[PATCH 1/2] nouveau/compiler: fix trivial compiler warnings

[PATCH 1/2] nouveau/compiler: fix trivial compiler warnings

[Tobias Klausmann]

nouveau_compiler.c: In function ‘main’:
nouveau_compiler.c:216:27: warning: ‘code’ may be used uninitialized in
this function [-Wmaybe-uninitialized]
       printf("%08x ", code[i / 4]);
                           ^
nouveau_compiler.c:215:4: warning: ‘size’ may be used uninitialized in
this function [-Wmaybe-uninitialized]
    for (i = 0; i < size; i += 4) {

Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
---
 src/gallium/drivers/nouveau/nouveau_compiler.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/gallium/drivers/nouveau/nouveau_compiler.c b/src/gallium/drivers/nouveau/nouveau_compiler.c
index 8660498..ca128b5 100644
--- a/src/gallium/drivers/nouveau/nouveau_compiler.c
+++ b/src/gallium/drivers/nouveau/nouveau_compiler.c
@@ -144,7 +144,7 @@ main(int argc, char *argv[])
    const char *filename = NULL;
    FILE *f;
    char text[65536] = {0};
-   unsigned size, *code;
+   unsigned size = 0, *code = NULL;
 
    for (i = 1; i < argc; i++) {
       if (!strcmp(argv[i], "-a"))
-- 
2.4.5

_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

[PATCH 2/2] nv50/ir: fix a compiler warning with debug-only code

[Tobias Klausmann]

codegen/nv50_ir_emit_nv50.cpp: In member function
‘void nv50_ir::CodeEmitterNV50::emitLOAD(const nv50_ir::Instruction*)’:
codegen/nv50_ir_emit_nv50.cpp:620:12: warning: unused variable ‘offset’
 [-Wunused-variable]
    int32_t offset = i->getSrc(0)->reg.data.offset;

Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
---
 src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
index 67ea6df..86b16f2 100644
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
@@ -616,8 +616,11 @@ CodeEmitterNV50::emitLoadStoreSizeCS(DataType ty)
 void
 CodeEmitterNV50::emitLOAD(const Instruction *i)
 {
-   DataFile sf = i->src(0).getFile();
+#ifdef DEBUG
    int32_t offset = i->getSrc(0)->reg.data.offset;
+#endif
+
+   DataFile sf = i->src(0).getFile();
 
    switch (sf) {
    case FILE_SHADER_INPUT:
-- 
2.4.5

_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 2/2] nv50/ir: fix a compiler warning with debug-only code

[Emil Velikov]

On 8 July 2015 at 19:27, Tobias Klausmann
<tobias.johannes.klausmann@mni.thm.de> wrote:
> codegen/nv50_ir_emit_nv50.cpp: In member function
> ‘void nv50_ir::CodeEmitterNV50::emitLOAD(const nv50_ir::Instruction*)’:
> codegen/nv50_ir_emit_nv50.cpp:620:12: warning: unused variable ‘offset’
>  [-Wunused-variable]
>     int32_t offset = i->getSrc(0)->reg.data.offset;
>
> Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
> ---
>  src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
> index 67ea6df..86b16f2 100644
> --- a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
> +++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
> @@ -616,8 +616,11 @@ CodeEmitterNV50::emitLoadStoreSizeCS(DataType ty)
>  void
>  CodeEmitterNV50::emitLOAD(const Instruction *i)
>  {
> -   DataFile sf = i->src(0).getFile();
> +#ifdef DEBUG
>     int32_t offset = i->getSrc(0)->reg.data.offset;
> +#endif
> +
assert is (normally) guarded by NDEBUG. Mesa/gallium has an in-house
replacement, which (not 100% sure) should be fine as well.

-Emil
_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 2/2] nv50/ir: fix a compiler warning with debug-only code

[Tobias Klausmann]

On 08.07.2015 21:34, Emil Velikov wrote:
> On 8 July 2015 at 19:27, Tobias Klausmann
> <tobias.johannes.klausmann@mni.thm.de> wrote:
>> codegen/nv50_ir_emit_nv50.cpp: In member function
>> ‘void nv50_ir::CodeEmitterNV50::emitLOAD(const nv50_ir::Instruction*)’:
>> codegen/nv50_ir_emit_nv50.cpp:620:12: warning: unused variable ‘offset’
>>   [-Wunused-variable]
>>      int32_t offset = i->getSrc(0)->reg.data.offset;
>>
>> Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
>> ---
>>   src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp | 5 ++++-
>>   1 file changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
>> index 67ea6df..86b16f2 100644
>> --- a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
>> +++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
>> @@ -616,8 +616,11 @@ CodeEmitterNV50::emitLoadStoreSizeCS(DataType ty)
>>   void
>>   CodeEmitterNV50::emitLOAD(const Instruction *i)
>>   {
>> -   DataFile sf = i->src(0).getFile();
>> +#ifdef DEBUG
>>      int32_t offset = i->getSrc(0)->reg.data.offset;
>> +#endif
>> +
> assert is (normally) guarded by NDEBUG. Mesa/gallium has an in-house
> replacement, which (not 100% sure) should be fine as well.
>
> -Emil
As far as i can see it in u_debug.h assert (debug_assert) is guarded by 
DEBUG as the above change...
_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 2/2] nv50/ir: fix a compiler warning with debug-only code

[Ilia Mirkin]

I suspect the issue is actually that u_debug.h isn't included. It
defines assert to be debug_assert, which in turn is

#define debug_assert(expr) (void)(0 && (expr))

which should cause the relevant var to be seen as used.


On Wed, Jul 8, 2015 at 3:40 PM, Tobias Klausmann
<tobias.johannes.klausmann@mni.thm.de> wrote:
>
>
> On 08.07.2015 21:34, Emil Velikov wrote:
>>
>> On 8 July 2015 at 19:27, Tobias Klausmann
>> <tobias.johannes.klausmann@mni.thm.de> wrote:
>>>
>>> codegen/nv50_ir_emit_nv50.cpp: In member function
>>> ‘void nv50_ir::CodeEmitterNV50::emitLOAD(const nv50_ir::Instruction*)’:
>>> codegen/nv50_ir_emit_nv50.cpp:620:12: warning: unused variable ‘offset’
>>>   [-Wunused-variable]
>>>      int32_t offset = i->getSrc(0)->reg.data.offset;
>>>
>>> Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
>>> ---
>>>   src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp | 5 ++++-
>>>   1 file changed, 4 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
>>> b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
>>> index 67ea6df..86b16f2 100644
>>> --- a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
>>> +++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
>>> @@ -616,8 +616,11 @@ CodeEmitterNV50::emitLoadStoreSizeCS(DataType ty)
>>>   void
>>>   CodeEmitterNV50::emitLOAD(const Instruction *i)
>>>   {
>>> -   DataFile sf = i->src(0).getFile();
>>> +#ifdef DEBUG
>>>      int32_t offset = i->getSrc(0)->reg.data.offset;
>>> +#endif
>>> +
>>
>> assert is (normally) guarded by NDEBUG. Mesa/gallium has an in-house
>> replacement, which (not 100% sure) should be fine as well.
>>
>> -Emil
>
> As far as i can see it in u_debug.h assert (debug_assert) is guarded by
> DEBUG as the above change...
>
> _______________________________________________
> Nouveau mailing list
> Nouveau@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/nouveau
_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 1/2] nouveau/compiler: fix trivial compiler warnings

[Ilia Mirkin]

Compiler is wrong.

On Wed, Jul 8, 2015 at 2:27 PM, Tobias Klausmann
<tobias.johannes.klausmann@mni.thm.de> wrote:
> nouveau_compiler.c: In function ‘main’:
> nouveau_compiler.c:216:27: warning: ‘code’ may be used uninitialized in
> this function [-Wmaybe-uninitialized]
>        printf("%08x ", code[i / 4]);
>                            ^
> nouveau_compiler.c:215:4: warning: ‘size’ may be used uninitialized in
> this function [-Wmaybe-uninitialized]
>     for (i = 0; i < size; i += 4) {
>
> Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
> ---
>  src/gallium/drivers/nouveau/nouveau_compiler.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/gallium/drivers/nouveau/nouveau_compiler.c b/src/gallium/drivers/nouveau/nouveau_compiler.c
> index 8660498..ca128b5 100644
> --- a/src/gallium/drivers/nouveau/nouveau_compiler.c
> +++ b/src/gallium/drivers/nouveau/nouveau_compiler.c
> @@ -144,7 +144,7 @@ main(int argc, char *argv[])
>     const char *filename = NULL;
>     FILE *f;
>     char text[65536] = {0};
> -   unsigned size, *code;
> +   unsigned size = 0, *code = NULL;
>
>     for (i = 1; i < argc; i++) {
>        if (!strcmp(argv[i], "-a"))
> --
> 2.4.5
>
> _______________________________________________
> Nouveau mailing list
> Nouveau@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/nouveau
_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 1/2] nouveau/compiler: fix trivial compiler warnings

[Tobias Klausmann]

On 08.07.2015 20:38, Ilia Mirkin wrote:
> Compiler is wrong.

So just "nouveau: ..." then? Anyway, change it to your liking.

> On Wed, Jul 8, 2015 at 2:27 PM, Tobias Klausmann
> <tobias.johannes.klausmann@mni.thm.de> wrote:
>> nouveau_compiler.c: In function ‘main’:
>> nouveau_compiler.c:216:27: warning: ‘code’ may be used uninitialized in
>> this function [-Wmaybe-uninitialized]
>>         printf("%08x ", code[i / 4]);
>>                             ^
>> nouveau_compiler.c:215:4: warning: ‘size’ may be used uninitialized in
>> this function [-Wmaybe-uninitialized]
>>      for (i = 0; i < size; i += 4) {
>>
>> Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
>> ---
>>   src/gallium/drivers/nouveau/nouveau_compiler.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/src/gallium/drivers/nouveau/nouveau_compiler.c b/src/gallium/drivers/nouveau/nouveau_compiler.c
>> index 8660498..ca128b5 100644
>> --- a/src/gallium/drivers/nouveau/nouveau_compiler.c
>> +++ b/src/gallium/drivers/nouveau/nouveau_compiler.c
>> @@ -144,7 +144,7 @@ main(int argc, char *argv[])
>>      const char *filename = NULL;
>>      FILE *f;
>>      char text[65536] = {0};
>> -   unsigned size, *code;
>> +   unsigned size = 0, *code = NULL;
>>
>>      for (i = 1; i < argc; i++) {
>>         if (!strcmp(argv[i], "-a"))
>> --
>> 2.4.5
>>
>> _______________________________________________
>> Nouveau mailing list
>> Nouveau@lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/nouveau

_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 1/2] nouveau/compiler: fix trivial compiler warnings

[Emil Velikov]

Curious how many times this needs to come along, for you to change
your mind. :-)

On 8 July 2015 at 19:38, Ilia Mirkin <imirkin@alum.mit.edu> wrote:
> Compiler is wrong.
>
> On Wed, Jul 8, 2015 at 2:27 PM, Tobias Klausmann
> <tobias.johannes.klausmann@mni.thm.de> wrote:
>> nouveau_compiler.c: In function ‘main’:
>> nouveau_compiler.c:216:27: warning: ‘code’ may be used uninitialized in
>> this function [-Wmaybe-uninitialized]
>>        printf("%08x ", code[i / 4]);
>>                            ^
>> nouveau_compiler.c:215:4: warning: ‘size’ may be used uninitialized in
>> this function [-Wmaybe-uninitialized]
>>     for (i = 0; i < size; i += 4) {
>>
>> Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
>> ---
>>  src/gallium/drivers/nouveau/nouveau_compiler.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/src/gallium/drivers/nouveau/nouveau_compiler.c b/src/gallium/drivers/nouveau/nouveau_compiler.c
>> index 8660498..ca128b5 100644
>> --- a/src/gallium/drivers/nouveau/nouveau_compiler.c
>> +++ b/src/gallium/drivers/nouveau/nouveau_compiler.c
>> @@ -144,7 +144,7 @@ main(int argc, char *argv[])
>>     const char *filename = NULL;
>>     FILE *f;
>>     char text[65536] = {0};
>> -   unsigned size, *code;
>> +   unsigned size = 0, *code = NULL;
>>
>>     for (i = 1; i < argc; i++) {
>>        if (!strcmp(argv[i], "-a"))
>> --
>> 2.4.5
>>
>> _______________________________________________
>> Nouveau mailing list
>> Nouveau@lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/nouveau
> _______________________________________________
> Nouveau mailing list
> Nouveau@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/nouveau
_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 1/2] nouveau/compiler: fix trivial compiler warnings

[Tobias Klausmann]

Mh i'm not aware of me ever changed the "nouveau_compiler". But i'm 
happy to see this made you laugh, so it has something positive at 
least... :/

On 08.07.2015 21:31, Emil Velikov wrote:
> Curious how many times this needs to come along, for you to change
> your mind. :-)
>
> On 8 July 2015 at 19:38, Ilia Mirkin <imirkin@alum.mit.edu> wrote:
>> Compiler is wrong.
>>
>> On Wed, Jul 8, 2015 at 2:27 PM, Tobias Klausmann
>> <tobias.johannes.klausmann@mni.thm.de> wrote:
>>> nouveau_compiler.c: In function ‘main’:
>>> nouveau_compiler.c:216:27: warning: ‘code’ may be used uninitialized in
>>> this function [-Wmaybe-uninitialized]
>>>         printf("%08x ", code[i / 4]);
>>>                             ^
>>> nouveau_compiler.c:215:4: warning: ‘size’ may be used uninitialized in
>>> this function [-Wmaybe-uninitialized]
>>>      for (i = 0; i < size; i += 4) {
>>>
>>> Signed-off-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
>>> ---
>>>   src/gallium/drivers/nouveau/nouveau_compiler.c | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/src/gallium/drivers/nouveau/nouveau_compiler.c b/src/gallium/drivers/nouveau/nouveau_compiler.c
>>> index 8660498..ca128b5 100644
>>> --- a/src/gallium/drivers/nouveau/nouveau_compiler.c
>>> +++ b/src/gallium/drivers/nouveau/nouveau_compiler.c
>>> @@ -144,7 +144,7 @@ main(int argc, char *argv[])
>>>      const char *filename = NULL;
>>>      FILE *f;
>>>      char text[65536] = {0};
>>> -   unsigned size, *code;
>>> +   unsigned size = 0, *code = NULL;
>>>
>>>      for (i = 1; i < argc; i++) {
>>>         if (!strcmp(argv[i], "-a"))
>>> --
>>> 2.4.5
>>>
>>> _______________________________________________
>>> Nouveau mailing list
>>> Nouveau@lists.freedesktop.org
>>> http://lists.freedesktop.org/mailman/listinfo/nouveau
>> _______________________________________________
>> Nouveau mailing list
>> Nouveau@lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/nouveau

_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 1/2] nouveau/compiler: fix trivial compiler warnings

[Emil Velikov]

On 8 July 2015 at 20:34, Tobias Klausmann
<tobias.johannes.klausmann@mni.thm.de> wrote:
> Mh i'm not aware of me ever changed the "nouveau_compiler". But i'm happy to
> see this made you laugh, so it has something positive at least... :/
>
Story time:
This particular compiler warning has been brought up (incl here) four
or five times. Each time, Ilia feels reluctant about the fix as the
(gcc) compiler gets it wrong.

Personally I do not see a problem with explicitly initialising the
variable at this instance, yet I'm curious for how long Ilia will say
"no" to this (type of) patch(es) :-P

No offence, I just find it funny.
Emil
_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 1/2] nouveau/compiler: fix trivial compiler warnings

[Tobias Klausmann]

On 08.07.2015 21:42, Emil Velikov wrote:
> On 8 July 2015 at 20:34, Tobias Klausmann
> <tobias.johannes.klausmann@mni.thm.de> wrote:
>> Mh i'm not aware of me ever changed the "nouveau_compiler". But i'm happy to
>> see this made you laugh, so it has something positive at least... :/
>>
> Story time:
> This particular compiler warning has been brought up (incl here) four
> or five times. Each time, Ilia feels reluctant about the fix as the
> (gcc) compiler gets it wrong.
>
> Personally I do not see a problem with explicitly initialising the
> variable at this instance, yet I'm curious for how long Ilia will say
> "no" to this (type of) patch(es) :-P
>
> No offence, I just find it funny.
> Emil
Oh i did even answer in a thread for a patch from Martin where he 
propose the same change (even with the same prefix :D). Ilia maybe you 
should take this after all, as it seems you are haunted by this :P
_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: [PATCH 1/2] nouveau/compiler: fix trivial compiler warnings

[Ilia Mirkin]

I don't mind telling people that the compiler is wrong :)

On Wed, Jul 8, 2015 at 3:53 PM, Tobias Klausmann
<tobias.johannes.klausmann@mni.thm.de> wrote:
>
>
> On 08.07.2015 21:42, Emil Velikov wrote:
>>
>> On 8 July 2015 at 20:34, Tobias Klausmann
>> <tobias.johannes.klausmann@mni.thm.de> wrote:
>>>
>>> Mh i'm not aware of me ever changed the "nouveau_compiler". But i'm happy
>>> to
>>> see this made you laugh, so it has something positive at least... :/
>>>
>> Story time:
>> This particular compiler warning has been brought up (incl here) four
>> or five times. Each time, Ilia feels reluctant about the fix as the
>> (gcc) compiler gets it wrong.
>>
>> Personally I do not see a problem with explicitly initialising the
>> variable at this instance, yet I'm curious for how long Ilia will say
>> "no" to this (type of) patch(es) :-P
>>
>> No offence, I just find it funny.
>> Emil
>
> Oh i did even answer in a thread for a patch from Martin where he propose
> the same change (even with the same prefix :D). Ilia maybe you should take
> this after all, as it seems you are haunted by this :P
_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/nouveau