On 5/4/22 14:32, Alexander Monakov wrote:
> On Wed, 4 May 2022, Martin Liška wrote:
> 
>> The patch is a follow-up of the discussion we've got in:
>> https://gcc.gnu.org/pipermail/gcc-patches/2022-May/593901.html
>>
>> Mold linker would appreciate knowing in advance if get_symbols_v3 is supported
>> by a GCC plug-in or not.
>>
>> Ready to be installed?
> 
> Quick note: if the linker is supposed to check for presence of this symbol
> via dlsym(), I expect it won't work as-is since lto-plugin.map hides every
> symbol except 'onload'.

Oh, good point!

Changing that, I get now:

$ nm ./lto-plugin/.libs/liblto_plugin.so | grep v3
0000000000015008 D supports_get_symbols_v3

> 
> (also it might be nicer to reword the comment to say 'Presence of the following
> symbols is used for ...', because you're leaving the value as 'false').

Sure. I'm also changing it's default value.

Sending v2.

Cheers,
Martin

> 
> Alexander
> 
>> Thanks,
>> Martin
>>
>> lto-plugin/ChangeLog:
>>
>> 	* lto-plugin.c (supports_get_symbols_v3): Add symbol.
>> ---
>>  lto-plugin/lto-plugin.c | 5 +++++
>>  1 file changed, 5 insertions(+)
>>
>> diff --git a/lto-plugin/lto-plugin.c b/lto-plugin/lto-plugin.c
>> index 47378435612..049f3841d5b 100644
>> --- a/lto-plugin/lto-plugin.c
>> +++ b/lto-plugin/lto-plugin.c
>> @@ -1554,3 +1554,8 @@ onload (struct ld_plugin_tv *tv)
>>  
>>    return LDPS_OK;
>>  }
>> +
>> +/* The following symbols are used for dynamic detection of plug-in features
>> +   from linker side.  */
>> +
>> +bool supports_get_symbols_v3;