Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
---
 target/tricore/translate.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/target/tricore/translate.c b/target/tricore/translate.c
index 7cd26d8eaba..ad959f3b0a1 100644
--- a/target/tricore/translate.c
+++ b/target/tricore/translate.c
@@ -24,6 +24,7 @@
 #include "tcg/tcg-op.h"
 #include "accel/tcg/cpu-ldst.h"
 #include "qemu/qemu-print.h"
+#include "qemu/target_info.h"
 
 #include "exec/helper-proto.h"
 #include "exec/helper-gen.h"
@@ -5922,6 +5923,7 @@ static void decode_rr_logical_shift(DisasContext *ctx)
 {
     uint32_t op2;
     int r3, r2, r1;
+    unsigned long_bits = target_long_bits();
 
     r3 = MASK_OP_RR_D(ctx->opcode);
     r2 = MASK_OP_RR_S2(ctx->opcode);
@@ -5937,7 +5939,7 @@ static void decode_rr_logical_shift(DisasContext *ctx)
         break;
     case OPC2_32_RR_CLO:
         tcg_gen_not_tl(cpu_gpr_d[r3], cpu_gpr_d[r1]);
-        tcg_gen_clzi_tl(cpu_gpr_d[r3], cpu_gpr_d[r3], TARGET_LONG_BITS);
+        tcg_gen_clzi_tl(cpu_gpr_d[r3], cpu_gpr_d[r3], long_bits);
         break;
     case OPC2_32_RR_CLO_H:
         gen_helper_clo_h(cpu_gpr_d[r3], cpu_gpr_d[r1]);
@@ -5949,7 +5951,7 @@ static void decode_rr_logical_shift(DisasContext *ctx)
         gen_helper_cls_h(cpu_gpr_d[r3], cpu_gpr_d[r1]);
         break;
     case OPC2_32_RR_CLZ:
-        tcg_gen_clzi_tl(cpu_gpr_d[r3], cpu_gpr_d[r1], TARGET_LONG_BITS);
+        tcg_gen_clzi_tl(cpu_gpr_d[r3], cpu_gpr_d[r1], long_bits);
         break;
     case OPC2_32_RR_CLZ_H:
         gen_helper_clz_h(cpu_gpr_d[r3], cpu_gpr_d[r1]);
-- 
2.47.1

On 4/3/25 16:49, Philippe Mathieu-Daudé wrote:
> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> ---
>   target/tricore/translate.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)
> 

The temptation is good, but please do not touch any target code at this 
point. We want to focus on defining the API first, and we can perform 
codebase changes as a second step, without letting any occurrences of 
the old macros/functions, instead of just adding "another way to do it".

On 4/4/25 18:48, Pierrick Bouvier wrote:
> On 4/3/25 16:49, Philippe Mathieu-Daudé wrote:
>> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
>> ---
>>   target/tricore/translate.c | 6 ++++--
>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>
> 
> The temptation is good, but please do not touch any target code at this 
> point. We want to focus on defining the API first, and we can perform 
> codebase changes as a second step, without letting any occurrences of 
> the old macros/functions, instead of just adding "another way to do it".

I meant to remove these patch before posting, to focus on ARM, but
apparently forgot to do so...

On 4/4/25 19:53, Philippe Mathieu-Daudé wrote:
> On 4/4/25 18:48, Pierrick Bouvier wrote:
>> On 4/3/25 16:49, Philippe Mathieu-Daudé wrote:
>>> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
>>> ---
>>>   target/tricore/translate.c | 6 ++++--
>>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>>
>>
>> The temptation is good, but please do not touch any target code at 
>> this point. We want to focus on defining the API first, and we can 
>> perform codebase changes as a second step, without letting any 
>> occurrences of the old macros/functions, instead of just adding 
>> "another way to do it".
> 
> I meant to remove these patch before posting, to focus on ARM, but
> apparently forgot to do so...

In this particular case TARGET_LONG_BITS is a constant 32 (and 32 is 
present many times in the code), so it can be replaced as a cleanup.

Paolo

On 17/4/25 20:00, Paolo Bonzini wrote:
> On 4/4/25 19:53, Philippe Mathieu-Daudé wrote:
>> On 4/4/25 18:48, Pierrick Bouvier wrote:
>>> On 4/3/25 16:49, Philippe Mathieu-Daudé wrote:
>>>> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
>>>> ---
>>>>   target/tricore/translate.c | 6 ++++--
>>>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>>>
>>>
>>> The temptation is good, but please do not touch any target code at 
>>> this point. We want to focus on defining the API first, and we can 
>>> perform codebase changes as a second step, without letting any 
>>> occurrences of the old macros/functions, instead of just adding 
>>> "another way to do it".
>>
>> I meant to remove these patch before posting, to focus on ARM, but
>> apparently forgot to do so...
> 
> In this particular case TARGET_LONG_BITS is a constant 32 (and 32 is 
> present many times in the code), so it can be replaced as a cleanup.

Indeed. Richard suggested a similar cleanup (~ s/tl/i32/) for TriCore:
https://lore.kernel.org/qemu-devel/7f6fc0ae-6cdd-4996-b411-762019efb003@linaro.org/