* System V IPC for psim try #2
@ 2008-11-14 18:01 Joel Sherrill
2008-11-17 20:31 ` Doug Evans
0 siblings, 1 reply; 7+ messages in thread
From: Joel Sherrill @ 2008-11-14 18:01 UTC (permalink / raw)
To: gdb-patches
[-- Attachment #1: Type: text/plain, Size: 793 bytes --]
Hi,
I think I have taken Doug's comments into account
with this revised version of the patch. Hopefully
it is ready to commit now.
Comments appreciated. Thanks.
2008-11-14 Joel Sherrill <joel.sherrill@oarcorp.com>
* Makefile.in, configure.ac, debug.c, debug.h, hw_sem.c, hw_shm.c:
Add support for System V shared memory and semaphores to PSIM where
available on the host. Shared memory can be mapped into the
simulated processor's address space and protected using a semaphore.
* configure: Regenerated.
--
Joel Sherrill, Ph.D. Director of Research & Development
joel.sherrill@OARcorp.com On-Line Applications Research
Ask me about RTEMS: a free RTOS Huntsville AL 35805
Support Available (256) 722-9985
[-- Attachment #2: psim-sysv-20081114.diff --]
[-- Type: text/x-patch, Size: 19520 bytes --]
Index: Makefile.in
===================================================================
RCS file: /cvs/src/src/sim/ppc/Makefile.in,v
retrieving revision 1.24
diff -u -r1.24 Makefile.in
--- Makefile.in 11 Jul 2008 02:39:11 -0000 1.24
+++ Makefile.in 14 Nov 2008 15:50:56 -0000
@@ -843,6 +843,8 @@
hw_pal.o: hw_pal.c $(DEVICE_TABLE_H) $(CPU_H)
hw_phb.o: hw_phb.c $(DEVICE_TABLE_H) $(HW_PHB_H) $(COREFILE_H)
hw_register.o: hw_register.c $(DEVICE_TABLE_H) $(PSIM_H)
+hw_sem.o: hw_sem.c $(DEVICE_TABLE_H) $(PSIM_H)
+hw_shm.o: hw_shm.c $(DEVICE_TABLE_H) $(PSIM_H)
hw_trace.o: hw_trace.c $(DEVICE_TABLE_H)
hw_vm.o: hw_vm.c $(DEVICE_TABLE_H) $(CPU_H)
# ignore this line, it stops make from getting confused
Index: configure.ac
===================================================================
RCS file: /cvs/src/src/sim/ppc/configure.ac,v
retrieving revision 1.9
diff -u -r1.9 configure.ac
--- configure.ac 11 Jul 2008 02:39:11 -0000 1.9
+++ configure.ac 14 Nov 2008 15:50:56 -0000
@@ -209,10 +209,105 @@
esac
])dnl
+AC_CACHE_CHECK([if union semun defined],
+ ac_cv_HAS_UNION_SEMUN,
+ [AC_TRY_COMPILE([
+#include <sys/types.h>
+#include <sys/ipc.h>
+#include <sys/sem.h>],
+[union semun arg ;],
+[ac_cv_has_union_semun="yes"],
+[ac_cv_has_union_semun="no"])
+AC_MSG_RESULT($ac_cv_has_union_semun)
+])
+
+
+if test "$ac_cv_has_union_semun" = "yes"; then
+ AC_CACHE_CHECK(whether System V semaphores are supported,
+ ac_cv_sysv_sem,
+ [
+ AC_TRY_RUN(
+ [
+ #include <sys/types.h>
+ #include <sys/ipc.h>
+ #include <sys/sem.h>
+ int main () {
+ union semun arg ;
+
+ int id=semget(IPC_PRIVATE,1,IPC_CREAT|0400);
+ if (id == -1)
+ exit(1);
+ arg.val = 0; /* avoid implicit type cast to union */
+ if (semctl(id, 0, IPC_RMID, arg) == -1)
+ exit(1);
+ exit(0);
+ }
+ ],
+ ac_cv_sysv_sem="yes", ac_cv_sysv_sem="no", :)
+ ])
+else # semun is not defined
+ AC_CACHE_CHECK(whether System V semaphores are supported,
+ ac_cv_sysv_sem,
+ [
+ AC_TRY_RUN(
+ [
+ #include <sys/types.h>
+ #include <sys/ipc.h>
+ #include <sys/sem.h>
+ union semun {
+ int val;
+ struct semid_ds *buf;
+ ushort *array;
+ };
+ int main () {
+ union semun arg ;
+
+ int id=semget(IPC_PRIVATE,1,IPC_CREAT|0400);
+ if (id == -1)
+ exit(1);
+ arg.val = 0; /* avoid implicit type cast to union */
+ if (semctl(id, 0, IPC_RMID, arg) == -1)
+ exit(1);
+ exit(0);
+ }
+ ],
+ ac_cv_sysv_sem="yes", ac_cv_sysv_sem="no", :)
+ ])
+fi
+
+AC_CACHE_CHECK(whether System V shared memory is supported,
+ac_cv_sysv_shm,
+[
+AC_TRY_RUN([
+#include <sys/types.h>
+#include <sys/ipc.h>
+#include <sys/shm.h>
+int main () {
+ int id=shmget(IPC_PRIVATE,1,IPC_CREAT|0400);
+ if (id == -1)
+ exit(1);
+ if (shmctl(id, IPC_RMID, 0) == -1)
+ exit(1);
+ exit(0);
+}
+],
+ac_cv_sysv_shm="yes", ac_cv_sysv_shm="no", :)
+])
+
+if test x"$ac_cv_sysv_shm" = x"yes" -a x"$ac_cv_sysv_sem" = x"yes" ; then
+ sim_sysv_ipc_hw=",sem,shm";
+else
+ sim_sysv_ipc_hw="";
+fi
+
+if test x"$ac_cv_has_union_semun" = x"yes" -a x"$ac_cv_sysv_sem" = x"yes" ; then
+ sim_hwflags="-DHAS_UNION_SEMUN";
+fi
+
AC_ARG_ENABLE(sim-hardware,
[ --enable-sim-hardware=list Specify the hardware to be included in the build.],
-[hardware="cpu,memory,nvram,iobus,htab,disk,trace,register,vm,init,core,pal,com,eeprom,opic,glue,phb,ide"
+[hardware="cpu,memory,nvram,iobus,htab,disk,trace,register,vm,init,core,pal,com,eeprom,opic,glue,phb,ide${sim_sysv_ipc_hw}"
case "${enableval}" in
yes) ;;
no) AC_MSG_ERROR("List of hardware must be specified for --enable-sim-hardware"); hardware="";;
@@ -224,14 +319,13 @@
sim_hw_obj=`echo $sim_hw_src | sed -e 's/\.c/.o/g'`
if test x"$silent" != x"yes" && test x"$hardware" != x""; then
echo "Setting hardware to $sim_hw_src, $sim_hw_obj"
-fi],[hardware="cpu,memory,nvram,iobus,htab,disk,trace,register,vm,init,core,pal,com,eeprom,opic,glue,phb,ide"
+fi],[hardware="cpu,memory,nvram,iobus,htab,disk,trace,register,vm,init,core,pal,com,eeprom,opic,glue,phb,ide${sim_sysv_ipc_hw}"
sim_hw_src=`echo $hardware | sed -e 's/,/.c hw_/g' -e 's/^/hw_/' -e s'/$/.c/'`
sim_hw_obj=`echo $sim_hw_src | sed -e 's/\.c/.o/g'`
if test x"$silent" != x"yes"; then
echo "Setting hardware to $sim_hw_src, $sim_hw_obj"
fi])dnl
-
AC_ARG_ENABLE(sim-hostbitsize,
[ --enable-sim-hostbitsize=32|64 Specify host bitsize (32 or 64).],
[case "${enableval}" in
Index: debug.c
===================================================================
RCS file: /cvs/src/src/sim/ppc/debug.c,v
retrieving revision 1.1.1.1
diff -u -r1.1.1.1 debug.c
--- debug.c 16 Apr 1999 01:35:08 -0000 1.1.1.1
+++ debug.c 14 Nov 2008 15:50:56 -0000
@@ -28,6 +28,9 @@
#ifdef HAVE_STDLIB_H
#include <stdlib.h>
#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
int ppc_trace[nr_trace_options];
@@ -70,6 +73,8 @@
{ trace_pass_device, "pass-device" },
{ trace_phb_device, "phb-device" },
{ trace_register_device, "register-device", "Device initializing registers" },
+ { trace_sem_device, "sem-device" },
+ { trace_shm_device, "shm-device" },
{ trace_stack_device, "stack-device" },
{ trace_vm_device, "vm-device" },
/* packages */
Index: debug.h
===================================================================
RCS file: /cvs/src/src/sim/ppc/debug.h,v
retrieving revision 1.1.1.1
diff -u -r1.1.1.1 debug.h
--- debug.h 16 Apr 1999 01:35:08 -0000 1.1.1.1
+++ debug.h 14 Nov 2008 15:50:56 -0000
@@ -51,6 +51,8 @@
trace_pal_device,
trace_pass_device,
trace_phb_device,
+ trace_sem_device,
+ trace_shm_device,
trace_stack_device,
trace_register_device,
trace_vm_device,
Index: hw_sem.c
===================================================================
RCS file: hw_sem.c
diff -N hw_sem.c
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ hw_sem.c 14 Nov 2008 15:50:56 -0000
@@ -0,0 +1,289 @@
+/* This file is part of the program psim.
+
+ Copyright (C) 1997,2008, Joel Sherrill <joel@OARcorp.com>
+
+ This program 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 of the License, or
+ (at your option) any later version.
+
+ This program 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., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+ */
+
+
+#ifndef _HW_SEM_C_
+#define _HW_SEM_C_
+
+#include "device_table.h"
+
+#ifdef HAVE_STRING_H
+#include <string.h>
+#else
+#ifdef HAVE_STRINGS_H
+#include <strings.h>
+#endif
+#endif
+
+#include <sys/ipc.h>
+#include <sys/sem.h>
+
+#include <errno.h>
+
+/* DEVICE
+
+
+ sem - provide access to a unix semaphore
+
+
+ DESCRIPTION
+
+
+ This device implements an interface to a unix semaphore.
+
+
+ PROPERTIES
+
+
+ reg = <address> <size> (required)
+
+ Determine where the memory lives in the parents address space.
+
+ key = <integer> (required)
+
+ This is the key of the unix semaphore.
+
+ EXAMPLES
+
+
+ Enable tracing of the sem:
+
+ | bash$ psim -t sem-device \
+
+
+ Configure a UNIX semaphore using key 0x12345678 mapped into psim
+ address space at 0xfff00000:
+
+ | -o '/sem@0xfff00000/reg 0xfff00000 0x80000' \
+ | -o '/sem@0xfff00000/key 0x12345678' \
+
+ sim/ppc/run -o '/#address-cells 1' \
+ -o '/sem@0xfff00000/reg 0xfff00000 12' \
+ -o '/sem@0xfff00000/key 0x12345678' ../psim-hello/hello
+
+ REGISTERS
+
+ offset 0 - lock count
+ offset 4 - lock operation
+ offset 8 - unlock operation
+
+ All reads return the current or resulting count.
+
+ BUGS
+
+ None known.
+
+ */
+
+typedef struct _hw_sem_device {
+ unsigned_word physical_address;
+ key_t key;
+ int id;
+ int initial;
+ int count;
+} hw_sem_device;
+
+#if !HAS_UNION_SEMUN
+union semun {
+ int val;
+ struct semid_ds *buf;
+ unsigned short int *array;
+#if defined(__linux__)
+ struct seminfo *__buf;
+#endif
+};
+#endif
+
+static void
+hw_sem_init_data(device *me)
+{
+ hw_sem_device *sem = (hw_sem_device*)device_data(me);
+ const device_unit *d;
+ int status;
+ union semun help;
+
+ /* initialize the properties of the sem */
+
+ if (device_find_property(me, "key") == NULL)
+ error("sem_init_data() required key property is missing\n");
+
+ if (device_find_property(me, "value") == NULL)
+ error("sem_init_data() required value property is missing\n");
+
+ sem->key = (key_t) device_find_integer_property(me, "key");
+ DTRACE(sem, ("semaphore key (%d)\n", sem->key) );
+
+ sem->initial = (int) device_find_integer_property(me, "value");
+ DTRACE(sem, ("semaphore initial value (%d)\n", sem->initial) );
+
+ d = device_unit_address(me);
+ sem->physical_address = d->cells[ d->nr_cells-1 ];
+ DTRACE(sem, ("semaphore physical_address=0x%x\n", sem->physical_address));
+
+ /* Now to initialize the semaphore */
+
+ if ( sem->initial != -1 ) {
+
+ sem->id = semget(sem->key, 1, IPC_CREAT | 0660);
+ if (sem->id == -1)
+ error("hw_sem_init_data() semget failed\n");
+
+ help.val = sem->initial;
+ status = semctl( sem->id, 0, SETVAL, help );
+ if (status == -1)
+ error("hw_sem_init_data() semctl -- set value failed\n");
+
+ } else {
+ sem->id = semget(sem->key, 1, 0660);
+ if (sem->id == -1)
+ error("hw_sem_init_data() semget failed\n");
+ }
+
+ sem->count = semctl( sem->id, 0, GETVAL, help );
+ if (sem->count == -1)
+ error("hw_sem_init_data() semctl -- get value failed\n");
+ DTRACE(sem, ("semaphore OS value (%d)\n", sem->count) );
+}
+
+static void
+hw_sem_attach_address_callback(device *me,
+ attach_type attach,
+ int space,
+ unsigned_word addr,
+ unsigned nr_bytes,
+ access_type access,
+ device *client) /*callback/default*/
+{
+ hw_sem_device *sem = (hw_sem_device*)device_data(me);
+
+ if (space != 0)
+ error("sem_attach_address_callback() invalid address space\n");
+
+ if (nr_bytes == 12)
+ error("sem_attach_address_callback() invalid size\n");
+
+ sem->physical_address = addr;
+ DTRACE(sem, ("semaphore physical_address=0x%x\n", addr));
+}
+
+static unsigned
+hw_sem_io_read_buffer(device *me,
+ void *dest,
+ int space,
+ unsigned_word addr,
+ unsigned nr_bytes,
+ cpu *processor,
+ unsigned_word cia)
+{
+ hw_sem_device *sem = (hw_sem_device*)device_data(me);
+ struct sembuf sb;
+ int status;
+ unsigned32 u32;
+ union semun help;
+
+ /* do we need to worry about out of range addresses? */
+
+ DTRACE(sem, ("semaphore read addr=0x%x length=%d\n", addr, nr_bytes));
+
+ if (!(addr >= sem->physical_address && addr <= sem->physical_address + 11))
+ error("hw_sem_io_read_buffer() invalid address - out of range\n");
+
+ if ((addr % 4) != 0)
+ error("hw_sem_io_read_buffer() invalid address - alignment\n");
+
+ if (nr_bytes != 4)
+ error("hw_sem_io_read_buffer() invalid length\n");
+
+ switch ( (addr - sem->physical_address) / 4 ) {
+
+ case 0: /* OBTAIN CURRENT VALUE */
+ break;
+
+ case 1: /* LOCK */
+ sb.sem_num = 0;
+ sb.sem_op = -1;
+ sb.sem_flg = 0;
+
+ status = semop(sem->id, &sb, 1);
+ if (status == -1) {
+ perror( "hw_sem.c: lock" );
+ error("hw_sem_io_read_buffer() sem lock\n");
+ }
+
+ DTRACE(sem, ("semaphore lock %d\n", sem->count));
+ break;
+
+ case 2: /* UNLOCK */
+ sb.sem_num = 0;
+ sb.sem_op = 1;
+ sb.sem_flg = 0;
+
+ status = semop(sem->id, &sb, 1);
+ if (status == -1) {
+ perror( "hw_sem.c: unlock" );
+ error("hw_sem_io_read_buffer() sem unlock\n");
+ }
+ DTRACE(sem, ("semaphore unlock %d\n", sem->count));
+ break;
+
+ default:
+ error("hw_sem_io_read_buffer() invalid address - unknown error\n");
+ break;
+ }
+
+ /* assume target is big endian */
+ u32 = H2T_4(semctl( sem->id, 0, GETVAL, help ));
+
+ DTRACE(sem, ("semaphore OS value (%d)\n", u32) );
+ if (u32 == 0xffffffff) {
+ perror( "hw_sem.c: getval" );
+ error("hw_sem_io_read_buffer() semctl -- get value failed\n");
+ }
+
+ memcpy(dest, &u32, nr_bytes);
+ return nr_bytes;
+
+}
+
+static device_callbacks const hw_sem_callbacks = {
+ { generic_device_init_address, hw_sem_init_data },
+ { hw_sem_attach_address_callback, }, /* address */
+ { hw_sem_io_read_buffer, NULL }, /* IO */
+ { NULL, }, /* DMA */
+ { NULL, }, /* interrupt */
+ { NULL, }, /* unit */
+ NULL,
+};
+
+static void *
+hw_sem_create(const char *name,
+ const device_unit *unit_address,
+ const char *args)
+{
+ hw_sem_device *sem = ZALLOC(hw_sem_device);
+ return sem;
+}
+
+const device_descriptor hw_sem_device_descriptor[] = {
+ { "sem", hw_sem_create, &hw_sem_callbacks },
+ { NULL },
+};
+
+#endif /* _HW_SEM_C_ */
Index: hw_shm.c
===================================================================
RCS file: hw_shm.c
diff -N hw_shm.c
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ hw_shm.c 14 Nov 2008 15:50:56 -0000
@@ -0,0 +1,236 @@
+/* This file is part of the program psim.
+
+ Copyright (C) 1997,2008, Joel Sherrill <joel@OARcorp.com>
+
+ This program 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 of the License, or
+ (at your option) any later version.
+
+ This program 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., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+ */
+
+
+#ifndef _HW_SHM_C_
+#define _HW_SHM_C_
+
+#include "device_table.h"
+
+#ifdef HAVE_STRING_H
+#include <string.h>
+#else
+#ifdef HAVE_STRINGS_H
+#include <strings.h>
+#endif
+#endif
+
+#include <sys/ipc.h>
+#include <sys/shm.h>
+
+
+/* DEVICE
+
+
+ shm - map unix shared memory into psim address space
+
+
+ DESCRIPTION
+
+
+ This device implements an area of memory which is mapped into UNIX
+ shared memory.
+
+
+ PROPERTIES
+
+
+ reg = <address> <size> (required)
+
+ Determine where the memory lives in the parents address space.
+ The SHM area is assumed to be of the same length.
+
+ key = <integer> (required)
+
+ This is the key of the unix shared memory area.
+
+ EXAMPLES
+
+
+ Enable tracing of the shm:
+
+ | bash$ psim -t shm-device \
+
+
+ Configure a 512 kilobytes of UNIX shared memory with the key 0x12345678
+ mapped into psim address space at 0x0c000000.
+
+ | -o '/shm@0x0c000000/reg 0x0c000000 0x80000' \
+ | -o '/shm@0x0c000000/key 0x12345678' \
+
+ sim/ppc/run -o '/#address-cells 1' \
+ -o '/shm@0x0c000000/reg 0x0c000000 0x80000' \
+ -o '/shm@0x0c000000/key 0x12345678' ../psim-hello/hello
+
+ BUGS
+
+ None known.
+
+ */
+
+typedef struct _hw_shm_device {
+ unsigned_word physical_address;
+ char *shm_address;
+ unsigned sizeof_memory;
+ key_t key;
+ int id;
+} hw_shm_device;
+
+static void
+hw_shm_init_data(device *me)
+{
+ hw_shm_device *shm = (hw_shm_device*)device_data(me);
+ const device_unit *d;
+ reg_property_spec reg;
+ int i;
+
+ /* Obtain the Key Value */
+ if (device_find_property(me, "key") == NULL)
+ error("shm_init_data() required key property is missing\n");
+
+ shm->key = (key_t) device_find_integer_property(me, "key");
+ DTRACE(shm, ("shm key (0x%08x)\n", shm->key) );
+
+ /* Figure out where this memory is in address space and how long it is */
+ if ( !device_find_reg_array_property(me, "reg", 0, ®) )
+ error("hw_shm_init_data() no address registered\n");
+
+ /* Determine the address and length being as paranoid as possible */
+ shm->physical_address = 0xffffffff;
+ shm->sizeof_memory = 0xffffffff;
+
+ for ( i=0 ; i<reg.address.nr_cells; i++ ) {
+ if (reg.address.cells[0] == 0 && reg.size.cells[0] == 0)
+ continue;
+
+ if ( shm->physical_address != 0xffffffff )
+ device_error(me, "Only single celled address ranges supported\n");
+
+ shm->physical_address = reg.address.cells[i];
+ DTRACE(shm, ("shm physical_address=0x%x\n", shm->physical_address));
+
+ shm->sizeof_memory = reg.size.cells[i];
+ DTRACE(shm, ("shm length=0x%x\n", shm->sizeof_memory));
+ }
+
+ if ( shm->physical_address == 0xffffffff )
+ device_error(me, "Address not specified\n" );
+
+ if ( shm->sizeof_memory == 0xffffffff )
+ device_error(me, "Length not specified\n" );
+
+ /* Now actually attach to or create the shared memory area */
+ shm->id = shmget(shm->key, shm->sizeof_memory, IPC_CREAT | 0660);
+ if (shm->id == -1)
+ error("hw_shm_init_data() shmget failed\n");
+
+ shm->shm_address = shmat(shm->id, (char *)0, SHM_RND);
+ if (shm->shm_address == (void *)-1)
+ error("hw_shm_init_data() shmat failed\n");
+}
+
+static void
+hw_shm_attach_address_callback(device *me,
+ attach_type attach,
+ int space,
+ unsigned_word addr,
+ unsigned nr_bytes,
+ access_type access,
+ device *client) /*callback/default*/
+{
+ hw_shm_device *shm = (hw_shm_device*)device_data(me);
+
+ if (space != 0)
+ error("shm_attach_address_callback() invalid address space\n");
+
+ if (nr_bytes == 0)
+ error("shm_attach_address_callback() invalid size\n");
+}
+
+
+static unsigned
+hw_shm_io_read_buffer(device *me,
+ void *dest,
+ int space,
+ unsigned_word addr,
+ unsigned nr_bytes,
+ cpu *processor,
+ unsigned_word cia)
+{
+ hw_shm_device *shm = (hw_shm_device*)device_data(me);
+
+ /* do we need to worry about out of range addresses? */
+
+ DTRACE(shm, ("read %p %x %x %x\n", \
+ shm->shm_address, shm->physical_address, addr, nr_bytes) );
+
+ memcpy(dest, &shm->shm_address[addr - shm->physical_address], nr_bytes);
+ return nr_bytes;
+}
+
+
+static unsigned
+hw_shm_io_write_buffer(device *me,
+ const void *source,
+ int space,
+ unsigned_word addr,
+ unsigned nr_bytes,
+ cpu *processor,
+ unsigned_word cia)
+{
+ hw_shm_device *shm = (hw_shm_device*)device_data(me);
+
+ /* do we need to worry about out of range addresses? */
+
+ DTRACE(shm, ("write %p %x %x %x\n", \
+ shm->shm_address, shm->physical_address, addr, nr_bytes) );
+
+ memcpy(&shm->shm_address[addr - shm->physical_address], source, nr_bytes);
+ return nr_bytes;
+}
+
+static device_callbacks const hw_shm_callbacks = {
+ { generic_device_init_address, hw_shm_init_data },
+ { hw_shm_attach_address_callback, }, /* address */
+ { hw_shm_io_read_buffer,
+ hw_shm_io_write_buffer }, /* IO */
+ { NULL, }, /* DMA */
+ { NULL, }, /* interrupt */
+ { NULL, }, /* unit */
+ NULL,
+};
+
+static void *
+hw_shm_create(const char *name,
+ const device_unit *unit_address,
+ const char *args)
+{
+ hw_shm_device *shm = ZALLOC(hw_shm_device);
+ return shm;
+}
+
+
+
+const device_descriptor hw_shm_device_descriptor[] = {
+ { "shm", hw_shm_create, &hw_shm_callbacks },
+ { NULL },
+};
+
+#endif /* _HW_SHM_C_ */
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: System V IPC for psim try #2
2008-11-14 18:01 System V IPC for psim try #2 Joel Sherrill
@ 2008-11-17 20:31 ` Doug Evans
2008-11-18 21:23 ` Joel Sherrill
0 siblings, 1 reply; 7+ messages in thread
From: Doug Evans @ 2008-11-17 20:31 UTC (permalink / raw)
To: Joel Sherrill; +Cc: gdb-patches
On Fri, Nov 14, 2008 at 8:03 AM, Joel Sherrill
<joel.sherrill@oarcorp.com> wrote:
> Hi,
>
> I think I have taken Doug's comments into account
> with this revised version of the patch. Hopefully
> it is ready to commit now.
>
> Comments appreciated. Thanks.
>
> 2008-11-14 Joel Sherrill <joel.sherrill@oarcorp.com>
>
> * Makefile.in, configure.ac, debug.c, debug.h, hw_sem.c, hw_shm.c:
> Add support for System V shared memory and semaphores to PSIM where
> available on the host. Shared memory can be mapped into the
> simulated processor's address space and protected using a semaphore.
> * configure: Regenerated.
Hi. Thanks for the making the changes.
nit: I think the convention is one *-line per file, though I see
sim/ppc/ChangeLog doesn't follow that rule precisely. I like your
original changelog entry better (fwiw).
2008-09-08 Joel Sherrill <joel.sherrill@oarcorp.com>
* configure: Regenerated.
* configure.ac: Add test for System V shared memory and semaphore.
* debug.c, debug.h: Add trace support for new devices.
* hw_sem.c, hw_shm.c: New files.
* Makefile.in: Add hw_sem.c and hw_shm.c.
Regarding:
+if test "$ac_cv_has_union_semun" = "yes"; then
+ AC_CACHE_CHECK(whether System V semaphores are supported,
+ ac_cv_sysv_sem,
+ [
+ AC_TRY_RUN(
+ [
Apologies for not noting this before. I see ppc/configure.ac already
uses AC_TRY_RUN, so it's ok with me. The problem with AC_TRY_RUN is
that it doesn't work in cross compilation environments. In this case
all that will happen is that the sim won't support these extra
features so no big deal (And apologies if all of this has already gone
through your mind ... :-))
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: System V IPC for psim try #2
2008-11-17 20:31 ` Doug Evans
@ 2008-11-18 21:23 ` Joel Sherrill
2008-11-18 21:32 ` Doug Evans
0 siblings, 1 reply; 7+ messages in thread
From: Joel Sherrill @ 2008-11-18 21:23 UTC (permalink / raw)
To: Doug Evans; +Cc: gdb-patches
Doug Evans wrote:
> On Fri, Nov 14, 2008 at 8:03 AM, Joel Sherrill
> <joel.sherrill@oarcorp.com> wrote:
>
>> Hi,
>>
>> I think I have taken Doug's comments into account
>> with this revised version of the patch. Hopefully
>> it is ready to commit now.
>>
>> Comments appreciated. Thanks.
>>
>> 2008-11-14 Joel Sherrill <joel.sherrill@oarcorp.com>
>>
>> * Makefile.in, configure.ac, debug.c, debug.h, hw_sem.c, hw_shm.c:
>> Add support for System V shared memory and semaphores to PSIM where
>> available on the host. Shared memory can be mapped into the
>> simulated processor's address space and protected using a semaphore.
>> * configure: Regenerated.
>>
>
> Hi. Thanks for the making the changes.
>
Thank you for reviewing it. Is is OK to commit now
with the ChangeLog you preferred?
> nit: I think the convention is one *-line per file, though I see
> sim/ppc/ChangeLog doesn't follow that rule precisely. I like your
> original changelog entry better (fwiw).
>
> 2008-09-08 Joel Sherrill <joel.sherrill@oarcorp.com>
>
> * configure: Regenerated.
> * configure.ac: Add test for System V shared memory and semaphore.
> * debug.c, debug.h: Add trace support for new devices.
> * hw_sem.c, hw_shm.c: New files.
> * Makefile.in: Add hw_sem.c and hw_shm.c.
>
>
That is a more informative log.
> Regarding:
>
> +if test "$ac_cv_has_union_semun" = "yes"; then
> + AC_CACHE_CHECK(whether System V semaphores are supported,
> + ac_cv_sysv_sem,
> + [
> + AC_TRY_RUN(
> + [
>
> Apologies for not noting this before. I see ppc/configure.ac already
> uses AC_TRY_RUN, so it's ok with me. The problem with AC_TRY_RUN is
> that it doesn't work in cross compilation environments. In this case
> all that will happen is that the sim won't support these extra
> features so no big deal (And apologies if all of this has already gone
> through your mind ... :-))
>
No it hadn't gone through my mind. I just assumed
that those who had come before me had good reasons to
use it.
--joel
--
Joel Sherrill, Ph.D. Director of Research & Development
joel.sherrill@OARcorp.com On-Line Applications Research
Ask me about RTEMS: a free RTOS Huntsville AL 35805
Support Available (256) 722-9985
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: System V IPC for psim try #2
2008-11-18 21:23 ` Joel Sherrill
@ 2008-11-18 21:32 ` Doug Evans
2008-11-18 21:32 ` Joel Sherrill
0 siblings, 1 reply; 7+ messages in thread
From: Doug Evans @ 2008-11-18 21:32 UTC (permalink / raw)
To: Joel Sherrill; +Cc: gdb-patches
On Tue, Nov 18, 2008 at 8:06 AM, Joel Sherrill
<joel.sherrill@oarcorp.com> wrote:
> Thank you for reviewing it. Is is OK to commit now
> with the ChangeLog you preferred?
>>
>> nit: I think the convention is one *-line per file, though I see
>> sim/ppc/ChangeLog doesn't follow that rule precisely. I like your
>> original changelog entry better (fwiw).
>>
>> 2008-09-08 Joel Sherrill <joel.sherrill@oarcorp.com>
>>
>> * configure: Regenerated.
>> * configure.ac: Add test for System V shared memory and semaphore.
>> * debug.c, debug.h: Add trace support for new devices.
>> * hw_sem.c, hw_shm.c: New files.
>> * Makefile.in: Add hw_sem.c and hw_shm.c.
Sure!
[modulo any concerns that I don't have official approval approval. :-)]
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: System V IPC for psim try #2
2008-11-18 21:32 ` Doug Evans
@ 2008-11-18 21:32 ` Joel Sherrill
2008-11-18 21:44 ` Daniel Jacobowitz
0 siblings, 1 reply; 7+ messages in thread
From: Joel Sherrill @ 2008-11-18 21:32 UTC (permalink / raw)
To: Doug Evans; +Cc: gdb-patches
Doug Evans wrote:
> On Tue, Nov 18, 2008 at 8:06 AM, Joel Sherrill
> <joel.sherrill@oarcorp.com> wrote:
>
>> Thank you for reviewing it. Is is OK to commit now
>> with the ChangeLog you preferred?
>>
>>> nit: I think the convention is one *-line per file, though I see
>>> sim/ppc/ChangeLog doesn't follow that rule precisely. I like your
>>> original changelog entry better (fwiw).
>>>
>>> 2008-09-08 Joel Sherrill <joel.sherrill@oarcorp.com>
>>>
>>> * configure: Regenerated.
>>> * configure.ac: Add test for System V shared memory and semaphore.
>>> * debug.c, debug.h: Add trace support for new devices.
>>> * hw_sem.c, hw_shm.c: New files.
>>> * Makefile.in: Add hw_sem.c and hw_shm.c.
>>>
>
> Sure!
> [modulo any concerns that I don't have official approval approval. :-)]
>
Well neither do I but I am ok with it and we actually use psim. :)
Can someone else confirm it is OK to commit now?
--joel
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: System V IPC for psim try #2
2008-11-18 21:32 ` Joel Sherrill
@ 2008-11-18 21:44 ` Daniel Jacobowitz
2008-11-19 12:41 ` Joel Sherrill
0 siblings, 1 reply; 7+ messages in thread
From: Daniel Jacobowitz @ 2008-11-18 21:44 UTC (permalink / raw)
To: Joel Sherrill; +Cc: Doug Evans, gdb-patches
On Tue, Nov 18, 2008 at 11:19:31AM -0600, Joel Sherrill wrote:
>> Sure!
>> [modulo any concerns that I don't have official approval approval. :-)]
>>
> Well neither do I but I am ok with it and we actually use psim. :)
>
> Can someone else confirm it is OK to commit now?
If Doug's happy with it, so am I. Go ahead.
--
Daniel Jacobowitz
CodeSourcery
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: System V IPC for psim try #2
2008-11-18 21:44 ` Daniel Jacobowitz
@ 2008-11-19 12:41 ` Joel Sherrill
0 siblings, 0 replies; 7+ messages in thread
From: Joel Sherrill @ 2008-11-19 12:41 UTC (permalink / raw)
To: Joel Sherrill, Doug Evans, gdb-patches
Daniel Jacobowitz wrote:
> On Tue, Nov 18, 2008 at 11:19:31AM -0600, Joel Sherrill wrote:
>
>>> Sure!
>>> [modulo any concerns that I don't have official approval approval. :-)]
>>>
>>>
>> Well neither do I but I am ok with it and we actually use psim. :)
>>
>> Can someone else confirm it is OK to commit now?
>>
>
> If Doug's happy with it, so am I. Go ahead.
>
>
Committed now. Hopefully all ok.
Thanks for the review guys.
--joel
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2008-11-18 21:32 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2008-11-14 18:01 System V IPC for psim try #2 Joel Sherrill
2008-11-17 20:31 ` Doug Evans
2008-11-18 21:23 ` Joel Sherrill
2008-11-18 21:32 ` Doug Evans
2008-11-18 21:32 ` Joel Sherrill
2008-11-18 21:44 ` Daniel Jacobowitz
2008-11-19 12:41 ` Joel Sherrill
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox