[Qemu-devel] [PATCH] Some fixes for TCG debugging

[Qemu-devel] [PATCH] Some fixes for TCG debugging

[Laurent Desnogues]

[-- Attachment #1: Type: text/plain, Size: 301 bytes --]

Hello,

this fixes a few things after Paul's improvements for TCG debugging:

   - change TCGv_i64 field name to something different from
     TCGv_i32
   - fix things in tcg that the above change made visible.

I am currently compiling all targets to see if this reveals some other
errors.


Laurent

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: tcg-debug.patch --]
[-- Type: text/x-patch; name=tcg-debug.patch, Size: 1847 bytes --]

Index: tcg/tcg-op.h
===================================================================
--- tcg/tcg-op.h	(revision 5733)
+++ tcg/tcg-op.h	(working copy)
@@ -132,7 +132,7 @@
                                        TCGArg offset)
 {
     *gen_opc_ptr++ = opc;
-    *gen_opparam_ptr++ = GET_TCGV_I32(val);
+    *gen_opparam_ptr++ = GET_TCGV_I64(val);
     *gen_opparam_ptr++ = GET_TCGV_PTR(base);
     *gen_opparam_ptr++ = offset;
 }
@@ -166,7 +166,7 @@
 }
 
 static inline void tcg_gen_op4_i64(int opc, TCGv_i64 arg1, TCGv_i64 arg2,
-                                   TCGv_i64 arg3, TCGv_i32 arg4)
+                                   TCGv_i64 arg3, TCGv_i64 arg4)
 {
     *gen_opc_ptr++ = opc;
     *gen_opparam_ptr++ = GET_TCGV_I64(arg1);
Index: tcg/tcg.c
===================================================================
--- tcg/tcg.c	(revision 5733)
+++ tcg/tcg.c	(working copy)
@@ -746,7 +746,7 @@
 
 char *tcg_get_arg_str_i64(TCGContext *s, char *buf, int buf_size, TCGv_i64 arg)
 {
-    return tcg_get_arg_str_idx(s, buf, buf_size, GET_TCGV_I32(arg));
+    return tcg_get_arg_str_idx(s, buf, buf_size, GET_TCGV_I64(arg));
 }
 
 static int helper_cmp(const void *p1, const void *p2)
Index: tcg/tcg.h
===================================================================
--- tcg/tcg.h	(revision 5733)
+++ tcg/tcg.h	(working copy)
@@ -131,7 +131,7 @@
 
 typedef struct
 {
-    int n;
+    int n2;
 } TCGv_i64;
 
 #define MAKE_TCGV_I32(i) __extension__                  \
@@ -139,7 +139,7 @@
 #define MAKE_TCGV_I64(i) __extension__                  \
     ({ TCGv_i64 make_tcgv_tmp = {i}; make_tcgv_tmp;})
 #define GET_TCGV_I32(t) ((t).n)
-#define GET_TCGV_I64(t) ((t).n)
+#define GET_TCGV_I64(t) ((t).n2)
 #if TCG_TARGET_REG_BITS == 32
 #define TCGV_LOW(t) MAKE_TCGV_I32(GET_TCGV_I64(t))
 #define TCGV_HIGH(t) MAKE_TCGV_I32(GET_TCGV_I64(t) + 1)

[Qemu-devel] Re: [PATCH] Some fixes for TCG debugging

[Laurent Desnogues]

On Tue, Nov 18, 2008 at 4:02 PM, Laurent Desnogues
<laurent.desnogues@gmail.com> wrote:
> Hello,
>
> this fixes a few things after Paul's improvements for TCG debugging:
>
>   - change TCGv_i64 field name to something different from
>     TCGv_i32
>   - fix things in tcg that the above change made visible.

No comment about that patch?  Except that it lacks this:

Signed-off-by: Laurent Desnogues <laurent.desnogues@gmail.com>

Thanks,

Laurent

[Qemu-devel] Re: [PATCH] Some fixes for TCG debugging

[Laurent Desnogues]

On Thu, Nov 20, 2008 at 1:11 PM, Laurent Desnogues
<laurent.desnogues@gmail.com> wrote:
> On Tue, Nov 18, 2008 at 4:02 PM, Laurent Desnogues
> <laurent.desnogues@gmail.com> wrote:
>> Hello,
>>
>> this fixes a few things after Paul's improvements for TCG debugging:
>>
>>   - change TCGv_i64 field name to something different from
>>     TCGv_i32
>>   - fix things in tcg that the above change made visible.
>
> No comment about that patch?  Except that it lacks this:
>
> Signed-off-by: Laurent Desnogues <laurent.desnogues@gmail.com>

Still no comment for something that obvious?


Laurent

Re: [Qemu-devel] Re: [PATCH] Some fixes for TCG debugging

[Blue Swirl]

On 11/29/08, Laurent Desnogues <laurent.desnogues@gmail.com> wrote:
> On Thu, Nov 20, 2008 at 1:11 PM, Laurent Desnogues
>
> <laurent.desnogues@gmail.com> wrote:
>  > On Tue, Nov 18, 2008 at 4:02 PM, Laurent Desnogues
>  > <laurent.desnogues@gmail.com> wrote:
>  >> Hello,
>  >>
>  >> this fixes a few things after Paul's improvements for TCG debugging:
>  >>
>  >>   - change TCGv_i64 field name to something different from
>  >>     TCGv_i32
>  >>   - fix things in tcg that the above change made visible.
>  >
>  > No comment about that patch?  Except that it lacks this:
>  >
>  > Signed-off-by: Laurent Desnogues <laurent.desnogues@gmail.com>
>
>
> Still no comment for something that obvious?

Looks OK. Anybody mind if I apply this?

Re: [Qemu-devel] Re: [PATCH] Some fixes for TCG debugging

[Paul Brook]

On Saturday 29 November 2008, Blue Swirl wrote:
> On 11/29/08, Laurent Desnogues <laurent.desnogues@gmail.com> wrote:
> > On Thu, Nov 20, 2008 at 1:11 PM, Laurent Desnogues
> >
> > <laurent.desnogues@gmail.com> wrote:
> >  > On Tue, Nov 18, 2008 at 4:02 PM, Laurent Desnogues
> >  >
> >  > <laurent.desnogues@gmail.com> wrote:
> >  >> Hello,
> >  >>
> >  >> this fixes a few things after Paul's improvements for TCG debugging:
> >  >>
> >  >>   - change TCGv_i64 field name to something different from
> >  >>     TCGv_i32
> >  >>   - fix things in tcg that the above change made visible.
> >  >
> >  > No comment about that patch?  Except that it lacks this:
> >  >
> >  > Signed-off-by: Laurent Desnogues <laurent.desnogues@gmail.com>
> >
> > Still no comment for something that obvious?
>
> Looks OK. Anybody mind if I apply this?

IMHO it would make sense to call them i32/i64, thather than i/i2.

Paul

Re: [Qemu-devel] Re: [PATCH] Some fixes for TCG debugging

[Laurent Desnogues]

[-- Attachment #1: Type: text/plain, Size: 263 bytes --]

On Sat, Nov 29, 2008 at 5:47 PM, Paul Brook <paul@codesourcery.com> wrote:
> IMHO it would make sense to call them i32/i64, thather than i/i2.

I think the same.  Here is an updated patch.


Laurent

Signed-off-by: Laurent Desnogues <laurent.desnogues@gmail.com>

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: tcg-debug2.patch --]
[-- Type: text/x-patch; name=tcg-debug2.patch, Size: 1981 bytes --]

Index: tcg/tcg-op.h
===================================================================
--- tcg/tcg-op.h	(revision 5809)
+++ tcg/tcg-op.h	(working copy)
@@ -132,7 +132,7 @@
                                        TCGArg offset)
 {
     *gen_opc_ptr++ = opc;
-    *gen_opparam_ptr++ = GET_TCGV_I32(val);
+    *gen_opparam_ptr++ = GET_TCGV_I64(val);
     *gen_opparam_ptr++ = GET_TCGV_PTR(base);
     *gen_opparam_ptr++ = offset;
 }
@@ -166,7 +166,7 @@
 }
 
 static inline void tcg_gen_op4_i64(int opc, TCGv_i64 arg1, TCGv_i64 arg2,
-                                   TCGv_i64 arg3, TCGv_i32 arg4)
+                                   TCGv_i64 arg3, TCGv_i64 arg4)
 {
     *gen_opc_ptr++ = opc;
     *gen_opparam_ptr++ = GET_TCGV_I64(arg1);
Index: tcg/tcg.c
===================================================================
--- tcg/tcg.c	(revision 5809)
+++ tcg/tcg.c	(working copy)
@@ -746,7 +746,7 @@
 
 char *tcg_get_arg_str_i64(TCGContext *s, char *buf, int buf_size, TCGv_i64 arg)
 {
-    return tcg_get_arg_str_idx(s, buf, buf_size, GET_TCGV_I32(arg));
+    return tcg_get_arg_str_idx(s, buf, buf_size, GET_TCGV_I64(arg));
 }
 
 static int helper_cmp(const void *p1, const void *p2)
Index: tcg/tcg.h
===================================================================
--- tcg/tcg.h	(revision 5809)
+++ tcg/tcg.h	(working copy)
@@ -126,20 +126,20 @@
 
 typedef struct
 {
-    int n;
+    int i32;
 } TCGv_i32;
 
 typedef struct
 {
-    int n;
+    int i64;
 } TCGv_i64;
 
 #define MAKE_TCGV_I32(i) __extension__                  \
     ({ TCGv_i32 make_tcgv_tmp = {i}; make_tcgv_tmp;})
 #define MAKE_TCGV_I64(i) __extension__                  \
     ({ TCGv_i64 make_tcgv_tmp = {i}; make_tcgv_tmp;})
-#define GET_TCGV_I32(t) ((t).n)
-#define GET_TCGV_I64(t) ((t).n)
+#define GET_TCGV_I32(t) ((t).i32)
+#define GET_TCGV_I64(t) ((t).i64)
 #if TCG_TARGET_REG_BITS == 32
 #define TCGV_LOW(t) MAKE_TCGV_I32(GET_TCGV_I64(t))
 #define TCGV_HIGH(t) MAKE_TCGV_I32(GET_TCGV_I64(t) + 1)

Re: [Qemu-devel] Re: [PATCH] Some fixes for TCG debugging

[Laurent Desnogues]

I take the opportunity of the clean up week-end to ask what has
happened to this proposed patch :-)


Laurent

Re: [Qemu-devel] Re: [PATCH] Some fixes for TCG debugging

[Blue Swirl]

On 12/7/08, Laurent Desnogues <laurent.desnogues@gmail.com> wrote:
> I take the opportunity of the clean up week-end to ask what has
>  happened to this proposed patch :-)

Simple: it has been forgotten, sorry. I'll apply it now, thank you for
the patch.