Re: [Qemu-arm] [PATCH 02/23] target/arm: Add SVE decode skeleton

[Richard Henderson]

On 01/11/2018 10:20 AM, Peter Maydell wrote:
> On 18 December 2017 at 17:45, Richard Henderson
> <address@hidden> wrote:
>> Including only 4, as-yet unimplemented, instruction patterns
>> so that the whole thing compiles.
>>
>> Signed-off-by: Richard Henderson <address@hidden>
>> ---
>>  target/arm/translate-a64.h | 111 
>> +++++++++++++++++++++++++++++++++++++++++++++
>>  target/arm/translate-a64.c |  91 +++++++------------------------------
>>  target/arm/translate-sve.c |  48 ++++++++++++++++++++
>>  .gitignore                 |   1 +
>>  target/arm/Makefile.objs   |  11 +++++
>>  target/arm/sve.def         |  45 ++++++++++++++++++
>>  6 files changed, 233 insertions(+), 74 deletions(-)
>>  create mode 100644 target/arm/translate-a64.h
>>  create mode 100644 target/arm/translate-sve.c
>>  create mode 100644 target/arm/sve.def
> 
> This will be easier to review if you split the stuff that's
> purely code motion from the .c file to the .h into its own
> patch.

Ok.

>> +target/arm/decode-sve.inc.c: $(SRC_PATH)/target/arm/sve.def $(DECODETREE)
>> +       $(call quiet-command,\
>> +         $(PYTHON) $(DECODETREE) -o $@ --decode disas_sve \
>> +               $(SRC_PATH)/target/arm/sve.def || rm -f $@, \
>> +               "GEN", $@)
>> +
>> +target/arm/translate-sve.o: target/arm/decode-sve.inc.c
>> +obj-$(TARGET_AARCH64) += translate-sve.o
> 
> If we're serious about the idea that this decoder script is
> general purpose, we should have a rules.mak rune for
> generally invoking it to create a decode-foo.inc.c from a foo.def.

I didn't want to attempt to generalize this until we have two users.
Particularly if we wind up with extra options to the script to change other
behavior.


r~