scheme_mutex_t *m; {
s_thread_t self = s_thread_self();
iptr count;
INT status;
if ((count = m->count) > 0 && s_thread_equal(m->owner, self)) {
if (count == most_positive_fixnum)
S_error1("mutex-acquire", "recursion limit exceeded for ~s", m);
m->count = count + 1;
return;
}
if ((status = s_thread_mutex_lock(&m->pmutex)) != 0)
S_error1("mutex-acquire", "failed: ~a", S_strerror(status));
m->owner = self;
m->count = 1;
}
INT S_mutex_tryacquire(m) scheme_mutex_t *m; {
s_thread_t self = s_thread_self();
iptr count;
INT status;
if ((count = m->count) > 0 && s_thread_equal(m->owner, self)) {
if (count == most_positive_fixnum)
S_error1("mutex-acquire", "recursion limit exceeded for ~s", m);
m->count = count + 1;
return 0;
}
status = s_thread_mutex_trylock(&m->pmutex);
if (status == 0) {
m->owner = self;
m->count = 1;
} else if (status != EBUSY) {
S_error1("mutex-acquire", "failed: ~a", S_strerror(status));
}
return status;
}
void S_mutex_release(m) scheme_mutex_t *m; {
s_thread_t self = s_thread_self();
iptr count;
INT status;
if ((count = m->count) == 0 || !s_thread_equal(m->owner, self))
S_error1("mutex-release", "thread does not own mutex ~s", m);
if ((m->count = count - 1) == 0)
if ((status = s_thread_mutex_unlock(&m->pmutex)) != 0)
S_error1("mutex-release", "failed: ~a", S_strerror(status));
}
s_thread_cond_t *S_make_condition() {
s_thread_cond_t *c;
c = (s_thread_cond_t *)malloc(sizeof(s_thread_cond_t));
if (c == (s_thread_cond_t *)0)
S_error("make-condition", "unable to malloc condition");
s_thread_cond_init(c);
return c;
}
void S_condition_free(c) s_thread_cond_t *c; {
s_thread_cond_destroy(c);
free(c);
}
#ifdef FEATURE_WINDOWS
static inline int s_thread_cond_timedwait(s_thread_cond_t *cond, s_thread_mutex_t *mutex, int typeno, long sec, long nsec) {
if (typeno == time_utc) {
struct timespec now;
S_gettime(time_utc, &now);
sec -= (long)now.tv_sec;
nsec -= now.tv_nsec;
if (nsec < 0) {
sec -= 1;
nsec += 1000000000;
}
}
if (sec < 0) {
sec = 0;
nsec = 0;
}
if (SleepConditionVariableCS(cond, mutex, sec*1000 + nsec/1000000)) {
return 0;
} else if (GetLastError() == ERROR_TIMEOUT) {
return ETIMEDOUT;
} else {
return EINVAL;
}
}
#else /* FEATURE_WINDOWS */
static inline int s_thread_cond_timedwait(s_thread_cond_t *cond, s_thread_mutex_t *mutex, int typeno, long sec, long nsec) {
struct timespec t;
if (typeno == time_duration) {
struct timespec now;
S_gettime(time_utc, &now);
t.tv_sec = now.tv_sec + sec;
t.tv_nsec = now.tv_nsec + nsec;
if (t.tv_nsec >= 1000000000) {
t.tv_sec += 1;
t.tv_nsec -= 1000000000;
}
} else {
t.tv_sec = sec;
t.tv_nsec = nsec;
}
return pthread_cond_timedwait(cond, mutex, &t);
}
#endif /* FEATURE_WINDOWS */
#define Srecord_ref(x,i) (((ptr *)((uptr)(x)+record_data_disp))[i])
IBOOL S_condition_wait(c, m, t) s_thread_cond_t *c; scheme_mutex_t *m; ptr t; {
ptr tc = get_thread_context();
s_thread_t self = s_thread_self();
iptr count;
INT typeno;
long sec;
long nsec;
INT status;
if ((count = m->count) == 0 || !s_thread_equal(m->owner, self))
S_error1("condition-wait", "thread does not own mutex ~s", m);
if (count != 1)
S_error1("condition-wait", "mutex ~s is recursively locked", m);
if (t != Sfalse) {
/* Keep in sync with ts record in s/date.ss */
typeno = Sinteger32_value(Srecord_ref(t,0));
sec = Sinteger32_value(Scar(Srecord_ref(t,1)));
nsec = Sinteger32_value(Scdr(Srecord_ref(t,1)));
}
if (c == &S_collect_cond || DISABLECOUNT(tc) == 0) {
deactivate_thread(tc)
}
m->count = 0;
status = (t == Sfalse) ? s_thread_cond_wait(c, &m->pmutex) :
s_thread_cond_timedwait(c, &m->pmutex, typeno, sec, nsec);
m->owner = self;
m->count = 1;
if (c == &S_collect_cond || DISABLECOUNT(tc) == 0) {
reactivate_thread(tc)
}
if (status == 0) {
return 1;
} else if (status == ETIMEDOUT) {
return 0;
} else {
S_error1("condition-wait", "failed: ~a", S_strerror(status));
return 0;
}
}
#endif /* PTHREADS */