diff options
Diffstat (limited to 'gcell/src/lib/runtime')
26 files changed, 4583 insertions, 0 deletions
diff --git a/gcell/src/lib/runtime/Makefile.am b/gcell/src/lib/runtime/Makefile.am new file mode 100644 index 000000000..3f2077c08 --- /dev/null +++ b/gcell/src/lib/runtime/Makefile.am @@ -0,0 +1,76 @@ +# +# Copyright 2007,2008 Free Software Foundation, Inc. +# +# This file is part of GNU Radio +# +# GNU Radio is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3, or (at your option) +# any later version. +# +# GNU Radio is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License along +# with this program; if not, write to the Free Software Foundation, Inc., +# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. +# + +include $(top_srcdir)/Makefile.common + +# SUBDIRS = spu . + +IBM_PPU_SYNC_INCLUDES = -I$(top_srcdir)/gcell/src/ibm/sync/ppu_source + + +AM_CPPFLAGS = $(DEFINES) $(OMNITHREAD_INCLUDES) $(MBLOCK_INCLUDES) $(CPPUNIT_INCLUDES) \ + $(GCELL_INCLUDES) $(IBM_PPU_SYNC_INCLUDES) $(WITH_INCLUDES) + + +noinst_LTLIBRARIES = libruntime.la libruntime-qa.la + +libruntime_la_SOURCES = \ + gc_job_manager.cc \ + gc_job_manager_impl.cc \ + gc_jd_queue.c \ + gc_jd_stack.c \ + gc_proc_def_utils.cc + +libruntime_qa_la_SOURCES = \ + qa_lib.cc \ + qa_jd_queue.cc \ + qa_jd_stack.cc \ + qa_job_manager.cc + + +gcellinclude_HEADERS = \ + gc_job_manager.h + +noinst_HEADERS = \ + gc_client_thread_info.h \ + gc_job_manager_impl.h \ + gc_proc_def_utils.h \ + qa_jd_queue.h \ + qa_jd_stack.h \ + qa_job_manager.h \ + qa_lib.h + + +# This kruft is required to link the QA SPU executable into the PPE shared lib w/o warnings +gcell_qa.lo: ../spu/gcell_qa + ppu-embedspu -m32 -fpic gcell_qa ../spu/gcell_qa .libs/gcell_qa.o + @rm -f gcell_qa.lo + @echo "# gcell_qa.lo - a libtool object file" >> gcell_qa.lo + @echo "# Generated by ltmain.sh - GNU libtool 1.5.22 (1.1220.2.365 2005/12/18 22:14:06)" >> gcell_qa.lo + @echo "#" >> gcell_qa.lo + @echo "# Please DO NOT delete this file!" >> gcell_qa.lo + @echo "# It is necessary for linking the library." >> gcell_qa.lo + @echo "" >> gcell_qa.lo + @echo "pic_object='.libs/gcell_qa.o'" >> gcell_qa.lo + @echo "non_pic_object=none" >> gcell_qa.lo + +libruntime_qa_la_LIBADD = \ + gcell_qa.lo \ + libruntime.la diff --git a/gcell/src/lib/runtime/gc_client_thread_info.h b/gcell/src/lib/runtime/gc_client_thread_info.h new file mode 100644 index 000000000..fbb35d966 --- /dev/null +++ b/gcell/src/lib/runtime/gc_client_thread_info.h @@ -0,0 +1,81 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef INCLUDED_GC_CLIENT_THREAD_INFO_H +#define INCLUDED_GC_CLIENT_THREAD_INFO_H + +#include <omnithread.h> +#include <boost/utility.hpp> + +enum gc_ct_state { + CT_NOT_WAITING, + CT_WAIT_ALL, + CT_WAIT_ANY, +}; + +/* + * \brief per client-thread data used by gc_job_manager + * + * "Client threads" are any threads that invoke methods on + * gc_job_manager. We use pthread_set_specific to store a pointer to + * one of these for each thread that comes our way. + */ +class gc_client_thread_info : boost::noncopyable { +public: + gc_client_thread_info() : + d_free(1), d_cond(&d_mutex), d_state(CT_NOT_WAITING), + d_jobs_done(0), d_njobs_waiting_for(0), + d_jobs_waiting_for(0){ } + + ~gc_client_thread_info() { + d_free = 1; + d_state = CT_NOT_WAITING; + d_jobs_done = 0; + d_njobs_waiting_for = 0; + d_jobs_waiting_for = 0; + } + + //! is this cti free? (1->free, 0->in use) + uint32_t d_free; + + //! which client info are we? + uint16_t d_client_id; + + //! hold this mutex to manipulate anything below here + omni_mutex d_mutex; + + //! signaled by event handler to wake client thread up + omni_condition d_cond; + + //! Is this client waiting? + gc_ct_state d_state; + + //! Jobs that have finished and not yet been waited for (bitvector) + unsigned long *d_jobs_done; + + //! # of jobs we're waiting for + unsigned int d_njobs_waiting_for; + + //! Jobs that client thread is waiting for + gc_job_desc **d_jobs_waiting_for; + +}; + +#endif /* INCLUDED_GC_CLIENT_THREAD_INFO_H */ diff --git a/gcell/src/lib/runtime/gc_jd_queue.c b/gcell/src/lib/runtime/gc_jd_queue.c new file mode 100644 index 000000000..b5cdcac9b --- /dev/null +++ b/gcell/src/lib/runtime/gc_jd_queue.c @@ -0,0 +1,78 @@ +/* -*- c -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include "gc_jd_queue.h" +#include "memory_barrier.h" +#include <mutex_init.h> +#include <mutex_lock.h> +#include <mutex_unlock.h> + +void +gc_jd_queue_init(gc_jd_queue_t *q) +{ + _mutex_init(ptr_to_ea(&q->mutex)); + q->head = 0; + q->tail = 0; + smp_wmb(); +} + +void +gc_jd_queue_enqueue(gc_jd_queue_t *q, gc_job_desc_t *item) +{ + item->sys.next = 0; + _mutex_lock(ptr_to_ea(&q->mutex)); + smp_rmb(); // import barrier + + if (q->tail == 0){ // currently empty + q->tail = q->head = jdp_to_ea(item); + } + else { // not empty, append + ea_to_jdp(q->tail)->sys.next = jdp_to_ea(item); + q->tail = jdp_to_ea(item); + } + + smp_wmb(); // orders stores above before clearing of mutex + _mutex_unlock(ptr_to_ea(&q->mutex)); +} + +gc_job_desc_t * +gc_jd_queue_dequeue(gc_jd_queue_t *q) +{ + _mutex_lock(ptr_to_ea(&q->mutex)); + smp_rmb(); // import barrier + + gc_eaddr_t item_ea = q->head; + if (item_ea == 0){ // empty + _mutex_unlock(ptr_to_ea(&q->mutex)); + return 0; + } + + q->head = ea_to_jdp(item_ea)->sys.next; + if (q->head == 0) // now emtpy + q->tail = 0; + + gc_job_desc_t *item = ea_to_jdp(item_ea); + item->sys.next = 0; + + smp_wmb(); // orders stores above before clearing of mutex + _mutex_unlock(ptr_to_ea(&q->mutex)); + return item; +} diff --git a/gcell/src/lib/runtime/gc_jd_stack.c b/gcell/src/lib/runtime/gc_jd_stack.c new file mode 100644 index 000000000..0fffc0d1f --- /dev/null +++ b/gcell/src/lib/runtime/gc_jd_stack.c @@ -0,0 +1,108 @@ +/* -*- c -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include "gc_jd_stack.h" +#include "memory_barrier.h" + + +void +gc_jd_stack_init(gc_jd_stack_t *stack) +{ + stack->top = 0; +} + + +#ifdef __powerpc64__ // 64-bit mode + +void +gc_jd_stack_push(gc_jd_stack_t *stack, gc_job_desc_t *item) +{ + gc_eaddr_t top; + gc_eaddr_t item_ea = ptr_to_ea(item); + unsigned int done; + + do { + top = __ldarx(&stack->top); + item->sys.next = top; + smp_wmb(); // order store of item->next before store of stack->top + done = __stdcx(&stack->top, item_ea); + } while (unlikely(done == 0)); +} + +gc_job_desc_t * +gc_jd_stack_pop(gc_jd_stack_t *stack) +{ + gc_eaddr_t s; + gc_eaddr_t t; + unsigned int done; + + do { + s = __ldarx(&stack->top); + if (s == 0) /* stack's empty */ + return 0; + t = ((gc_job_desc_t *) ea_to_ptr(s))->sys.next; + done = __stdcx(&stack->top, t); + } while (unlikely(done == 0)); + + return ea_to_ptr(s); +} + +#else // 32-bit mode + +/* + * In 32-bit mode, gc_eaddr's will have the top 32-bits zero. + * The ldarx/stdcx instructions aren't available in 32-bit mode, + * thus we use lwarx/stwcx on the low 32-bits of the 64-bit addresses. + * Since we're big-endian, the low 32-bits are at word offset 1. + */ +void +gc_jd_stack_push(gc_jd_stack_t *stack, gc_job_desc_t *item) +{ + gc_eaddr_t top; + unsigned int done; + + do { + top = __lwarx((int32_t *)(&stack->top) + 1); + item->sys.next = top; + smp_wmb(); // order store of item->sys.next before store of stack->top + done = __stwcx((int32_t *)(&stack->top) + 1, item); + } while (unlikely(done == 0)); +} + +gc_job_desc_t * +gc_jd_stack_pop(gc_jd_stack_t *stack) +{ + gc_eaddr_t s; + gc_eaddr_t t; + unsigned int done; + + do { + s = __lwarx((int32_t *)(&stack->top) + 1); + if (s == 0) /* stack's empty */ + return 0; + t = ((gc_job_desc_t *) ea_to_ptr(s))->sys.next; + done = __stwcx((int32_t *)(&stack->top) + 1, (uint32_t) t); + } while (unlikely(done == 0)); + + return ea_to_ptr(s); +} + +#endif diff --git a/gcell/src/lib/runtime/gc_job_manager.cc b/gcell/src/lib/runtime/gc_job_manager.cc new file mode 100644 index 000000000..edb52a5b7 --- /dev/null +++ b/gcell/src/lib/runtime/gc_job_manager.cc @@ -0,0 +1,54 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif +#include "gc_job_manager.h" +#include "gc_job_manager_impl.h" + +gc_job_manager * +gc_make_job_manager(const gc_jm_options *options) +{ + return new gc_job_manager_impl(options); +} + +gc_job_manager::gc_job_manager(const gc_jm_options *options) +{ + // nop +} + +gc_job_manager::~gc_job_manager() +{ + // nop +} + +void +gc_job_manager::set_debug(int debug) +{ + // nop +} + +int +gc_job_manager::debug() +{ + return 0; +} diff --git a/gcell/src/lib/runtime/gc_job_manager.h b/gcell/src/lib/runtime/gc_job_manager.h new file mode 100644 index 000000000..9c8e70bf8 --- /dev/null +++ b/gcell/src/lib/runtime/gc_job_manager.h @@ -0,0 +1,171 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007,2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifndef INCLUDED_GC_JOB_MANAGER_H +#define INCLUDED_GC_JOB_MANAGER_H + +#include <boost/utility.hpp> +#include <vector> +#include <string> +#include <libspe2.h> +#include "gc_job_desc.h" + +class gc_job_manager; + +enum gc_wait_mode { + GC_WAIT_ANY, + GC_WAIT_ALL, +}; + +/* + * \brief Options that configure the job_manager. + * The default values are reasonable. + */ +struct gc_jm_options { + unsigned int max_jobs; // max # of job descriptors in system + unsigned int max_client_threads; // max # of client threads of job manager + unsigned int nspes; // how many SPEs shall we use? 0 -> all of them + bool gang_schedule; // shall we gang schedule? + bool use_affinity; // shall we try for affinity (FIXME not implmented) + bool enable_logging; // shall we log SPE events? + uint32_t log2_nlog_entries; // log2 of number of log entries (default is 12 == 4k) + spe_program_handle_t *program_handle; // program to load into SPEs + + gc_jm_options() : + max_jobs(0), max_client_threads(0), nspes(0), + gang_schedule(true), use_affinity(false), + enable_logging(false), log2_nlog_entries(12), + program_handle(0) + { + } +}; + + +/* + * \brief Create an instance of the job manager + */ +gc_job_manager * +gc_make_job_manager(const gc_jm_options *options = 0); + + +/*! + * \brief Abstract class that manages SPE jobs. + * + * There is typically a single instance derived from this class. + * It is safe to call its methods from any thread. + */ +class gc_job_manager : boost::noncopyable +{ +public: + gc_job_manager(const gc_jm_options *options = 0); + + virtual ~gc_job_manager(); + + /*! + * Stop accepting new jobs. Wait for existing jobs to complete. + * Return all managed SPE's to the system. + */ + virtual bool shutdown() = 0; + + /*! + * \brief Return number of SPE's currently allocated to job manager. + */ + virtual int nspes() const = 0; + + /*! + * \brief Return a pointer to a properly aligned job descriptor, + * or zero if none are available. + */ + virtual gc_job_desc *alloc_job_desc() = 0; + + /* + *! Free a job descriptor previously allocated with alloc_job_desc() + * + * \param[in] jd pointer to job descriptor to free. + */ + virtual void free_job_desc(gc_job_desc *jd) = 0; + + /*! + * \brief Submit a job for asynchronous processing on an SPE. + * + * \param[in] jd pointer to job description + * + * The caller must not read or write the job description + * or any of the memory associated with any indirect arguments + * until after calling wait_job. + * + * \returns true iff the job was successfully enqueued. + * If submit_job returns false, check jd->status for additional info. + */ + virtual bool submit_job(gc_job_desc *jd) = 0; + + /*! + * \brief Wait for job to complete. + * + * A thread may only wait for jobs which it submitted. + * + * \returns true if sucessful, else false. + */ + virtual bool + wait_job(gc_job_desc *jd) = 0; + + /*! + * \brief wait for 1 or more jobs to complete. + * + * \param[input] njobs is the length of arrays \p jd and \p done. + * \param[input] jd are the jobs that are to be waited for. + * \param[output] done indicates whether the corresponding job is complete. + * \param[input] mode indicates whether to wait for ALL or ANY of the jobs + * in \p jd to complete. + * + * A thread may only wait for jobs which it submitted. + * + * \returns number of jobs completed, or -1 if error. + */ + virtual int + wait_jobs(unsigned int njobs, + gc_job_desc *jd[], bool done[], gc_wait_mode mode) = 0; + + /*! + * Return the maximum number of bytes of EA arguments that may be + * copied to or from the SPE in a single job. The limit applies + * independently to the "get" and "put" args. + * \sa gc_job_desc_t, gc_job_ea_args_t + */ + virtual int ea_args_maxsize() = 0; + + /*! + * Return gc_proc_id_t associated with spu procedure \p proc_name if one + * exists, otherwise return GCP_UNKNOWN_PROC. + */ + virtual gc_proc_id_t lookup_proc(const std::string &proc_name) = 0; + + /*! + * Return a vector of all known spu procedure names. + */ + virtual std::vector<std::string> proc_names() = 0; + + virtual void set_debug(int debug); + virtual int debug(); +}; + + +#endif /* INCLUDED_GC_JOB_MANAGER_H */ diff --git a/gcell/src/lib/runtime/gc_job_manager_impl.cc b/gcell/src/lib/runtime/gc_job_manager_impl.cc new file mode 100644 index 000000000..13d8f0d97 --- /dev/null +++ b/gcell/src/lib/runtime/gc_job_manager_impl.cc @@ -0,0 +1,1274 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif +#include <gc_job_manager_impl.h> +#include <gc_mbox.h> +#include <gc_proc_def_utils.h> + +#include <stdio.h> +#include <stdexcept> +#include <stdlib.h> +#include <atomic_dec_if_positive.h> +#include <memory_barrier.h> +#include <unistd.h> +#include <sys/mman.h> +#include <sys/types.h> +#include <sys/stat.h> +#include <fcntl.h> + + +static const size_t CACHE_LINE_SIZE = 128; + +static const unsigned int DEFAULT_MAX_JOBS = 128; +static const unsigned int DEFAULT_MAX_CLIENT_THREADS = 64; + +// FIXME this really depends on the SPU code... +static const unsigned int MAX_TOTAL_INDIRECT_LENGTH = 16 * 1024; + + +static bool s_key_initialized = false; +static pthread_key_t s_client_key; + +static int s_worker_debug = 0; + +// custom deleter of gang_contexts for use with boost::shared_ptr +class gang_deleter { +public: + void operator()(spe_gang_context_ptr_t ctx) { + if (ctx){ + int r = spe_gang_context_destroy(ctx); + if (r != 0){ + perror("spe_gang_context_destroy"); + } + } + } +}; + +// custom deleter +class spe_program_handle_deleter { +public: + void operator()(spe_program_handle_t *program) { + if (program){ + int r = spe_image_close(program); + if (r != 0){ + perror("spe_image_close"); + } + } + } +}; + + +// custom deleter of anything that can be freed with "free" +class free_deleter { +public: + void operator()(void *p) { + free(p); + } +}; + + +/* + * Called when client thread is destroyed. + * We mark our client info free. + */ +static void +client_key_destructor(void *p) +{ + ((gc_client_thread_info *) p)->d_free = 1; +} + +/* + * Return pointer to cache-aligned chunk of storage of size size bytes. + * Throw if can't allocate memory. The storage should be freed + * with "free" when done. The memory is initialized to zero. + */ +static void * +aligned_alloc(size_t size, size_t alignment = CACHE_LINE_SIZE) +{ + void *p = 0; + if (posix_memalign(&p, alignment, size) != 0){ + perror("posix_memalign"); + throw std::runtime_error("memory"); + } + memset(p, 0, size); // zero the memory + return p; +} + +static bool +is_power_of_2(uint32_t x) +{ + return (x != 0) && !(x & (x - 1)); +} + +//////////////////////////////////////////////////////////////////////// + + +gc_job_manager_impl::gc_job_manager_impl(const gc_jm_options *options) + : d_debug(0), d_spu_args(0), + d_eh_cond(&d_eh_mutex), d_eh_thread(0), d_eh_state(EHS_INIT), + d_shutdown_requested(false), + d_client_thread(0), d_ea_args_maxsize(0), + d_proc_def(0), d_proc_def_ls_addr(0), d_nproc_defs(0) +{ + if (!s_key_initialized){ + int r = pthread_key_create(&s_client_key, client_key_destructor); + if (r != 0) + throw std::runtime_error("pthread_key_create"); + s_key_initialized = true; + } + + // ensure it's zero + pthread_setspecific(s_client_key, 0); + + if (options != 0) + d_options = *options; + + // provide the real default for those indicated with a zero + if (d_options.max_jobs == 0) + d_options.max_jobs = DEFAULT_MAX_JOBS; + if (d_options.max_client_threads == 0) + d_options.max_client_threads = DEFAULT_MAX_CLIENT_THREADS; + + if (d_options.program_handle == 0){ + fprintf(stderr, "gc_job_manager: options->program_handle must be non-zero\n"); + throw std::runtime_error("gc_job_manager: options->program_handle must be non-zero"); + } + + int ncpu_nodes = spe_cpu_info_get(SPE_COUNT_PHYSICAL_CPU_NODES, -1); + int nusable_spes = spe_cpu_info_get(SPE_COUNT_USABLE_SPES, -1); + + if (debug()){ + printf("cpu_nodes = %d\n", ncpu_nodes); + for (int i = 0; i < ncpu_nodes; i++){ + printf("node[%d].physical_spes = %2d\n", i, + spe_cpu_info_get(SPE_COUNT_PHYSICAL_SPES, i)); + printf("node[%d].usable_spes = %2d\n", i, + spe_cpu_info_get(SPE_COUNT_USABLE_SPES, i)); + } + } + + // clamp nspes + d_options.nspes = std::min(d_options.nspes, (unsigned int) MAX_SPES); + nusable_spes = std::min(nusable_spes, (int) MAX_SPES); + + // + // sanity check requested number of spes. + // + if (d_options.nspes == 0) // use all of them + d_options.nspes = nusable_spes; + else { + if (d_options.nspes > (unsigned int) nusable_spes){ + fprintf(stderr, + "gc_job_manager: warning: caller requested %d spes. There are only %d available.\n", + d_options.nspes, nusable_spes); + if (d_options.gang_schedule){ + // If we're gang scheduling we'll never get scheduled if we + // ask for more than are available. + throw std::out_of_range("gang_scheduling: not enough spes available"); + } + else { // FIXME clamp to usable. problem on PS3 when overcommited + fprintf(stderr, "gc_job_manager: clamping nspes to %d\n", nusable_spes); + d_options.nspes = nusable_spes; + } + } + } + + if (d_options.use_affinity){ + printf("gc_job_manager: warning: affinity request was ignored\n"); + } + + if (d_options.gang_schedule){ + d_gang = spe_gang_context_sptr(spe_gang_context_create(0), gang_deleter()); + if (!d_gang){ + perror("gc_job_manager_impl[spe_gang_context_create]"); + throw std::runtime_error("spe_gang_context_create"); + } + } + + // ---------------------------------------------------------------- + // initalize the job queue + + d_queue = (gc_jd_queue_t *) aligned_alloc(sizeof(gc_jd_queue_t)); + _d_queue_boost = + boost::shared_ptr<void>((void *) d_queue, free_deleter()); + gc_jd_queue_init(d_queue); + + + // ---------------------------------------------------------------- + // create the spe contexts + + // 1 spu_arg struct for each SPE + assert(sizeof(gc_spu_args_t) % 16 == 0); + d_spu_args = + (gc_spu_args_t *) aligned_alloc(MAX_SPES * sizeof(gc_spu_args_t), 16); + _d_spu_args_boost = + boost::shared_ptr<void>((void *) d_spu_args, free_deleter()); + + // 2 completion info structs for each SPE (we double buffer them) + assert(sizeof(gc_comp_info_t) % CACHE_LINE_SIZE == 0); + d_comp_info = + (gc_comp_info_t *) aligned_alloc(2 * MAX_SPES * sizeof(gc_comp_info_t), + CACHE_LINE_SIZE); + _d_comp_info_boost = + boost::shared_ptr<void>((void *) d_comp_info, free_deleter()); + + + // get a handle to the spe program + + spe_program_handle_t *spe_image = d_options.program_handle; + + // fish proc_def table out of SPE ELF file + + if (!gcpd_find_table(spe_image, &d_proc_def, &d_nproc_defs, &d_proc_def_ls_addr)){ + fprintf(stderr, "gc_job_manager_impl: couldn't find gc_proc_defs in SPE ELF file.\n"); + throw std::runtime_error("no gc_proc_defs"); + } + // fprintf(stderr, "d_proc_def_ls_addr = 0x%0x\n", d_proc_def_ls_addr); + + int spe_flags = (SPE_EVENTS_ENABLE + | SPE_CFG_SIGNOTIFY1_OR + | SPE_CFG_SIGNOTIFY2_OR); + + for (unsigned int i = 0; i < d_options.nspes; i++){ + // FIXME affinity stuff goes here + d_worker[i].spe_ctx = spe_context_create(spe_flags, d_gang.get());; + if (d_worker[i].spe_ctx == 0){ + perror("spe_context_create"); + throw std::runtime_error("spe_context_create"); + } + d_worker[i].spe_idx = i; + d_worker[i].spu_args = &d_spu_args[i]; + d_worker[i].spu_args->queue = ptr_to_ea(d_queue); + d_worker[i].spu_args->comp_info[0] = ptr_to_ea(&d_comp_info[2*i+0]); + d_worker[i].spu_args->comp_info[1] = ptr_to_ea(&d_comp_info[2*i+1]); + d_worker[i].spu_args->spu_idx = i; + d_worker[i].spu_args->nspus = d_options.nspes; + d_worker[i].spu_args->proc_def_ls_addr = d_proc_def_ls_addr; + d_worker[i].spu_args->nproc_defs = d_nproc_defs; + d_worker[i].spu_args->log.base = 0; + d_worker[i].spu_args->log.nentries = 0; + d_worker[i].state = WS_INIT; + + int r = spe_program_load(d_worker[i].spe_ctx, spe_image); + if (r != 0){ + perror("spe_program_load"); + throw std::runtime_error("spe_program_load"); + } + } + + setup_logfiles(); + + // ---------------------------------------------------------------- + // initalize the free list of job descriptors + + d_free_list = (gc_jd_stack_t *) aligned_alloc(sizeof(gc_jd_stack_t)); + // This ensures that the memory associated with d_free_list is + // automatically freed in the destructor or if an exception occurs + // here in the constructor. + _d_free_list_boost = + boost::shared_ptr<void>((void *) d_free_list, free_deleter()); + gc_jd_stack_init(d_free_list); + + if (debug()){ + printf("sizeof(d_jd[0]) = %d (0x%x)\n", sizeof(d_jd[0]), sizeof(d_jd[0])); + printf("max_jobs = %u\n", d_options.max_jobs); + } + + // Initialize the array of job descriptors. + d_jd = (gc_job_desc_t *) aligned_alloc(sizeof(d_jd[0]) * d_options.max_jobs); + _d_jd_boost = boost::shared_ptr<void>((void *) d_jd, free_deleter()); + + + // set unique job_id + for (int i = 0; i < (int) d_options.max_jobs; i++) + d_jd[i].sys.job_id = i; + + // push them onto the free list + for (int i = d_options.max_jobs - 1; i >= 0; i--) + free_job_desc(&d_jd[i]); + + // ---------------------------------------------------------------- + // initialize d_client_thread + + { + gc_client_thread_info_sa cti( + new gc_client_thread_info[d_options.max_client_threads]); + + d_client_thread.swap(cti); + + for (unsigned int i = 0; i < d_options.max_client_threads; i++) + d_client_thread[i].d_client_id = i; + } + + // ---------------------------------------------------------------- + // initialize bitvectors + + // initialize d_bvlen, the number of longs in job related bitvectors. + int bits_per_long = sizeof(unsigned long) * 8; + d_bvlen = (d_options.max_jobs + bits_per_long - 1) / bits_per_long; + + // allocate all bitvectors in a single cache-aligned chunk + size_t nlongs = d_bvlen * d_options.max_client_threads; + void *p = aligned_alloc(nlongs * sizeof(unsigned long)); + _d_all_bitvectors = boost::shared_ptr<void>(p, free_deleter()); + + // Now point the gc_client_thread_info bitvectors into this storage + unsigned long *v = (unsigned long *) p; + + for (unsigned int i = 0; i < d_options.max_client_threads; i++, v += d_bvlen) + d_client_thread[i].d_jobs_done = v; + + + // ---------------------------------------------------------------- + // create the spe event handler & worker (SPE) threads + + create_event_handler(); + +} + +//////////////////////////////////////////////////////////////////////// + +gc_job_manager_impl::~gc_job_manager_impl() +{ + shutdown(); + + d_jd = 0; // handled via _d_jd_boost + d_free_list = 0; // handled via _d_free_list_boost + d_queue = 0; // handled via _d_queue_boost + + // clear cti, since we've deleted the underlying data + pthread_setspecific(s_client_key, 0); + + unmap_logfiles(); +} + +bool +gc_job_manager_impl::shutdown() +{ + omni_mutex_lock l(d_eh_mutex); + + d_shutdown_requested = true; // set flag for event handler thread + + // should only happens during early QA code + if (d_eh_thread == 0 && d_eh_state == EHS_INIT) + return false; + + while (d_eh_state != EHS_DEAD) // wait for it to finish + d_eh_cond.wait(); + + return true; +} + +int +gc_job_manager_impl::nspes() const +{ + return d_options.nspes; +} + +//////////////////////////////////////////////////////////////////////// + +void +gc_job_manager_impl::bv_zero(unsigned long *bv) +{ + memset(bv, 0, sizeof(unsigned long) * d_bvlen); +} + +inline void +gc_job_manager_impl::bv_clr(unsigned long *bv, unsigned int bitno) +{ + unsigned int wi = bitno / (sizeof (unsigned long) * 8); + unsigned int bi = bitno & ((sizeof (unsigned long) * 8) - 1); + bv[wi] &= ~(1UL << bi); +} + +inline void +gc_job_manager_impl::bv_set(unsigned long *bv, unsigned int bitno) +{ + unsigned int wi = bitno / (sizeof (unsigned long) * 8); + unsigned int bi = bitno & ((sizeof (unsigned long) * 8) - 1); + bv[wi] |= (1UL << bi); +} + +inline bool +gc_job_manager_impl::bv_isset(unsigned long *bv, unsigned int bitno) +{ + unsigned int wi = bitno / (sizeof (unsigned long) * 8); + unsigned int bi = bitno & ((sizeof (unsigned long) * 8) - 1); + return (bv[wi] & (1UL << bi)) != 0; +} + +inline bool +gc_job_manager_impl::bv_isclr(unsigned long *bv, unsigned int bitno) +{ + unsigned int wi = bitno / (sizeof (unsigned long) * 8); + unsigned int bi = bitno & ((sizeof (unsigned long) * 8) - 1); + return (bv[wi] & (1UL << bi)) == 0; +} + +//////////////////////////////////////////////////////////////////////// + +gc_job_desc * +gc_job_manager_impl::alloc_job_desc() +{ + // stack is lock free, thus safe to call from any thread + return gc_jd_stack_pop(d_free_list); +} + +void +gc_job_manager_impl::free_job_desc(gc_job_desc *jd) +{ + // stack is lock free, thus safe to call from any thread + if (jd != 0) + gc_jd_stack_push(d_free_list, jd); +} + +//////////////////////////////////////////////////////////////////////// + +/* + * We check as much as we can here on the PPE side, so that the SPE + * doesn't have to. + */ +static bool +check_direct_args(gc_job_desc *jd, gc_job_direct_args *args) +{ + if (args->nargs > MAX_ARGS_DIRECT){ + jd->status = JS_BAD_N_DIRECT; + return false; + } + + return true; +} + +static bool +check_ea_args(gc_job_desc *jd, gc_job_ea_args *p) +{ + if (p->nargs > MAX_ARGS_EA){ + jd->status = JS_BAD_N_EA; + return false; + } + + uint32_t dir_union = 0; + + for (unsigned int i = 0; i < p->nargs; i++){ + dir_union |= p->arg[i].direction; + switch(p->arg[i].direction){ + case GCJD_DMA_GET: + case GCJD_DMA_PUT: + break; + + default: + jd->status = JS_BAD_DIRECTION; + return false; + } + } + + if (p->nargs > 1){ + unsigned int common_eah = (p->arg[0].ea_addr) >> 32; + for (unsigned int i = 1; i < p->nargs; i++){ + if ((p->arg[i].ea_addr >> 32) != common_eah){ + jd->status = JS_BAD_EAH; + return false; + } + } + } + + jd->sys.direction_union = dir_union; + return true; +} + +bool +gc_job_manager_impl::submit_job(gc_job_desc *jd) +{ + if (unlikely(d_shutdown_requested)){ + jd->status = JS_SHUTTING_DOWN; + return false; + } + + // Ensure it's one of our job descriptors + + if (jd < d_jd || jd >= &d_jd[d_options.max_jobs]){ + jd->status = JS_BAD_JOB_DESC; + return false; + } + + // Ensure we've got a client_thread_info assigned to this thread. + + gc_client_thread_info *cti = + (gc_client_thread_info *) pthread_getspecific(s_client_key); + if (unlikely(cti == 0)){ + if ((cti = alloc_cti()) == 0){ + fprintf(stderr, "gc_job_manager_impl::submit_job: Too many client threads.\n"); + jd->status = JS_TOO_MANY_CLIENTS; + return false; + } + int r = pthread_setspecific(s_client_key, cti); + if (r != 0){ + jd->status = JS_BAD_JUJU; + fprintf(stderr, "pthread_setspecific failed (return = %d)\n", r); + return false; + } + } + + if (jd->proc_id == GCP_UNKNOWN_PROC){ + jd->status = JS_UNKNOWN_PROC; + return false; + } + + if (!check_direct_args(jd, &jd->input)) + return false; + + if (!check_direct_args(jd, &jd->output)) + return false; + + if (!check_ea_args(jd, &jd->eaa)) + return false; + + jd->status = JS_OK; + jd->sys.client_id = cti->d_client_id; + + // FIXME keep count of jobs in progress? + + gc_jd_queue_enqueue(d_queue, jd); + return true; +} + +bool +gc_job_manager_impl::wait_job(gc_job_desc *jd) +{ + bool done; + return wait_jobs(1, &jd, &done, GC_WAIT_ANY) == 1; +} + +int +gc_job_manager_impl::wait_jobs(unsigned int njobs, + gc_job_desc *jd[], + bool done[], + gc_wait_mode mode) +{ + unsigned int i; + + gc_client_thread_info *cti = + (gc_client_thread_info *) pthread_getspecific(s_client_key); + if (unlikely(cti == 0)) + return -1; + + for (i = 0; i < njobs; i++){ + done[i] = false; + if (unlikely(jd[i]->sys.client_id != cti->d_client_id)){ + fprintf(stderr, "gc_job_manager_impl::wait_jobs: can't wait for a job you didn't submit\n"); + return -1; + } + } + + { + omni_mutex_lock l(cti->d_mutex); + + // setup info for event handler + cti->d_state = (mode == GC_WAIT_ANY) ? CT_WAIT_ANY : CT_WAIT_ALL; + cti->d_njobs_waiting_for = njobs; + cti->d_jobs_waiting_for = jd; + assert(cti->d_jobs_done != 0); + + unsigned int ndone = 0; + + // wait for jobs to complete + + while (1){ + ndone = 0; + for (i= 0; i < njobs; i++){ + if (done[i]) + ndone++; + else if (bv_isset(cti->d_jobs_done, jd[i]->sys.job_id)){ + bv_clr(cti->d_jobs_done, jd[i]->sys.job_id); + done[i] = true; + ndone++; + } + } + + if (mode == GC_WAIT_ANY && ndone > 0) + break; + + if (mode == GC_WAIT_ALL && ndone == njobs) + break; + + // FIXME what happens when somebody calls shutdown? + + cti->d_cond.wait(); // wait for event handler to wake us up + } + + cti->d_state = CT_NOT_WAITING; + cti->d_njobs_waiting_for = 0; // tidy up (not reqd) + cti->d_jobs_waiting_for = 0; // tidy up (not reqd) + return ndone; + } +} + +//////////////////////////////////////////////////////////////////////// + +bool +gc_job_manager_impl::send_all_spes(uint32_t msg) +{ + bool ok = true; + + for (unsigned int i = 0; i < d_options.nspes; i++) + ok &= send_spe(i, msg); + + return ok; +} + +bool +gc_job_manager_impl::send_spe(unsigned int spe, uint32_t msg) +{ + if (spe >= d_options.nspes) + return false; + + int r = spe_in_mbox_write(d_worker[spe].spe_ctx, &msg, 1, + SPE_MBOX_ALL_BLOCKING); + if (r < 0){ + perror("spe_in_mbox_write"); + return false; + } + + return r == 1; +} + +//////////////////////////////////////////////////////////////////////// + +static void +pthread_create_failure_msg(int r, const char *which) +{ + char buf[256]; + char *s = 0; + + switch (r){ + case EAGAIN: s = "EAGAIN"; break; + case EINVAL: s = "EINVAL"; break; + case EPERM: s = "EPERM"; break; + default: + snprintf(buf, sizeof(buf), "Unknown error %d", r); + s = buf; + break; + } + fprintf(stderr, "pthread_create[%s] failed: %s\n", which, s); +} + + +static bool +start_thread(pthread_t *thread, + void *(*start_routine)(void *), void *arg, + const char *msg) +{ + pthread_attr_t attr; + pthread_attr_init(&attr); + pthread_attr_setdetachstate(&attr, PTHREAD_CREATE_DETACHED); + + // FIXME save sigprocmask + // FIXME set sigprocmask + + int r = pthread_create(thread, &attr, start_routine, arg); + + // FIXME restore sigprocmask + + if (r != 0){ + pthread_create_failure_msg(r, msg); + return false; + } + return true; +} + + +//////////////////////////////////////////////////////////////////////// + +static void *start_worker(void *arg); + +static void * +start_event_handler(void *arg) +{ + gc_job_manager_impl *p = (gc_job_manager_impl *) arg; + p->event_handler_loop(); + return 0; +} + +void +gc_job_manager_impl::create_event_handler() +{ + // create the SPE event handler and register our interest in events + + d_spe_event_handler.ptr = spe_event_handler_create(); + if (d_spe_event_handler.ptr == 0){ + perror("spe_event_handler_create"); + throw std::runtime_error("spe_event_handler_create"); + } + + for (unsigned int i = 0; i < d_options.nspes; i++){ + spe_event_unit_t eu; + memset(&eu, 0, sizeof(eu)); + eu.events = SPE_EVENT_OUT_INTR_MBOX | SPE_EVENT_SPE_STOPPED; + eu.spe = d_worker[i].spe_ctx; + eu.data.u32 = i; // set in events returned by spe_event_wait + + if (spe_event_handler_register(d_spe_event_handler.ptr, &eu) != 0){ + perror("spe_event_handler_register"); + throw std::runtime_error("spe_event_handler_register"); + } + } + + // create our event handling thread + + if (!start_thread(&d_eh_thread, start_event_handler, this, "event_handler")){ + throw std::runtime_error("pthread_create"); + } + + // create the SPE worker threads + + bool ok = true; + for (unsigned int i = 0; ok && i < d_options.nspes; i++){ + char name[256]; + snprintf(name, sizeof(name), "worker[%d]", i); + ok &= start_thread(&d_worker[i].thread, start_worker, + &d_worker[i], name); + } + + if (!ok){ + // + // FIXME Clean up the mess. Need to terminate event handler and all workers. + // + // this should cause the workers to exit, unless they're seriously broken + send_all_spes(MK_MBOX_MSG(OP_EXIT, 0)); + + shutdown(); + + throw std::runtime_error("pthread_create"); + } +} + +//////////////////////////////////////////////////////////////////////// + +void +gc_job_manager_impl::set_eh_state(evt_handler_state s) +{ + omni_mutex_lock l(d_eh_mutex); + d_eh_state = s; + d_eh_cond.broadcast(); +} + +void +gc_job_manager_impl::set_ea_args_maxsize(int maxsize) +{ + omni_mutex_lock l(d_eh_mutex); + d_ea_args_maxsize = maxsize; + d_eh_cond.broadcast(); +} + +void +gc_job_manager_impl::print_event(spe_event_unit_t *evt) +{ + printf("evt: spe = %d events = (0x%x)", evt->data.u32, evt->events); + + if (evt->events & SPE_EVENT_OUT_INTR_MBOX) + printf(" OUT_INTR_MBOX"); + + if (evt->events & SPE_EVENT_IN_MBOX) + printf(" IN_MBOX"); + + if (evt->events & SPE_EVENT_TAG_GROUP) + printf(" TAG_GROUP"); + + if (evt->events & SPE_EVENT_SPE_STOPPED) + printf(" SPE_STOPPED"); + + printf("\n"); +} + +struct job_client_info { + uint16_t job_id; + uint16_t client_id; +}; + +static int +compare_jci_clients(const void *va, const void *vb) +{ + const job_client_info *a = (job_client_info *) va; + const job_client_info *b = (job_client_info *) vb; + + return a->client_id - b->client_id; +} + +void +gc_job_manager_impl::notify_clients_jobs_are_done(unsigned int spe_num, + unsigned int completion_info_idx) +{ + const char *msg = "gc_job_manager_impl::notify_client_job_is_done (INTERNAL ERROR)"; + + smp_rmb(); // order reads so we know that data sent from SPE is here + + gc_comp_info_t *ci = &d_comp_info[2 * spe_num + (completion_info_idx & 0x1)]; + + if (ci->ncomplete == 0){ // never happens, but ensures code below is correct + ci->in_use = 0; + return; + } + + if (0){ + static int total_jobs; + static int total_msgs; + total_msgs++; + total_jobs += ci->ncomplete; + printf("ppe: tj = %6d tm = %6d\n", total_jobs, total_msgs); + } + + job_client_info gci[GC_CI_NJOBS]; + + /* + * Make one pass through and sanity check everything while filling in gci + */ + for (unsigned int i = 0; i < ci->ncomplete; i++){ + unsigned int job_id = ci->job_id[i]; + + if (job_id >= d_options.max_jobs){ + // internal error, shouldn't happen + fprintf(stderr,"%s: invalid job_id = %d\n", msg, job_id); + ci->in_use = 0; // clear flag so SPE knows we're done with it + return; + } + gc_job_desc *jd = &d_jd[job_id]; + + if (jd->sys.client_id >= d_options.max_client_threads){ + // internal error, shouldn't happen + fprintf(stderr, "%s: invalid client_id = %d\n", msg, jd->sys.client_id); + ci->in_use = 0; // clear flag so SPE knows we're done with it + return; + } + + gci[i].job_id = job_id; + gci[i].client_id = jd->sys.client_id; + } + + // sort by client_id so we only have to lock & signal once / client + + if (ci->ncomplete > 1) + qsort(gci, ci->ncomplete, sizeof(gci[0]), compare_jci_clients); + + // "wind-in" + + gc_client_thread_info *last_cti = &d_client_thread[gci[0].client_id]; + last_cti->d_mutex.lock(); + bv_set(last_cti->d_jobs_done, gci[0].job_id); // mark job done + + for (unsigned int i = 1; i < ci->ncomplete; i++){ + + gc_client_thread_info *cti = &d_client_thread[gci[i].client_id]; + + if (cti != last_cti){ // new client? + + // yes. signal old client, unlock old, lock new + + // FIXME we could distinguish between CT_WAIT_ALL & CT_WAIT_ANY + + if (last_cti->d_state == CT_WAIT_ANY || last_cti->d_state == CT_WAIT_ALL) + last_cti->d_cond.signal(); // wake client thread up + + last_cti->d_mutex.unlock(); + cti->d_mutex.lock(); + last_cti = cti; + } + + // mark job done + bv_set(cti->d_jobs_done, gci[i].job_id); + } + + // "wind-out" + + if (last_cti->d_state == CT_WAIT_ANY || last_cti->d_state == CT_WAIT_ALL) + last_cti->d_cond.signal(); // wake client thread up + last_cti->d_mutex.unlock(); + + ci->in_use = 0; // clear flag so SPE knows we're done with it +} + +void +gc_job_manager_impl::handle_event(spe_event_unit_t *evt) +{ + // print_event(evt); + + int spe_num = evt->data.u32; + + // only a single event type can be signaled at a time + + if (evt->events == SPE_EVENT_OUT_INTR_MBOX) { // SPE sent us 1 or more msgs + static const int NMSGS = 32; + unsigned int msg[NMSGS]; + int n = spe_out_intr_mbox_read(evt->spe, msg, NMSGS, SPE_MBOX_ANY_BLOCKING); + // printf("spe_out_intr_mbox_read = %d\n", n); + if (n < 0){ + perror("spe_out_intr_mbox_read"); + } + else { + for (int i = 0; i < n; i++){ + switch(MBOX_MSG_OP(msg[i])){ + case OP_JOBS_DONE: + if (debug()) + printf("eh: job_done (0x%08x) from spu[%d]\n", msg[i], spe_num); + notify_clients_jobs_are_done(spe_num, MBOX_MSG_ARG(msg[i])); + break; + + case OP_SPU_BUFSIZE: + set_ea_args_maxsize(MBOX_MSG_ARG(msg[i])); + break; + + case OP_EXIT: + default: + printf("eh: Unexpected msg (0x%08x) from spu[%d]\n", msg[i], spe_num); + break; + } + } + } + } + else if (evt->events == SPE_EVENT_SPE_STOPPED){ // the SPE stopped + spe_stop_info_t si; + int r = spe_stop_info_read(evt->spe, &si); + if (r < 0){ + perror("spe_stop_info_read"); + } + else { + switch (si.stop_reason){ + case SPE_EXIT: + if (debug()){ + printf("eh: spu[%d] SPE_EXIT w/ exit_code = %d\n", + spe_num, si.result.spe_exit_code); + } + break; + case SPE_STOP_AND_SIGNAL: + printf("eh: spu[%d] SPE_STOP_AND_SIGNAL w/ spe_signal_code = 0x%x\n", + spe_num, si.result.spe_signal_code); + break; + case SPE_RUNTIME_ERROR: + printf("eh: spu[%d] SPE_RUNTIME_ERROR w/ spe_runtime_error = 0x%x\n", + spe_num, si.result.spe_runtime_error); + break; + case SPE_RUNTIME_EXCEPTION: + printf("eh: spu[%d] SPE_RUNTIME_EXCEPTION w/ spe_runtime_exception = 0x%x\n", + spe_num, si.result.spe_runtime_exception); + break; + case SPE_RUNTIME_FATAL: + printf("eh: spu[%d] SPE_RUNTIME_FATAL w/ spe_runtime_fatal = 0x%x\n", + spe_num, si.result.spe_runtime_fatal); + break; + case SPE_CALLBACK_ERROR: + printf("eh: spu[%d] SPE_CALLBACK_ERROR w/ spe_callback_error = 0x%x\n", + spe_num, si.result.spe_callback_error); + break; + case SPE_ISOLATION_ERROR: + printf("eh: spu[%d] SPE_ISOLATION_ERROR w/ spe_isolation_error = 0x%x\n", + spe_num, si.result.spe_isolation_error); + break; + default: + printf("eh: spu[%d] UNKNOWN STOP REASON (%d) w/ spu_status = 0x%x\n", + spe_num, si.stop_reason, si.spu_status); + break; + } + } + } +#if 0 // not enabled + else if (evt->events == SPE_EVENT_IN_MBOX){ // there's room to write to SPE + // spe_in_mbox_write (ignore) + } + else if (evt->events == SPE_EVENT_TAG_GROUP){ // our DMA completed + // spe_mfcio_tag_status_read + } +#endif + else { + fprintf(stderr, "handle_event: unexpected evt->events = 0x%x\n", evt->events); + return; + } +} + +// +// This is the "main program" of the event handling thread +// +void +gc_job_manager_impl::event_handler_loop() +{ + static const int MAX_EVENTS = 16; + static const int TIMEOUT = 20; // how long to block in milliseconds + + spe_event_unit_t events[MAX_EVENTS]; + + if (d_debug) + printf("event_handler_loop: starting\n"); + + set_eh_state(EHS_RUNNING); + + // ask the first spe for its max bufsize + send_spe(0, MK_MBOX_MSG(OP_GET_SPU_BUFSIZE, 0)); + + while (1){ + switch(d_eh_state){ + + case EHS_RUNNING: // normal stuff + if (d_shutdown_requested) { + set_eh_state(EHS_SHUTTING_DOWN); + } + break; + + case EHS_SHUTTING_DOWN: + + // FIXME wait until job queue is empty, then tell them to exit + + send_all_spes(MK_MBOX_MSG(OP_EXIT, 0)); + set_eh_state(EHS_WAITING_FOR_WORKERS_TO_DIE); + break; + + case EHS_WAITING_FOR_WORKERS_TO_DIE: + { + bool all_dead = true; + for (unsigned int i = 0; i < d_options.nspes; i++) + all_dead &= d_worker[i].state == WS_DEAD; + + if (all_dead){ + set_eh_state(EHS_DEAD); + if (d_debug) + printf("event_handler_loop: exiting\n"); + return; + } + } + break; + + default: + set_eh_state(EHS_DEAD); + printf("event_handler_loop(default): exiting\n"); + return; + } + + // block waiting for events... + int nevents = spe_event_wait(d_spe_event_handler.ptr, + events, MAX_EVENTS, TIMEOUT); + if (nevents < 0){ + perror("spe_wait_event"); + // FIXME bail? + } + for (int i = 0; i < nevents; i++){ + handle_event(&events[i]); + } + } +} + +//////////////////////////////////////////////////////////////////////// +// This is the top of the SPE worker threads + +static void * +start_worker(void *arg) +{ + worker_ctx *w = (worker_ctx *) arg; + spe_stop_info_t si; + + w->state = WS_RUNNING; + if (s_worker_debug) + printf("worker[%d]: WS_RUNNING\n", w->spe_idx); + + unsigned int entry = SPE_DEFAULT_ENTRY; + int r = spe_context_run(w->spe_ctx, &entry, 0, w->spu_args, 0, &si); + + if (r < 0){ // error + char buf[64]; + snprintf(buf, sizeof(buf), "worker[%d]: spe_context_run", w->spe_idx); + perror(buf); + } + else if (r == 0){ + // spe program called exit. + if (s_worker_debug) + printf("worker[%d]: SPE_EXIT w/ exit_code = %d\n", + w->spe_idx, si.result.spe_exit_code); + } + else { + // called stop_and_signal + // + // I'm not sure we'll ever get here. I think the event + // handler will catch this... + printf("worker[%d]: SPE_STOP_AND_SIGNAL w/ spe_signal_code = 0x%x\n", + w->spe_idx, si.result.spe_signal_code); + } + + // in any event, we're committing suicide now ;) + if (s_worker_debug) + printf("worker[%d]: WS_DEAD\n", w->spe_idx); + + w->state = WS_DEAD; + return 0; +} + +//////////////////////////////////////////////////////////////////////// + +gc_client_thread_info * +gc_job_manager_impl::alloc_cti() +{ + for (unsigned int i = 0; i < d_options.max_client_threads; i++){ + if (d_client_thread[i].d_free){ + // try to atomically grab it + if (_atomic_dec_if_positive(ptr_to_ea(&d_client_thread[i].d_free)) == 0){ + // got it... + gc_client_thread_info *cti = &d_client_thread[i]; + cti->d_state = CT_NOT_WAITING; + bv_zero(cti->d_jobs_done); + cti->d_njobs_waiting_for = 0; + cti->d_jobs_waiting_for = 0; + + return cti; + } + } + } + return 0; +} + +void +gc_job_manager_impl::free_cti(gc_client_thread_info *cti) +{ + assert((size_t) (cti - d_client_thread.get()) < d_options.max_client_threads); + cti->d_free = 1; +} + +int +gc_job_manager_impl::ea_args_maxsize() +{ + omni_mutex_lock l(d_eh_mutex); + + while (d_ea_args_maxsize == 0) // wait for it to be initialized + d_eh_cond.wait(); + + return d_ea_args_maxsize; +} + +void +gc_job_manager_impl::set_debug(int debug) +{ + d_debug = debug; + s_worker_debug = debug; +} + +int +gc_job_manager_impl::debug() +{ + return d_debug; +} + +//////////////////////////////////////////////////////////////////////// + +void +gc_job_manager_impl::setup_logfiles() +{ + if (!d_options.enable_logging) + return; + + if (d_options.log2_nlog_entries == 0) + d_options.log2_nlog_entries = 12; + + // must end up a multiple of the page size + + size_t pagesize = getpagesize(); + size_t s = (1 << d_options.log2_nlog_entries) * sizeof(gc_log_entry_t); + s = ((s + pagesize - 1) / pagesize) * pagesize; + size_t nentries = s / sizeof(gc_log_entry_t); + assert(is_power_of_2(nentries)); + + for (unsigned int i = 0; i < d_options.nspes; i++){ + char filename[100]; + snprintf(filename, sizeof(filename), "spu_log.%02d", i); + int fd = open(filename, O_CREAT|O_TRUNC|O_RDWR, 0664); + if (fd == -1){ + perror(filename); + return; + } + lseek(fd, s - 1, SEEK_SET); + write(fd, "\0", 1); + void *p = mmap(0, s, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0); + if (p == MAP_FAILED){ + perror("gc_job_manager_impl::setup_logfiles: mmap"); + close(fd); + return; + } + close(fd); + memset(p, 0, s); + d_spu_args[i].log.base = ptr_to_ea(p); + d_spu_args[i].log.nentries = nentries; + } +} + +void +gc_job_manager_impl::sync_logfiles() +{ + for (unsigned int i = 0; i < d_options.nspes; i++){ + if (d_spu_args[i].log.base) + msync(ea_to_ptr(d_spu_args[i].log.base), + d_spu_args[i].log.nentries * sizeof(gc_log_entry_t), + MS_ASYNC); + } +} + +void +gc_job_manager_impl::unmap_logfiles() +{ + for (unsigned int i = 0; i < d_options.nspes; i++){ + if (d_spu_args[i].log.base) + munmap(ea_to_ptr(d_spu_args[i].log.base), + d_spu_args[i].log.nentries * sizeof(gc_log_entry_t)); + } +} + +//////////////////////////////////////////////////////////////////////// +// +// lookup proc names in d_proc_def table + +gc_proc_id_t +gc_job_manager_impl::lookup_proc(const std::string &proc_name) +{ + for (int i = 0; i < d_nproc_defs; i++) + if (proc_name == d_proc_def[i].name) + return i; + + return GCP_UNKNOWN_PROC; +} + +std::vector<std::string> +gc_job_manager_impl::proc_names() +{ + std::vector<std::string> r; + for (int i = 0; i < d_nproc_defs; i++) + r.push_back(d_proc_def[i].name); + + return r; +} + +//////////////////////////////////////////////////////////////////////// + +worker_ctx::~worker_ctx() +{ + if (spe_ctx){ + int r = spe_context_destroy(spe_ctx); + if (r != 0){ + perror("spe_context_destroy"); + } + spe_ctx = 0; + } + state = WS_FREE; +} diff --git a/gcell/src/lib/runtime/gc_job_manager_impl.h b/gcell/src/lib/runtime/gc_job_manager_impl.h new file mode 100644 index 000000000..46897848c --- /dev/null +++ b/gcell/src/lib/runtime/gc_job_manager_impl.h @@ -0,0 +1,253 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifndef INCLUDED_GC_JOB_MANAGER_IMPL_H +#define INCLUDED_GC_JOB_MANAGER_IMPL_H + +#include "gc_job_manager.h" +#include "gc_client_thread_info.h" +#include "gc_jd_stack.h" +#include "gc_jd_queue.h" +#include "gc_spu_args.h" +#include <libspe2.h> +#include <vector> +#include <boost/shared_ptr.hpp> +#include <boost/scoped_array.hpp> + +typedef boost::shared_ptr<spe_gang_context> spe_gang_context_sptr; +typedef boost::shared_ptr<spe_program_handle_t> spe_program_handle_sptr; +typedef boost::scoped_array<gc_client_thread_info> gc_client_thread_info_sa; + + +enum worker_state { + WS_FREE, // not in use + WS_INIT, // allocated and being initialized + WS_RUNNING, // the thread is running + WS_DEAD, // the thread is dead +}; + +struct worker_ctx { + volatile worker_state state; + unsigned int spe_idx; // [0, nspes-1] + spe_context_ptr_t spe_ctx; + pthread_t thread; + gc_spu_args_t *spu_args; // pointer to 16-byte aligned struct + + worker_ctx() + : state(WS_FREE), spe_idx(0), spe_ctx(0), + thread(0), spu_args(0) {} + ~worker_ctx(); +}; + +enum evt_handler_state { + EHS_INIT, // being initialized + EHS_RUNNING, // thread is running + EHS_SHUTTING_DOWN, // in process of shutting down everything + EHS_WAITING_FOR_WORKERS_TO_DIE, + EHS_DEAD, // thread is dead +}; + +struct spe_event_handler { + spe_event_handler_ptr_t ptr; + + spe_event_handler() : ptr(0) {} + ~spe_event_handler(){ + if (ptr){ + if (spe_event_handler_destroy(ptr) != 0){ + perror("spe_event_handler_destroy"); + } + } + } +}; + + +/*! + * \brief Concrete class that manages SPE jobs. + * + * This class contains all the implementation details. + */ +class gc_job_manager_impl : public gc_job_manager +{ + enum { MAX_SPES = 16 }; + + int d_debug; + gc_jm_options d_options; + spe_program_handle_sptr d_spe_image; + spe_gang_context_sptr d_gang; // boost::shared_ptr + + worker_ctx d_worker[MAX_SPES]; // SPE ctx, thread, etc + gc_spu_args_t *d_spu_args; // 16-byte aligned structs + boost::shared_ptr<void> _d_spu_args_boost; // hack for automatic storage mgmt + + gc_comp_info_t *d_comp_info; // 128-byte aligned structs + boost::shared_ptr<void> _d_comp_info_boost; // hack for automatic storage mgmt + + // used to coordinate communication w/ the event handling thread + omni_mutex d_eh_mutex; + omni_condition d_eh_cond; + pthread_t d_eh_thread; // the event handler thread + volatile evt_handler_state d_eh_state; + volatile bool d_shutdown_requested; + spe_event_handler d_spe_event_handler; + + + // All of the job descriptors are hung off of here. + // We allocate them all in a single cache aligned chunk. + gc_job_desc_t *d_jd; // [options.max_jobs] + boost::shared_ptr<void> _d_jd_boost; // hack for automatic storage mgmt + + gc_client_thread_info_sa d_client_thread; // [options.max_client_threads] + + // We use bitvectors to represent the completing state of a job. Each + // bitvector is d_bvlen longs in length. + int d_bvlen; // bit vector length in longs + + // This contains the storage for all the bitvectors used by the job + // manager. There's 1 for each client thread, in the d_jobs_done + // field. We allocate them all in a single cache aligned chunk. + boost::shared_ptr<void> _d_all_bitvectors; // hack for automatic storage mgmt + + // Lock free stack where we keep track of the free job descriptors. + gc_jd_stack_t *d_free_list; // stack of free job descriptors + boost::shared_ptr<void> _d_free_list_boost; // hack for automatic storage mgmt + + // The PPE inserts jobs here; SPEs pull jobs from here. + gc_jd_queue_t *d_queue; // job queue + boost::shared_ptr<void> _d_queue_boost; // hack for automatic storage mgmt + + int d_ea_args_maxsize; + + struct gc_proc_def *d_proc_def; // the SPE procedure table + uint32_t d_proc_def_ls_addr; // the LS address of the table + int d_nproc_defs; // number of proc_defs in table + + gc_client_thread_info *alloc_cti(); + void free_cti(gc_client_thread_info *cti); + + void create_event_handler(); + void set_eh_state(evt_handler_state s); + void set_ea_args_maxsize(int maxsize); + + void notify_clients_jobs_are_done(unsigned int spe_num, + unsigned int completion_info_idx); + +public: + void event_handler_loop(); // really private + +private: + bool send_all_spes(uint32_t msg); + bool send_spe(unsigned int spe, uint32_t msg); + void print_event(spe_event_unit_t *evt); + void handle_event(spe_event_unit_t *evt); + + // bitvector ops + void bv_zero(unsigned long *bv); + void bv_clr(unsigned long *bv, unsigned int bitno); + void bv_set(unsigned long *bv, unsigned int bitno); + bool bv_isset(unsigned long *bv, unsigned int bitno); + bool bv_isclr(unsigned long *bv, unsigned int bitno); + + void setup_logfiles(); + void sync_logfiles(); + void unmap_logfiles(); + + friend gc_job_manager *gc_make_job_manager(const gc_jm_options *options); + + gc_job_manager_impl(const gc_jm_options *options = 0); + +public: + virtual ~gc_job_manager_impl(); + + /*! + * Stop accepting new jobs. Wait for existing jobs to complete. + * Return all managed SPE's to the system. + */ + virtual bool shutdown(); + + /*! + * \brief Return number of SPE's currently allocated to job manager. + */ + virtual int nspes() const; + + /*! + * \brief Return a pointer to a properly aligned job descriptor, + * or zero if none are available. + */ + virtual gc_job_desc *alloc_job_desc(); + + /* + *! Return a job descriptor previously allocated with alloc_job_desc() + * + * \param[in] jd pointer to job descriptor to free. + */ + virtual void free_job_desc(gc_job_desc *jd); + + /*! + * \brief Submit a job for asynchronous processing on an SPE. + * + * \param[in] jd pointer to job description + * + * The caller must not read or write the job description + * or any of the memory associated with any indirect arguments + * until after calling wait_job. + * + * \returns true iff the job was successfully enqueued. + * If submit_job returns false, check jd->status for additional info. + */ + virtual bool submit_job(gc_job_desc *jd); + + /*! + * \brief Wait for job to complete. + * + * A thread may only wait for jobs which it submitted. + * + * \returns true if sucessful, else false. + */ + virtual bool + wait_job(gc_job_desc *jd); + + /*! + * \brief wait for 1 or more jobs to complete. + * + * \param[input] njobs is the length of arrays \p jd and \p done. + * \param[input] jd are the jobs that are to be waited for. + * \param[output] done indicates whether the corresponding job is complete. + * \param[input] mode indicates whether to wait for ALL or ANY of the jobs + * in \p jd to complete. + * + * A thread may only wait for jobs which it submitted. + * + * \returns number of jobs completed, or -1 if error. + */ + virtual int + wait_jobs(unsigned int njobs, + gc_job_desc *jd[], bool done[], gc_wait_mode mode); + + virtual int ea_args_maxsize(); + + virtual gc_proc_id_t lookup_proc(const std::string &name); + virtual std::vector<std::string> proc_names(); + + virtual void set_debug(int debug); + virtual int debug(); +}; + +#endif /* INCLUDED_GC_JOB_MANAGER_IMPL_H */ diff --git a/gcell/src/lib/runtime/gc_proc_def_utils.cc b/gcell/src/lib/runtime/gc_proc_def_utils.cc new file mode 100644 index 000000000..c5b9848fc --- /dev/null +++ b/gcell/src/lib/runtime/gc_proc_def_utils.cc @@ -0,0 +1,123 @@ +/* -*- c++ -*- */ +/* + * Copyright 2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#include <gc_proc_def_utils.h> +#include <gc_declare_proc.h> +#include <elf.h> +#include <stdio.h> +#include <string.h> + +static const unsigned char expected[EI_PAD] = { + ELFMAG0, + ELFMAG1, + ELFMAG2, + ELFMAG3, + ELFCLASS32, + ELFDATA2MSB, + EV_CURRENT, + ELFOSABI_SYSV, + 0 +}; + + +/* + * Basically we're going to find the GC_PROC_DEF_SECTION section + * in the ELF file and return a pointer to it. The only things in that + * section are gc_proc_def's + */ +bool +gcpd_find_table(spe_program_handle_t *handle, + struct gc_proc_def **table, int *nentries, uint32_t *ls_addr) +{ + if (!handle || !table || !nentries) + return false; + + *table = 0; + *nentries = 0; + + Elf32_Ehdr *ehdr = (Elf32_Ehdr *)handle->elf_image; + if (!ehdr){ + fprintf(stderr, "gcpd: No ELF image has been loaded\n"); + return false; + } + + // quick check that we're looking at a SPE EXEC object + + if (memcmp(ehdr->e_ident, expected, EI_PAD) != 0){ + fprintf(stderr, "gcpd: invalid ELF header\n"); + return false; + } + + if (ehdr->e_machine != 0x17){ // confirm machine type (EM_SPU) + fprintf(stderr, "gcpd: not an SPE ELF object\n"); + return false; + } + + if (ehdr->e_type != ET_EXEC){ + fprintf(stderr, "gcpd: invalid SPE ELF type.\n"); + fprintf(stderr, "gcpd: SPE type %d != %d\n", ehdr->e_type, ET_EXEC); + return false; + } + + // find the section header table + + Elf32_Shdr *shdr; + Elf32_Shdr *sh; + + if (ehdr->e_shentsize != sizeof (*shdr)){ + fprintf(stderr, "gcpd: invalid section header format.\n"); + return false; + } + + if (ehdr->e_shnum == 0){ + fprintf(stderr, "gcpd: no section headers in file.\n"); + return false; + } + + shdr = (Elf32_Shdr *) ((char *)ehdr + ehdr->e_shoff); + char *str_table = (char *)ehdr + shdr[ehdr->e_shstrndx].sh_offset; + + // traverse the sections looking for GC_PROC_DEF_SECTION + + for (sh = shdr; sh < &shdr[ehdr->e_shnum]; sh++){ + if (0){ + fprintf(stderr, "section name: %s (start: 0x%04x, size: 0x%04x)\n", + str_table + sh->sh_name, sh->sh_offset, sh->sh_size); + } + + if (strcmp(GC_PROC_DEF_SECTION, str_table+sh->sh_name) == 0){ + *table = (struct gc_proc_def *)((char *)ehdr + sh->sh_offset); + if (sh->sh_size % (sizeof(struct gc_proc_def)) != 0){ + fprintf(stderr, "gcpd: %s section has invalid format\n", GC_PROC_DEF_SECTION); + return false; + } + *nentries = sh->sh_size / sizeof(struct gc_proc_def); + *ls_addr = sh->sh_addr; + return true; + } + } + + return false; +} diff --git a/gcell/src/lib/runtime/gc_proc_def_utils.h b/gcell/src/lib/runtime/gc_proc_def_utils.h new file mode 100644 index 000000000..c59e472af --- /dev/null +++ b/gcell/src/lib/runtime/gc_proc_def_utils.h @@ -0,0 +1,42 @@ +/* -*- c++ -*- */ +/* + * Copyright 2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef INCLUDED_GC_PROC_DEF_UTILS_H +#define INCLUDED_GC_PROC_DEF_UTILS_H + +#include <gc_declare_proc.h> +#include <libspe2.h> + +/*! + * \brief find the gc_proc_def table in the SPE program + * + * \param[in] program is the handle to the loaded SPE program + * \param[out] table points to the table, if it's found + * \param[out] nentries is set to the number of entries in the table. + * \param[out] ls_addr is set to the Local Store address of the table + * + * \returns true if successful, else false + */ +bool +gcpd_find_table(spe_program_handle_t *program, + struct gc_proc_def **table, int *nentries, uint32_t *ls_addr); + + +#endif /* INCLUDED_GC_PROC_DEF_UTILS_H */ diff --git a/gcell/src/lib/runtime/qa_jd_queue.cc b/gcell/src/lib/runtime/qa_jd_queue.cc new file mode 100644 index 000000000..267d4cfb7 --- /dev/null +++ b/gcell/src/lib/runtime/qa_jd_queue.cc @@ -0,0 +1,78 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include "qa_jd_queue.h" +#include <cppunit/TestAssert.h> +#include "gc_jd_queue.h" +#include <stdio.h> + + + +static const int NJDS = 16; +static gc_jd_queue_t queue; +static gc_job_desc_t jds[NJDS]; + +// no brainer, single threaded basic checkout +void +qa_jd_queue::t1() +{ + // N.B., queue allocated stuff doesn't obey ((aligned (N))) attributes + //const int NJDS = 8; + //gc_jd_queue_t queue; + //gc_job_desc_t jds[NJDS]; + + //printf("&queue = %p\n", &queue); + //printf("&jds[0] = %p\n", &jds[0]); + //printf("&jds[1] = %p\n", &jds[1]); + + CPPUNIT_ASSERT(((uintptr_t) &queue & 0x7f) == 0); + CPPUNIT_ASSERT(((uintptr_t) &jds[0] & 0x7f) == 0); + CPPUNIT_ASSERT(((uintptr_t) &jds[1] & 0x7f) == 0); + + gc_jd_queue_init(&queue); + + CPPUNIT_ASSERT(gc_jd_queue_dequeue(&queue) == 0); + + gc_jd_queue_enqueue(&queue, &jds[0]); + CPPUNIT_ASSERT_EQUAL(&jds[0], gc_jd_queue_dequeue(&queue)); + + CPPUNIT_ASSERT(gc_jd_queue_dequeue(&queue) == 0); + + for (int i = 0; i < NJDS; i++) + gc_jd_queue_enqueue(&queue, &jds[i]); + + for (int i = 0; i < NJDS; i++) + CPPUNIT_ASSERT_EQUAL(&jds[i], gc_jd_queue_dequeue(&queue)); + + CPPUNIT_ASSERT(gc_jd_queue_dequeue(&queue) == 0); +} + +// FIXME multithreaded (running on PPE) +void +qa_jd_queue::t2() +{ +} + +// FIXME multithreaded (running on PPE & SPE) +void +qa_jd_queue::t3() +{ +} diff --git a/gcell/src/lib/runtime/qa_jd_queue.h b/gcell/src/lib/runtime/qa_jd_queue.h new file mode 100644 index 000000000..5e1aab84f --- /dev/null +++ b/gcell/src/lib/runtime/qa_jd_queue.h @@ -0,0 +1,42 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef INCLUDED_QA_JD_QUEUE_H +#define INCLUDED_QA_JD_QUEUE_H + +#include <cppunit/extensions/HelperMacros.h> +#include <cppunit/TestCase.h> + +class qa_jd_queue : public CppUnit::TestCase { + + CPPUNIT_TEST_SUITE(qa_jd_queue); + CPPUNIT_TEST(t1); + CPPUNIT_TEST(t2); + CPPUNIT_TEST(t3); + CPPUNIT_TEST_SUITE_END(); + + private: + void t1(); + void t2(); + void t3(); +}; + + +#endif /* INCLUDED_QA_JD_QUEUE_H */ diff --git a/gcell/src/lib/runtime/qa_jd_stack.cc b/gcell/src/lib/runtime/qa_jd_stack.cc new file mode 100644 index 000000000..67e97c9a3 --- /dev/null +++ b/gcell/src/lib/runtime/qa_jd_stack.cc @@ -0,0 +1,67 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include "qa_jd_stack.h" +#include <cppunit/TestAssert.h> +#include "gc_jd_stack.h" +#include <stdio.h> + + + +static const int NJDS = 8; +static gc_jd_stack_t stack; +static gc_job_desc_t jds[NJDS]; + +// no brainer, single threaded basic checkout +void +qa_jd_stack::t1() +{ + // N.B., stack allocated stuff doesn't obey ((aligned (N))) attributes + //const int NJDS = 8; + //gc_jd_stack_t stack; + //gc_job_desc_t jds[NJDS]; + + //printf("&stack = %p\n", &stack); + //printf("&jds[0] = %p\n", &jds[0]); + //printf("&jds[1] = %p\n", &jds[1]); + + CPPUNIT_ASSERT(((uintptr_t) &stack & 0x7f) == 0); + CPPUNIT_ASSERT(((uintptr_t) &jds[0] & 0x7f) == 0); + CPPUNIT_ASSERT(((uintptr_t) &jds[1] & 0x7f) == 0); + + gc_jd_stack_init(&stack); + + CPPUNIT_ASSERT(gc_jd_stack_pop(&stack) == 0); + + for (int i = 0; i < NJDS; i++) + gc_jd_stack_push(&stack, &jds[i]); + + for (int i = 0; i < NJDS; i++) + CPPUNIT_ASSERT_EQUAL(&jds[NJDS - i - 1], gc_jd_stack_pop(&stack)); + + CPPUNIT_ASSERT(gc_jd_stack_pop(&stack) == 0); +} + +// FIXME multithreaded (running on PPE) +void +qa_jd_stack::t2() +{ +} diff --git a/gcell/src/lib/runtime/qa_jd_stack.h b/gcell/src/lib/runtime/qa_jd_stack.h new file mode 100644 index 000000000..1546bbf9f --- /dev/null +++ b/gcell/src/lib/runtime/qa_jd_stack.h @@ -0,0 +1,42 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef INCLUDED_QA_JD_STACK_H +#define INCLUDED_QA_JD_STACK_H + +#include <cppunit/extensions/HelperMacros.h> +#include <cppunit/TestCase.h> + +class qa_jd_stack : public CppUnit::TestCase { + + CPPUNIT_TEST_SUITE(qa_jd_stack); + CPPUNIT_TEST(t1); + CPPUNIT_TEST(t2); + CPPUNIT_TEST_SUITE_END(); + + private: + void t1(); + void t2(); + +}; + + + +#endif /* INCLUDED_QA_JD_STACK_H */ diff --git a/gcell/src/lib/runtime/qa_job_manager.cc b/gcell/src/lib/runtime/qa_job_manager.cc new file mode 100644 index 000000000..3f2780c52 --- /dev/null +++ b/gcell/src/lib/runtime/qa_job_manager.cc @@ -0,0 +1,798 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007,2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include "qa_job_manager.h" +#include <cppunit/TestAssert.h> +#include "gc_job_manager.h" +#include <stdexcept> +#include <stdio.h> +#include <time.h> +#include <errno.h> + +#include <malloc.h> + +extern spe_program_handle_t gcell_qa; // handle to embedded SPU executable w/ QA routines + +#if 0 +static void +gc_msleep(unsigned long millisecs) +{ + int r; + struct timespec tv; + tv.tv_sec = millisecs / 1000; + tv.tv_nsec = (millisecs - (tv.tv_sec * 1000)) * 1000000; + + while (1){ + r = nanosleep(&tv, &tv); + if (r == 0) + return; + if (r == -1 && errno == EINTR) + continue; + perror("nanosleep"); + return; + } +} +#endif + +void +qa_job_manager::leak_check(test_t t, const std::string &name) +{ + struct mallinfo before, after; + + before = mallinfo(); + (this->*t)(); + after = mallinfo(); + + size_t delta = after.uordblks - before.uordblks; + if (delta != 0){ + std::cout << name << " leaked memory\n"; + printf(" before.uordblks = %6d\n", before.uordblks); + printf(" after.uordblks = %6d\n", after.uordblks); + printf(" delta = %d\n", after.uordblks - before.uordblks); + } +} + +void +qa_job_manager::t0() +{ + //leak_check(&qa_job_manager::t1_body, "t1-0"); +} + +void +qa_job_manager::t1() +{ + t1_body(); // leaks 800 bytes first time, could be one-time inits + leak_check(&qa_job_manager::t1_body, "t1"); +} + +void +qa_job_manager::t2() +{ + leak_check(&qa_job_manager::t2_body, "t2"); +} + +void +qa_job_manager::t3() +{ + t3_body(); // leaks first time only, could be cppunit + leak_check(&qa_job_manager::t3_body, "t3"); +} + +void +qa_job_manager::t4() +{ + leak_check(&qa_job_manager::t4_body, "t4"); +} + +void +qa_job_manager::t5() +{ + leak_check(&qa_job_manager::t5_body, "t5"); +} + +void +qa_job_manager::t6() +{ + leak_check(&qa_job_manager::t6_body, "t6"); +} + +void +qa_job_manager::t7() +{ + leak_check(&qa_job_manager::t7_body, "t7"); +} + +void +qa_job_manager::t8() +{ + leak_check(&qa_job_manager::t8_body, "t8"); +} + +void +qa_job_manager::t9() +{ + leak_check(&qa_job_manager::t9_body, "t9"); +} + +void +qa_job_manager::t10() +{ + leak_check(&qa_job_manager::t10_body, "t10"); +} + +void +qa_job_manager::t11() +{ + leak_check(&qa_job_manager::t11_body, "t11"); +} + +void +qa_job_manager::t12() +{ + leak_check(&qa_job_manager::t12_body, "t12"); +} + +void +qa_job_manager::t13() +{ + leak_check(&qa_job_manager::t13_body, "t13"); +} + +void +qa_job_manager::t14() +{ + leak_check(&qa_job_manager::t14_body, "t14"); +} + +void +qa_job_manager::t15() +{ + leak_check(&qa_job_manager::t15_body, "t15"); +} + +// ---------------------------------------------------------------- + +void +qa_job_manager::t1_body() +{ + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + mgr = gc_make_job_manager(&opts); + delete mgr; +} + +void +qa_job_manager::t2_body() +{ + gc_job_manager *mgr = 0; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 100; + opts.gang_schedule = false; + mgr = gc_make_job_manager(&opts); + delete mgr; +} + +void +qa_job_manager::t3_body() +{ + // This leaks memory the first time it's invoked, but I'm not sure + // if it's us or the underlying exception handling mechanism, or + // cppunit. cppunit is the prime suspect. + +#if 0 + gc_job_manager *mgr = 0; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 100; + opts.gang_schedule = true; + CPPUNIT_ASSERT_THROW(mgr = gc_make_job_manager(&opts), std::out_of_range); + delete mgr; +#endif +} + +static void +init_jd(gc_job_desc *jd, gc_proc_id_t proc_id) +{ + jd->proc_id = proc_id; + jd->input.nargs = 0; + jd->output.nargs = 0; + jd->eaa.nargs = 0; +} + +void +qa_job_manager::t4_body() +{ + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + //mgr->set_debug(-1); + static const int NJOBS = 32; + gc_job_desc *jds[NJOBS]; + bool done[NJOBS]; + + gc_proc_id_t gcp_no_such = mgr->lookup_proc("--no-such-proc-name--"); + CPPUNIT_ASSERT_EQUAL(GCP_UNKNOWN_PROC, gcp_no_such); + + gc_proc_id_t gcp_qa_nop = mgr->lookup_proc("qa_nop"); + CPPUNIT_ASSERT(gcp_qa_nop != GCP_UNKNOWN_PROC); + + for (int i = 0; i < NJOBS; i++){ + jds[i] = mgr->alloc_job_desc(); + init_jd(jds[i], gcp_qa_nop); + } + + for (int i = 0; i < NJOBS; i++){ + if (!mgr->submit_job(jds[i])){ + printf("%d: submit_job(jds[%d]) failed, status = %d\n", + __LINE__, i, jds[i]->status); + } + } + + int n = mgr->wait_jobs(NJOBS, jds, done, GC_WAIT_ALL); + CPPUNIT_ASSERT_EQUAL(NJOBS, n); + + for (int i = 0; i < NJOBS; i++){ + mgr->free_job_desc(jds[i]); + } + + delete mgr; +} + +void +qa_job_manager::t5_body() +{ + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 0; // use them all + mgr = gc_make_job_manager(&opts); + //mgr->set_debug(-1); + static const int NJOBS = 32; + gc_job_desc *jds[NJOBS]; + bool done[NJOBS]; + + gc_proc_id_t gcp_qa_nop = mgr->lookup_proc("qa_nop"); + + for (int i = 0; i < NJOBS; i++){ + jds[i] = mgr->alloc_job_desc(); + init_jd(jds[i], gcp_qa_nop); + } + + for (int i = 0; i < NJOBS; i++){ + if (!mgr->submit_job(jds[i])){ + printf("%d: submit_job(jds[%d]) failed, status = %d\n", + __LINE__, i, jds[i]->status); + } + } + + int n = mgr->wait_jobs(NJOBS, jds, done, GC_WAIT_ALL); + CPPUNIT_ASSERT_EQUAL(NJOBS, n); + + for (int i = 0; i < NJOBS; i++){ + mgr->free_job_desc(jds[i]); + } + + delete mgr; +} + +void +qa_job_manager::t6_body() +{ + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + gc_proc_id_t gcp_qa_nop = mgr->lookup_proc("qa_nop"); + gc_job_desc *jd = mgr->alloc_job_desc(); + + + // test for success with gcp_qa_nop procedure + init_jd(jd, gcp_qa_nop); + if (!mgr->submit_job(jd)){ + printf("%d: submit_job(jd) failed, status = %d\n", __LINE__, jd->status); + } + else { + mgr->wait_job(jd); + CPPUNIT_ASSERT_EQUAL(JS_OK, jd->status); + } + + // test for JS_UNKNOWN_PROC with bogus procedure + init_jd(jd, -2); + if (!mgr->submit_job(jd)){ + printf("%d: submit_job(jd) failed, status = %d\n", __LINE__, jd->status); + } + else { + mgr->wait_job(jd); + CPPUNIT_ASSERT_EQUAL(JS_UNKNOWN_PROC, jd->status); + } + + mgr->free_job_desc(jd); + delete mgr; +} + +static int +sum_shorts(short *p, int nshorts) +{ + int total = 0; + for (int i = 0; i < nshorts; i++) + total += p[i]; + + return total; +} + +static void +test_sum_shorts(gc_job_manager *mgr, short *buf, int nshorts) +{ + gc_job_desc *jd = mgr->alloc_job_desc(); + gc_proc_id_t gcp_qa_sum_shorts = mgr->lookup_proc("qa_sum_shorts"); + + init_jd(jd, gcp_qa_sum_shorts); + jd->eaa.nargs = 1; + jd->eaa.arg[0].ea_addr = ptr_to_ea(buf); + jd->eaa.arg[0].direction = GCJD_DMA_GET; + jd->eaa.arg[0].get_size = nshorts * sizeof(short); + + + if (!mgr->submit_job(jd)){ + printf("%d: submit_job(jd) failed, status = %d\n", __LINE__, jd->status); + } + else { + mgr->wait_job(jd); + CPPUNIT_ASSERT_EQUAL(JS_OK, jd->status); + int expected = sum_shorts(buf, nshorts); + int actual = jd->output.arg[0].s32; + CPPUNIT_ASSERT_EQUAL(expected, actual); + } + + mgr->free_job_desc(jd); +} + +static const int NS = 32768; +static short short_buf[NS] _AL128; // for known alignment + +// +// test all "get" alignments and sizes +// +void +qa_job_manager::t7_body() +{ + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + + int ea_args_maxsize = mgr->ea_args_maxsize(); + + for (int i = 0; i < NS; i++) // init buffer with known qty + short_buf[i] = 0x1234 + i; + + for (int offset = 0; offset <= 128; offset++){ + for (int len = 0; len <= 128; len++){ + test_sum_shorts(mgr, &short_buf[offset], len); + } + } + + // confirm maximum length + for (int offset = 0; offset <= 64; offset++){ + test_sum_shorts(mgr, &short_buf[offset], ea_args_maxsize/sizeof(short)); + } + + delete mgr; +} + +// +// test "get" args too long +// +void +qa_job_manager::t8_body() +{ + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + gc_job_desc *jd = mgr->alloc_job_desc(); + gc_proc_id_t gcp_qa_sum_shorts = mgr->lookup_proc("qa_sum_shorts"); + + init_jd(jd, gcp_qa_sum_shorts); + jd->eaa.nargs = 1; + jd->eaa.arg[0].ea_addr = 0; + jd->eaa.arg[0].direction = GCJD_DMA_GET; + jd->eaa.arg[0].get_size = 1 << 20; + + if (!mgr->submit_job(jd)){ + printf("%d: submit_job(jd) failed, status = %d\n", __LINE__, jd->status); + } + else { + mgr->wait_job(jd); + CPPUNIT_ASSERT_EQUAL(JS_ARGS_TOO_LONG, jd->status); + } + + mgr->free_job_desc(jd); + delete mgr; +} + +// +// test MAX_ARGS_EA "get" case +// +void +qa_job_manager::t9_body() +{ + static const int N = 127; + static const int M = 201; + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + gc_job_desc *jd = mgr->alloc_job_desc(); + gc_proc_id_t gcp_qa_sum_shorts = mgr->lookup_proc("qa_sum_shorts"); + + init_jd(jd, gcp_qa_sum_shorts); + jd->eaa.nargs = MAX_ARGS_EA; + for (int i = 0; i < MAX_ARGS_EA; i++){ + jd->eaa.arg[i].direction = GCJD_DMA_GET; + jd->eaa.arg[i].ea_addr = ptr_to_ea(&short_buf[i * M]); + jd->eaa.arg[i].get_size = N * sizeof(short); + } + + if (!mgr->submit_job(jd)){ + printf("%d: submit_job(jd) failed, status = %d\n", __LINE__, jd->status); + } + else { + mgr->wait_job(jd); + CPPUNIT_ASSERT_EQUAL(JS_OK, jd->status); + for (int i = 0; i < MAX_ARGS_EA; i++){ + int expected = sum_shorts(&short_buf[i * M], N); + int actual = jd->output.arg[i].s32; + CPPUNIT_ASSERT_EQUAL(expected, actual); + } + } + + mgr->free_job_desc(jd); + delete mgr; +} + +static bool +confirm_const(const unsigned char *buf, size_t len, unsigned char v) +{ + bool ok = true; + + for (size_t i = 0; i < len; i++){ + if (buf[i] != v){ + ok = false; + printf("confirm_const: buf[%6d] = 0x%02x, expected = 0x%02x\n", + i, buf[i], v); + } + } + + return ok; +} + +static bool +confirm_seq(const unsigned char *buf, size_t len, unsigned char v) +{ + bool ok = true; + + for (size_t i = 0; i < len; i++, v++){ + if (buf[i] != v){ + ok = false; + printf("confirm_seq: buf[%6d] = 0x%02x, expected = 0x%02x\n", + i, buf[i], v); + } + } + + return ok; +} + +static void +test_put_seq(gc_job_manager *mgr, int offset, int len, int starting_val) +{ + gc_job_desc *jd = mgr->alloc_job_desc(); + gc_proc_id_t gcp_qa_put_seq = mgr->lookup_proc("qa_put_seq"); + + unsigned char *buf = (unsigned char *) short_buf; + size_t buf_len = sizeof(short_buf); + memset(buf, 0xff, buf_len); + + // two cache lines into the buffer, so we can check before and after + int fixed_offset = 256; + + init_jd(jd, gcp_qa_put_seq); + jd->input.nargs = 1; + jd->input.arg[0].s32 = starting_val; + jd->eaa.nargs = 1; + jd->eaa.arg[0].ea_addr = ptr_to_ea(buf + fixed_offset + offset); + jd->eaa.arg[0].direction = GCJD_DMA_PUT; + jd->eaa.arg[0].put_size = len; + + if (!mgr->submit_job(jd)){ + printf("%d: submit_job(jd) failed, status = %d\n", __LINE__, jd->status); + } + else { + mgr->wait_job(jd); + CPPUNIT_ASSERT_EQUAL(JS_OK, jd->status); + + // check before + CPPUNIT_ASSERT(confirm_const(&buf[0], fixed_offset + offset, 0xff)); + + // check sequence + CPPUNIT_ASSERT(confirm_seq(&buf[fixed_offset + offset], len, starting_val)); + + // check after + CPPUNIT_ASSERT(confirm_const(&buf[fixed_offset + offset + len], + buf_len - fixed_offset - offset - len, 0xff)); + } + mgr->free_job_desc(jd); +} + +// +// Test all "put" alignments and sizes +// +void +qa_job_manager::t10_body() +{ + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + + int starting_val = 13; + + for (int offset = 0; offset <= 128; offset++){ + for (int len = 0; len <= 128; len++){ + test_put_seq(mgr, offset, len, starting_val++); + } + } + + int ea_args_maxsize = mgr->ea_args_maxsize(); + + // confirm maximum length + for (int offset = 0; offset <= 64; offset++){ + test_put_seq(mgr, offset, ea_args_maxsize, starting_val++); + } + + delete mgr; +} + +// +// test "put" args too long +// +void +qa_job_manager::t11_body() +{ + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + gc_job_desc *jd = mgr->alloc_job_desc(); + gc_proc_id_t gcp_qa_put_seq = mgr->lookup_proc("qa_put_seq"); + + init_jd(jd, gcp_qa_put_seq); + jd->input.nargs = 1; + jd->input.arg[0].s32 = 0; + jd->eaa.nargs = 1; + jd->eaa.arg[0].ea_addr = 0; + jd->eaa.arg[0].direction = GCJD_DMA_PUT; + jd->eaa.arg[0].put_size = 1 << 20; + + if (!mgr->submit_job(jd)){ + printf("%d: submit_job(jd) failed, status = %d\n", __LINE__, jd->status); + } + else { + mgr->wait_job(jd); + CPPUNIT_ASSERT_EQUAL(JS_ARGS_TOO_LONG, jd->status); + } + + mgr->free_job_desc(jd); + delete mgr; +} + +// +// test MAX_ARGS_EA "put" case +// +void +qa_job_manager::t12_body() +{ + static const int N = 127; + static const int M = 201; + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + gc_job_desc *jd = mgr->alloc_job_desc(); + gc_proc_id_t gcp_qa_put_seq = mgr->lookup_proc("qa_put_seq"); + + unsigned char *buf = (unsigned char *) short_buf; + size_t buf_len = sizeof(short_buf); + memset(buf, 0xff, buf_len); + + // two cache lines into the buffer, so we can check before and after + int fixed_offset = 256; + + int starting_val = 13; + + init_jd(jd, gcp_qa_put_seq); + jd->input.nargs = 1; + jd->input.arg[0].s32 = starting_val; + jd->eaa.nargs = MAX_ARGS_EA; + for (int i = 0; i < MAX_ARGS_EA; i++){ + jd->eaa.arg[i].direction = GCJD_DMA_PUT; + jd->eaa.arg[i].ea_addr = ptr_to_ea(&buf[i * M + fixed_offset]); + jd->eaa.arg[i].put_size = N; + } + + if (!mgr->submit_job(jd)){ + printf("%d: submit_job(jd) failed, status = %d\n", __LINE__, jd->status); + } + else { + mgr->wait_job(jd); + CPPUNIT_ASSERT_EQUAL(JS_OK, jd->status); + for (int i = 0; i < MAX_ARGS_EA; i++){ + CPPUNIT_ASSERT(confirm_seq(&buf[i * M + fixed_offset], N, starting_val)); + starting_val += N; + } + } + + mgr->free_job_desc(jd); + delete mgr; +} + +// +// test qa_copy primitive +// +void +qa_job_manager::t13_body() +{ + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + + memset(short_buf, 0, sizeof(short_buf)); + for (int i = 0; i < NS/2; i++) // init buffer with known qty + short_buf[i] = 0x1234 + i; + + int nshorts = NS/2; + + gc_job_desc *jd = mgr->alloc_job_desc(); + gc_proc_id_t gcp_qa_copy = mgr->lookup_proc("qa_copy"); + +#if 0 + printf("gcq_qa_copy = %d\n", gcp_qa_copy); + std::vector<std::string> procs = mgr->proc_names(); + for (unsigned int i = 0; i < procs.size(); ++i) + std::cout << procs[i] << std::endl; +#endif + + init_jd(jd, gcp_qa_copy); + jd->eaa.nargs = 2; + jd->eaa.arg[0].ea_addr = ptr_to_ea(&short_buf[nshorts]); + jd->eaa.arg[0].direction = GCJD_DMA_PUT; + jd->eaa.arg[0].put_size = nshorts * sizeof(short); + + jd->eaa.arg[1].ea_addr = ptr_to_ea(&short_buf[0]); + jd->eaa.arg[1].direction = GCJD_DMA_GET; + jd->eaa.arg[1].get_size = nshorts * sizeof(short); + + + if (!mgr->submit_job(jd)){ + printf("%d: submit_job(jd) failed, status = %d\n", __LINE__, jd->status); + } + else { + mgr->wait_job(jd); + CPPUNIT_ASSERT_EQUAL(JS_OK, jd->status); + CPPUNIT_ASSERT_EQUAL(0, jd->output.arg[0].s32); + + bool ok = true; + for (int i = 0; i < nshorts; i++){ + if (short_buf[i] != short_buf[i + nshorts]) + ok = false; + } + CPPUNIT_ASSERT(ok); + } + mgr->free_job_desc(jd); + + delete mgr; +} + +/* + * Parallel submission of NJOBS "put" jobs will test double buffered puts. + */ +void +qa_job_manager::t14_body() +{ + //return; + + //static const int NJOBS = 64; + static const int NJOBS = 128; + static const int LEN_PER_JOB = 1021; + unsigned char buf[NJOBS * LEN_PER_JOB]; + gc_job_desc_t *jd[NJOBS]; + bool done[NJOBS]; + + static const int STARTING_VAL = 13; + + memset(buf, 0xff, LEN_PER_JOB * NJOBS); + + gc_job_manager *mgr; + gc_jm_options opts; + opts.program_handle = &gcell_qa; + opts.nspes = 1; + mgr = gc_make_job_manager(&opts); + + + gc_proc_id_t gcp_qa_put_seq = mgr->lookup_proc("qa_put_seq"); + + // do all the initialization up front + + for (int i = 0, val = STARTING_VAL; i < NJOBS; i++, val += 3){ + jd[i] = mgr->alloc_job_desc(); + init_jd(jd[i], gcp_qa_put_seq); + jd[i]->input.nargs = 1; + jd[i]->input.arg[0].s32 = val; + jd[i]->eaa.nargs = 1; + jd[i]->eaa.arg[0].ea_addr = ptr_to_ea(&buf[i * LEN_PER_JOB]); + jd[i]->eaa.arg[0].direction = GCJD_DMA_PUT; + jd[i]->eaa.arg[0].put_size = LEN_PER_JOB; + } + + // submit them all + + for (int i = 0; i < NJOBS; i++){ + if (!mgr->submit_job(jd[i])){ + printf("%d: submit_job(jd[%2d]) failed, status = %d\n", __LINE__, i, jd[i]->status); + } + } + + // wait for them all + + int n = mgr->wait_jobs(NJOBS, jd, done, GC_WAIT_ALL); + CPPUNIT_ASSERT_EQUAL(NJOBS, n); + + // check results + + for (int i = 0, val = STARTING_VAL; i < NJOBS; i++, val += 3){ + CPPUNIT_ASSERT_EQUAL(JS_OK, jd[i]->status); + CPPUNIT_ASSERT(confirm_seq(&buf[i * LEN_PER_JOB], LEN_PER_JOB, val)); + } + + // cleanup + for (int i = 0; i < NJOBS; i++) + mgr->free_job_desc(jd[i]); + + delete mgr; +} + +void +qa_job_manager::t15_body() +{ +} diff --git a/gcell/src/lib/runtime/qa_job_manager.h b/gcell/src/lib/runtime/qa_job_manager.h new file mode 100644 index 000000000..ab3325b73 --- /dev/null +++ b/gcell/src/lib/runtime/qa_job_manager.h @@ -0,0 +1,89 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef INCLUDED_QA_JOB_MANAGER_H +#define INCLUDED_QA_JOB_MANAGER_H + +#include <cppunit/extensions/HelperMacros.h> +#include <cppunit/TestCase.h> + +class qa_job_manager; +typedef void (qa_job_manager::*test_t)(); + + +class qa_job_manager : public CppUnit::TestCase { + + CPPUNIT_TEST_SUITE(qa_job_manager); + CPPUNIT_TEST(t0); + CPPUNIT_TEST(t1); + CPPUNIT_TEST(t2); + CPPUNIT_TEST(t3); + CPPUNIT_TEST(t4); + CPPUNIT_TEST(t5); + CPPUNIT_TEST(t6); + CPPUNIT_TEST(t7); + CPPUNIT_TEST(t8); + CPPUNIT_TEST(t9); + CPPUNIT_TEST(t10); + CPPUNIT_TEST(t11); + CPPUNIT_TEST(t12); + CPPUNIT_TEST(t13); + CPPUNIT_TEST(t14); + CPPUNIT_TEST(t15); + CPPUNIT_TEST_SUITE_END(); + + private: + void leak_check(test_t t, const std::string &name); + + void t0(); + void t1(); + void t1_body(); + void t2(); + void t2_body(); + void t3(); + void t3_body(); + void t4(); + void t4_body(); + void t5(); + void t5_body(); + void t6(); + void t6_body(); + void t7(); + void t7_body(); + void t8(); + void t8_body(); + void t9(); + void t9_body(); + void t10(); + void t10_body(); + void t11(); + void t11_body(); + void t12(); + void t12_body(); + void t13(); + void t13_body(); + void t14(); + void t14_body(); + void t15(); + void t15_body(); + +}; + +#endif /* INCLUDED_QA_JOB_MANAGER_H */ diff --git a/gcell/src/lib/runtime/qa_lib.cc b/gcell/src/lib/runtime/qa_lib.cc new file mode 100644 index 000000000..d8a8960c6 --- /dev/null +++ b/gcell/src/lib/runtime/qa_lib.cc @@ -0,0 +1,43 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +/* + * This class gathers together all the test cases for the lib + * directory into a single test suite. As you create new test cases, + * add them here. + */ + +#include <qa_lib.h> +#include <qa_jd_stack.h> +#include <qa_jd_queue.h> +#include <qa_job_manager.h> + +CppUnit::TestSuite * +qa_lib::suite() +{ + CppUnit::TestSuite *s = new CppUnit::TestSuite("lib"); + + s->addTest(qa_jd_stack::suite()); + s->addTest(qa_jd_queue::suite()); + s->addTest(qa_job_manager::suite()); + + return s; +} diff --git a/gcell/src/lib/runtime/qa_lib.h b/gcell/src/lib/runtime/qa_lib.h new file mode 100644 index 000000000..594efcdc8 --- /dev/null +++ b/gcell/src/lib/runtime/qa_lib.h @@ -0,0 +1,35 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef INCLUDED_QA_LIB_H +#define INCLUDED_QA_LIB_H + +#include <cppunit/TestSuite.h> + +//! collect all the tests for the lib directory + +class qa_lib { +public: + //! return suite of tests + static CppUnit::TestSuite *suite(); +}; + + +#endif /* INCLUDED_QA_LIB_H */ diff --git a/gcell/src/lib/runtime/spu/gc_delay.c b/gcell/src/lib/runtime/spu/gc_delay.c new file mode 100644 index 000000000..21ee587f8 --- /dev/null +++ b/gcell/src/lib/runtime/spu/gc_delay.c @@ -0,0 +1,58 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include "gc_delay.h" +#include <compiler.h> + +inline static void +gc_udelay_1us(void) +{ + unsigned int i = 158; + + do { // 20 clocks per iteration of the loop + asm ("nop $127; nop $127; nop $127; nop $127; nop $127"); + asm ("nop $127; nop $127; nop $127; nop $127; nop $127"); + asm ("nop $127; nop $127; nop $127; nop $127; nop $127"); + asm ("nop $127; nop $127"); + } while(--i != 0); +} + +void +gc_udelay(unsigned int usecs) +{ + unsigned int i; + for (i = 0; i < usecs; i++) + gc_udelay_1us(); +} + +void +gc_cdelay(unsigned int cpu_cycles) +{ + if (cpu_cycles < 40) // roughly the amount of overhead + return; + + cpu_cycles >>= 2; // about 4 cycles / loop + + while (cpu_cycles-- != 0){ + asm ("nop $127"); // keeps compiler from removing the loop + } +} + diff --git a/gcell/src/lib/runtime/spu/gc_logging.c b/gcell/src/lib/runtime/spu/gc_logging.c new file mode 100644 index 000000000..65a537504 --- /dev/null +++ b/gcell/src/lib/runtime/spu/gc_logging.c @@ -0,0 +1,77 @@ +/* -*- c++ -*- */ +/* + * Copyright 2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <gc_logging.h> +#include <spu_intrinsics.h> +#include <spu_mfcio.h> +#include <gc_spu_args.h> + +static gc_eaddr_t log_base_ea; // base address of log entries in EA +static uint32_t log_idx_mask; // nentries - 1 +static uint32_t log_idx; // current log entry index +static uint32_t log_seqno; + +static int log_tags; // two consecutive tags +static int tmp_buffer_busy; // bitmask: buffer busy state +static int tmp_buffer_idx; // 0 or 1 +static gc_log_entry_t tmp_buffer[2]; + +void +_gc_log_init(gc_log_t info) +{ + spu_write_decrementer(~0); + + log_base_ea = info.base; + log_idx_mask = info.nentries - 1; + log_idx = 0; + log_seqno = 0; + + log_tags = mfc_multi_tag_reserve(2); + tmp_buffer_busy = 0; + tmp_buffer_idx = 0; + + gc_log_write0(GCL_SS_SYS, 0); +} + +void +_gc_log_write(gc_log_entry_t entry) +{ + if (log_base_ea == 0) + return; + + entry.seqno = log_seqno++; + entry.timestamp = spu_read_decrementer(); + + if (tmp_buffer_busy & (1 << tmp_buffer_idx)){ + mfc_write_tag_mask(1 << (log_tags + tmp_buffer_idx)); + mfc_read_tag_status_all(); + } + + tmp_buffer[tmp_buffer_idx] = entry; // save local copy + + mfc_put(&tmp_buffer[tmp_buffer_idx], + log_base_ea + log_idx * sizeof(entry), sizeof(entry), + log_tags + tmp_buffer_idx, 0, 0); + + tmp_buffer_busy |= (1 << tmp_buffer_idx); + tmp_buffer_idx ^= 0x1; + log_idx = (log_idx + 1) & log_idx_mask; +} diff --git a/gcell/src/lib/runtime/spu/gc_main.c b/gcell/src/lib/runtime/spu/gc_main.c new file mode 100644 index 000000000..96381fa66 --- /dev/null +++ b/gcell/src/lib/runtime/spu/gc_main.c @@ -0,0 +1,682 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007,2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define ENABLE_GC_LOGGING // define to enable logging + +#include <spu_intrinsics.h> +#include <spu_mfcio.h> +#include <sync_utils.h> +#include "gc_spu_config.h" +#include "gc_spu_args.h" +#include "gc_job_desc.h" +#include "gc_mbox.h" +#include "gc_jd_queue.h" +#include "gc_delay.h" +#include "gc_declare_proc.h" +#include "spu_buffers.h" +#include <string.h> +#include <assert.h> +#include <stdio.h> + + +#define MIN(a,b) ((a) < (b) ? (a) : (b)) +#define MAX(a,b) ((a) > (b) ? (a) : (b)) + +//! round x down to p2 boundary (p2 must be a power-of-2) +#define ROUND_DN(x, p2) ((x) & ~((p2)-1)) + +//! round x up to p2 boundary (p2 must be a power-of-2) +#define ROUND_UP(x, p2) (((x)+((p2)-1)) & ~((p2)-1)) + + +#define USE_LLR_LOST_EVENT 0 // define to 0 or 1 + +int gc_sys_tag; // tag for misc DMA operations +static gc_spu_args_t spu_args; + +static struct gc_proc_def *gc_proc_def; // procedure entry points + +// ------------------------------------------------------------------------ + +// state for DMA'ing arguments in and out + +static int get_tag; // 1 tag for job arg gets +static int put_tags; // 2 tags for job arg puts + +static int pb_idx = 0; // current put buffer index (0 or 1) + +// bitmask (bit per put buffer): bit is set if DMA is started but not complete +static int put_in_progress = 0; +#define PBI_MASK(_pbi_) (1 << (_pbi_)) + +// ------------------------------------------------------------------------ + +// our working copy of the completion info +static gc_comp_info_t comp_info = { + .in_use = 1, + .ncomplete = 0 +}; + +static int ci_idx = 0; // index of current comp_info +static int ci_tags; // two consecutive dma tags + +// ------------------------------------------------------------------------ + +/* + * Wait until EA copy of comp_info[idx].in_use is 0 + */ +static void +wait_for_ppe_to_be_done_with_comp_info(int idx) +{ + char _tmp[256]; + char *buf = (char *) ALIGN(_tmp, 128); // get cache-aligned buffer + gc_comp_info_t *p = (gc_comp_info_t *) buf; + + assert(sizeof(gc_comp_info_t) == 128); + + do { + mfc_get(buf, spu_args.comp_info[idx], 128, gc_sys_tag, 0, 0); + mfc_write_tag_mask(1 << gc_sys_tag); + mfc_read_tag_status_all(); + if (p->in_use == 0) + return; + + gc_udelay(5); + + } while (1); +} + +static void +flush_completion_info(void) +{ + // events: 0x3X + + static int total_complete = 0; + + if (comp_info.ncomplete == 0) + return; + + // ensure that PPE is done with the buffer we're about to overwrite + wait_for_ppe_to_be_done_with_comp_info(ci_idx); + + // dma the comp_info out to PPE + int tag = ci_tags + ci_idx; + mfc_put(&comp_info, spu_args.comp_info[ci_idx], sizeof(gc_comp_info_t), tag, 0, 0); + + // we need to wait for the completion info to finish, as well as + // any EA argument puts. + + int tag_mask = 1 << tag; // the comp_info tag + if (put_in_progress & PBI_MASK(0)) + tag_mask |= (1 << (put_tags + 0)); + if (put_in_progress & PBI_MASK(1)) + tag_mask |= (1 << (put_tags + 1)); + + gc_log_write2(GCL_SS_SYS, 0x30, put_in_progress, tag_mask); + + mfc_write_tag_mask(tag_mask); // the tags we're interested in + mfc_read_tag_status_all(); // wait for DMA to complete + put_in_progress = 0; // mark them all complete + + total_complete += comp_info.ncomplete; + gc_log_write4(GCL_SS_SYS, 0x31, + put_in_progress, ci_idx, comp_info.ncomplete, total_complete); + + // send PPE a message + spu_writech(SPU_WrOutIntrMbox, MK_MBOX_MSG(OP_JOBS_DONE, ci_idx)); + + ci_idx ^= 0x1; // switch buffers + comp_info.in_use = 1; + comp_info.ncomplete = 0; +} + +// ------------------------------------------------------------------------ + +static unsigned int backoff; // current backoff value in clock cycles +static unsigned int _backoff_start; +static unsigned int _backoff_cap; + +/* + * For 3.2 GHz SPE + * + * 12 4095 cycles 1.3 us + * 13 8191 cycles 2.6 us + * 14 16383 cycles 5.1 us + * 15 32767 cycles 10.2 us + * 16 20.4 us + * 17 40.8 us + * 18 81.9 us + * 19 163.8 us + * 20 327.7 us + * 21 655.4 us + */ +static unsigned char log2_backoff_start[16] = { +// 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 +// ------------------------------------------------------------- + 12, 12, 12, 13, 13, 13, 13, 14, 14, 14, 14, 15, 15, 15, 16, 16 +}; + +static unsigned char log2_backoff_cap[16] = { +// 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 +// ------------------------------------------------------------- + 17, 17, 17, 18, 18, 18, 18, 19, 19, 19, 19, 20, 20, 20, 21, 21 +}; + +static void +backoff_init(void) +{ + _backoff_cap = (1 << (log2_backoff_cap[(spu_args.nspus - 1) & 0xf])) - 1; + _backoff_start = (1 << (log2_backoff_start[(spu_args.nspus - 1) & 0xf])) - 1; + + backoff = _backoff_start; +} + +static void +backoff_reset(void) +{ + backoff = _backoff_start; +} + +static void +backoff_delay(void) +{ + gc_cdelay(backoff); + + // capped exponential backoff + backoff = ((backoff << 1) + 1) & _backoff_cap; +} + +// ------------------------------------------------------------------------ + +static inline unsigned int +make_mask(int nbits) +{ + return ~(~0 << nbits); +} + +static unsigned int dc_work; +static int dc_put_tag; +static unsigned char *dc_ls_base; +static gc_eaddr_t dc_ea_base; + +// divide and conquer +static void +d_and_c(unsigned int offset, unsigned int len) +{ + unsigned int mask = make_mask(len) << offset; + unsigned int t = mask & dc_work; + if (t == 0) // nothing to do + return; + if (t == mask){ // got a match, generate dma + mfc_put(dc_ls_base + offset, dc_ea_base + offset, len, dc_put_tag, 0, 0); + } + else { // bisect + len >>= 1; + d_and_c(offset, len); + d_and_c(offset + len, len); + } +} + +// Handle the nasty case of a dma xfer that's less than 16 bytes long. +// len is guaranteed to be in [1, 15] + +static void +handle_slow_and_tedious_dma(gc_eaddr_t ea, unsigned char *ls, + unsigned int len, int put_tag) +{ + // Set up for divide and conquer + unsigned int alignment = ((uintptr_t) ls) & 0x7; + dc_work = make_mask(len) << alignment; + dc_ls_base = (unsigned char *) ROUND_DN((uintptr_t) ls, 8); + dc_ea_base = ROUND_DN(ea, (gc_eaddr_t) 8); + dc_put_tag = put_tag; + + d_and_c( 0, 8); + d_and_c( 8, 8); + d_and_c(16, 8); +} + + +static void +process_job(gc_eaddr_t jd_ea, gc_job_desc_t *jd) +{ + // events: 0x2X + + jd->status = JS_OK; // assume success + + if (jd->proc_id >= spu_args.nproc_defs) + jd->status = JS_UNKNOWN_PROC; + + else { + + if (jd->eaa.nargs == 0) + (*gc_proc_def[jd->proc_id].proc)(&jd->input, &jd->output, &jd->eaa); + + else { // handle EA args that must be DMA'd in/out + + gc_job_ea_args_t *eaa = &jd->eaa; + + int NELMS = + MAX(MAX_ARGS_EA, + (GC_SPU_BUFSIZE + MFC_MAX_DMA_SIZE - 1) / MFC_MAX_DMA_SIZE); + + mfc_list_element_t dma_get_list[NELMS]; + //mfc_list_element_t dma_put_list[NELMS]; + + memset(dma_get_list, 0, sizeof(dma_get_list)); + //memset(dma_put_list, 0, sizeof(dma_put_list)); + + int gli = 0; // get list index + //int pli = 0; // put list index + + unsigned char *get_base = _gci_getbuf[0]; + unsigned char *get_t = get_base; + unsigned int total_get_dma_len = 0; + + unsigned char *put_base = _gci_putbuf[pb_idx]; + unsigned char *put_t = put_base; + unsigned int total_put_alloc = 0; + int put_tag = put_tags + pb_idx; + + // Do we have any "put" args? If so ensure that previous + // dma from this buffer is complete + + gc_log_write2(GCL_SS_SYS, 0x24, put_in_progress, jd->sys.direction_union); + + if ((jd->sys.direction_union & GCJD_DMA_PUT) + && (put_in_progress & PBI_MASK(pb_idx))){ + + gc_log_write2(GCL_SS_SYS, 0x25, put_in_progress, 1 << put_tag); + + mfc_write_tag_mask(1 << put_tag); // the tag we're interested in + mfc_read_tag_status_all(); // wait for DMA to complete + put_in_progress &= ~(PBI_MASK(pb_idx)); + + gc_log_write1(GCL_SS_SYS, 0x26, put_in_progress); + } + + + // for now, all EA's must have the same high 32-bits + gc_eaddr_t common_ea = eaa->arg[0].ea_addr; + + + // assign LS addresses for buffers + + for (unsigned int i = 0; i < eaa->nargs; i++){ + + gc_eaddr_t ea_base = 0; + unsigned char *ls_base; + int offset; + unsigned int dma_len; + + if (eaa->arg[i].direction == GCJD_DMA_GET){ + ea_base = ROUND_DN(eaa->arg[i].ea_addr, (gc_eaddr_t) CACHE_LINE_SIZE); + offset = eaa->arg[i].ea_addr & (CACHE_LINE_SIZE-1); + dma_len = ROUND_UP(eaa->arg[i].get_size + offset, CACHE_LINE_SIZE); + total_get_dma_len += dma_len; + + if (total_get_dma_len > GC_SPU_BUFSIZE){ + jd->status = JS_ARGS_TOO_LONG; + goto wrap_up; + } + + ls_base = get_t; + get_t += dma_len; + eaa->arg[i].ls_addr = ls_base + offset; + + if (0){ + assert((mfc_ea2l(eaa->arg[i].ea_addr) & 0x7f) == ((intptr_t)eaa->arg[i].ls_addr & 0x7f)); + assert((ea_base & 0x7f) == 0); + assert(((intptr_t)ls_base & 0x7f) == 0); + assert((dma_len & 0x7f) == 0); + assert((eaa->arg[i].get_size <= dma_len) + && dma_len <= (eaa->arg[i].get_size + offset + CACHE_LINE_SIZE - 1)); + } + + // add to dma get list + // FIXME (someday) the dma list is where the JS_BAD_EAH limitation comes from + + while (dma_len != 0){ + int n = MIN(dma_len, MFC_MAX_DMA_SIZE); + dma_get_list[gli].size = n; + dma_get_list[gli].eal = mfc_ea2l(ea_base); + dma_len -= n; + ea_base += n; + gli++; + } + } + + else if (eaa->arg[i].direction == GCJD_DMA_PUT){ + // + // This case is a trickier than the PUT case since we can't + // write outside of the bounds of the user provided buffer. + // We still align the buffers to 128-bytes for good performance + // in the middle portion of the xfers. + // + ea_base = ROUND_DN(eaa->arg[i].ea_addr, (gc_eaddr_t) CACHE_LINE_SIZE); + offset = eaa->arg[i].ea_addr & (CACHE_LINE_SIZE-1); + + uint32_t ls_alloc_len = + ROUND_UP(eaa->arg[i].put_size + offset, CACHE_LINE_SIZE); + + total_put_alloc += ls_alloc_len; + + if (total_put_alloc > GC_SPU_BUFSIZE){ + jd->status = JS_ARGS_TOO_LONG; + goto wrap_up; + } + + ls_base = put_t; + put_t += ls_alloc_len; + eaa->arg[i].ls_addr = ls_base + offset; + + if (1){ + assert((mfc_ea2l(eaa->arg[i].ea_addr) & 0x7f) + == ((intptr_t)eaa->arg[i].ls_addr & 0x7f)); + assert((ea_base & 0x7f) == 0); + assert(((intptr_t)ls_base & 0x7f) == 0); + } + } + + else + assert(0); + } + + // fire off the dma to fetch the args and wait for it to complete + mfc_getl(get_base, common_ea, dma_get_list, gli*sizeof(dma_get_list[0]), get_tag, 0, 0); + mfc_write_tag_mask(1 << get_tag); // the tag we're interested in + mfc_read_tag_status_all(); // wait for DMA to complete + + // do the work + (*gc_proc_def[jd->proc_id].proc)(&jd->input, &jd->output, &jd->eaa); + + + // Do we have any "put" args? If so copy them out + if (jd->sys.direction_union & GCJD_DMA_PUT){ + + // Do the copy out using single DMA xfers. The LS ranges + // aren't generally contiguous. + + bool started_dma = false; + + for (unsigned int i = 0; i < eaa->nargs; i++){ + if (eaa->arg[i].direction == GCJD_DMA_PUT && eaa->arg[i].put_size != 0){ + + started_dma = true; + + gc_eaddr_t ea; + unsigned char *ls; + unsigned int len; + + ea = eaa->arg[i].ea_addr; + ls = (unsigned char *) eaa->arg[i].ls_addr; + len = eaa->arg[i].put_size; + + if (len < 16) + handle_slow_and_tedious_dma(ea, ls, len, put_tag); + + else { + if ((ea & 0xf) != 0){ + + // printf("1: ea = 0x%x len = %5d\n", (int) ea, len); + + // handle the "pre-multiple-of-16" portion + // do 1, 2, 4, or 8 byte xfers as required + + if (ea & 0x1){ // do a 1-byte xfer + mfc_put(ls, ea, 1, put_tag, 0, 0); + ea += 1; + ls += 1; + len -= 1; + } + if (ea & 0x2){ // do a 2-byte xfer + mfc_put(ls, ea, 2, put_tag, 0, 0); + ea += 2; + ls += 2; + len -= 2; + } + if (ea & 0x4){ // do a 4-byte xfer + mfc_put(ls, ea, 4, put_tag, 0, 0); + ea += 4; + ls += 4; + len -= 4; + } + if (ea & 0x8){ // do an 8-byte xfer + mfc_put(ls, ea, 8, put_tag, 0, 0); + ea += 8; + ls += 8; + len -= 8; + } + } + + if (1){ + // printf("2: ea = 0x%x len = %5d\n", (int) ea, len); + assert((ea & 0xf) == 0); + assert((((intptr_t) ls) & 0xf) == 0); + } + + // handle the "multiple-of-16" portion + + int aligned_len = ROUND_DN(len, 16); + len = len & (16 - 1); + + while (aligned_len != 0){ + int dma_len = MIN(aligned_len, MFC_MAX_DMA_SIZE); + mfc_put(ls, ea, dma_len, put_tag, 0, 0); + ea += dma_len; + ls += dma_len; + aligned_len -= dma_len; + } + + if (1){ + // printf("3: ea = 0x%x len = %5d\n", (int)ea, len); + assert((ea & 0xf) == 0); + assert((((intptr_t) ls) & 0xf) == 0); + } + + // handle "post-multiple-of-16" portion + + if (len != 0){ + + if (len >= 8){ // do an 8-byte xfer + mfc_put(ls, ea, 8, put_tag, 0, 0); + ea += 8; + ls += 8; + len -= 8; + } + if (len >= 4){ // do a 4-byte xfer + mfc_put(ls, ea, 4, put_tag, 0, 0); + ea += 4; + ls += 4; + len -= 4; + } + if (len >= 2){ // do a 2-byte xfer + mfc_put(ls, ea, 2, put_tag, 0, 0); + ea += 2; + ls += 2; + len -= 2; + } + if (len >= 1){ // do a 1-byte xfer + mfc_put(ls, ea, 1, put_tag, 0, 0); + ea += 1; + ls += 1; + len -= 1; + } + if (1) + assert(len == 0); + } + } + } + } + if (started_dma){ + put_in_progress |= PBI_MASK(pb_idx); // note it's running + gc_log_write2(GCL_SS_SYS, 0x27, put_in_progress, pb_idx); + pb_idx ^= 1; // toggle current buffer + } + } + } + } + + wrap_up:; // semicolon creates null statement for C99 compliance + + // Copy job descriptor back out to EA. + // (The dma will be waited on in flush_completion_info) + int tag = ci_tags + ci_idx; // use the current completion tag + mfc_put(jd, jd_ea, sizeof(*jd), tag, 0, 0); + + mfc_sync(tag); // FIXME this makes it work, but is expensive + + + // Tell PPE we're done with the job. + // + // We queue these up until we run out of room, or until we can send + // the info to the PPE w/o blocking. The blocking check is in + // main_loop + + comp_info.job_id[comp_info.ncomplete++] = jd->sys.job_id; + + if (comp_info.ncomplete == GC_CI_NJOBS){ + gc_log_write0(GCL_SS_SYS, 0x28); + flush_completion_info(); + } +} + +static void +main_loop(void) +{ + // events: 0x1X + + static gc_job_desc_t jd; // static gets us proper alignment + gc_eaddr_t jd_ea; + int total_jobs = 0; + + // setup events + spu_writech(SPU_WrEventMask, MFC_LLR_LOST_EVENT); + gc_jd_queue_getllar(spu_args.queue); // get a line reservation on the queue + + while (1){ + +#if (USE_LLR_LOST_EVENT) + + if (unlikely(spu_readchcnt(SPU_RdEventStat))){ + // + // execute standard event handling prologue + // + int status = spu_readch(SPU_RdEventStat); + int mask = spu_readch(SPU_RdEventMask); + spu_writech(SPU_WrEventMask, mask & ~status); // disable active events + spu_writech(SPU_WrEventAck, status); // ack active events + + // execute per-event actions + + if (status & MFC_LLR_LOST_EVENT){ + // + // We've lost a line reservation. This is most likely caused + // by somebody doing something to the queue. Go look and see + // if there's anything for us. + // + if (gc_jd_queue_dequeue(spu_args.queue, &jd_ea, &jd)) + process_job(jd_ea, &jd); + + gc_jd_queue_getllar(spu_args.queue); // get a new reservation + } + + // + // execute standard event handling epilogue + // + spu_writech(SPU_WrEventMask, mask); // restore event mask + } + +#else + + // try to get a job from the job queue + if (gc_jd_queue_dequeue(spu_args.queue, &jd_ea, &jd)){ + total_jobs++; + gc_log_write2(GCL_SS_SYS, 0x10, jd.sys.job_id, total_jobs); + + process_job(jd_ea, &jd); + + gc_log_write2(GCL_SS_SYS, 0x11, jd.sys.job_id, total_jobs); + backoff_reset(); + } + else + backoff_delay(); + +#endif + + // any msgs for us? + + if (unlikely(spu_readchcnt(SPU_RdInMbox))){ + int msg = spu_readch(SPU_RdInMbox); + // printf("spu[%d] mbox_msg: 0x%08x\n", spu_args.spu_idx, msg); + if (MBOX_MSG_OP(msg) == OP_EXIT){ + flush_completion_info(); + return; + } + if (MBOX_MSG_OP(msg) == OP_GET_SPU_BUFSIZE){ + spu_writech(SPU_WrOutIntrMbox, MK_MBOX_MSG(OP_SPU_BUFSIZE, GC_SPU_BUFSIZE_BASE)); + } + } + + // If we've got job completion info for the PPE and we can send a + // message without blocking, do it. + + if (comp_info.ncomplete != 0 && spu_readchcnt(SPU_WrOutIntrMbox) != 0){ + gc_log_write0(GCL_SS_SYS, 0x12); + flush_completion_info(); + } + } +} + + +int +main(unsigned long long spe_id __attribute__((unused)), + unsigned long long argp, + unsigned long long envp __attribute__((unused))) +{ + gc_sys_tag = mfc_tag_reserve(); // allocate a tag for our misc DMA operations + get_tag = mfc_tag_reserve(); + ci_tags = mfc_multi_tag_reserve(2); + put_tags = mfc_multi_tag_reserve(2); + +#if 0 + printf("gc_sys_tag = %d\n", gc_sys_tag); + printf("get_tag = %d\n", get_tag); + printf("ci_tags = %d\n", ci_tags); + printf("put_tags = %d\n", put_tags); +#endif + + // dma the args in + mfc_get(&spu_args, argp, sizeof(spu_args), gc_sys_tag, 0, 0); + mfc_write_tag_mask(1 << gc_sys_tag); // the tag we're interested in + mfc_read_tag_status_all(); // wait for DMA to complete + + // initialize pointer to procedure entry table + gc_proc_def = (gc_proc_def_t *) spu_args.proc_def_ls_addr; + + // initialize logging + _gc_log_init(spu_args.log); + + backoff_init(); // initialize backoff parameters + + main_loop(); + return 0; +} diff --git a/gcell/src/lib/runtime/spu/gc_spu_config.h b/gcell/src/lib/runtime/spu/gc_spu_config.h new file mode 100644 index 000000000..997645e68 --- /dev/null +++ b/gcell/src/lib/runtime/spu/gc_spu_config.h @@ -0,0 +1,33 @@ +/* -*- c++ -*- */ +/* + * Copyright 2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef INCLUDED_GC_SPU_CONFIG_H +#define INCLUDED_GC_SPU_CONFIG_H + +#include <gc_job_desc.h> + +#define CACHE_LINE_SIZE 128 // in bytes +#define GC_SPU_BUFSIZE_BASE (32 * 1024) // must be multiple of CACHE_LINE_SIZE +#define GC_SPU_BUFSIZE (GC_SPU_BUFSIZE_BASE + MAX_ARGS_EA * CACHE_LINE_SIZE) + +#define NGETBUFS 1 // single buffer job arg gets +#define NPUTBUFS 2 // double buffer job arg puts + +#endif /* INCLUDED_GC_SPU_CONFIG_H */ diff --git a/gcell/src/lib/runtime/spu/gc_spu_jd_queue.c b/gcell/src/lib/runtime/spu/gc_spu_jd_queue.c new file mode 100644 index 000000000..ba4a1b9d2 --- /dev/null +++ b/gcell/src/lib/runtime/spu/gc_spu_jd_queue.c @@ -0,0 +1,103 @@ +/* -*- c++ -*- */ +/* + * Copyright 2007 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include "gc_jd_queue.h" +#include "mutex_lock.h" +#include "mutex_unlock.h" + +extern int gc_sys_tag; + +bool +gc_jd_queue_dequeue(gc_eaddr_t q, gc_eaddr_t *item_ea, gc_job_desc_t *item) +{ + gc_jd_queue_t local_q; + + // Before aquiring the lock, see if it's possible that there's + // something in the queue. Checking in this way makes it easier + // for the PPE to insert things, since we're not contending for + // the lock unless there is something in the queue. + + // copy in the queue structure + mfc_get(&local_q, q, sizeof(gc_jd_queue_t), gc_sys_tag, 0, 0); + mfc_write_tag_mask(1 << gc_sys_tag); // the tag we're interested in + mfc_read_tag_status_all(); // wait for DMA to complete + + if (local_q.head == 0){ // empty + return false; + } + + // When we peeked, head was non-zero. Now grab the + // lock and do it for real. + + _mutex_lock(q + offsetof(gc_jd_queue_t, mutex)); + + // copy in the queue structure + mfc_get(&local_q, q, sizeof(gc_jd_queue_t), gc_sys_tag, 0, 0); + mfc_write_tag_mask(1 << gc_sys_tag); // the tag we're interested in + mfc_read_tag_status_all(); // wait for DMA to complete + + if (local_q.head == 0){ // empty + _mutex_unlock(q + offsetof(gc_jd_queue_t, mutex)); + return false; + } + + // copy in job descriptor at head of queue + *item_ea = local_q.head; + mfc_get(item, local_q.head, sizeof(gc_job_desc_t), gc_sys_tag, 0, 0); + mfc_write_tag_mask(1 << gc_sys_tag); // the tag we're interested in + mfc_read_tag_status_all(); // wait for DMA to complete + + local_q.head = item->sys.next; + item->sys.next = 0; + if (local_q.head == 0) // now empty? + local_q.tail = 0; + + + // copy the queue structure back out + mfc_put(&local_q, q, sizeof(gc_jd_queue_t), gc_sys_tag, 0, 0); + mfc_write_tag_mask(1 << gc_sys_tag); // the tag we're interested in + mfc_read_tag_status_all(); // wait for DMA to complete + + // Q: FIXME do we need to order stores in EA or can we just clear the + // local copy of the mutex above and blast it out, removing the need + // for this explicit unlock? + // + // A: Manual says it's better to use an atomic op rather than + // a normal DMA, and that a putlluc is better than a putllc if + // you can use it. + + _mutex_unlock(q + offsetof(gc_jd_queue_t, mutex)); + return true; +} + + +void +gc_jd_queue_getllar(gc_eaddr_t q) +{ + // get reservation that includes the tail of the queue + gc_eaddr_t tail = q + offsetof(gc_jd_queue_t, tail); + + char _tmp[256]; + char *buf = (char *) ALIGN(_tmp, 128); // get cache-aligned buffer + + mfc_getllar(buf, ALIGN128_EA(tail), 0, 0); + spu_readch(MFC_RdAtomicStat); +} diff --git a/gcell/src/lib/runtime/spu/gcell_qa.c b/gcell/src/lib/runtime/spu/gcell_qa.c new file mode 100644 index 000000000..51bf38a6a --- /dev/null +++ b/gcell/src/lib/runtime/spu/gcell_qa.c @@ -0,0 +1,109 @@ +/* -*- c++ -*- */ +/* + * Copyright 2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <gc_delay.h> +#include <gc_declare_proc.h> +#include <string.h> + + +#define _UNUSED __attribute__((unused)) + +// FIXME move these out of here; only for QA usage + +static void +qa_nop(const gc_job_direct_args_t *input _UNUSED, + gc_job_direct_args_t *output _UNUSED, + const gc_job_ea_args_t *eaa _UNUSED) +{ +} + +GC_DECLARE_PROC(qa_nop, "qa_nop"); + +static int +sum_shorts(short *p, int nshorts) +{ + int total = 0; + for (int i = 0; i < nshorts; i++) + total += p[i]; + + return total; +} + +static void +qa_sum_shorts(const gc_job_direct_args_t *input _UNUSED, + gc_job_direct_args_t *output, + const gc_job_ea_args_t *eaa) +{ + for (unsigned int i = 0; i < eaa->nargs; i++){ + short *p = eaa->arg[i].ls_addr; + int n = eaa->arg[i].get_size / sizeof(short); + output->arg[i].s32 = sum_shorts(p, n); + //printf("qa_sum_shorts(%p, %d) = %d\n", p, n, output->arg[i].s32); + } +} + +GC_DECLARE_PROC(qa_sum_shorts, "qa_sum_shorts"); + +static void +write_seq(unsigned char *p, int nbytes, int counter) +{ + for (int i = 0; i < nbytes; i++) + p[i] = counter++; +} + +static void +qa_put_seq(const gc_job_direct_args_t *input, + gc_job_direct_args_t *output _UNUSED, + const gc_job_ea_args_t *eaa) +{ + int counter = input->arg[0].s32; + + for (unsigned int i = 0; i < eaa->nargs; i++){ + unsigned char *p = eaa->arg[i].ls_addr; + int n = eaa->arg[i].put_size; + write_seq(p, n, counter); + counter += n; + } +} + +GC_DECLARE_PROC(qa_put_seq, "qa_put_seq"); + +static void +qa_copy(const gc_job_direct_args_t *input _UNUSED, + gc_job_direct_args_t *output, + const gc_job_ea_args_t *eaa) +{ + if (eaa->nargs != 2 + || eaa->arg[0].direction != GCJD_DMA_PUT + || eaa->arg[1].direction != GCJD_DMA_GET){ + output->arg[0].s32 = -1; + return; + } + + output->arg[0].s32 = 0; + unsigned n = eaa->arg[0].put_size; + if (eaa->arg[1].get_size < n) + n = eaa->arg[1].get_size; + + memcpy(eaa->arg[0].ls_addr, eaa->arg[1].ls_addr, n); +} + +GC_DECLARE_PROC(qa_copy, "qa_copy"); diff --git a/gcell/src/lib/runtime/spu/spu_buffers.c b/gcell/src/lib/runtime/spu/spu_buffers.c new file mode 100644 index 000000000..58b405931 --- /dev/null +++ b/gcell/src/lib/runtime/spu/spu_buffers.c @@ -0,0 +1,35 @@ +/* -*- c++ -*- */ +/* + * Copyright 2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <spu_buffers.h> +#include <compiler.h> + +static unsigned char _getbuf[NGETBUFS][GC_SPU_BUFSIZE] _AL128; +static unsigned char _putbuf[NPUTBUFS][GC_SPU_BUFSIZE] _AL128; + +unsigned char *_gci_getbuf[NGETBUFS] = { + _getbuf[0] +}; + +unsigned char *_gci_putbuf[NPUTBUFS] = { + _putbuf[0], + _putbuf[1] +}; diff --git a/gcell/src/lib/runtime/spu/spu_buffers.h b/gcell/src/lib/runtime/spu/spu_buffers.h new file mode 100644 index 000000000..24811dc38 --- /dev/null +++ b/gcell/src/lib/runtime/spu/spu_buffers.h @@ -0,0 +1,32 @@ +/* -*- c++ -*- */ +/* + * Copyright 2008 Free Software Foundation, Inc. + * + * This file is part of GNU Radio + * + * GNU Radio is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3, or (at your option) + * any later version. + * + * GNU Radio is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef INCLUDED_SPU_BUFFERS_H +#define INCLUDED_SPU_BUFFERS_H + +#include "gc_spu_config.h" + +//! pointer to input buffer +extern unsigned char *_gci_getbuf[NGETBUFS]; + +//! pointers to output buffers +extern unsigned char *_gci_putbuf[NPUTBUFS]; + +#endif /* INCLUDED_SPU_BUFFERS_H */ |