URL
https://opencores.org/ocsvn/openrisc/openrisc/trunk
Subversion Repositories openrisc
[/] [openrisc/] [trunk/] [rtos/] [ecos-2.0/] [packages/] [kernel/] [v2_0/] [tests/] [smp.cxx] - Rev 665
Go to most recent revision | Compare with Previous | Blame | View Log
//========================================================================== // // smp.cxx // // SMP tests // //========================================================================== //####ECOSGPLCOPYRIGHTBEGIN#### // ------------------------------------------- // This file is part of eCos, the Embedded Configurable Operating System. // Copyright (C) 1998, 1999, 2000, 2001, 2002 Red Hat, Inc. // // eCos 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 2 or (at your option) any later version. // // eCos 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 eCos; if not, write to the Free Software Foundation, Inc., // 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. // // As a special exception, if other files instantiate templates or use macros // or inline functions from this file, or you compile this file and link it // with other works to produce a work based on this file, this file does not // by itself cause the resulting work to be covered by the GNU General Public // License. However the source code for this file must still be made available // in accordance with section (3) of the GNU General Public License. // // This exception does not invalidate any other reasons why a work based on // this file might be covered by the GNU General Public License. // // Alternative licenses for eCos may be arranged by contacting Red Hat, Inc. // at http://sources.redhat.com/ecos/ecos-license/ // ------------------------------------------- //####ECOSGPLCOPYRIGHTEND#### //========================================================================== //#####DESCRIPTIONBEGIN#### // // Author(s): nickg // Contributors: nickg // Date: 2001-06-18 // Description: Some basic SMP tests. // //####DESCRIPTIONEND#### //========================================================================== #include <pkgconf/kernel.h> #include <pkgconf/hal.h> #if 1 #include <cyg/kernel/sched.hxx> #include <cyg/kernel/thread.hxx> #include <cyg/kernel/thread.inl> #include <cyg/kernel/mutex.hxx> #include <cyg/kernel/sema.hxx> #include <cyg/kernel/sched.inl> #include <cyg/kernel/clock.hxx> #include <cyg/kernel/clock.inl> #endif #include <cyg/kernel/kapi.h> #include <cyg/infra/testcase.h> #include <cyg/infra/diag.h> //========================================================================== #if defined(CYGPKG_KERNEL_SMP_SUPPORT) && \ defined(CYGFUN_KERNEL_API_C) && \ defined(CYGSEM_KERNEL_SCHED_MLQUEUE) && \ defined(CYGVAR_KERNEL_COUNTERS_CLOCK) && \ !defined(CYGPKG_HAL_I386_LINUX) && \ !defined(CYGDBG_INFRA_DIAG_USE_DEVICE) && \ (CYGNUM_KERNEL_SCHED_PRIORITIES > 12) //========================================================================== #define NTHREADS 1 #include "testaux.hxx" #define STACK_SIZE CYGNUM_HAL_STACK_SIZE_TYPICAL #define NTHREADS_MAX (CYGNUM_KERNEL_CPU_MAX*3) static int ncpus = CYGNUM_KERNEL_CPU_MAX; static int nthread = NTHREADS_MAX; static char stacks[NTHREADS_MAX][STACK_SIZE]; static cyg_thread test_threads[NTHREADS_MAX]; static cyg_handle_t threads[NTHREADS_MAX]; static volatile cyg_uint32 cpu_run[CYGNUM_KERNEL_CPU_MAX]; static volatile int failed = false; static volatile cyg_uint32 cpu_thread[CYGNUM_KERNEL_CPU_MAX]; static volatile cyg_uint32 slicerun[NTHREADS_MAX][CYGNUM_KERNEL_CPU_MAX]; static cyg_mutex_t mx; //========================================================================== // Compute a name for a thread char * thread_name(char *basename, int indx) { return "<<NULL>>"; // Not currently used } //========================================================================== void test_thread_cpu(CYG_ADDRESS id) { for(;;) cpu_run[CYG_KERNEL_CPU_THIS()] = true; } //========================================================================== // First test: just run as many threads as CPUs and check that we // get to run on each CPU. void run_smp_test_cpus() { int i; CYG_TEST_INFO( "CPU Test: Check CPUs functional"); // Init flags. for (i = 0; i < ncpus; i++) cpu_run[i] = false; // Set my priority higher than any I plan to create cyg_thread_set_priority(cyg_thread_self(), 2); for (i = 0; i < ncpus; i++) { cyg_thread_create(10, // Priority - just a number test_thread_cpu, // entry i, // index thread_name("thread", i), // Name &stacks[i][0], // Stack STACK_SIZE, // Size &threads[i], // Handle &test_threads[i] // Thread data structure ); cyg_thread_resume( threads[i]); } // Just wait a while, until the threads have all run for a bit. cyg_thread_delay( 10 ); // Delete all the threads for (i = 0; i < ncpus; i++) { cyg_thread_delete(threads[i]); } // And check that a thread ran on each CPU for (i = 0; i < ncpus; i++) { // CYG_TEST_CHECK( cpu_run[i], "CPU didn't run"); if( !cpu_run[i] ) { CYG_TEST_INFO( "CPU didn't run" ); failed++; } } CYG_TEST_INFO( "CPU Test: done"); } //========================================================================== void test_thread_pri(CYG_ADDRESS id) { for(;;) { cpu_thread[CYG_KERNEL_CPU_THIS()] = id; } } //========================================================================== // Second test: Run a thread on each CPU and then by manipulating the // priorities, get the current thread to migrate to each CPU in turn. void run_smp_test_pri() { int i; CYG_TEST_INFO( "Pri Test: Check set_priority functionality"); // Init flags. for (i = 0; i < ncpus; i++) cpu_run[i] = false; // Set my priority higher than any I plan to creat cyg_thread_set_priority(cyg_thread_self(), 2); for (i = 0; i < ncpus; i++) { cyg_thread_create(10, // Priority - just a number test_thread_pri, // entry i, // index thread_name("thread", i), // Name &stacks[i][0], // Stack STACK_SIZE, // Size &threads[i], // Handle &test_threads[i] // Thread data structure ); cyg_thread_resume( threads[i]); } cyg_thread_delay( 2 ); cyg_handle_t cthread = threads[0]; cyg_thread_set_priority(cthread, 25); // Just wait a while, until the threads have all run for a bit. cyg_thread_delay( 2 ); for (i = 0; i < ncpus*500; i++) { HAL_SMP_CPU_TYPE cpu = i % CYG_KERNEL_CPU_COUNT(); if( cpu != CYG_KERNEL_CPU_THIS() ) { // At this point we have the current thread running on a // CPU at priority 2, ncpus-1 threads running at priority // 10 and the last thread (cthread) in the run queue at // priority 25. // Pick a thread on a different CPU cyg_handle_t dthread; do { dthread = threads[cpu_thread[cpu]]; } while( dthread == cthread ); // Change the priority of the victim thread to 20. It is // still higher priority than cthread so it will continue // running. cyg_thread_set_priority(dthread, 20); // Now change our priority to 15. We are still higher // priority that cthread so we will still run. cyg_thread_set_priority(cyg_thread_self(), 15); // Finally change the priority of cthread to 10. This will // cause it to preempt us on the current CPU. In turn we // will preempt dthread on its CPU. // NOTE: This relies somewhat on the SMP scheduler doing // what we expect here. Specifically, that it will preempt // the current thread with cthread locally. A more // sophisticated scheduler might decide that the most // efficient thing to do is to preempt dthread with // cthread remotely, leaving the current thread where it // is. If we ever bother to implement this, then this test // will need to change. cyg_thread_set_priority(cthread, 10); // Spin here a while until the scheduler sorts itself out. for( int j = 0; j < 100000; j++ ); // Indicate that we have run on this CPU cpu_run[CYG_KERNEL_CPU_THIS()]++; // Restore our priority to 2 and depress dthread to 25 and // make it the new cthread. cyg_thread_set_priority(cyg_thread_self(), 2); cyg_thread_set_priority(dthread, 25); cthread = dthread; } } // Delete all the threads for (i = 0; i < ncpus; i++) { cyg_thread_delete(threads[i]); } // And check that a thread ran on each CPU for (i = 0; i < ncpus; i++) { // CYG_TEST_CHECK( cpu_run[i], "CPU didn't run"); if( !cpu_run[i] ) { CYG_TEST_INFO( "CPU didn't run" ); failed++; } } CYG_TEST_INFO( "PRI Test: done"); } //========================================================================== void test_thread_timeslice(CYG_ADDRESS id) { for(;;) slicerun[id][CYG_KERNEL_CPU_THIS()]++; } //========================================================================== // First test: just run as many threads as CPUs and check that we // get to run on each CPU. void run_smp_test_timeslice() { int i; CYG_TEST_INFO( "Timeslice Test: Check timeslicing works"); // Init flags. for (i = 0; i < nthread; i++) for( int j = 0; j < ncpus; j++ ) slicerun[i][j] = 0; // Set my priority higher than any I plan to create cyg_thread_set_priority(cyg_thread_self(), 2); for (i = 0; i < nthread; i++) { cyg_thread_create(10, // Priority - just a number test_thread_timeslice, // entry i, // index thread_name("thread", i), // Name &stacks[i][0], // Stack STACK_SIZE, // Size &threads[i], // Handle &test_threads[i] // Thread data structure ); cyg_thread_resume( threads[i]); } // Just wait a while, until the threads have all run for a bit. cyg_thread_delay( 200 ); // Delete all the threads for (i = 0; i < nthread; i++) { cyg_thread_suspend(threads[i]); } // And check that a thread ran on each CPU cyg_uint32 cpu_total[ncpus]; cyg_uint32 cpu_threads[ncpus]; cyg_uint32 thread_total[nthread]; diag_printf(" Thread "); for( int j = 0; j < ncpus; j++ ) { cpu_total[j] = 0; cpu_threads[j] = 0; diag_printf(" CPU %2d",j); } diag_printf(" Total\n"); for (i = 0; i < nthread; i++) { thread_total[i] = 0; diag_printf(" %2d ",i); for( int j = 0; j < ncpus; j++ ) { thread_total[i] += slicerun[i][j]; cpu_total[j] += slicerun[i][j]; if( slicerun[i][j] > 0 ) cpu_threads[j]++; diag_printf(" %8d",slicerun[i][j]); } diag_printf("%8d\n",thread_total[i]); } diag_printf(" Total "); for( int j = 0; j < ncpus; j++ ) diag_printf(" %8d",cpu_total[j]); diag_printf("\n"); diag_printf("Threads "); for( int j = 0; j < ncpus; j++ ) { diag_printf(" %8d",cpu_threads[j]); if( cpu_threads[j] < 2 ) failed++; } diag_printf("\n"); // Delete all the threads for (i = 0; i < nthread; i++) { cyg_thread_delete(threads[i]); } CYG_TEST_INFO( "Timeslice Test: done"); } //========================================================================== void run_smp_tests(CYG_ADDRESS id) { cyg_mutex_init( &mx ); for( int i = 0; i < 100; i++ ) { run_smp_test_cpus(); run_smp_test_pri(); run_smp_test_timeslice(); } if( failed ) CYG_TEST_FAIL_FINISH("SMP tests failed\n"); CYG_TEST_PASS_FINISH("SMP tests OK"); } //========================================================================== void smp_main( void ) { CYG_TEST_INIT(); // Work out how many CPUs we actually have. ncpus = CYG_KERNEL_CPU_COUNT(); new_thread(run_smp_tests, 0); cyg_scheduler_start(); } //========================================================================== #ifdef CYGSEM_HAL_STOP_CONSTRUCTORS_ON_FLAG externC void cyg_hal_invoke_constructors(); #endif externC void cyg_start( void ) { #ifdef CYGSEM_HAL_STOP_CONSTRUCTORS_ON_FLAG cyg_hal_invoke_constructors(); #endif smp_main(); } //========================================================================== #else // CYGPKG_KERNEL_SMP_SUPPORT etc. externC void cyg_start( void ) { CYG_TEST_INIT(); CYG_TEST_NA("SMP test requires:\n" "CYGPKG_KERNEL_SMP_SUPPORT &&\n" "CYGFUN_KERNEL_API_C && \n" "CYGSEM_KERNEL_SCHED_MLQUEUE &&\n" "CYGVAR_KERNEL_COUNTERS_CLOCK &&\n" "!CYGPKG_HAL_I386_LINUX &&\n" "!CYGDBG_INFRA_DIAG_USE_DEVICE &&\n" "(CYGNUM_KERNEL_SCHED_PRIORITIES > 12)\n"); } #endif // CYGPKG_KERNEL_SMP_SUPPORT etc. //========================================================================== // EOF tm_basic.cxx
Go to most recent revision | Compare with Previous | Blame | View Log