* [PATCH] tracepoint: add new trace command "printf"[1] gdbserver
@ 2011-01-03 16:29 Hui Zhu
2011-01-28 6:41 ` Hui Zhu
0 siblings, 1 reply; 8+ messages in thread
From: Hui Zhu @ 2011-01-03 16:29 UTC (permalink / raw)
To: gdb-patches
[-- Attachment #1: Type: text/plain, Size: 247 bytes --]
This patch is for the gdbserver.
2011-01-04 Hui Zhu <teawater@gmail.com>
* tracepoint.c (gdb_agent_op): Add gdb_agent_op_printf.
(gdb_agent_op_names): Add "printf".
(tp_printf): New function.
(eval_agent_expr): Handle gdb_agent_op_printf.
[-- Attachment #2: tp_print_server.txt --]
[-- Type: text/plain, Size: 1465 bytes --]
---
gdbserver/tracepoint.c | 28 ++++++++++++++++++++++++++++
1 file changed, 28 insertions(+)
--- a/gdbserver/tracepoint.c
+++ b/gdbserver/tracepoint.c
@@ -515,6 +515,7 @@ enum gdb_agent_op
gdb_agent_op_setv = 0x2d,
gdb_agent_op_tracev = 0x2e,
gdb_agent_op_trace16 = 0x30,
+ gdb_agent_op_printf = 0x31,
gdb_agent_op_last
};
@@ -569,6 +570,7 @@ static const char *gdb_agent_op_names [g
"tracev",
"?undef?",
"trace16",
+ "printf",
};
struct agent_expr
@@ -4288,6 +4290,16 @@ unparse_agent_expr (struct agent_expr *a
#endif
+int
+tp_printf(const char *format, ...)
+{
+ va_list ap;
+ va_start (ap, format);
+ vprintf (format, ap);
+ va_end(ap);
+ return 0;
+}
+
/* The agent expression evaluator, as specified by the GDB docs. It
returns 0 if everything went OK, and a nonzero error code
otherwise. */
@@ -4630,6 +4642,22 @@ eval_agent_expr (struct tracepoint_hit_c
agent_tsv_read (tframe, arg);
break;
+ case gdb_agent_op_printf:
+ {
+ void *argv;
+ arg = aexpr->bytes[pc++];
+ argv = (void *)top;
+ if (--sp >= 0)
+ top = stack[sp];
+
+ if (arg)
+ tp_printf ((char *)(aexpr->bytes + pc), argv);
+ else
+ tp_printf ((char *)(aexpr->bytes + pc));
+ pc += strlen ((char *)aexpr->bytes + pc) + 1;
+ }
+ break;
+
/* GDB never (currently) generates any of these ops. */
case gdb_agent_op_float:
case gdb_agent_op_ref_float:
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tracepoint: add new trace command "printf"[1] gdbserver
2011-01-03 16:29 [PATCH] tracepoint: add new trace command "printf"[1] gdbserver Hui Zhu
@ 2011-01-28 6:41 ` Hui Zhu
2011-02-04 16:01 ` Hui Zhu
0 siblings, 1 reply; 8+ messages in thread
From: Hui Zhu @ 2011-01-28 6:41 UTC (permalink / raw)
To: gdb-patches
[-- Attachment #1: Type: text/plain, Size: 630 bytes --]
On Tue, Jan 4, 2011 at 00:29, Hui Zhu <teawater@gmail.com> wrote:
> This patch is for the gdbserver.
>
> 2011-01-04 Hui Zhu <teawater@gmail.com>
>
> * tracepoint.c (gdb_agent_op): Add gdb_agent_op_printf.
> (gdb_agent_op_names): Add "printf".
> (tp_printf): New function.
> (eval_agent_expr): Handle gdb_agent_op_printf.
>
Fix a build error in i386.
Thanks,
Hui
2011-01-28 Hui Zhu <teawater@gmail.com>
* tracepoint.c (gdb_agent_op): Add gdb_agent_op_printf.
(gdb_agent_op_names): Add "printf".
(tp_printf): New function.
(eval_agent_expr): Handle gdb_agent_op_printf.
[-- Attachment #2: tp_print_server.txt --]
[-- Type: text/plain, Size: 1480 bytes --]
---
gdbserver/tracepoint.c | 28 ++++++++++++++++++++++++++++
1 file changed, 28 insertions(+)
--- a/gdbserver/tracepoint.c
+++ b/gdbserver/tracepoint.c
@@ -517,6 +517,7 @@ enum gdb_agent_op
gdb_agent_op_setv = 0x2d,
gdb_agent_op_tracev = 0x2e,
gdb_agent_op_trace16 = 0x30,
+ gdb_agent_op_printf = 0x31,
gdb_agent_op_last
};
@@ -571,6 +572,7 @@ static const char *gdb_agent_op_names [g
"tracev",
"?undef?",
"trace16",
+ "printf",
};
struct agent_expr
@@ -4292,6 +4294,16 @@ unparse_agent_expr (struct agent_expr *a
#endif
+int
+tp_printf(const char *format, ...)
+{
+ va_list ap;
+ va_start (ap, format);
+ vprintf (format, ap);
+ va_end(ap);
+ return 0;
+}
+
/* The agent expression evaluator, as specified by the GDB docs. It
returns 0 if everything went OK, and a nonzero error code
otherwise. */
@@ -4634,6 +4646,22 @@ eval_agent_expr (struct tracepoint_hit_c
agent_tsv_read (tframe, arg);
break;
+ case gdb_agent_op_printf:
+ {
+ void *argv;
+ arg = aexpr->bytes[pc++];
+ argv = (void *)(unsigned long)top;
+ if (--sp >= 0)
+ top = stack[sp];
+
+ if (arg)
+ tp_printf ((char *)(aexpr->bytes + pc), argv);
+ else
+ tp_printf ((char *)(aexpr->bytes + pc));
+ pc += strlen ((char *)aexpr->bytes + pc) + 1;
+ }
+ break;
+
/* GDB never (currently) generates any of these ops. */
case gdb_agent_op_float:
case gdb_agent_op_ref_float:
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tracepoint: add new trace command "printf"[1] gdbserver
2011-01-28 6:41 ` Hui Zhu
@ 2011-02-04 16:01 ` Hui Zhu
2011-02-11 21:31 ` Tom Tromey
0 siblings, 1 reply; 8+ messages in thread
From: Hui Zhu @ 2011-02-04 16:01 UTC (permalink / raw)
To: gdb-patches
[-- Attachment #1: Type: text/plain, Size: 261 bytes --]
To make the gdbserver support %s.
Thanks,
Hui
2011-01-28 Hui Zhu <teawater@gmail.com>
* tracepoint.c (gdb_agent_op): Add gdb_agent_op_printf.
(gdb_agent_op_names): Add "printf".
(tp_printf): New function.
(eval_agent_expr): Handle gdb_agent_op_printf.
[-- Attachment #2: tp_print_server.txt --]
[-- Type: text/plain, Size: 1874 bytes --]
---
gdbserver/tracepoint.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 46 insertions(+)
--- a/gdbserver/tracepoint.c
+++ b/gdbserver/tracepoint.c
@@ -517,6 +517,7 @@ enum gdb_agent_op
gdb_agent_op_setv = 0x2d,
gdb_agent_op_tracev = 0x2e,
gdb_agent_op_trace16 = 0x30,
+ gdb_agent_op_printf = 0x31,
gdb_agent_op_last
};
@@ -571,6 +572,7 @@ static const char *gdb_agent_op_names [g
"tracev",
"?undef?",
"trace16",
+ "printf",
};
struct agent_expr
@@ -4292,6 +4294,16 @@ unparse_agent_expr (struct agent_expr *a
#endif
+int
+tp_printf(const char *format, ...)
+{
+ va_list ap;
+ va_start (ap, format);
+ vprintf (format, ap);
+ va_end(ap);
+ return 0;
+}
+
/* The agent expression evaluator, as specified by the GDB docs. It
returns 0 if everything went OK, and a nonzero error code
otherwise. */
@@ -4634,6 +4646,40 @@ eval_agent_expr (struct tracepoint_hit_c
agent_tsv_read (tframe, arg);
break;
+ case gdb_agent_op_printf:
+ {
+ void *argv;
+ arg = aexpr->bytes[pc++];
+ argv = (void *)(unsigned long)top;
+ if (--sp >= 0)
+ top = stack[sp];
+
+ if (arg)
+ {
+ if (strstr ((char *)(aexpr->bytes + pc), "%s"))
+ {
+ int i;
+ unsigned char buf[100];
+
+ for (i = 0; i < 100; i++)
+ {
+ agent_mem_read (tframe, buf + i,
+ (CORE_ADDR) ((unsigned long)argv + i),
+ 1);
+ if (!buf[i])
+ break;
+ }
+ tp_printf ((char *)(aexpr->bytes + pc), buf);
+ }
+ else
+ tp_printf ((char *)(aexpr->bytes + pc), argv);
+ }
+ else
+ tp_printf ((char *)(aexpr->bytes + pc));
+ pc += strlen ((char *)aexpr->bytes + pc) + 1;
+ }
+ break;
+
/* GDB never (currently) generates any of these ops. */
case gdb_agent_op_float:
case gdb_agent_op_ref_float:
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tracepoint: add new trace command "printf"[1] gdbserver
2011-02-04 16:01 ` Hui Zhu
@ 2011-02-11 21:31 ` Tom Tromey
2011-02-17 9:54 ` Hui Zhu
0 siblings, 1 reply; 8+ messages in thread
From: Tom Tromey @ 2011-02-11 21:31 UTC (permalink / raw)
To: Hui Zhu; +Cc: gdb-patches
>>>>> ">" == Hui Zhu <teawater@gmail.com> writes:
>> To make the gdbserver support %s.
Oops, I missed this when reviewing your earlier patch.
Sorry about that.
>> +int
>> +tp_printf(const char *format, ...)
Space before open paren.
>> + va_end(ap);
Likewise.
>> + argv = (void *)(unsigned long)top;
Spacing.
There are several more spacing bugs, please fix them all.
I didn't go read the other patch to try to deduce the format of the
printf opcode to see if it all makes sense. Documentation would help.
Tom
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tracepoint: add new trace command "printf"[1] gdbserver
2011-02-11 21:31 ` Tom Tromey
@ 2011-02-17 9:54 ` Hui Zhu
2011-02-21 8:18 ` Hui Zhu
0 siblings, 1 reply; 8+ messages in thread
From: Hui Zhu @ 2011-02-17 9:54 UTC (permalink / raw)
To: Tom Tromey; +Cc: gdb-patches
[-- Attachment #1: Type: text/plain, Size: 960 bytes --]
On Sat, Feb 12, 2011 at 05:31, Tom Tromey <tromey@redhat.com> wrote:
>>>>>> ">" == Hui Zhu <teawater@gmail.com> writes:
>
>>> To make the gdbserver support %s.
>
> Oops, I missed this when reviewing your earlier patch.
> Sorry about that.
>
>>> +int
>>> +tp_printf(const char *format, ...)
>
> Space before open paren.
>
>>> + va_end(ap);
>
> Likewise.
>
>>> + argv = (void *)(unsigned long)top;
>
> Spacing.
>
> There are several more spacing bugs, please fix them all.
>
> I didn't go read the other patch to try to deduce the format of the
> printf opcode to see if it all makes sense. Documentation would help.
>
> Tom
>
Hi Tom,
I make a new patch according to your comments.
Thanks,
Hui
2011-02-17 Hui Zhu <teawater@gmail.com>
* tracepoint.c (gdb_agent_op): Add gdb_agent_op_printf.
(gdb_agent_op_names): Add "printf".
(tp_printf): New function.
(eval_agent_expr): Handle gdb_agent_op_printf.
[-- Attachment #2: tp_print_server.txt --]
[-- Type: text/plain, Size: 1883 bytes --]
---
gdbserver/tracepoint.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 46 insertions(+)
--- a/gdbserver/tracepoint.c
+++ b/gdbserver/tracepoint.c
@@ -517,6 +517,7 @@ enum gdb_agent_op
gdb_agent_op_setv = 0x2d,
gdb_agent_op_tracev = 0x2e,
gdb_agent_op_trace16 = 0x30,
+ gdb_agent_op_printf = 0x31,
gdb_agent_op_last
};
@@ -571,6 +572,7 @@ static const char *gdb_agent_op_names [g
"tracev",
"?undef?",
"trace16",
+ "printf",
};
struct agent_expr
@@ -4292,6 +4294,16 @@ unparse_agent_expr (struct agent_expr *a
#endif
+int
+tp_printf (const char *format, ...)
+{
+ va_list ap;
+ va_start (ap, format);
+ vprintf (format, ap);
+ va_end (ap);
+ return 0;
+}
+
/* The agent expression evaluator, as specified by the GDB docs. It
returns 0 if everything went OK, and a nonzero error code
otherwise. */
@@ -4634,6 +4646,40 @@ eval_agent_expr (struct tracepoint_hit_c
agent_tsv_read (tframe, arg);
break;
+ case gdb_agent_op_printf:
+ {
+ void *argv;
+ arg = aexpr->bytes[pc++];
+ argv = (void *) (unsigned long) top;
+ if (--sp >= 0)
+ top = stack[sp];
+
+ if (arg)
+ {
+ if (strstr ((char *) (aexpr->bytes + pc), "%s"))
+ {
+ int i;
+ unsigned char buf[100];
+
+ for (i = 0; i < 100; i++)
+ {
+ agent_mem_read (tframe, buf + i,
+ (CORE_ADDR) ((unsigned long)argv + i),
+ 1);
+ if (!buf[i])
+ break;
+ }
+ tp_printf ((char *) (aexpr->bytes + pc), buf);
+ }
+ else
+ tp_printf ((char *) (aexpr->bytes + pc), argv);
+ }
+ else
+ tp_printf ((char *) (aexpr->bytes + pc));
+ pc += strlen ((char *) aexpr->bytes + pc) + 1;
+ }
+ break;
+
/* GDB never (currently) generates any of these ops. */
case gdb_agent_op_float:
case gdb_agent_op_ref_float:
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tracepoint: add new trace command "printf"[1] gdbserver
2011-02-17 9:54 ` Hui Zhu
@ 2011-02-21 8:18 ` Hui Zhu
2011-02-24 7:46 ` Joel Brobecker
0 siblings, 1 reply; 8+ messages in thread
From: Hui Zhu @ 2011-02-21 8:18 UTC (permalink / raw)
To: Tom Tromey; +Cc: gdb-patches
[-- Attachment #1: Type: text/plain, Size: 1278 bytes --]
On Thu, Feb 17, 2011 at 16:16, Hui Zhu <teawater@gmail.com> wrote:
> On Sat, Feb 12, 2011 at 05:31, Tom Tromey <tromey@redhat.com> wrote:
>>>>>>> ">" == Hui Zhu <teawater@gmail.com> writes:
>>
>>>> To make the gdbserver support %s.
>>
>> Oops, I missed this when reviewing your earlier patch.
>> Sorry about that.
>>
>>>> +int
>>>> +tp_printf(const char *format, ...)
>>
>> Space before open paren.
>>
>>>> + va_end(ap);
>>
>> Likewise.
>>
>>>> + argv = (void *)(unsigned long)top;
>>
>> Spacing.
>>
>> There are several more spacing bugs, please fix them all.
>>
>> I didn't go read the other patch to try to deduce the format of the
>> printf opcode to see if it all makes sense. Documentation would help.
>>
>> Tom
>>
>
> Hi Tom,
>
> I make a new patch according to your comments.
>
> Thanks,
> Hui
>
> 2011-02-17 Hui Zhu <teawater@gmail.com>
>
> * tracepoint.c (gdb_agent_op): Add gdb_agent_op_printf.
> (gdb_agent_op_names): Add "printf".
> (tp_printf): New function.
> (eval_agent_expr): Handle gdb_agent_op_printf.
>
Checked in.
Thanks,
Hui
2011-02-21 Hui Zhu <teawater@gmail.com>
* tracepoint.c (tp_printf): New function.
(eval_agent_expr): Handle gdb_agent_op_printf.
[-- Attachment #2: tp_print_server.txt --]
[-- Type: text/plain, Size: 1544 bytes --]
---
gdbserver/tracepoint.c | 44 ++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 44 insertions(+)
--- a/gdbserver/tracepoint.c
+++ b/gdbserver/tracepoint.c
@@ -4214,6 +4214,16 @@ gdb_agent_op_name (int op)
return gdb_agent_op_names[op];
}
+int
+tp_printf (const char *format, ...)
+{
+ va_list ap;
+ va_start (ap, format);
+ vprintf (format, ap);
+ va_end (ap);
+ return 0;
+}
+
/* The agent expression evaluator, as specified by the GDB docs. It
returns 0 if everything went OK, and a nonzero error code
otherwise. */
@@ -4573,6 +4583,40 @@ eval_agent_expr (struct tracepoint_hit_c
agent_tsv_read (tframe, arg);
break;
+ case gdb_agent_op_printf:
+ {
+ void *argv;
+ arg = aexpr->bytes[pc++];
+ argv = (void *) (unsigned long) top;
+ if (--sp >= 0)
+ top = stack[sp];
+
+ if (arg)
+ {
+ if (strstr ((char *) (aexpr->bytes + pc), "%s"))
+ {
+ int i;
+ unsigned char buf[100];
+
+ for (i = 0; i < 100; i++)
+ {
+ agent_mem_read (tframe, buf + i,
+ (CORE_ADDR) ((unsigned long)argv + i),
+ 1);
+ if (!buf[i])
+ break;
+ }
+ tp_printf ((char *) (aexpr->bytes + pc), buf);
+ }
+ else
+ tp_printf ((char *) (aexpr->bytes + pc), argv);
+ }
+ else
+ tp_printf ((char *) (aexpr->bytes + pc));
+ pc += strlen ((char *) aexpr->bytes + pc) + 1;
+ }
+ break;
+
/* GDB never (currently) generates any of these ops. */
case gdb_agent_op_float:
case gdb_agent_op_ref_float:
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tracepoint: add new trace command "printf"[1] gdbserver
2011-02-21 8:18 ` Hui Zhu
@ 2011-02-24 7:46 ` Joel Brobecker
2011-02-24 8:37 ` Hui Zhu
0 siblings, 1 reply; 8+ messages in thread
From: Joel Brobecker @ 2011-02-24 7:46 UTC (permalink / raw)
To: Hui Zhu; +Cc: gdb-patches
> * tracepoint.c (tp_printf): New function.
> (eval_agent_expr): Handle gdb_agent_op_printf.
For the record, this patch and the other two associated patches
have been temporarily reverted, as they hadn't actually been approved.
Hopefully we'll have a change to review them soon.
--
Joel
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tracepoint: add new trace command "printf"[1] gdbserver
2011-02-24 7:46 ` Joel Brobecker
@ 2011-02-24 8:37 ` Hui Zhu
0 siblings, 0 replies; 8+ messages in thread
From: Hui Zhu @ 2011-02-24 8:37 UTC (permalink / raw)
To: Joel Brobecker; +Cc: gdb-patches
Hi Joel,
Thanks for your help.
Best,
Hui
On Thu, Feb 24, 2011 at 15:44, Joel Brobecker <brobecker@adacore.com> wrote:
>> * tracepoint.c (tp_printf): New function.
>> (eval_agent_expr): Handle gdb_agent_op_printf.
>
> For the record, this patch and the other two associated patches
> have been temporarily reverted, as they hadn't actually been approved.
> Hopefully we'll have a change to review them soon.
>
> --
> Joel
>
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2011-02-24 7:46 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-01-03 16:29 [PATCH] tracepoint: add new trace command "printf"[1] gdbserver Hui Zhu
2011-01-28 6:41 ` Hui Zhu
2011-02-04 16:01 ` Hui Zhu
2011-02-11 21:31 ` Tom Tromey
2011-02-17 9:54 ` Hui Zhu
2011-02-21 8:18 ` Hui Zhu
2011-02-24 7:46 ` Joel Brobecker
2011-02-24 8:37 ` Hui Zhu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).