On 1/23/21 7:13 PM, Paolo Bonzini wrote:
> On 22/01/21 14:30, Philippe Mathieu-Daudé wrote:
>> Users might want to enable all features, without realizing some
>> features have negative effect. Rename '--enable-tcg-interpreter'
>> as '--disable-native-tcg' to avoid user selecting this feature
>> without understanding it. '--enable-tcg-interpreter' is kept in
>> for backward compability with scripts.
>>
>> Suggested-by: Thomas Huth <th...@redhat.com>
>> Reviewed-by: Thomas Huth <th...@redhat.com>
>> Signed-off-by: Philippe Mathieu-Daudé <phi...@redhat.com>
>> ---
>>   configure | 5 +++--
>>   1 file changed, 3 insertions(+), 2 deletions(-)
>>
>> diff --git a/configure b/configure
>> index 48bd6f48d7a..5e5ff779a69 100755
>> --- a/configure
>> +++ b/configure
>> @@ -1121,7 +1121,8 @@ for opt do
>>     ;;
>>     --disable-tcg-interpreter) tcg_interpreter="no"
>>     ;;
>> -  --enable-tcg-interpreter) tcg_interpreter="yes"
>> +  --enable-tcg-interpreter) # backward compatibility
> 
> Do you really want to break the old option?

I am not breaking it, I'm keeping it, but not listed
in the help. Maybe you meant "not break the old option",
in that case you suggest me to add a warning such "this
option is deprecated and has been replaced by
--disable-native-tcg"?

> 
>> +  --disable-native-tcg) tcg_interpreter="yes"
>>     ;;
>>     --disable-cap-ng)  cap_ng="disabled"
>>     ;;
>> @@ -1753,7 +1754,7 @@ Advanced options (experts only):
>>     --with-trace-file=NAME   Full PATH,NAME of file to store traces
>>                              Default:trace-<pid>
>>     --disable-slirp          disable SLIRP userspace network connectivity
>> -  --enable-tcg-interpreter enable TCI (TCG with bytecode interpreter,
>> experimental and slow)
>> +  --disable-native-tcg     enable TCI (TCG with bytecode interpreter,
>> experimental and slow)
>>     --enable-malloc-trim     enable libc malloc_trim() for memory
>> optimization
>>     --oss-lib                path to OSS library
>>     --cpu=CPU                Build for host CPU [$cpu]
>>
> 
> The problem here is that for some CPUs there is no native TCG...  I
> mean, what's unclear in "exprimental and slow"?

OK, we can skip this patch then.

> 
> Paolo
> 


Reply via email to