* [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2)
@ 2005-01-14 9:54 Orjan Friberg
2005-01-30 4:54 ` Daniel Jacobowitz
0 siblings, 1 reply; 12+ messages in thread
From: Orjan Friberg @ 2005-01-14 9:54 UTC (permalink / raw)
To: gdb-patches
As promised, here is part 2 which contains the actual gdbserver code for
CRIS/CRISv32.
Ok to commit?
2005-01-15 Orjan Friberg <orjanf@axis.com>
* linux-cris-low.c: New file with support for CRIS and CRISv32 targets.
* Makefile.in (SFILES): Add linux-cris-low.c.
(clean): Add reg-cris-c and reg-crisv32.c.
Add linux-cris-low.o, reg-cris.o, reg-cris.c, reg-crisv32.o, and
reg-crisv32.c to make rules.
* configure.srv: Add cris-*-linux* and crisv32-*-linux* to list of
recognized targets.
New file linux-cris-low.c:
/* GNU/Linux/CRIS specific low level interface, for the remote server for GDB.
Copyright 1995, 1996, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005
Free Software Foundation, Inc.
This file is part of GDB.
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. */
#include "server.h"
#include "linux-low.h"
#include <sys/ptrace.h>
#ifdef __arch_v32
/* CRISv32 */
#define cris_num_regs 49
/* Note: Ignoring USP (having the stack pointer in two locations causes trouble
without any significant gain). */
/* Locations need to match <include/asm/arch/ptrace.h>. */
static int cris_regmap[] = {
1*4, 2*4, 3*4, 4*4,
5*4, 6*4, 7*4, 8*4,
9*4, 10*4, 11*4, 12*4,
13*4, 14*4, 24*4, 15*4,
-1, -1, -1, 16*4,
-1, 22*4, 23*4, 17*4,
-1, -1, 21*4, 20*4,
-1, 19*4, -1, 18*4,
25*4,
26*4, -1, -1, 29*4,
30*4, 31*4, 32*4, 33*4,
34*4, 35*4, 36*4, 37*4,
38*4, 39*4, 40*4, -1
};
#else
/* CRISv10 */
#define cris_num_regs 32
/* Locations need to match <include/asm/arch/ptrace.h>. */
static int cris_regmap[] = {
15*4, 14*4, 13*4, 12*4,
11*4, 10*4, 9*4, 8*4,
7*4, 6*4, 5*4, 4*4,
3*4, 2*4, 23*4, 19*4,
-1, -1, -1, -1,
-1, 17*4, -1, 16*4,
-1, -1, -1, 18*4,
-1, 17*4, -1, -1
};
#endif /* __arch_v32 */
static int
cris_cannot_store_register (int regno)
{
if (cris_regmap[regno] == -1)
return 1;
return (regno >= cris_num_regs);
}
static int
cris_cannot_fetch_register (int regno)
{
if (cris_regmap[regno] == -1)
return 1;
return (regno >= cris_num_regs);
}
extern int debug_threads;
static CORE_ADDR
cris_get_pc ()
{
unsigned long pc;
collect_register_by_name ("pc", &pc);
if (debug_threads)
fprintf (stderr, "stop pc is %08lx\n", pc);
return pc;
}
static void
cris_set_pc (CORE_ADDR pc)
{
unsigned long newpc = pc;
supply_register_by_name ("pc", &newpc);
}
static const unsigned long cris_breakpoint = 0xe938;
#define cris_breakpoint_len 2
static int
cris_breakpoint_at (CORE_ADDR where)
{
unsigned long insn;
(*the_target->read_memory) (where, (char *) &insn, 4);
if (insn == cris_breakpoint)
return 1;
/* If necessary, recognize more trap instructions here. GDB only uses the
one. */
return 0;
}
/* We only place breakpoints in empty marker functions, and thread locking
is outside of the function. So rather than importing software single-step,
we can just run until exit. */
static CORE_ADDR
cris_reinsert_addr ()
{
unsigned long pc;
collect_register_by_name ("srp", &pc);
return pc;
}
#ifdef __arch_v32
void
cris_write_data_breakpoint (int bp, unsigned long start, unsigned long end)
{
switch (bp)
{
case 0:
supply_register_by_name("s3", &start);
supply_register_by_name("s4", &end);
break;
case 1:
supply_register_by_name("s5", &start);
supply_register_by_name("s6", &end);
break;
case 2:
supply_register_by_name("s7", &start);
supply_register_by_name("s8", &end);
break;
case 3:
supply_register_by_name("s9", &start);
supply_register_by_name("s10", &end);
break;
case 4:
supply_register_by_name("s11", &start);
supply_register_by_name("s12", &end);
break;
case 5:
supply_register_by_name("s13", &start);
supply_register_by_name("s14", &end);
break;
}
}
int
cris_insert_watchpoint(char type, CORE_ADDR addr, int len)
{
int bp;
unsigned long bp_ctrl;
unsigned long start, end;
unsigned long ccs;
/* Breakpoint/watchpoint types (GDB terminology):
0 = memory breakpoint for instructions
(not supported; done via memory write instead)
1 = hardware breakpoint for instructions (not supported)
2 = write watchpoint (supported)
3 = read watchpoint (supported)
4 = access watchpoint (supported). */
if (type < '2' || type > '4') {
/* Unsupported. */
return 1;
}
/* Read watchpoints are set as access watchpoints, because of GDB's
inability to deal with pure read watchpoints. */
if (type == '3')
type = '4';
/* Get the configuration register. */
collect_register_by_name("s0", &bp_ctrl);
/* The watchpoint allocation scheme is the simplest possible.
For example, if a region is watched for read and
a write watch is requested, a new watchpoint will
be used. Also, if a watch for a region that is already
covered by one or more existing watchpoints, a new
watchpoint will be used. */
/* First, find a free data watchpoint. */
for (bp = 0; bp < 6; bp++) {
/* Each data watchpoint's control registers occupy 2 bits
(hence the 3), starting at bit 2 for D0 (hence the 2)
with 4 bits between for each watchpoint (yes, the 4). */
if (!(bp_ctrl & (0x3 << (2 + (bp * 4))))) {
break;
}
}
if (bp > 5) {
/* We're out of watchpoints. */
return -1;
}
/* Configure the control register first. */
if (type == '3' || type == '4') {
/* Trigger on read. */
bp_ctrl |= (1 << (2 + bp * 4));
}
if (type == '2' || type == '4') {
/* Trigger on write. */
bp_ctrl |= (2 << (2 + bp * 4));
}
/* Setup the configuration register. */
supply_register_by_name("s0", &bp_ctrl);
/* Setup the range. */
start = addr;
end = addr + len - 1;
/* Configure the watchpoint register. */
cris_write_data_breakpoint (bp, start, end);
collect_register_by_name("ccs", &ccs);
/* Set the S1 flag to enable watchpoints. */
ccs |= (1 << 19);
supply_register_by_name("ccs", &ccs);
return 0;
}
int
cris_remove_watchpoint(char type, CORE_ADDR addr, int len)
{
int bp;
unsigned long bp_ctrl;
unsigned long start, end;
/* Breakpoint/watchpoint types:
0 = memory breakpoint for instructions
(not supported; done via memory write instead)
1 = hardware breakpoint for instructions (not supported)
2 = write watchpoint (supported)
3 = read watchpoint (supported)
4 = access watchpoint (supported). */
if (type < '2' || type > '4') {
return -1;
}
/* Read watchpoints are set as access watchpoints, because of GDB's
inability to deal with pure read watchpoints. */
if (type == '3')
type = '4';
/* Get the configuration register. */
collect_register_by_name("s0", &bp_ctrl);
/* Try to find a watchpoint that is configured for the
specified range, then check that read/write also matches. */
/* Ugly pointer arithmetic, since I cannot rely on a
single switch (addr) as there may be several watchpoints with
the same start address for example. */
unsigned long bp_d_regs[12];
/* Get all range registers to simplify search. */
collect_register_by_name("s3", &bp_d_regs[0]);
collect_register_by_name("s4", &bp_d_regs[1]);
collect_register_by_name("s5", &bp_d_regs[2]);
collect_register_by_name("s6", &bp_d_regs[3]);
collect_register_by_name("s7", &bp_d_regs[4]);
collect_register_by_name("s8", &bp_d_regs[5]);
collect_register_by_name("s9", &bp_d_regs[6]);
collect_register_by_name("s10", &bp_d_regs[7]);
collect_register_by_name("s11", &bp_d_regs[8]);
collect_register_by_name("s12", &bp_d_regs[9]);
collect_register_by_name("s13", &bp_d_regs[10]);
collect_register_by_name("s14", &bp_d_regs[11]);
for (bp = 0; bp < 6; bp++) {
if (bp_d_regs[bp * 2] == addr
&& bp_d_regs[bp * 2 + 1] == (addr + len - 1)) {
/* Matching range. */
int bitpos = 2 + bp * 4;
int rw_bits;
/* Read/write bits for this BP. */
rw_bits = (bp_ctrl & (0x3 << bitpos)) >> bitpos;
if ((type == '3' && rw_bits == 0x1)
|| (type == '2' && rw_bits == 0x2)
|| (type == '4' && rw_bits == 0x3)) {
/* Read/write matched. */
break;
}
}
}
if (bp > 5) {
/* No watchpoint matched. */
return -1;
}
/* Found a matching watchpoint. Now, deconfigure it by
both disabling read/write in bp_ctrl and zeroing its
start/end addresses. */
bp_ctrl &= ~(3 << (2 + (bp * 4)));
/* Setup the configuration register. */
supply_register_by_name("s0", &bp_ctrl);
start = end = 0;
/* Configure the watchpoint register. */
cris_write_data_breakpoint (bp, start, end);
/* Note that we don't clear the S1 flag here. It's done when continuing. */
return 0;
}
int
cris_stopped_by_watchpoint ()
{
unsigned long exs;
collect_register_by_name ("exs", &exs);
return (((exs & 0xff00) >> 8) == 0xc);
}
CORE_ADDR
cris_stopped_data_address ()
{
unsigned long eda;
collect_register_by_name ("eda", &eda);
/* FIXME: Possibly adjust to match watched range. */
return eda;
}
#endif /* __arch_v32 */
#ifdef __arch_v32
static void
cris_fill_gregset (void *buf)
{
int i;
for (i = 0; i < cris_num_regs; i++)
if (cris_regmap[i] != -1)
collect_register (i, ((char *) buf) + cris_regmap[i]);
}
static void
cris_store_gregset (const void *buf)
{
int i;
for (i = 0; i < cris_num_regs; i++)
if (cris_regmap[i] != -1)
supply_register (i, ((char *) buf) + cris_regmap[i]);
}
typedef unsigned long elf_gregset_t[cris_num_regs];
struct regset_info target_regsets[] = {
{ PTRACE_GETREGS, PTRACE_SETREGS, sizeof (elf_gregset_t),
GENERAL_REGS, cris_fill_gregset, cris_store_gregset },
{ 0, 0, -1, -1, NULL, NULL }
};
#endif /* __arch_v32 */
struct linux_target_ops the_low_target = {
cris_num_regs,
cris_regmap,
cris_cannot_fetch_register,
cris_cannot_store_register,
cris_get_pc,
cris_set_pc,
(const char *) &cris_breakpoint,
cris_breakpoint_len,
cris_reinsert_addr,
0,
cris_breakpoint_at,
#ifdef __arch_v32
cris_insert_watchpoint,
cris_remove_watchpoint,
cris_stopped_by_watchpoint,
cris_stopped_data_address,
#else
0,
0,
0,
0,
#endif /* __arch_v32 */
};
Index: Makefile.in
===================================================================
RCS file: /cvs/src/src/gdb/gdbserver/Makefile.in,v
retrieving revision 1.27
diff -u -p -r1.27 Makefile.in
--- Makefile.in 16 Oct 2004 16:18:54 -0000 1.27
+++ Makefile.in 13 Jan 2005 16:31:45 -0000
@@ -118,8 +118,8 @@ SFILES= $(srcdir)/gdbreplay.c $(srcdir)/
$(srcdir)/mem-break.c $(srcdir)/proc-service.c $(srcdir)/regcache.c \
$(srcdir)/remote-utils.c $(srcdir)/server.c $(srcdir)/target.c \
$(srcdir)/thread-db.c $(srcdir)/utils.c \
- $(srcdir)/linux-arm-low.c $(srcdir)/linux-i386-low.c \
- $(srcdir)/i387-fp.c \
+ $(srcdir)/linux-arm-low.c $(srcdir)/linux-cris-low.c \
+ $(srcdir)/linux-i386-low.c $(srcdir)/i387-fp.c \
$(srcdir)/linux-ia64-low.c $(srcdir)/linux-low.c \
$(srcdir)/linux-m68k-low.c $(srcdir)/linux-mips-low.c \
$(srcdir)/linux-ppc-low.c $(srcdir)/linux-s390-low.c \
@@ -200,6 +200,7 @@ clean:
rm -f gdbserver gdbreplay core make.log
rm -f reg-arm.c reg-i386.c reg-ia64.c reg-m68k.c reg-mips.c
rm -f reg-ppc.c reg-sh.c reg-x86-64.c reg-i386-linux.c
+ rm -f reg-cris.c reg-crisv32.c
maintainer-clean realclean distclean: clean
rm -f nm.h tm.h xm.h config.status config.h stamp-h config.log
@@ -261,6 +262,7 @@ linux-low.o: linux-low.c $(linux_low_h)
$(CC) -c $(CPPFLAGS) $(INTERNAL_CFLAGS) $< @USE_THREAD_DB@
linux-arm-low.o: linux-arm-low.c $(linux_low_h) $(server_h)
+linux-cris-low.o: linux-cris-low.c $(linux_low_h) $(server_h)
linux-i386-low.o: linux-i386-low.c $(linux_low_h) $(server_h)
linux-ia64-low.o: linux-ia64-low.c $(linux_low_h) $(server_h)
linux-mips-low.o: linux-mips-low.c $(linux_low_h) $(server_h)
@@ -272,6 +274,12 @@ linux-x86-64-low.o: linux-x86-64-low.c $
reg-arm.o : reg-arm.c $(regdef_h)
reg-arm.c : $(srcdir)/../regformats/reg-arm.dat $(regdat_sh)
sh $(regdat_sh) $(srcdir)/../regformats/reg-arm.dat reg-arm.c
+reg-cris.o : reg-cris.c $(regdef_h)
+reg-cris.c : $(srcdir)/../regformats/reg-cris.dat $(regdat_sh)
+ sh $(regdat_sh) $(srcdir)/../regformats/reg-cris.dat reg-cris.c
+reg-crisv32.o : reg-crisv32.c $(regdef_h)
+reg-crisv32.c : $(srcdir)/../regformats/reg-crisv32.dat $(regdat_sh)
+ sh $(regdat_sh) $(srcdir)/../regformats/reg-crisv32.dat reg-crisv32.c
reg-i386.o : reg-i386.c $(regdef_h)
reg-i386.c : $(srcdir)/../regformats/reg-i386.dat $(regdat_sh)
sh $(regdat_sh) $(srcdir)/../regformats/reg-i386.dat reg-i386.c
Index: configure.srv
===================================================================
RCS file: /cvs/src/src/gdb/gdbserver/configure.srv,v
retrieving revision 1.8
diff -u -p -r1.8 configure.srv
--- configure.srv 21 Nov 2004 03:09:39 -0000 1.8
+++ configure.srv 13 Jan 2005 16:31:45 -0000
@@ -23,6 +23,17 @@ case "${target}" in
srv_linux_usrregs=yes
srv_linux_thread_db=yes
;;
+ crisv32-*-linux*) srv_regobj=reg-crisv32.o
+ srv_tgtobj="linux-low.o linux-cris-low.o"
+ srv_linux_usrregs=yes
+ srv_linux_regsets=yes
+ srv_linux_thread_db=yes
+ ;;
+ cris-*-linux*) srv_regobj=reg-cris.o
+ srv_tgtobj="linux-low.o linux-cris-low.o"
+ srv_linux_usrregs=yes
+ srv_linux_thread_db=yes
+ ;;
i[34567]86-*-linux*) srv_regobj=reg-i386-linux.o
srv_tgtobj="linux-low.o linux-i386-low.o i387-fp.o"
srv_linux_usrregs=yes
--
Orjan Friberg
Axis Communications
^ permalink raw reply [flat|nested] 12+ messages in thread* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-01-14 9:54 [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) Orjan Friberg @ 2005-01-30 4:54 ` Daniel Jacobowitz 2005-02-01 14:31 ` Orjan Friberg 0 siblings, 1 reply; 12+ messages in thread From: Daniel Jacobowitz @ 2005-01-30 4:54 UTC (permalink / raw) To: Orjan Friberg; +Cc: gdb-patches On Fri, Jan 14, 2005 at 10:53:35AM +0100, Orjan Friberg wrote: > As promised, here is part 2 which contains the actual gdbserver code for > CRIS/CRISv32. > > Ok to commit? > > > 2005-01-15 Orjan Friberg <orjanf@axis.com> > > * linux-cris-low.c: New file with support for CRIS and CRISv32 > targets. > * Makefile.in (SFILES): Add linux-cris-low.c. > (clean): Add reg-cris-c and reg-crisv32.c. > Add linux-cris-low.o, reg-cris.o, reg-cris.c, reg-crisv32.o, and > reg-crisv32.c to make rules. > * configure.srv: Add cris-*-linux* and crisv32-*-linux* to list of > recognized targets. A couple of comments: - cris and crisv32 appear to have, from point of view of gdbserver, not much more in common than cris and mips. How about having linux-crisv32-low.c instead? There can be additional file for common code if it becomes necessary in the future. > /* We only place breakpoints in empty marker functions, and thread locking > is outside of the function. So rather than importing software > single-step, > we can just run until exit. */ > static CORE_ADDR > cris_reinsert_addr () > { > unsigned long pc; > collect_register_by_name ("srp", &pc); > return pc; > } I gather that you don't have PTRACE_SINGLESTEP on cris; not even on crisv32? (Ew. In fact, PTRACE_SINGLESTEP on cris will act as PTRACE_CONT in the kernel sources I'm looking at. That's nasty!) > > #ifdef __arch_v32 > void > cris_write_data_breakpoint (int bp, unsigned long start, unsigned long end) > { > switch (bp) > { > case 0: > supply_register_by_name("s3", &start); > supply_register_by_name("s4", &end); > break; What to do about the threaded case? GDB is having so much trouble getting it right that I'd like to make sure gdbserver does from the start. I guess this can be handled completely in the linux-low wrapper functions by switching the current thread and setting watchpoints for each. I think that is correct for all Linux targets using hardware watchpoints, because of the nature of Linux threading. -- Daniel Jacobowitz ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-01-30 4:54 ` Daniel Jacobowitz @ 2005-02-01 14:31 ` Orjan Friberg 2005-02-14 16:17 ` Orjan Friberg 2005-02-24 20:57 ` Daniel Jacobowitz 0 siblings, 2 replies; 12+ messages in thread From: Orjan Friberg @ 2005-02-01 14:31 UTC (permalink / raw) To: Daniel Jacobowitz; +Cc: gdb-patches [-- Attachment #1: Type: text/plain, Size: 2186 bytes --] Daniel Jacobowitz wrote: > > - cris and crisv32 appear to have, from point of view of gdbserver, not > much more in common than cris and mips. How about having > linux-crisv32-low.c instead? There can be additional file for common > code if it becomes necessary in the future. Good idea; I split them up as suggested. >>static CORE_ADDR >>cris_reinsert_addr () >>{ >> unsigned long pc; >> collect_register_by_name ("srp", &pc); >> return pc; >>} > > > I gather that you don't have PTRACE_SINGLESTEP on cris; not even on > crisv32? On CRISv32 I do; thanks for catching this (I removed reinsert_addr for CRISv32). >>#ifdef __arch_v32 >>void >>cris_write_data_breakpoint (int bp, unsigned long start, unsigned long end) >>{ >> switch (bp) >> { >> case 0: >> supply_register_by_name("s3", &start); >> supply_register_by_name("s4", &end); >> break; > > > What to do about the threaded case? GDB is having so much trouble > getting it right that I'd like to make sure gdbserver does from the > start. > > I guess this can be handled completely in the linux-low wrapper > functions by switching the current thread and setting watchpoints for > each. I think that is correct for all Linux targets using hardware > watchpoints, because of the nature of Linux threading. For CRISv32, there might be an additional problem lurking in the shadows: we treat the watchpoint registers as global (i.e., only one process at a time can use them, and thus they are not saved/restored on context switches etc). Anyway, is this something you want resolved before I commit the CRIS/CRISv32 support? (Here's an updated ChangeLog entry and updated patches if not.) 2005-02-01 Orjan Friberg <orjanf@axis.com> * linux-cris-low.c: New file with support for CRIS. * linux-crisv32-low.c: Ditto for CRISv32. * Makefile.in (SFILES): Add linux-cris-low.c, linux-crisv32-low.c. (clean): Add reg-cris.c and reg-crisv32.c. Add linux-cris-low.o, linux-crisv32-low.o, reg-cris.o, reg-cris.c, reg-crisv32.o, and reg-crisv32.c to make rules. * configure.srv: Add cris-*-linux* and crisv32-*-linux* to list of recognized targets. -- Orjan Friberg Axis Communications [-- Attachment #2: linux-cris-low.c --] [-- Type: text/x-csrc, Size: 2878 bytes --] /* GNU/Linux/CRIS specific low level interface, for the remote server for GDB. Copyright 1995, 1996, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc. This file is part of GDB. 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. */ #include "server.h" #include "linux-low.h" #include <sys/ptrace.h> /* CRISv10 */ #define cris_num_regs 32 /* Locations need to match <include/asm/arch/ptrace.h>. */ static int cris_regmap[] = { 15*4, 14*4, 13*4, 12*4, 11*4, 10*4, 9*4, 8*4, 7*4, 6*4, 5*4, 4*4, 3*4, 2*4, 23*4, 19*4, -1, -1, -1, -1, -1, 17*4, -1, 16*4, -1, -1, -1, 18*4, -1, 17*4, -1, -1 }; static int cris_cannot_store_register (int regno) { if (cris_regmap[regno] == -1) return 1; return (regno >= cris_num_regs); } static int cris_cannot_fetch_register (int regno) { if (cris_regmap[regno] == -1) return 1; return (regno >= cris_num_regs); } extern int debug_threads; static CORE_ADDR cris_get_pc () { unsigned long pc; collect_register_by_name ("pc", &pc); if (debug_threads) fprintf (stderr, "stop pc is %08lx\n", pc); return pc; } static void cris_set_pc (CORE_ADDR pc) { unsigned long newpc = pc; supply_register_by_name ("pc", &newpc); } static const unsigned long cris_breakpoint = 0xe938; #define cris_breakpoint_len 2 static int cris_breakpoint_at (CORE_ADDR where) { unsigned long insn; (*the_target->read_memory) (where, (char *) &insn, 4); if (insn == cris_breakpoint) return 1; /* If necessary, recognize more trap instructions here. GDB only uses the one. */ return 0; } /* We only place breakpoints in empty marker functions, and thread locking is outside of the function. So rather than importing software single-step, we can just run until exit. */ static CORE_ADDR cris_reinsert_addr () { unsigned long pc; collect_register_by_name ("srp", &pc); return pc; } struct linux_target_ops the_low_target = { cris_num_regs, cris_regmap, cris_cannot_fetch_register, cris_cannot_store_register, cris_get_pc, cris_set_pc, (const char *) &cris_breakpoint, cris_breakpoint_len, cris_reinsert_addr, 0, cris_breakpoint_at, 0, 0, 0, 0, }; [-- Attachment #3: linux-crisv32-low.c --] [-- Type: text/x-csrc, Size: 9614 bytes --] /* GNU/Linux/CRIS specific low level interface, for the remote server for GDB. Copyright 1995, 1996, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc. This file is part of GDB. 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. */ #include "server.h" #include "linux-low.h" #include <sys/ptrace.h> /* CRISv32 */ #define cris_num_regs 49 /* Note: Ignoring USP (having the stack pointer in two locations causes trouble without any significant gain). */ /* Locations need to match <include/asm/arch/ptrace.h>. */ static int cris_regmap[] = { 1*4, 2*4, 3*4, 4*4, 5*4, 6*4, 7*4, 8*4, 9*4, 10*4, 11*4, 12*4, 13*4, 14*4, 24*4, 15*4, -1, -1, -1, 16*4, -1, 22*4, 23*4, 17*4, -1, -1, 21*4, 20*4, -1, 19*4, -1, 18*4, 25*4, 26*4, -1, -1, 29*4, 30*4, 31*4, 32*4, 33*4, 34*4, 35*4, 36*4, 37*4, 38*4, 39*4, 40*4, -1 }; static int cris_cannot_store_register (int regno) { if (cris_regmap[regno] == -1) return 1; return (regno >= cris_num_regs); } static int cris_cannot_fetch_register (int regno) { if (cris_regmap[regno] == -1) return 1; return (regno >= cris_num_regs); } extern int debug_threads; static CORE_ADDR cris_get_pc () { unsigned long pc; collect_register_by_name ("pc", &pc); if (debug_threads) fprintf (stderr, "stop pc is %08lx\n", pc); return pc; } static void cris_set_pc (CORE_ADDR pc) { unsigned long newpc = pc; supply_register_by_name ("pc", &newpc); } static const unsigned long cris_breakpoint = 0xe938; #define cris_breakpoint_len 2 static int cris_breakpoint_at (CORE_ADDR where) { unsigned long insn; (*the_target->read_memory) (where, (char *) &insn, 4); if (insn == cris_breakpoint) return 1; /* If necessary, recognize more trap instructions here. GDB only uses the one. */ return 0; } void cris_write_data_breakpoint (int bp, unsigned long start, unsigned long end) { switch (bp) { case 0: supply_register_by_name("s3", &start); supply_register_by_name("s4", &end); break; case 1: supply_register_by_name("s5", &start); supply_register_by_name("s6", &end); break; case 2: supply_register_by_name("s7", &start); supply_register_by_name("s8", &end); break; case 3: supply_register_by_name("s9", &start); supply_register_by_name("s10", &end); break; case 4: supply_register_by_name("s11", &start); supply_register_by_name("s12", &end); break; case 5: supply_register_by_name("s13", &start); supply_register_by_name("s14", &end); break; } } int cris_insert_watchpoint(char type, CORE_ADDR addr, int len) { int bp; unsigned long bp_ctrl; unsigned long start, end; unsigned long ccs; /* Breakpoint/watchpoint types (GDB terminology): 0 = memory breakpoint for instructions (not supported; done via memory write instead) 1 = hardware breakpoint for instructions (not supported) 2 = write watchpoint (supported) 3 = read watchpoint (supported) 4 = access watchpoint (supported). */ if (type < '2' || type > '4') { /* Unsupported. */ return 1; } /* Read watchpoints are set as access watchpoints, because of GDB's inability to deal with pure read watchpoints. */ if (type == '3') type = '4'; /* Get the configuration register. */ collect_register_by_name("s0", &bp_ctrl); /* The watchpoint allocation scheme is the simplest possible. For example, if a region is watched for read and a write watch is requested, a new watchpoint will be used. Also, if a watch for a region that is already covered by one or more existing watchpoints, a new watchpoint will be used. */ /* First, find a free data watchpoint. */ for (bp = 0; bp < 6; bp++) { /* Each data watchpoint's control registers occupy 2 bits (hence the 3), starting at bit 2 for D0 (hence the 2) with 4 bits between for each watchpoint (yes, the 4). */ if (!(bp_ctrl & (0x3 << (2 + (bp * 4))))) { break; } } if (bp > 5) { /* We're out of watchpoints. */ return -1; } /* Configure the control register first. */ if (type == '3' || type == '4') { /* Trigger on read. */ bp_ctrl |= (1 << (2 + bp * 4)); } if (type == '2' || type == '4') { /* Trigger on write. */ bp_ctrl |= (2 << (2 + bp * 4)); } /* Setup the configuration register. */ supply_register_by_name("s0", &bp_ctrl); /* Setup the range. */ start = addr; end = addr + len - 1; /* Configure the watchpoint register. */ cris_write_data_breakpoint (bp, start, end); collect_register_by_name("ccs", &ccs); /* Set the S1 flag to enable watchpoints. */ ccs |= (1 << 19); supply_register_by_name("ccs", &ccs); return 0; } int cris_remove_watchpoint(char type, CORE_ADDR addr, int len) { int bp; unsigned long bp_ctrl; unsigned long start, end; /* Breakpoint/watchpoint types: 0 = memory breakpoint for instructions (not supported; done via memory write instead) 1 = hardware breakpoint for instructions (not supported) 2 = write watchpoint (supported) 3 = read watchpoint (supported) 4 = access watchpoint (supported). */ if (type < '2' || type > '4') { return -1; } /* Read watchpoints are set as access watchpoints, because of GDB's inability to deal with pure read watchpoints. */ if (type == '3') type = '4'; /* Get the configuration register. */ collect_register_by_name("s0", &bp_ctrl); /* Try to find a watchpoint that is configured for the specified range, then check that read/write also matches. */ /* Ugly pointer arithmetic, since I cannot rely on a single switch (addr) as there may be several watchpoints with the same start address for example. */ unsigned long bp_d_regs[12]; /* Get all range registers to simplify search. */ collect_register_by_name("s3", &bp_d_regs[0]); collect_register_by_name("s4", &bp_d_regs[1]); collect_register_by_name("s5", &bp_d_regs[2]); collect_register_by_name("s6", &bp_d_regs[3]); collect_register_by_name("s7", &bp_d_regs[4]); collect_register_by_name("s8", &bp_d_regs[5]); collect_register_by_name("s9", &bp_d_regs[6]); collect_register_by_name("s10", &bp_d_regs[7]); collect_register_by_name("s11", &bp_d_regs[8]); collect_register_by_name("s12", &bp_d_regs[9]); collect_register_by_name("s13", &bp_d_regs[10]); collect_register_by_name("s14", &bp_d_regs[11]); for (bp = 0; bp < 6; bp++) { if (bp_d_regs[bp * 2] == addr && bp_d_regs[bp * 2 + 1] == (addr + len - 1)) { /* Matching range. */ int bitpos = 2 + bp * 4; int rw_bits; /* Read/write bits for this BP. */ rw_bits = (bp_ctrl & (0x3 << bitpos)) >> bitpos; if ((type == '3' && rw_bits == 0x1) || (type == '2' && rw_bits == 0x2) || (type == '4' && rw_bits == 0x3)) { /* Read/write matched. */ break; } } } if (bp > 5) { /* No watchpoint matched. */ return -1; } /* Found a matching watchpoint. Now, deconfigure it by both disabling read/write in bp_ctrl and zeroing its start/end addresses. */ bp_ctrl &= ~(3 << (2 + (bp * 4))); /* Setup the configuration register. */ supply_register_by_name("s0", &bp_ctrl); start = end = 0; /* Configure the watchpoint register. */ cris_write_data_breakpoint (bp, start, end); /* Note that we don't clear the S1 flag here. It's done when continuing. */ return 0; } int cris_stopped_by_watchpoint () { unsigned long exs; collect_register_by_name ("exs", &exs); return (((exs & 0xff00) >> 8) == 0xc); } CORE_ADDR cris_stopped_data_address () { unsigned long eda; collect_register_by_name ("eda", &eda); /* FIXME: Possibly adjust to match watched range. */ return eda; } static void cris_fill_gregset (void *buf) { int i; for (i = 0; i < cris_num_regs; i++) if (cris_regmap[i] != -1) collect_register (i, ((char *) buf) + cris_regmap[i]); } static void cris_store_gregset (const void *buf) { int i; for (i = 0; i < cris_num_regs; i++) if (cris_regmap[i] != -1) supply_register (i, ((char *) buf) + cris_regmap[i]); } typedef unsigned long elf_gregset_t[cris_num_regs]; struct regset_info target_regsets[] = { { PTRACE_GETREGS, PTRACE_SETREGS, sizeof (elf_gregset_t), GENERAL_REGS, cris_fill_gregset, cris_store_gregset }, { 0, 0, -1, -1, NULL, NULL } }; struct linux_target_ops the_low_target = { cris_num_regs, cris_regmap, cris_cannot_fetch_register, cris_cannot_store_register, cris_get_pc, cris_set_pc, (const char *) &cris_breakpoint, cris_breakpoint_len, 0, 0, cris_breakpoint_at, cris_insert_watchpoint, cris_remove_watchpoint, cris_stopped_by_watchpoint, cris_stopped_data_address, }; [-- Attachment #4: patch --] [-- Type: text/plain, Size: 3271 bytes --] Index: Makefile.in =================================================================== RCS file: /cvs/src/src/gdb/gdbserver/Makefile.in,v retrieving revision 1.27 diff -u -p -r1.27 Makefile.in --- Makefile.in 16 Oct 2004 16:18:54 -0000 1.27 +++ Makefile.in 1 Feb 2005 14:27:15 -0000 @@ -118,7 +118,8 @@ SFILES= $(srcdir)/gdbreplay.c $(srcdir)/ $(srcdir)/mem-break.c $(srcdir)/proc-service.c $(srcdir)/regcache.c \ $(srcdir)/remote-utils.c $(srcdir)/server.c $(srcdir)/target.c \ $(srcdir)/thread-db.c $(srcdir)/utils.c \ - $(srcdir)/linux-arm-low.c $(srcdir)/linux-i386-low.c \ + $(srcdir)/linux-arm-low.c $(srcdir)/linux-cris-low.c \ + $(srcdir)/linux-crisv32-low.c $(srcdir)/linux-i386-low.c \ $(srcdir)/i387-fp.c \ $(srcdir)/linux-ia64-low.c $(srcdir)/linux-low.c \ $(srcdir)/linux-m68k-low.c $(srcdir)/linux-mips-low.c \ @@ -200,6 +201,7 @@ clean: rm -f gdbserver gdbreplay core make.log rm -f reg-arm.c reg-i386.c reg-ia64.c reg-m68k.c reg-mips.c rm -f reg-ppc.c reg-sh.c reg-x86-64.c reg-i386-linux.c + rm -f reg-cris.c reg-crisv32.c maintainer-clean realclean distclean: clean rm -f nm.h tm.h xm.h config.status config.h stamp-h config.log @@ -261,6 +263,8 @@ linux-low.o: linux-low.c $(linux_low_h) $(CC) -c $(CPPFLAGS) $(INTERNAL_CFLAGS) $< @USE_THREAD_DB@ linux-arm-low.o: linux-arm-low.c $(linux_low_h) $(server_h) +linux-cris-low.o: linux-cris-low.c $(linux_low_h) $(server_h) +linux-crisv32-low.o: linux-crisv32-low.c $(linux_low_h) $(server_h) linux-i386-low.o: linux-i386-low.c $(linux_low_h) $(server_h) linux-ia64-low.o: linux-ia64-low.c $(linux_low_h) $(server_h) linux-mips-low.o: linux-mips-low.c $(linux_low_h) $(server_h) @@ -272,6 +276,12 @@ linux-x86-64-low.o: linux-x86-64-low.c $ reg-arm.o : reg-arm.c $(regdef_h) reg-arm.c : $(srcdir)/../regformats/reg-arm.dat $(regdat_sh) sh $(regdat_sh) $(srcdir)/../regformats/reg-arm.dat reg-arm.c +reg-cris.o : reg-cris.c $(regdef_h) +reg-cris.c : $(srcdir)/../regformats/reg-cris.dat $(regdat_sh) + sh $(regdat_sh) $(srcdir)/../regformats/reg-cris.dat reg-cris.c +reg-crisv32.o : reg-crisv32.c $(regdef_h) +reg-crisv32.c : $(srcdir)/../regformats/reg-crisv32.dat $(regdat_sh) + sh $(regdat_sh) $(srcdir)/../regformats/reg-crisv32.dat reg-crisv32.c reg-i386.o : reg-i386.c $(regdef_h) reg-i386.c : $(srcdir)/../regformats/reg-i386.dat $(regdat_sh) sh $(regdat_sh) $(srcdir)/../regformats/reg-i386.dat reg-i386.c Index: configure.srv =================================================================== RCS file: /cvs/src/src/gdb/gdbserver/configure.srv,v retrieving revision 1.8 diff -u -p -r1.8 configure.srv --- configure.srv 21 Nov 2004 03:09:39 -0000 1.8 +++ configure.srv 1 Feb 2005 14:27:16 -0000 @@ -23,6 +23,17 @@ case "${target}" in srv_linux_usrregs=yes srv_linux_thread_db=yes ;; + crisv32-*-linux*) srv_regobj=reg-crisv32.o + srv_tgtobj="linux-low.o linux-crisv32-low.o" + srv_linux_usrregs=yes + srv_linux_regsets=yes + srv_linux_thread_db=yes + ;; + cris-*-linux*) srv_regobj=reg-cris.o + srv_tgtobj="linux-low.o linux-cris-low.o" + srv_linux_usrregs=yes + srv_linux_thread_db=yes + ;; i[34567]86-*-linux*) srv_regobj=reg-i386-linux.o srv_tgtobj="linux-low.o linux-i386-low.o i387-fp.o" srv_linux_usrregs=yes ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-02-01 14:31 ` Orjan Friberg @ 2005-02-14 16:17 ` Orjan Friberg 2005-02-24 20:57 ` Daniel Jacobowitz 1 sibling, 0 replies; 12+ messages in thread From: Orjan Friberg @ 2005-02-14 16:17 UTC (permalink / raw) To: Daniel Jacobowitz; +Cc: gdb-patches Orjan Friberg wrote: >> >> What to do about the threaded case? GDB is having so much trouble >> getting it right that I'd like to make sure gdbserver does from the >> start. >> >> I guess this can be handled completely in the linux-low wrapper >> functions by switching the current thread and setting watchpoints for >> each. I think that is correct for all Linux targets using hardware >> watchpoints, because of the nature of Linux threading. > > > For CRISv32, there might be an additional problem lurking in the > shadows: we treat the watchpoint registers as global (i.e., only one > process at a time can use them, and thus they are not saved/restored on > context switches etc). > > Anyway, is this something you want resolved before I commit the > CRIS/CRISv32 support? (Here's an updated ChangeLog entry and updated > patches if not.) Two week ping. (If you want the multi-threaded watchpoint support issue resolved before the commit please elaborate, if possible, on what needs to be done.) -- Orjan Friberg Axis Communications ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-02-01 14:31 ` Orjan Friberg 2005-02-14 16:17 ` Orjan Friberg @ 2005-02-24 20:57 ` Daniel Jacobowitz 2005-05-12 13:03 ` Orjan Friberg 1 sibling, 1 reply; 12+ messages in thread From: Daniel Jacobowitz @ 2005-02-24 20:57 UTC (permalink / raw) To: Orjan Friberg; +Cc: gdb-patches On Tue, Feb 01, 2005 at 03:30:12PM +0100, Orjan Friberg wrote: > For CRISv32, there might be an additional problem lurking in the shadows: > we treat the watchpoint registers as global (i.e., only one process at a > time can use them, and thus they are not saved/restored on context switches > etc). As I wrote in my other message, that's a pretty serious bug. > Anyway, is this something you want resolved before I commit the > CRIS/CRISv32 support? (Here's an updated ChangeLog entry and updated > patches if not.) Let's go ahead. The patch is mostly OK, but I have a last question or two... > 2005-02-01 Orjan Friberg <orjanf@axis.com> > > * linux-cris-low.c: New file with support for CRIS. > * linux-crisv32-low.c: Ditto for CRISv32. > * Makefile.in (SFILES): Add linux-cris-low.c, linux-crisv32-low.c. > (clean): Add reg-cris.c and reg-crisv32.c. > Add linux-cris-low.o, linux-crisv32-low.o, reg-cris.o, reg-cris.c, > reg-crisv32.o, and reg-crisv32.c to make rules. > * configure.srv: Add cris-*-linux* and crisv32-*-linux* to list of > recognized targets. > > -- > Orjan Friberg > Axis Communications > > static const unsigned long cris_breakpoint = 0xe938; > #define cris_breakpoint_len 2 > > static int > cris_breakpoint_at (CORE_ADDR where) > { > unsigned long insn; > > (*the_target->read_memory) (where, (char *) &insn, 4); > if (insn == cris_breakpoint) > return 1; > > /* If necessary, recognize more trap instructions here. GDB only uses the > one. */ > return 0; > } If the breakpoint length is two, why are you reading and testing four bytes? > void > cris_write_data_breakpoint (int bp, unsigned long start, unsigned long end) Please make new functions static where possible, which is pretty much everywhere. > int > cris_stopped_by_watchpoint () And use prototypes - that's not a prototype, because of the missing (void). > struct regset_info target_regsets[] = { > { PTRACE_GETREGS, PTRACE_SETREGS, sizeof (elf_gregset_t), > GENERAL_REGS, cris_fill_gregset, cris_store_gregset }, > { 0, 0, -1, -1, NULL, NULL } > }; Does the regset cover everything that that regmap covers, including debug registers? The current support for both PEEKUSER and regsets is purely for compatibility; it doesn't support getting some registers from one and some from another. So if all crisv32 kernels support PTRACE_GETREGS, your regmap is never getting used. I have code around here somewhere to support targets that need both, but it's gross. -- Daniel Jacobowitz CodeSourcery, LLC ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-02-24 20:57 ` Daniel Jacobowitz @ 2005-05-12 13:03 ` Orjan Friberg 2005-05-12 13:11 ` Orjan Friberg 2005-05-12 13:13 ` Daniel Jacobowitz 0 siblings, 2 replies; 12+ messages in thread From: Orjan Friberg @ 2005-05-12 13:03 UTC (permalink / raw) To: Daniel Jacobowitz; +Cc: gdb-patches [-- Attachment #1: Type: text/plain, Size: 2287 bytes --] Daniel Jacobowitz wrote: > > Let's go ahead. The patch is mostly OK, but I have a last question or > two... Thanks for your comments, and apologies for the long delay. (There are a couple of questions at the end also.) > If the breakpoint length is two, why are you reading and testing four > bytes? I really don't know why. Thanks for catching this. > Please make new functions static where possible, which is pretty much > everywhere. Taken care of. > And use prototypes - that's not a prototype, because of the missing > (void). Also fixed. >>struct regset_info target_regsets[] = { >> { PTRACE_GETREGS, PTRACE_SETREGS, sizeof (elf_gregset_t), >> GENERAL_REGS, cris_fill_gregset, cris_store_gregset }, >> { 0, 0, -1, -1, NULL, NULL } >>}; > > > Does the regset cover everything that that regmap covers, including > debug registers? The current support for both PEEKUSER and regsets is > purely for compatibility; it doesn't support getting some registers > from one and some from another. So if all crisv32 kernels support > PTRACE_GETREGS, your regmap is never getting used. You're right; I don't need to use PEEKUSER (configure.srv updated, cris_cannot_store_register/cris_cannot_fetch_register removed, and linux_target_ops updated). I kept the actual regmap though, since it's used in cris_fill_gregset and cris_store_gregset. Another thing: when removing cris_reinsert_addr, I got the same errors as described in http://sourceware.org/ml/gdb/2005-01/msg00071.html (i.e. "thread getmsg err: no event message for getmsg" and gdbserver getting a SIGSEGV). At the moment I have no idea what's going on, so as a temporary solution I reinstated cris_reinsert_addr but with a FIXME. For completeness: updated ChangeLog entry below and patch attached. 2005-05-12 Orjan Friberg <orjanf@axis.com> * linux-cris-low.c: New file with support for CRIS. * linux-crisv32-low.c: Ditto for CRISv32. * Makefile.in (SFILES): Add linux-cris-low.c, linux-crisv32-low.c. (clean): Add reg-cris.c and reg-crisv32.c. Add linux-cris-low.o, linux-crisv32-low.o, reg-cris.o, reg-cris.c, reg-crisv32.o, and reg-crisv32.c to make rules. * configure.srv: Add cris-*-linux* and crisv32-*-linux* to list of recognized targets. -- Orjan Friberg Axis Communications [-- Attachment #2: patch --] [-- Type: text/plain, Size: 3075 bytes --] Index: Makefile.in =================================================================== RCS file: /cvs/src/src/gdb/gdbserver/Makefile.in,v retrieving revision 1.29 diff -u -r1.29 Makefile.in --- Makefile.in 15 Apr 2005 00:24:51 -0000 1.29 +++ Makefile.in 12 May 2005 12:56:25 -0000 @@ -118,7 +118,8 @@ $(srcdir)/mem-break.c $(srcdir)/proc-service.c $(srcdir)/regcache.c \ $(srcdir)/remote-utils.c $(srcdir)/server.c $(srcdir)/target.c \ $(srcdir)/thread-db.c $(srcdir)/utils.c \ - $(srcdir)/linux-arm-low.c $(srcdir)/linux-i386-low.c \ + $(srcdir)/linux-arm-low.c $(srcdir)/linux-cris-low.c \ + $(srcdir)/linux-crisv32-low.c $(srcdir)/linux-i386-low.c \ $(srcdir)/i387-fp.c \ $(srcdir)/linux-ia64-low.c $(srcdir)/linux-low.c \ $(srcdir)/linux-m32r-low.c \ @@ -201,6 +202,7 @@ rm -f gdbserver gdbreplay core make.log rm -f reg-arm.c reg-i386.c reg-ia64.c reg-m32r.c reg-m68k.c reg-mips.c rm -f reg-ppc.c reg-sh.c reg-x86-64.c reg-i386-linux.c + rm -f reg-cris.c reg-crisv32.c maintainer-clean realclean distclean: clean rm -f nm.h tm.h xm.h config.status config.h stamp-h config.log @@ -262,6 +264,8 @@ $(CC) -c $(CPPFLAGS) $(INTERNAL_CFLAGS) $< @USE_THREAD_DB@ linux-arm-low.o: linux-arm-low.c $(linux_low_h) $(server_h) +linux-cris-low.o: linux-cris-low.c $(linux_low_h) $(server_h) +linux-crisv32-low.o: linux-crisv32-low.c $(linux_low_h) $(server_h) linux-i386-low.o: linux-i386-low.c $(linux_low_h) $(server_h) linux-ia64-low.o: linux-ia64-low.c $(linux_low_h) $(server_h) linux-m32r-low.o: linux-m32r-low.c $(linux_low_h) $(server_h) @@ -274,6 +278,12 @@ reg-arm.o : reg-arm.c $(regdef_h) reg-arm.c : $(srcdir)/../regformats/reg-arm.dat $(regdat_sh) sh $(regdat_sh) $(srcdir)/../regformats/reg-arm.dat reg-arm.c +reg-cris.o : reg-cris.c $(regdef_h) +reg-cris.c : $(srcdir)/../regformats/reg-cris.dat $(regdat_sh) + sh $(regdat_sh) $(srcdir)/../regformats/reg-cris.dat reg-cris.c +reg-crisv32.o : reg-crisv32.c $(regdef_h) +reg-crisv32.c : $(srcdir)/../regformats/reg-crisv32.dat $(regdat_sh) + sh $(regdat_sh) $(srcdir)/../regformats/reg-crisv32.dat reg-crisv32.c reg-i386.o : reg-i386.c $(regdef_h) reg-i386.c : $(srcdir)/../regformats/reg-i386.dat $(regdat_sh) sh $(regdat_sh) $(srcdir)/../regformats/reg-i386.dat reg-i386.c Index: configure.srv =================================================================== RCS file: /cvs/src/src/gdb/gdbserver/configure.srv,v retrieving revision 1.9 diff -u -r1.9 configure.srv --- configure.srv 15 Apr 2005 00:24:51 -0000 1.9 +++ configure.srv 12 May 2005 12:56:25 -0000 @@ -23,6 +23,16 @@ srv_linux_usrregs=yes srv_linux_thread_db=yes ;; + crisv32-*-linux*) srv_regobj=reg-crisv32.o + srv_tgtobj="linux-low.o linux-crisv32-low.o" + srv_linux_regsets=yes + srv_linux_thread_db=yes + ;; + cris-*-linux*) srv_regobj=reg-cris.o + srv_tgtobj="linux-low.o linux-cris-low.o" + srv_linux_usrregs=yes + srv_linux_thread_db=yes + ;; i[34567]86-*-linux*) srv_regobj=reg-i386-linux.o srv_tgtobj="linux-low.o linux-i386-low.o i387-fp.o" srv_linux_usrregs=yes [-- Attachment #3: linux-cris-low.c --] [-- Type: text/x-csrc, Size: 2904 bytes --] /* GNU/Linux/CRIS specific low level interface, for the remote server for GDB. Copyright 1995, 1996, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc. This file is part of GDB. 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. */ #include "server.h" #include "linux-low.h" #include <sys/ptrace.h> /* CRISv10 */ #define cris_num_regs 32 /* Locations need to match <include/asm/arch/ptrace.h>. */ static int cris_regmap[] = { 15*4, 14*4, 13*4, 12*4, 11*4, 10*4, 9*4, 8*4, 7*4, 6*4, 5*4, 4*4, 3*4, 2*4, 23*4, 19*4, -1, -1, -1, -1, -1, 17*4, -1, 16*4, -1, -1, -1, 18*4, -1, 17*4, -1, -1 }; static int cris_cannot_store_register (int regno) { if (cris_regmap[regno] == -1) return 1; return (regno >= cris_num_regs); } static int cris_cannot_fetch_register (int regno) { if (cris_regmap[regno] == -1) return 1; return (regno >= cris_num_regs); } extern int debug_threads; static CORE_ADDR cris_get_pc (void) { unsigned long pc; collect_register_by_name ("pc", &pc); if (debug_threads) fprintf (stderr, "stop pc is %08lx\n", pc); return pc; } static void cris_set_pc (CORE_ADDR pc) { unsigned long newpc = pc; supply_register_by_name ("pc", &newpc); } static const unsigned long cris_breakpoint = 0xe938; #define cris_breakpoint_len 2 static int cris_breakpoint_at (CORE_ADDR where) { unsigned long insn; (*the_target->read_memory) (where, (char *) &insn, cris_breakpoint_len); if (insn == cris_breakpoint) return 1; /* If necessary, recognize more trap instructions here. GDB only uses the one. */ return 0; } /* We only place breakpoints in empty marker functions, and thread locking is outside of the function. So rather than importing software single-step, we can just run until exit. */ static CORE_ADDR cris_reinsert_addr (void) { unsigned long pc; collect_register_by_name ("srp", &pc); return pc; } struct linux_target_ops the_low_target = { cris_num_regs, cris_regmap, cris_cannot_fetch_register, cris_cannot_store_register, cris_get_pc, cris_set_pc, (const char *) &cris_breakpoint, cris_breakpoint_len, cris_reinsert_addr, 0, cris_breakpoint_at, 0, 0, 0, 0, }; [-- Attachment #4: linux-crisv32-low.c --] [-- Type: text/x-csrc, Size: 9884 bytes --] /* GNU/Linux/CRIS specific low level interface, for the remote server for GDB. Copyright 1995, 1996, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc. This file is part of GDB. 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. */ #include "server.h" #include "linux-low.h" #include <sys/ptrace.h> /* CRISv32 */ #define cris_num_regs 49 /* Note: Ignoring USP (having the stack pointer in two locations causes trouble without any significant gain). */ /* Locations need to match <include/asm/arch/ptrace.h>. */ static int cris_regmap[] = { 1*4, 2*4, 3*4, 4*4, 5*4, 6*4, 7*4, 8*4, 9*4, 10*4, 11*4, 12*4, 13*4, 14*4, 24*4, 15*4, -1, -1, -1, 16*4, -1, 22*4, 23*4, 17*4, -1, -1, 21*4, 20*4, -1, 19*4, -1, 18*4, 25*4, 26*4, -1, -1, 29*4, 30*4, 31*4, 32*4, 33*4, 34*4, 35*4, 36*4, 37*4, 38*4, 39*4, 40*4, -1 }; extern int debug_threads; static CORE_ADDR cris_get_pc (void) { unsigned long pc; collect_register_by_name ("pc", &pc); if (debug_threads) fprintf (stderr, "stop pc is %08lx\n", pc); return pc; } static void cris_set_pc (CORE_ADDR pc) { unsigned long newpc = pc; supply_register_by_name ("pc", &newpc); } static const unsigned long cris_breakpoint = 0xe938; #define cris_breakpoint_len 2 static int cris_breakpoint_at (CORE_ADDR where) { unsigned long insn; (*the_target->read_memory) (where, (char *) &insn, cris_breakpoint_len); if (insn == cris_breakpoint) return 1; /* If necessary, recognize more trap instructions here. GDB only uses the one. */ return 0; } /* We only place breakpoints in empty marker functions, and thread locking is outside of the function. So rather than importing software single-step, we can just run until exit. */ /* FIXME: This function should not be needed, since we have PTRACE_SINGLESTEP for CRISv32. Without it, td_ta_event_getmsg in thread_db_create_event will fail when debugging multi-threaded applications. */ static CORE_ADDR cris_reinsert_addr (void) { unsigned long pc; collect_register_by_name ("srp", &pc); return pc; } static void cris_write_data_breakpoint (int bp, unsigned long start, unsigned long end) { switch (bp) { case 0: supply_register_by_name("s3", &start); supply_register_by_name("s4", &end); break; case 1: supply_register_by_name("s5", &start); supply_register_by_name("s6", &end); break; case 2: supply_register_by_name("s7", &start); supply_register_by_name("s8", &end); break; case 3: supply_register_by_name("s9", &start); supply_register_by_name("s10", &end); break; case 4: supply_register_by_name("s11", &start); supply_register_by_name("s12", &end); break; case 5: supply_register_by_name("s13", &start); supply_register_by_name("s14", &end); break; } } static int cris_insert_watchpoint(char type, CORE_ADDR addr, int len) { int bp; unsigned long bp_ctrl; unsigned long start, end; unsigned long ccs; /* Breakpoint/watchpoint types (GDB terminology): 0 = memory breakpoint for instructions (not supported; done via memory write instead) 1 = hardware breakpoint for instructions (not supported) 2 = write watchpoint (supported) 3 = read watchpoint (supported) 4 = access watchpoint (supported). */ if (type < '2' || type > '4') { /* Unsupported. */ return 1; } /* Read watchpoints are set as access watchpoints, because of GDB's inability to deal with pure read watchpoints. */ if (type == '3') type = '4'; /* Get the configuration register. */ collect_register_by_name("s0", &bp_ctrl); /* The watchpoint allocation scheme is the simplest possible. For example, if a region is watched for read and a write watch is requested, a new watchpoint will be used. Also, if a watch for a region that is already covered by one or more existing watchpoints, a new watchpoint will be used. */ /* First, find a free data watchpoint. */ for (bp = 0; bp < 6; bp++) { /* Each data watchpoint's control registers occupy 2 bits (hence the 3), starting at bit 2 for D0 (hence the 2) with 4 bits between for each watchpoint (yes, the 4). */ if (!(bp_ctrl & (0x3 << (2 + (bp * 4))))) { break; } } if (bp > 5) { /* We're out of watchpoints. */ return -1; } /* Configure the control register first. */ if (type == '3' || type == '4') { /* Trigger on read. */ bp_ctrl |= (1 << (2 + bp * 4)); } if (type == '2' || type == '4') { /* Trigger on write. */ bp_ctrl |= (2 << (2 + bp * 4)); } /* Setup the configuration register. */ supply_register_by_name("s0", &bp_ctrl); /* Setup the range. */ start = addr; end = addr + len - 1; /* Configure the watchpoint register. */ cris_write_data_breakpoint (bp, start, end); collect_register_by_name("ccs", &ccs); /* Set the S1 flag to enable watchpoints. */ ccs |= (1 << 19); supply_register_by_name("ccs", &ccs); return 0; } static int cris_remove_watchpoint(char type, CORE_ADDR addr, int len) { int bp; unsigned long bp_ctrl; unsigned long start, end; /* Breakpoint/watchpoint types: 0 = memory breakpoint for instructions (not supported; done via memory write instead) 1 = hardware breakpoint for instructions (not supported) 2 = write watchpoint (supported) 3 = read watchpoint (supported) 4 = access watchpoint (supported). */ if (type < '2' || type > '4') { return -1; } /* Read watchpoints are set as access watchpoints, because of GDB's inability to deal with pure read watchpoints. */ if (type == '3') type = '4'; /* Get the configuration register. */ collect_register_by_name("s0", &bp_ctrl); /* Try to find a watchpoint that is configured for the specified range, then check that read/write also matches. */ /* Ugly pointer arithmetic, since I cannot rely on a single switch (addr) as there may be several watchpoints with the same start address for example. */ unsigned long bp_d_regs[12]; /* Get all range registers to simplify search. */ collect_register_by_name("s3", &bp_d_regs[0]); collect_register_by_name("s4", &bp_d_regs[1]); collect_register_by_name("s5", &bp_d_regs[2]); collect_register_by_name("s6", &bp_d_regs[3]); collect_register_by_name("s7", &bp_d_regs[4]); collect_register_by_name("s8", &bp_d_regs[5]); collect_register_by_name("s9", &bp_d_regs[6]); collect_register_by_name("s10", &bp_d_regs[7]); collect_register_by_name("s11", &bp_d_regs[8]); collect_register_by_name("s12", &bp_d_regs[9]); collect_register_by_name("s13", &bp_d_regs[10]); collect_register_by_name("s14", &bp_d_regs[11]); for (bp = 0; bp < 6; bp++) { if (bp_d_regs[bp * 2] == addr && bp_d_regs[bp * 2 + 1] == (addr + len - 1)) { /* Matching range. */ int bitpos = 2 + bp * 4; int rw_bits; /* Read/write bits for this BP. */ rw_bits = (bp_ctrl & (0x3 << bitpos)) >> bitpos; if ((type == '3' && rw_bits == 0x1) || (type == '2' && rw_bits == 0x2) || (type == '4' && rw_bits == 0x3)) { /* Read/write matched. */ break; } } } if (bp > 5) { /* No watchpoint matched. */ return -1; } /* Found a matching watchpoint. Now, deconfigure it by both disabling read/write in bp_ctrl and zeroing its start/end addresses. */ bp_ctrl &= ~(3 << (2 + (bp * 4))); /* Setup the configuration register. */ supply_register_by_name("s0", &bp_ctrl); start = end = 0; /* Configure the watchpoint register. */ cris_write_data_breakpoint (bp, start, end); /* Note that we don't clear the S1 flag here. It's done when continuing. */ return 0; } static int cris_stopped_by_watchpoint (void) { unsigned long exs; collect_register_by_name ("exs", &exs); return (((exs & 0xff00) >> 8) == 0xc); } static CORE_ADDR cris_stopped_data_address (void) { unsigned long eda; collect_register_by_name ("eda", &eda); /* FIXME: Possibly adjust to match watched range. */ return eda; } static void cris_fill_gregset (void *buf) { int i; for (i = 0; i < cris_num_regs; i++) if (cris_regmap[i] != -1) collect_register (i, ((char *) buf) + cris_regmap[i]); } static void cris_store_gregset (const void *buf) { int i; for (i = 0; i < cris_num_regs; i++) if (cris_regmap[i] != -1) supply_register (i, ((char *) buf) + cris_regmap[i]); } typedef unsigned long elf_gregset_t[cris_num_regs]; struct regset_info target_regsets[] = { { PTRACE_GETREGS, PTRACE_SETREGS, sizeof (elf_gregset_t), GENERAL_REGS, cris_fill_gregset, cris_store_gregset }, { 0, 0, -1, -1, NULL, NULL } }; struct linux_target_ops the_low_target = { -1, NULL, NULL, NULL, cris_get_pc, cris_set_pc, (const char *) &cris_breakpoint, cris_breakpoint_len, cris_reinsert_addr, 0, cris_breakpoint_at, cris_insert_watchpoint, cris_remove_watchpoint, cris_stopped_by_watchpoint, cris_stopped_data_address, }; ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-05-12 13:03 ` Orjan Friberg @ 2005-05-12 13:11 ` Orjan Friberg 2005-05-12 13:21 ` Daniel Jacobowitz 2005-05-12 13:13 ` Daniel Jacobowitz 1 sibling, 1 reply; 12+ messages in thread From: Orjan Friberg @ 2005-05-12 13:11 UTC (permalink / raw) To: gdb-patches; +Cc: Daniel Jacobowitz Orjan Friberg wrote: > > Another thing: when removing cris_reinsert_addr, I got the same errors > as described in http://sourceware.org/ml/gdb/2005-01/msg00071.html (i.e. > "thread getmsg err: no event message for getmsg" and gdbserver getting a > SIGSEGV). It seems I didn't finish that thought. The reported fix for that problem (at http://www.linux-mips.org/archives/linux-mips/2005-02/msg00103.html) does not seem to be applicable to this case as our cache works on physical addresses. -- Orjan Friberg Axis Communications ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-05-12 13:11 ` Orjan Friberg @ 2005-05-12 13:21 ` Daniel Jacobowitz 0 siblings, 0 replies; 12+ messages in thread From: Daniel Jacobowitz @ 2005-05-12 13:21 UTC (permalink / raw) To: Orjan Friberg; +Cc: gdb-patches On Thu, May 12, 2005 at 03:05:53PM +0200, Orjan Friberg wrote: > Orjan Friberg wrote: > > > >Another thing: when removing cris_reinsert_addr, I got the same errors > >as described in http://sourceware.org/ml/gdb/2005-01/msg00071.html (i.e. > >"thread getmsg err: no event message for getmsg" and gdbserver getting a > >SIGSEGV). > > It seems I didn't finish that thought. The reported fix for that problem > (at http://www.linux-mips.org/archives/linux-mips/2005-02/msg00103.html) > does not seem to be applicable to this case as our cache works on physical > addresses. Ah, that bug. It could be any other problem causing incoherency between the inferior and the ptracing process; there's a number of other ways to trigger the error, too. (The segfault after the error is "expected". Gdbserver is not at all robust in the face of errors which shouldn't happen in correct operation.) -- Daniel Jacobowitz CodeSourcery, LLC ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-05-12 13:03 ` Orjan Friberg 2005-05-12 13:11 ` Orjan Friberg @ 2005-05-12 13:13 ` Daniel Jacobowitz 2005-05-12 14:59 ` Orjan Friberg 1 sibling, 1 reply; 12+ messages in thread From: Daniel Jacobowitz @ 2005-05-12 13:13 UTC (permalink / raw) To: Orjan Friberg; +Cc: gdb-patches On Thu, May 12, 2005 at 02:59:16PM +0200, Orjan Friberg wrote: > Daniel Jacobowitz wrote: > > > >Let's go ahead. The patch is mostly OK, but I have a last question or > >two... > > Thanks for your comments, and apologies for the long delay. (There are a > couple of questions at the end also.) I didn't see any questions? > Another thing: when removing cris_reinsert_addr, I got the same errors as > described in http://sourceware.org/ml/gdb/2005-01/msg00071.html (i.e. > "thread getmsg err: no event message for getmsg" and gdbserver getting a > SIGSEGV). At the moment I have no idea what's going on, so as a temporary > solution I reinstated cris_reinsert_addr but with a FIXME. Odd. Does your hardware have real single step support, or implement PTRACE_SINGLESTEP by managing breakpoints? Other than that, I'm not sure what could be wrong. I've got more nits for you though. > static const unsigned long cris_breakpoint = 0xe938; > #define cris_breakpoint_len 2 > > static int > cris_breakpoint_at (CORE_ADDR where) > { > unsigned long insn; > > (*the_target->read_memory) (where, (char *) &insn, cris_breakpoint_len); > if (insn == cris_breakpoint) > return 1; > > /* If necessary, recognize more trap instructions here. GDB only uses the > one. */ > return 0; > } This is presumably going to work for you, but it's seriously untidy - there's a hidden endianness dependency. Can you see it? Hint: the types of cris_breakpoint and insn are wrong. Same thing applies to linux-crisv32-low.c. I just realized that these files will require an entry in config/djgpp/fnchange.lst; their names are too similar. > static void > cris_write_data_breakpoint (int bp, unsigned long start, unsigned long end) > { > switch (bp) > { > case 0: > supply_register_by_name("s3", &start); > supply_register_by_name("s4", &end); > break; You were doing great on GNU formatting right up until here, and then you lost it. Everything from here down in this file needs to be checked. For instance: > static int > cris_insert_watchpoint(char type, CORE_ADDR addr, int len) Space. > if (type < '2' || type > '4') { > /* Unsupported. */ > return 1; > } Unnecessary braces; even if they were necessary they'd have to be on their own lines and indented. -- Daniel Jacobowitz CodeSourcery, LLC ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-05-12 13:13 ` Daniel Jacobowitz @ 2005-05-12 14:59 ` Orjan Friberg 2005-05-18 1:55 ` Daniel Jacobowitz 0 siblings, 1 reply; 12+ messages in thread From: Orjan Friberg @ 2005-05-12 14:59 UTC (permalink / raw) To: Daniel Jacobowitz; +Cc: gdb-patches [-- Attachment #1: Type: text/plain, Size: 2447 bytes --] Daniel Jacobowitz wrote: > > I didn't see any questions? They were badly disguised as statements ;). (I was referring to the "thread getmsg err" section.) >>Another thing: when removing cris_reinsert_addr, I got the same errors as >>described in http://sourceware.org/ml/gdb/2005-01/msg00071.html (i.e. >>"thread getmsg err: no event message for getmsg" and gdbserver getting a >>SIGSEGV). At the moment I have no idea what's going on, so as a temporary >>solution I reinstated cris_reinsert_addr but with a FIXME. > > > Odd. Does your hardware have real single step support, or implement > PTRACE_SINGLESTEP by managing breakpoints? Other than that, I'm not > sure what could be wrong. It's real hardware single step. > I've got more nits for you though. > > >>static const unsigned long cris_breakpoint = 0xe938; >>#define cris_breakpoint_len 2 >> >>static int >>cris_breakpoint_at (CORE_ADDR where) >>{ >> unsigned long insn; >> >> (*the_target->read_memory) (where, (char *) &insn, cris_breakpoint_len); >> if (insn == cris_breakpoint) >> return 1; >> >> /* If necessary, recognize more trap instructions here. GDB only uses the >> one. */ >> return 0; >>} > > > This is presumably going to work for you, but it's seriously untidy - > there's a hidden endianness dependency. Can you see it? Hint: the > types of cris_breakpoint and insn are wrong. Both changed to unsigned shorts; thanks. > Same thing applies to linux-crisv32-low.c. Also fixed. > I just realized that these files will require an entry in > config/djgpp/fnchange.lst; their names are too similar. Ok, I will submit that as a separate patch once this is in. > You were doing great on GNU formatting right up until here, and then > you lost it. Everything from here down in this file needs to be > checked. For instance: > > >>static int >>cris_insert_watchpoint(char type, CORE_ADDR addr, int len) > > > Space. Hm, I really did lose it. Corrected a bunch of them. >> if (type < '2' || type > '4') { >> /* Unsupported. */ >> return 1; >> } > > > Unnecessary braces; even if they were necessary they'd have to be on > their own lines and indented. I like to have braces when there's more than one line (even if one is a comment) so if it's ok to keep them (and I couldn't find anything in the coding standard saying otherwise) I'd like to. I fixed the formatting; thanks. -- Orjan Friberg Axis Communications [-- Attachment #2: linux-cris-low.c --] [-- Type: text/x-csrc, Size: 2906 bytes --] /* GNU/Linux/CRIS specific low level interface, for the remote server for GDB. Copyright 1995, 1996, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc. This file is part of GDB. 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. */ #include "server.h" #include "linux-low.h" #include <sys/ptrace.h> /* CRISv10 */ #define cris_num_regs 32 /* Locations need to match <include/asm/arch/ptrace.h>. */ static int cris_regmap[] = { 15*4, 14*4, 13*4, 12*4, 11*4, 10*4, 9*4, 8*4, 7*4, 6*4, 5*4, 4*4, 3*4, 2*4, 23*4, 19*4, -1, -1, -1, -1, -1, 17*4, -1, 16*4, -1, -1, -1, 18*4, -1, 17*4, -1, -1 }; static int cris_cannot_store_register (int regno) { if (cris_regmap[regno] == -1) return 1; return (regno >= cris_num_regs); } static int cris_cannot_fetch_register (int regno) { if (cris_regmap[regno] == -1) return 1; return (regno >= cris_num_regs); } extern int debug_threads; static CORE_ADDR cris_get_pc (void) { unsigned long pc; collect_register_by_name ("pc", &pc); if (debug_threads) fprintf (stderr, "stop pc is %08lx\n", pc); return pc; } static void cris_set_pc (CORE_ADDR pc) { unsigned long newpc = pc; supply_register_by_name ("pc", &newpc); } static const unsigned short cris_breakpoint = 0xe938; #define cris_breakpoint_len 2 static int cris_breakpoint_at (CORE_ADDR where) { unsigned short insn; (*the_target->read_memory) (where, (char *) &insn, cris_breakpoint_len); if (insn == cris_breakpoint) return 1; /* If necessary, recognize more trap instructions here. GDB only uses the one. */ return 0; } /* We only place breakpoints in empty marker functions, and thread locking is outside of the function. So rather than importing software single-step, we can just run until exit. */ static CORE_ADDR cris_reinsert_addr (void) { unsigned long pc; collect_register_by_name ("srp", &pc); return pc; } struct linux_target_ops the_low_target = { cris_num_regs, cris_regmap, cris_cannot_fetch_register, cris_cannot_store_register, cris_get_pc, cris_set_pc, (const char *) &cris_breakpoint, cris_breakpoint_len, cris_reinsert_addr, 0, cris_breakpoint_at, 0, 0, 0, 0, }; [-- Attachment #3: linux-crisv32-low.c --] [-- Type: text/x-csrc, Size: 9960 bytes --] /* GNU/Linux/CRIS specific low level interface, for the remote server for GDB. Copyright 1995, 1996, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc. This file is part of GDB. 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. */ #include "server.h" #include "linux-low.h" #include <sys/ptrace.h> /* CRISv32 */ #define cris_num_regs 49 /* Note: Ignoring USP (having the stack pointer in two locations causes trouble without any significant gain). */ /* Locations need to match <include/asm/arch/ptrace.h>. */ static int cris_regmap[] = { 1*4, 2*4, 3*4, 4*4, 5*4, 6*4, 7*4, 8*4, 9*4, 10*4, 11*4, 12*4, 13*4, 14*4, 24*4, 15*4, -1, -1, -1, 16*4, -1, 22*4, 23*4, 17*4, -1, -1, 21*4, 20*4, -1, 19*4, -1, 18*4, 25*4, 26*4, -1, -1, 29*4, 30*4, 31*4, 32*4, 33*4, 34*4, 35*4, 36*4, 37*4, 38*4, 39*4, 40*4, -1 }; extern int debug_threads; static CORE_ADDR cris_get_pc (void) { unsigned long pc; collect_register_by_name ("pc", &pc); if (debug_threads) fprintf (stderr, "stop pc is %08lx\n", pc); return pc; } static void cris_set_pc (CORE_ADDR pc) { unsigned long newpc = pc; supply_register_by_name ("pc", &newpc); } static const unsigned short cris_breakpoint = 0xe938; #define cris_breakpoint_len 2 static int cris_breakpoint_at (CORE_ADDR where) { unsigned short insn; (*the_target->read_memory) (where, (char *) &insn, cris_breakpoint_len); if (insn == cris_breakpoint) return 1; /* If necessary, recognize more trap instructions here. GDB only uses the one. */ return 0; } /* We only place breakpoints in empty marker functions, and thread locking is outside of the function. So rather than importing software single-step, we can just run until exit. */ /* FIXME: This function should not be needed, since we have PTRACE_SINGLESTEP for CRISv32. Without it, td_ta_event_getmsg in thread_db_create_event will fail when debugging multi-threaded applications. */ static CORE_ADDR cris_reinsert_addr (void) { unsigned long pc; collect_register_by_name ("srp", &pc); return pc; } static void cris_write_data_breakpoint (int bp, unsigned long start, unsigned long end) { switch (bp) { case 0: supply_register_by_name ("s3", &start); supply_register_by_name ("s4", &end); break; case 1: supply_register_by_name ("s5", &start); supply_register_by_name ("s6", &end); break; case 2: supply_register_by_name ("s7", &start); supply_register_by_name ("s8", &end); break; case 3: supply_register_by_name ("s9", &start); supply_register_by_name ("s10", &end); break; case 4: supply_register_by_name ("s11", &start); supply_register_by_name ("s12", &end); break; case 5: supply_register_by_name ("s13", &start); supply_register_by_name ("s14", &end); break; } } static int cris_insert_watchpoint (char type, CORE_ADDR addr, int len) { int bp; unsigned long bp_ctrl; unsigned long start, end; unsigned long ccs; /* Breakpoint/watchpoint types (GDB terminology): 0 = memory breakpoint for instructions (not supported; done via memory write instead) 1 = hardware breakpoint for instructions (not supported) 2 = write watchpoint (supported) 3 = read watchpoint (supported) 4 = access watchpoint (supported). */ if (type < '2' || type > '4') { /* Unsupported. */ return 1; } /* Read watchpoints are set as access watchpoints, because of GDB's inability to deal with pure read watchpoints. */ if (type == '3') type = '4'; /* Get the configuration register. */ collect_register_by_name ("s0", &bp_ctrl); /* The watchpoint allocation scheme is the simplest possible. For example, if a region is watched for read and a write watch is requested, a new watchpoint will be used. Also, if a watch for a region that is already covered by one or more existing watchpoints, a new watchpoint will be used. */ /* First, find a free data watchpoint. */ for (bp = 0; bp < 6; bp++) { /* Each data watchpoint's control registers occupy 2 bits (hence the 3), starting at bit 2 for D0 (hence the 2) with 4 bits between for each watchpoint (yes, the 4). */ if (!(bp_ctrl & (0x3 << (2 + (bp * 4))))) break; } if (bp > 5) { /* We're out of watchpoints. */ return -1; } /* Configure the control register first. */ if (type == '3' || type == '4') { /* Trigger on read. */ bp_ctrl |= (1 << (2 + bp * 4)); } if (type == '2' || type == '4') { /* Trigger on write. */ bp_ctrl |= (2 << (2 + bp * 4)); } /* Setup the configuration register. */ supply_register_by_name ("s0", &bp_ctrl); /* Setup the range. */ start = addr; end = addr + len - 1; /* Configure the watchpoint register. */ cris_write_data_breakpoint (bp, start, end); collect_register_by_name ("ccs", &ccs); /* Set the S1 flag to enable watchpoints. */ ccs |= (1 << 19); supply_register_by_name ("ccs", &ccs); return 0; } static int cris_remove_watchpoint (char type, CORE_ADDR addr, int len) { int bp; unsigned long bp_ctrl; unsigned long start, end; /* Breakpoint/watchpoint types: 0 = memory breakpoint for instructions (not supported; done via memory write instead) 1 = hardware breakpoint for instructions (not supported) 2 = write watchpoint (supported) 3 = read watchpoint (supported) 4 = access watchpoint (supported). */ if (type < '2' || type > '4') return -1; /* Read watchpoints are set as access watchpoints, because of GDB's inability to deal with pure read watchpoints. */ if (type == '3') type = '4'; /* Get the configuration register. */ collect_register_by_name ("s0", &bp_ctrl); /* Try to find a watchpoint that is configured for the specified range, then check that read/write also matches. */ /* Ugly pointer arithmetic, since I cannot rely on a single switch (addr) as there may be several watchpoints with the same start address for example. */ unsigned long bp_d_regs[12]; /* Get all range registers to simplify search. */ collect_register_by_name ("s3", &bp_d_regs[0]); collect_register_by_name ("s4", &bp_d_regs[1]); collect_register_by_name ("s5", &bp_d_regs[2]); collect_register_by_name ("s6", &bp_d_regs[3]); collect_register_by_name ("s7", &bp_d_regs[4]); collect_register_by_name ("s8", &bp_d_regs[5]); collect_register_by_name ("s9", &bp_d_regs[6]); collect_register_by_name ("s10", &bp_d_regs[7]); collect_register_by_name ("s11", &bp_d_regs[8]); collect_register_by_name ("s12", &bp_d_regs[9]); collect_register_by_name ("s13", &bp_d_regs[10]); collect_register_by_name ("s14", &bp_d_regs[11]); for (bp = 0; bp < 6; bp++) { if (bp_d_regs[bp * 2] == addr && bp_d_regs[bp * 2 + 1] == (addr + len - 1)) { /* Matching range. */ int bitpos = 2 + bp * 4; int rw_bits; /* Read/write bits for this BP. */ rw_bits = (bp_ctrl & (0x3 << bitpos)) >> bitpos; if ((type == '3' && rw_bits == 0x1) || (type == '2' && rw_bits == 0x2) || (type == '4' && rw_bits == 0x3)) { /* Read/write matched. */ break; } } } if (bp > 5) { /* No watchpoint matched. */ return -1; } /* Found a matching watchpoint. Now, deconfigure it by both disabling read/write in bp_ctrl and zeroing its start/end addresses. */ bp_ctrl &= ~(3 << (2 + (bp * 4))); /* Setup the configuration register. */ supply_register_by_name ("s0", &bp_ctrl); start = end = 0; /* Configure the watchpoint register. */ cris_write_data_breakpoint (bp, start, end); /* Note that we don't clear the S1 flag here. It's done when continuing. */ return 0; } static int cris_stopped_by_watchpoint (void) { unsigned long exs; collect_register_by_name ("exs", &exs); return (((exs & 0xff00) >> 8) == 0xc); } static CORE_ADDR cris_stopped_data_address (void) { unsigned long eda; collect_register_by_name ("eda", &eda); /* FIXME: Possibly adjust to match watched range. */ return eda; } static void cris_fill_gregset (void *buf) { int i; for (i = 0; i < cris_num_regs; i++) { if (cris_regmap[i] != -1) collect_register (i, ((char *) buf) + cris_regmap[i]); } } static void cris_store_gregset (const void *buf) { int i; for (i = 0; i < cris_num_regs; i++) { if (cris_regmap[i] != -1) supply_register (i, ((char *) buf) + cris_regmap[i]); } } typedef unsigned long elf_gregset_t[cris_num_regs]; struct regset_info target_regsets[] = { { PTRACE_GETREGS, PTRACE_SETREGS, sizeof (elf_gregset_t), GENERAL_REGS, cris_fill_gregset, cris_store_gregset }, { 0, 0, -1, -1, NULL, NULL } }; struct linux_target_ops the_low_target = { -1, NULL, NULL, NULL, cris_get_pc, cris_set_pc, (const char *) &cris_breakpoint, cris_breakpoint_len, cris_reinsert_addr, 0, cris_breakpoint_at, cris_insert_watchpoint, cris_remove_watchpoint, cris_stopped_by_watchpoint, cris_stopped_data_address, }; ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-05-12 14:59 ` Orjan Friberg @ 2005-05-18 1:55 ` Daniel Jacobowitz 2005-05-23 12:35 ` Orjan Friberg 0 siblings, 1 reply; 12+ messages in thread From: Daniel Jacobowitz @ 2005-05-18 1:55 UTC (permalink / raw) To: Orjan Friberg; +Cc: gdb-patches On Thu, May 12, 2005 at 04:35:40PM +0200, Orjan Friberg wrote: > >I just realized that these files will require an entry in > >config/djgpp/fnchange.lst; their names are too similar. > > Ok, I will submit that as a separate patch once this is in. These revised files are OK. Please post the fnchange.lst patch before checking them in, though. > I like to have braces when there's more than one line (even if one is a > comment) so if it's ok to keep them (and I couldn't find anything in the > coding standard saying otherwise) I'd like to. I fixed the formatting; > thanks. OK. -- Daniel Jacobowitz CodeSourcery, LLC ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) 2005-05-18 1:55 ` Daniel Jacobowitz @ 2005-05-23 12:35 ` Orjan Friberg 0 siblings, 0 replies; 12+ messages in thread From: Orjan Friberg @ 2005-05-23 12:35 UTC (permalink / raw) To: Daniel Jacobowitz; +Cc: gdb-patches Daniel Jacobowitz wrote: > On Thu, May 12, 2005 at 04:35:40PM +0200, Orjan Friberg wrote: > >>>I just realized that these files will require an entry in >>>config/djgpp/fnchange.lst; their names are too similar. >> >>Ok, I will submit that as a separate patch once this is in. > > > These revised files are OK. Please post the fnchange.lst patch before > checking them in, though. Posted separately. >>I like to have braces when there's more than one line (even if one is a >>comment) so if it's ok to keep them (and I couldn't find anything in the >>coding standard saying otherwise) I'd like to. I fixed the formatting; >>thanks. > > > OK. Thanks; committed. -- Orjan Friberg Axis Communications ^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2005-05-23 11:21 UTC | newest] Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2005-01-14 9:54 [gdbserver/rfa] CRIS/CRISv32 gdbserver support (part 2) Orjan Friberg 2005-01-30 4:54 ` Daniel Jacobowitz 2005-02-01 14:31 ` Orjan Friberg 2005-02-14 16:17 ` Orjan Friberg 2005-02-24 20:57 ` Daniel Jacobowitz 2005-05-12 13:03 ` Orjan Friberg 2005-05-12 13:11 ` Orjan Friberg 2005-05-12 13:21 ` Daniel Jacobowitz 2005-05-12 13:13 ` Daniel Jacobowitz 2005-05-12 14:59 ` Orjan Friberg 2005-05-18 1:55 ` Daniel Jacobowitz 2005-05-23 12:35 ` Orjan Friberg
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox