* [PATCH 01/11] sim/rx: define sim_memory_map
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 14:46 ` Mike Frysinger
2021-02-01 11:44 ` [PATCH 02/11] sim/rx: fix an issue where we try to modify a const string Andrew Burgess
` (10 subsequent siblings)
11 siblings, 1 reply; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
The rx simulator doesn't build due to sim_memory_map not being
defined. Define it now to return NULL, this can be extended later to
return an actual memory map if anyone wants this functionality.
sim/rx/ChangeLog:
* gdb-if.c (sim_memory_map): New function.
---
sim/rx/ChangeLog | 4 ++++
sim/rx/gdb-if.c | 8 ++++++++
2 files changed, 12 insertions(+)
diff --git a/sim/rx/gdb-if.c b/sim/rx/gdb-if.c
index 63705f9bb0e..6f8bfbd1c50 100644
--- a/sim/rx/gdb-if.c
+++ b/sim/rx/gdb-if.c
@@ -854,3 +854,11 @@ sim_complete_command (SIM_DESC sd, const char *text, const char *word)
{
return NULL;
}
+
+/* Stub this out for now. */
+
+char *
+sim_memory_map (SIM_DESC sd)
+{
+ return NULL;
+}
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* Re: [PATCH 01/11] sim/rx: define sim_memory_map
2021-02-01 11:44 ` [PATCH 01/11] sim/rx: define sim_memory_map Andrew Burgess
@ 2021-02-01 14:46 ` Mike Frysinger
2021-02-08 11:01 ` Andrew Burgess
0 siblings, 1 reply; 20+ messages in thread
From: Mike Frysinger @ 2021-02-01 14:46 UTC (permalink / raw)
To: Andrew Burgess; +Cc: gdb-patches
On 01 Feb 2021 11:44, Andrew Burgess wrote:
> The rx simulator doesn't build due to sim_memory_map not being
> defined. Define it now to return NULL, this can be extended later to
> return an actual memory map if anyone wants this functionality.
i build every sim port, and rx was/is building fine for me. nothing in
the sim refers to the sim_memory_map, so it failing to build is a bit
surprising. do you mean gdb is failing to link instead ? i don't always
try to build those for ports as not all have gdb ports.
-mike
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [PATCH 01/11] sim/rx: define sim_memory_map
2021-02-01 14:46 ` Mike Frysinger
@ 2021-02-08 11:01 ` Andrew Burgess
0 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-08 11:01 UTC (permalink / raw)
To: gdb-patches
* Mike Frysinger <vapier@gentoo.org> [2021-02-01 09:46:44 -0500]:
> On 01 Feb 2021 11:44, Andrew Burgess wrote:
> > The rx simulator doesn't build due to sim_memory_map not being
> > defined. Define it now to return NULL, this can be extended later to
> > return an actual memory map if anyone wants this functionality.
>
> i build every sim port, and rx was/is building fine for me. nothing in
> the sim refers to the sim_memory_map, so it failing to build is a bit
> surprising. do you mean gdb is failing to link instead ? i don't always
> try to build those for ports as not all have gdb ports.
You are correct of course.
I updated the wording of the commit message more inline with the patch
you submitted here:
https://sourceware.org/pipermail/gdb-patches/2021-February/175689.html
Thanks,
Andrew
^ permalink raw reply [flat|nested] 20+ messages in thread
* [PATCH 02/11] sim/rx: fix an issue where we try to modify a const string
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
2021-02-01 11:44 ` [PATCH 01/11] sim/rx: define sim_memory_map Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 11:44 ` [PATCH 03/11] sim/rx: fill in missing 'void' for empty argument lists Andrew Burgess
` (9 subsequent siblings)
11 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
While experimenting with switching on warnings for the rx simulator I
discovered this bug. In sim_do_command we get passed a 'const char *'
argument. We create a copy of this string to work with locally, but
then while processing this we accidentally switch back to reference
the original string.
sim/rx/ChangeLog:
* gdb-if.c (sim_do_command): Work with a copy of the command.
---
sim/rx/ChangeLog | 4 ++++
sim/rx/gdb-if.c | 4 ++--
2 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/sim/rx/gdb-if.c b/sim/rx/gdb-if.c
index 6f8bfbd1c50..55eb13d12ef 100644
--- a/sim/rx/gdb-if.c
+++ b/sim/rx/gdb-if.c
@@ -804,13 +804,13 @@ sim_do_command (SIM_DESC sd, const char *cmd)
p++;
/* Find the extent of the command word. */
- for (p = cmd; *p; p++)
+ for (; *p != '\0'; p++)
if (isspace (*p))
break;
/* Null-terminate the command word, and record the start of any
further arguments. */
- if (*p)
+ if (*p != '\0')
{
*p = '\0';
args = p + 1;
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* [PATCH 03/11] sim/rx: fill in missing 'void' for empty argument lists
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
2021-02-01 11:44 ` [PATCH 01/11] sim/rx: define sim_memory_map Andrew Burgess
2021-02-01 11:44 ` [PATCH 02/11] sim/rx: fix an issue where we try to modify a const string Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 11:44 ` [PATCH 04/11] sim/rx: mark some functions as static Andrew Burgess
` (8 subsequent siblings)
11 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
Ensure we have 'void' inside empty argument lists. This was causing
several warnings for the rx simulator.
sim/rx/ChangeLog:
* cpu.h (trace_register_changes): Add void parameter type.
* err.c (ee_overrides): Likewise.
* mem.c (mem_usage_stats): Likewise.
(e): Likewise.
* reg.c (stack_heap_stats): Likewise.
* rx.c (pop): Likewise.
(poppc): Likewise.
(decode_opcode): Likewise.
* syscalls.c (arg): Likewise.
---
sim/rx/ChangeLog | 12 ++++++++++++
sim/rx/cpu.h | 2 +-
sim/rx/err.c | 2 +-
sim/rx/mem.c | 4 ++--
sim/rx/reg.c | 2 +-
sim/rx/rx.c | 6 +++---
sim/rx/syscalls.c | 2 +-
7 files changed, 21 insertions(+), 9 deletions(-)
diff --git a/sim/rx/cpu.h b/sim/rx/cpu.h
index 700f2de3c2b..ef4503cc720 100644
--- a/sim/rx/cpu.h
+++ b/sim/rx/cpu.h
@@ -240,6 +240,6 @@ extern void reset_pipeline_stats (void);
extern void halt_pipeline_stats (void);
extern void pipeline_stats (void);
-extern void trace_register_changes ();
+extern void trace_register_changes (void);
extern void generate_access_exception (void);
extern jmp_buf decode_jmp_buf;
diff --git a/sim/rx/err.c b/sim/rx/err.c
index 3fc64d1d685..ce1cd571f59 100644
--- a/sim/rx/err.c
+++ b/sim/rx/err.c
@@ -28,7 +28,7 @@ static unsigned char ee_actions[SIM_ERR_NUM_ERRORS];
static enum execution_error last_error;
static void
-ee_overrides ()
+ee_overrides (void)
{
/* GCC may initialize a bitfield by reading the uninitialized byte,
masking in the bitfield, and writing the byte back out. */
diff --git a/sim/rx/mem.c b/sim/rx/mem.c
index 40ab0bd5e6a..7fbf08e6415 100644
--- a/sim/rx/mem.c
+++ b/sim/rx/mem.c
@@ -161,7 +161,7 @@ mcs (int isput, int bytes)
}
void
-mem_usage_stats ()
+mem_usage_stats (void)
{
int i, j;
int rstart = 0;
@@ -244,7 +244,7 @@ s (int address, char *dir)
#define S(d) if (trace) s(address, d)
static void
-e ()
+e (void)
{
if (!trace)
return;
diff --git a/sim/rx/reg.c b/sim/rx/reg.c
index c300f3ba6b2..00796a0c99a 100644
--- a/sim/rx/reg.c
+++ b/sim/rx/reg.c
@@ -149,7 +149,7 @@ get_reg64 (int id)
static int highest_sp = 0, lowest_sp = 0xffffff;
void
-stack_heap_stats ()
+stack_heap_stats (void)
{
if (heapbottom < heaptop)
printf ("heap: %08x - %08x (%d bytes)\n", heapbottom, heaptop,
diff --git a/sim/rx/rx.c b/sim/rx/rx.c
index beed6a14c81..e3737a56e98 100644
--- a/sim/rx/rx.c
+++ b/sim/rx/rx.c
@@ -682,7 +682,7 @@ pushpc(int val)
}
static int
-pop()
+pop (void)
{
int rv;
int rsp = get_reg (sp);
@@ -693,7 +693,7 @@ pop()
}
static int
-poppc()
+poppc (void)
{
int rv;
int rsp = get_reg (sp);
@@ -931,7 +931,7 @@ op_is_memory (const RX_Opcode_Decoded *rd, int i)
#define DO_RETURN(x) { longjmp (decode_jmp_buf, x); }
int
-decode_opcode ()
+decode_opcode (void)
{
unsigned int uma=0, umb=0;
int ma=0, mb=0;
diff --git a/sim/rx/syscalls.c b/sim/rx/syscalls.c
index 1a6f65fad16..2f89da17330 100644
--- a/sim/rx/syscalls.c
+++ b/sim/rx/syscalls.c
@@ -69,7 +69,7 @@ get_callbacks (void)
int argp, stackp;
static int
-arg ()
+arg (void)
{
int rv = 0;
argp++;
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* [PATCH 04/11] sim/rx: mark some functions as static
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
` (2 preceding siblings ...)
2021-02-01 11:44 ` [PATCH 03/11] sim/rx: fill in missing 'void' for empty argument lists Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 11:44 ` [PATCH 05/11] sim/rx: delete an unused function Andrew Burgess
` (7 subsequent siblings)
11 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
Some functions that should be marked static.
sim/rx/ChangeLog:
* fpu.c (check_exceptions): Make static.
* gdb-if.c (handle_step): Likewise.
* mem.c (mem_put_byte): Likewise.
---
sim/rx/ChangeLog | 6 ++++++
sim/rx/fpu.c | 2 +-
sim/rx/gdb-if.c | 2 +-
sim/rx/mem.c | 2 +-
4 files changed, 9 insertions(+), 3 deletions(-)
diff --git a/sim/rx/fpu.c b/sim/rx/fpu.c
index 00c3ab73f0b..f9e9007192a 100644
--- a/sim/rx/fpu.c
+++ b/sim/rx/fpu.c
@@ -354,7 +354,7 @@ static const char *ex_names[] = {
/* This checks for all exceptional cases (not all FP exceptions) and
returns TRUE if it is providing the result in *c. If it returns
FALSE, the caller should do the "normal" operation. */
-int
+static int
check_exceptions (FP_Parts *a, FP_Parts *b, fp_t *c,
FP_ExceptionCases ex_tab[5][5],
FP_ExceptionCases *case_ret)
diff --git a/sim/rx/gdb-if.c b/sim/rx/gdb-if.c
index 55eb13d12ef..3d052e62baa 100644
--- a/sim/rx/gdb-if.c
+++ b/sim/rx/gdb-if.c
@@ -688,7 +688,7 @@ rx_signal_to_gdb_signal (int rx)
/* Take a step return code RC and set up the variables consulted by
sim_stop_reason appropriately. */
-void
+static void
handle_step (int rc)
{
if (execution_error_get_last_error () != SIM_ERR_NONE)
diff --git a/sim/rx/mem.c b/sim/rx/mem.c
index 7fbf08e6415..fe8e08d1460 100644
--- a/sim/rx/mem.c
+++ b/sim/rx/mem.c
@@ -262,7 +262,7 @@ mtypec (int address)
#define E() if (trace) e()
-void
+static void
mem_put_byte (unsigned int address, unsigned char value)
{
unsigned char *m;
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* [PATCH 05/11] sim/rx: delete an unused function
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
` (3 preceding siblings ...)
2021-02-01 11:44 ` [PATCH 04/11] sim/rx: mark some functions as static Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 11:44 ` [PATCH 06/11] sim/rx: use puts instead of printf in a few places Andrew Burgess
` (6 subsequent siblings)
11 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
This function is not used.
sim/rx/ChangeLog:
* err.c (execution_error_exit_all): Delete.
---
sim/rx/ChangeLog | 4 ++++
sim/rx/err.c | 9 ---------
2 files changed, 4 insertions(+), 9 deletions(-)
diff --git a/sim/rx/err.c b/sim/rx/err.c
index ce1cd571f59..e72b6d3ddee 100644
--- a/sim/rx/err.c
+++ b/sim/rx/err.c
@@ -60,15 +60,6 @@ execution_error_init_debugger (void)
ee_overrides ();
}
-void
-execution_error_exit_all (void)
-{
- int i;
-
- for (i = 0; i < SIM_ERR_NUM_ERRORS; i++)
- ee_actions[i] = SIM_ERRACTION_EXIT;
-}
-
void
execution_error_warn_all (void)
{
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* [PATCH 06/11] sim/rx: use puts instead of printf in a few places
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
` (4 preceding siblings ...)
2021-02-01 11:44 ` [PATCH 05/11] sim/rx: delete an unused function Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 14:49 ` Mike Frysinger
2021-02-01 11:44 ` [PATCH 07/11] sim/rx: move some variable declarations to the start of the block Andrew Burgess
` (5 subsequent siblings)
11 siblings, 1 reply; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
Calling printf with the format being a non constant string results in
a GCC warning. Use puts instead in a few places.
sim/rx/ChangeLog:
* reg.c (fpsw2str): Replace use of printf with puts.
(trace_register_changes): Likewise.
---
sim/rx/ChangeLog | 5 +++++
sim/rx/reg.c | 8 ++++----
2 files changed, 9 insertions(+), 4 deletions(-)
diff --git a/sim/rx/reg.c b/sim/rx/reg.c
index 00796a0c99a..f1bdd4c9841 100644
--- a/sim/rx/reg.c
+++ b/sim/rx/reg.c
@@ -509,7 +509,7 @@ fpsw2str(int rpsw)
#define TRC(f,n) \
if (oldregs.f != regs.f) \
{ \
- if (tag) { printf (tag); tag = 0; } \
+ if (tag) { puts (tag); tag = 0; } \
printf(" %s %08x:%08x", n, \
(unsigned int)oldregs.f, \
(unsigned int)regs.f); \
@@ -531,7 +531,7 @@ trace_register_changes (void)
TRC (r_isp, "isp");
if (oldregs.r_psw != regs.r_psw)
{
- if (tag) { printf (tag); tag = 0; }
+ if (tag) { puts (tag); tag = 0; }
printf(" psw %s:", psw2str(oldregs.r_psw));
printf("%s", psw2str(regs.r_psw));
oldregs.r_psw = regs.r_psw;
@@ -539,7 +539,7 @@ trace_register_changes (void)
if (oldregs.r_fpsw != regs.r_fpsw)
{
- if (tag) { printf (tag); tag = 0; }
+ if (tag) { puts (tag); tag = 0; }
printf(" fpsw %s:", fpsw2str(oldregs.r_fpsw));
printf("%s", fpsw2str(regs.r_fpsw));
oldregs.r_fpsw = regs.r_fpsw;
@@ -547,9 +547,9 @@ trace_register_changes (void)
if (oldregs.r_acc != regs.r_acc)
{
- if (tag) { printf (tag); tag = 0; }
printf(" acc %016llx:", oldregs.r_acc);
printf("%016llx", regs.r_acc);
+ if (tag) { puts (tag); tag = 0; }
oldregs.r_acc = regs.r_acc;
}
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* Re: [PATCH 06/11] sim/rx: use puts instead of printf in a few places
2021-02-01 11:44 ` [PATCH 06/11] sim/rx: use puts instead of printf in a few places Andrew Burgess
@ 2021-02-01 14:49 ` Mike Frysinger
2021-02-08 10:58 ` Andrew Burgess
0 siblings, 1 reply; 20+ messages in thread
From: Mike Frysinger @ 2021-02-01 14:49 UTC (permalink / raw)
To: Andrew Burgess; +Cc: gdb-patches
On 01 Feb 2021 11:44, Andrew Burgess wrote:
> Calling printf with the format being a non constant string results in
> a GCC warning. Use puts instead in a few places.
puts() outputs the string with a trailing newline.
doesn't seem equiv in this code ?
i change to printf ("%s", ...) to avoid these warnings.
-mike
^ permalink raw reply [flat|nested] 20+ messages in thread* Re: [PATCH 06/11] sim/rx: use puts instead of printf in a few places
2021-02-01 14:49 ` Mike Frysinger
@ 2021-02-08 10:58 ` Andrew Burgess
0 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-08 10:58 UTC (permalink / raw)
To: gdb-patches
* Mike Frysinger <vapier@gentoo.org> [2021-02-01 09:49:00 -0500]:
> On 01 Feb 2021 11:44, Andrew Burgess wrote:
> > Calling printf with the format being a non constant string results in
> > a GCC warning. Use puts instead in a few places.
>
> puts() outputs the string with a trailing newline.
> doesn't seem equiv in this code ?
>
> i change to printf ("%s", ...) to avoid these warnings.
Good point, I think I got confused with fputs :( I made the change
back to printf inline with your suggestion.
Thanks,
Andrew
^ permalink raw reply [flat|nested] 20+ messages in thread
* [PATCH 07/11] sim/rx: move some variable declarations to the start of the block
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
` (5 preceding siblings ...)
2021-02-01 11:44 ` [PATCH 06/11] sim/rx: use puts instead of printf in a few places Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 11:44 ` [PATCH 08/11] sim/rx: use PRIx64 in printf format string Andrew Burgess
` (4 subsequent siblings)
11 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
For sim code variables still need to be declared at the start of the
enclosing block. This silences a few GCC warnings.
sim/rx/ChangeLog:
* syscalls.c (rx_syscall): Move declaration of some variables to
the start of the enclosing block.
* trace.c (load_file_and_line): Likewise.
---
sim/rx/ChangeLog | 6 ++++++
sim/rx/syscalls.c | 5 +++--
sim/rx/trace.c | 3 ++-
3 files changed, 11 insertions(+), 3 deletions(-)
diff --git a/sim/rx/syscalls.c b/sim/rx/syscalls.c
index 2f89da17330..22778b3df49 100644
--- a/sim/rx/syscalls.c
+++ b/sim/rx/syscalls.c
@@ -160,6 +160,7 @@ rx_syscall (int id)
case SYS_open:
{
+ int oflags, cflags;
int path = arg ();
/* The open function is defined as taking a variable number of arguments
because the third parameter to it is optional:
@@ -167,8 +168,8 @@ rx_syscall (int id)
Hence the oflags and cflags arguments will be on the stack and we need
to skip the (empty) argument registers r3 and r4. */
argp = 4;
- int oflags = arg ();
- int cflags = arg ();
+ oflags = arg ();
+ cflags = arg ();
read_target (buf, path, 256, 1);
if (trace)
diff --git a/sim/rx/trace.c b/sim/rx/trace.c
index 6f9294ad75a..9d9f2dfd98d 100644
--- a/sim/rx/trace.c
+++ b/sim/rx/trace.c
@@ -139,6 +139,7 @@ load_file_and_line (const char *filename, int lineno)
break;
if (!f)
{
+ FILE *file;
int i;
struct stat s;
const char *found_filename, *slash;
@@ -159,7 +160,7 @@ load_file_and_line (const char *filename, int lineno)
files = f;
f->filename = strdup (filename);
f->data = (char *) malloc (s.st_size + 2);
- FILE *file = fopen (found_filename, "rb");
+ file = fopen (found_filename, "rb");
fread (f->data, 1, s.st_size, file);
f->data[s.st_size] = 0;
fclose (file);
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* [PATCH 08/11] sim/rx: use PRIx64 in printf format string
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
` (6 preceding siblings ...)
2021-02-01 11:44 ` [PATCH 07/11] sim/rx: move some variable declarations to the start of the block Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 11:44 ` [PATCH 09/11] sim/rx: add some missing includes Andrew Burgess
` (3 subsequent siblings)
11 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
Silence a GCC compiler warning by using PRIx64 in printf format string
instead of hard coded "llx".
sim/rx/ChangeLog:
* reg.c (trace_register_changes): Use PRIx64 in printf format
string.
---
sim/rx/ChangeLog | 5 +++++
sim/rx/reg.c | 4 ++--
2 files changed, 7 insertions(+), 2 deletions(-)
diff --git a/sim/rx/reg.c b/sim/rx/reg.c
index f1bdd4c9841..b84507a7c9f 100644
--- a/sim/rx/reg.c
+++ b/sim/rx/reg.c
@@ -547,9 +547,9 @@ trace_register_changes (void)
if (oldregs.r_acc != regs.r_acc)
{
- printf(" acc %016llx:", oldregs.r_acc);
- printf("%016llx", regs.r_acc);
if (tag) { puts (tag); tag = 0; }
+ printf(" acc %016" PRIx64 ":", oldregs.r_acc);
+ printf("%016" PRIx64, regs.r_acc);
oldregs.r_acc = regs.r_acc;
}
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* [PATCH 09/11] sim/rx: add some missing includes
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
` (7 preceding siblings ...)
2021-02-01 11:44 ` [PATCH 08/11] sim/rx: use PRIx64 in printf format string Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 11:44 ` [PATCH 10/11] sim/rx: avoid pointer arithmetic on void * pointers Andrew Burgess
` (2 subsequent siblings)
11 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
In load.c there's some GCC warnings about undefined
functions (bfd_get_elf_phdr_upper_bound and bfd_get_elf_phdrs). To
get the declarations of these functions include 'elf-bfd.h'. This
headers also pulls in other elf related headers, like 'elf/internal.h'
and 'elf/common.h', so these no longer need to be explicitly included
from load.c.
In trace.c and include for trace.h is missing, again this results in
GCC warnings for missing function declarations.
sim/rx/ChangeLog:
* load.c: Replace 'elf/internal.h' and 'elf/common.h' includes
with 'elf-bfd.h' include.
* trace.c: Add 'trace.h' include.
---
sim/rx/ChangeLog | 6 ++++++
sim/rx/load.c | 3 +--
sim/rx/trace.c | 1 +
3 files changed, 8 insertions(+), 2 deletions(-)
diff --git a/sim/rx/load.c b/sim/rx/load.c
index d74f4ff3872..a8a473346c6 100644
--- a/sim/rx/load.c
+++ b/sim/rx/load.c
@@ -28,8 +28,7 @@ along with this program. If not, see <http://www.gnu.org/licenses/>. */
#include "cpu.h"
#include "mem.h"
#include "load.h"
-#include "elf/internal.h"
-#include "elf/common.h"
+#include "elf-bfd.h"
/* Helper function for invoking a GDB-specified printf. */
static void
diff --git a/sim/rx/trace.c b/sim/rx/trace.c
index 9d9f2dfd98d..e5db49b5ca9 100644
--- a/sim/rx/trace.c
+++ b/sim/rx/trace.c
@@ -34,6 +34,7 @@ along with this program. If not, see <http://www.gnu.org/licenses/>. */
#include "cpu.h"
#include "mem.h"
#include "load.h"
+#include "trace.h"
static int
sim_dis_read (bfd_vma memaddr, bfd_byte * ptr, unsigned int length,
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* [PATCH 10/11] sim/rx: avoid pointer arithmetic on void * pointers
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
` (8 preceding siblings ...)
2021-02-01 11:44 ` [PATCH 09/11] sim/rx: add some missing includes Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 11:44 ` [PATCH 11/11] sim/rx: enable build with warnings Andrew Burgess
2021-02-02 4:37 ` [PATCH 00/11] fix build warnings for rx simulator Mike Frysinger
11 siblings, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
Pointer arithmetic on void * pointers results in a GCC warning. Avoid
the warning by casting the pointer to its actual type earlier in the
function.
sim/rx/ChangeLog:
* mem.c (mem_put_blk): Rename parameter, add cast from parameter
type to local type. Remove cast later in the function.
(mem_get_blk): Likewise.
* mem.h (mem_put_blk): Rename parameter to match definition.
(mem_get_blk): Likewise.
---
sim/rx/ChangeLog | 8 ++++++++
sim/rx/mem.c | 12 ++++++++----
sim/rx/mem.h | 4 ++--
3 files changed, 18 insertions(+), 6 deletions(-)
diff --git a/sim/rx/mem.c b/sim/rx/mem.c
index fe8e08d1460..7e62bfb8953 100644
--- a/sim/rx/mem.c
+++ b/sim/rx/mem.c
@@ -434,13 +434,15 @@ mem_put_si (int address, unsigned long value)
}
void
-mem_put_blk (int address, void *bufptr, int nbytes)
+mem_put_blk (int address, void *bufptr_void, int nbytes)
{
+ unsigned char *bufptr = (unsigned char *) bufptr_void;
+
S ("<=");
if (enable_counting)
mem_counters[1][1] += nbytes;
while (nbytes--)
- mem_put_byte (address++, *(unsigned char *) bufptr++);
+ mem_put_byte (address++, *bufptr++);
E ();
}
@@ -567,13 +569,15 @@ mem_get_si (int address)
}
void
-mem_get_blk (int address, void *bufptr, int nbytes)
+mem_get_blk (int address, void *bufptr_void, int nbytes)
{
+ char *bufptr = (char *) bufptr_void;
+
S ("=>");
if (enable_counting)
mem_counters[0][1] += nbytes;
while (nbytes--)
- *(char *) bufptr++ = mem_get_byte (address++);
+ *bufptr++ = mem_get_byte (address++);
E ();
}
diff --git a/sim/rx/mem.h b/sim/rx/mem.h
index 21080f6e6e3..a4c7455a20d 100644
--- a/sim/rx/mem.h
+++ b/sim/rx/mem.h
@@ -53,7 +53,7 @@ void mem_put_hi (int address, unsigned short value);
void mem_put_psi (int address, unsigned long value);
void mem_put_si (int address, unsigned long value);
-void mem_put_blk (int address, void *bufptr, int nbytes);
+void mem_put_blk (int address, void *bufptr_void, int nbytes);
unsigned char mem_get_pc (int address);
@@ -62,7 +62,7 @@ unsigned short mem_get_hi (int address);
unsigned long mem_get_psi (int address);
unsigned long mem_get_si (int address);
-void mem_get_blk (int address, void *bufptr, int nbytes);
+void mem_get_blk (int address, void *bufptr_void, int nbytes);
int sign_ext (int v, int bits);
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* [PATCH 11/11] sim/rx: enable build with warnings
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
` (9 preceding siblings ...)
2021-02-01 11:44 ` [PATCH 10/11] sim/rx: avoid pointer arithmetic on void * pointers Andrew Burgess
@ 2021-02-01 11:44 ` Andrew Burgess
2021-02-01 14:50 ` Mike Frysinger
2021-02-02 4:37 ` [PATCH 00/11] fix build warnings for rx simulator Mike Frysinger
11 siblings, 1 reply; 20+ messages in thread
From: Andrew Burgess @ 2021-02-01 11:44 UTC (permalink / raw)
To: gdb-patches
The rx simulator now has no build warnings, so enable build with
warnings by default.
sim/rx/ChangeLog:
* configure: Regenerate.
* configure.ac (SIM_AC_OPTION_WARNINGS): Pass yes.
---
sim/rx/ChangeLog | 5 +++++
sim/rx/configure | 17 ++++++++++++++---
sim/rx/configure.ac | 2 +-
3 files changed, 20 insertions(+), 4 deletions(-)
diff --git a/sim/rx/configure b/sim/rx/configure
index 3969c670380..8501f84e727 100755
--- a/sim/rx/configure
+++ b/sim/rx/configure
@@ -6903,7 +6903,15 @@ fi
test -z "$AR" && AR=ar
if test -n "$plugin_option"; then
if $AR --help 2>&1 | grep -q "\--plugin"; then
- AR="$AR $plugin_option"
+ touch conftest.c
+ $AR $plugin_option rc conftest.a conftest.c
+ if test "$?" != 0; then
+ { $as_echo "$as_me:${as_lineno-$LINENO}: WARNING: Failed: $AR $plugin_option rc" >&5
+$as_echo "$as_me: WARNING: Failed: $AR $plugin_option rc" >&2;}
+ else
+ AR="$AR $plugin_option"
+ fi
+ rm -f conftest.*
fi
fi
test -z "$AR_FLAGS" && AR_FLAGS=cru
@@ -12927,7 +12935,7 @@ else
lt_dlunknown=0; lt_dlno_uscore=1; lt_dlneed_uscore=2
lt_status=$lt_dlunknown
cat > conftest.$ac_ext <<_LT_EOF
-#line 12930 "configure"
+#line 12938 "configure"
#include "confdefs.h"
#if HAVE_DLFCN_H
@@ -13033,7 +13041,7 @@ else
lt_dlunknown=0; lt_dlno_uscore=1; lt_dlneed_uscore=2
lt_status=$lt_dlunknown
cat > conftest.$ac_ext <<_LT_EOF
-#line 13036 "configure"
+#line 13044 "configure"
#include "confdefs.h"
#if HAVE_DLFCN_H
@@ -13600,6 +13608,9 @@ if test "${GCC}" = yes -a -z "${ERROR_ON_WARNING}" ; then
fi
WERROR_CFLAGS=""
+ if test "${ERROR_ON_WARNING}" = yes ; then
+ WERROR_CFLAGS="-Werror"
+ fi
build_warnings="-Wall -Wdeclaration-after-statement -Wpointer-arith \
-Wpointer-sign \
diff --git a/sim/rx/configure.ac b/sim/rx/configure.ac
index 6c14434f4af..b81c322d441 100644
--- a/sim/rx/configure.ac
+++ b/sim/rx/configure.ac
@@ -22,7 +22,7 @@ AC_INIT(Makefile.in)
sinclude(../common/acinclude.m4)
SIM_AC_COMMON
-SIM_AC_OPTION_WARNINGS(no)
+SIM_AC_OPTION_WARNINGS(yes)
AC_CHECK_HEADERS(getopt.h)
--
2.25.4
^ permalink raw reply [flat|nested] 20+ messages in thread* Re: [PATCH 11/11] sim/rx: enable build with warnings
2021-02-01 11:44 ` [PATCH 11/11] sim/rx: enable build with warnings Andrew Burgess
@ 2021-02-01 14:50 ` Mike Frysinger
2021-02-02 4:53 ` [PATCH] sim: erc32/m32c/rl78: add sim_memory_map stub for gdb Mike Frysinger
2021-02-08 10:57 ` [PATCH 11/11] sim/rx: enable build with warnings Andrew Burgess
0 siblings, 2 replies; 20+ messages in thread
From: Mike Frysinger @ 2021-02-01 14:50 UTC (permalink / raw)
To: Andrew Burgess; +Cc: gdb-patches
On 01 Feb 2021 11:44, Andrew Burgess wrote:
> The rx simulator now has no build warnings, so enable build with
> warnings by default.
>
> --- a/sim/rx/configure.ac
> +++ b/sim/rx/configure.ac
> @@ -22,7 +22,7 @@ AC_INIT(Makefile.in)
> sinclude(../common/acinclude.m4)
>
> SIM_AC_COMMON
> -SIM_AC_OPTION_WARNINGS(no)
> +SIM_AC_OPTION_WARNINGS(yes)
i would just delete the call while you're here. SIM_AC_OUTPUT will
call the func with (yes) as the arg for you.
-mike
^ permalink raw reply [flat|nested] 20+ messages in thread
* [PATCH] sim: erc32/m32c/rl78: add sim_memory_map stub for gdb
2021-02-01 14:50 ` Mike Frysinger
@ 2021-02-02 4:53 ` Mike Frysinger
2021-02-08 10:57 ` [PATCH 11/11] sim/rx: enable build with warnings Andrew Burgess
1 sibling, 0 replies; 20+ messages in thread
From: Mike Frysinger @ 2021-02-02 4:53 UTC (permalink / raw)
To: gdb-patches
These ports don't use the common sim core, so they weren't providing
a sim_memory_map for gdb, so they failed to link with the new memory
map logic added for the sim. Add stubs to fix.
erc32/
* interf.c (sim_memory_map): Define.
m32c/
* gdb-if.c (sim_memory_map): Define.
rl78/
* gdb-if.c (sim_memory_map): Define.
---
sim/erc32/interf.c | 6 ++++++
sim/m32c/gdb-if.c | 6 ++++++
sim/rl78/gdb-if.c | 6 ++++++
3 files changed, 18 insertions(+)
diff --git a/sim/erc32/interf.c b/sim/erc32/interf.c
index 5dd464b0ce76..ca2e9faea217 100644
--- a/sim/erc32/interf.c
+++ b/sim/erc32/interf.c
@@ -465,6 +465,12 @@ sim_complete_command (SIM_DESC sd, const char *text, const char *word)
return NULL;
}
+char *
+sim_memory_map (SIM_DESC sd)
+{
+ return NULL;
+}
+
#if 0 /* FIXME: These shouldn't exist. */
int
diff --git a/sim/m32c/gdb-if.c b/sim/m32c/gdb-if.c
index 5ada0b0bc17e..92e447f17faa 100644
--- a/sim/m32c/gdb-if.c
+++ b/sim/m32c/gdb-if.c
@@ -707,6 +707,12 @@ sim_complete_command (SIM_DESC sd, const char *text, const char *word)
return NULL;
}
+char *
+sim_memory_map (SIM_DESC sd)
+{
+ return NULL;
+}
+
void
sim_info (SIM_DESC sd, int verbose)
{
diff --git a/sim/rl78/gdb-if.c b/sim/rl78/gdb-if.c
index 320db34f436a..56717917e552 100644
--- a/sim/rl78/gdb-if.c
+++ b/sim/rl78/gdb-if.c
@@ -603,3 +603,9 @@ sim_complete_command (SIM_DESC sd, const char *text, const char *word)
{
return NULL;
}
+
+char *
+sim_memory_map (SIM_DESC sd)
+{
+ return NULL;
+}
--
2.30.0
^ permalink raw reply [flat|nested] 20+ messages in thread* Re: [PATCH 11/11] sim/rx: enable build with warnings
2021-02-01 14:50 ` Mike Frysinger
2021-02-02 4:53 ` [PATCH] sim: erc32/m32c/rl78: add sim_memory_map stub for gdb Mike Frysinger
@ 2021-02-08 10:57 ` Andrew Burgess
1 sibling, 0 replies; 20+ messages in thread
From: Andrew Burgess @ 2021-02-08 10:57 UTC (permalink / raw)
To: gdb-patches
* Mike Frysinger <vapier@gentoo.org> [2021-02-01 09:50:55 -0500]:
> On 01 Feb 2021 11:44, Andrew Burgess wrote:
> > The rx simulator now has no build warnings, so enable build with
> > warnings by default.
> >
> > --- a/sim/rx/configure.ac
> > +++ b/sim/rx/configure.ac
> > @@ -22,7 +22,7 @@ AC_INIT(Makefile.in)
> > sinclude(../common/acinclude.m4)
> >
> > SIM_AC_COMMON
> > -SIM_AC_OPTION_WARNINGS(no)
> > +SIM_AC_OPTION_WARNINGS(yes)
>
> i would just delete the call while you're here. SIM_AC_OUTPUT will
> call the func with (yes) as the arg for you.
Thanks, I made this change.
Andrew
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [PATCH 00/11] fix build warnings for rx simulator
2021-02-01 11:44 [PATCH 00/11] fix build warnings for rx simulator Andrew Burgess
` (10 preceding siblings ...)
2021-02-01 11:44 ` [PATCH 11/11] sim/rx: enable build with warnings Andrew Burgess
@ 2021-02-02 4:37 ` Mike Frysinger
11 siblings, 0 replies; 20+ messages in thread
From: Mike Frysinger @ 2021-02-02 4:37 UTC (permalink / raw)
To: Andrew Burgess; +Cc: gdb-patches
On 01 Feb 2021 11:44, Andrew Burgess wrote:
> Resolve all the build warnings for the rx simulator and enable build
> warnings by default.
all the ones i didn't comment on already look good!
-mike
^ permalink raw reply [flat|nested] 20+ messages in thread