[PATCH] include: sync with GCC

Pietro Monteiro pietro@sociotechnical.xyz
Tue Nov 17 22:37:18 GMT 2026


On Tue, Nov 17, 2026, at 8:04 AM, Nix wrote:
> On 8 Nov 2026, Pietro Monteiro stated:
>> diff --git a/include/ctf.h b/include/ctf.h
>> index e28a37d4658..02e10d7136a 100644
>> --- a/include/ctf.h
>> +++ b/include/ctf.h
>> @@ -423,6 +423,10 @@ union
>>  #define CTF_K_CONST	12	/* ctt_type is base type.  */
>>  #define CTF_K_RESTRICT	13	/* ctt_type is base type.  */
>>  #define CTF_K_SLICE	14	/* Variant data is a ctf_slice_t.  */
>> +#define CTF_K_DECL_TAG	15	/* Declaration tag.  Internal use only.
>> +				   Not valid externally until CTF V4.  */
>> +#define CTF_K_TYPE_TAG	16	/* Type tag.  Internal use only.
>> +				   Not valid externally until CTF V4.  */
>>  
>>  #define CTF_K_MAX	63	/* Maximum possible (V2) CTF_K_* value.  */
>
> Not relevant yet, but soon (once I get CTFv4 done and dusted in
> binutils) the ctf.h in binutils will be for CTFv4, while the one in the
> GCC tree will still be for v3 (binutils can still read CTFv3 and
> upconvert it to v4, so this is compatible).
>
> I guess the right fix there is to pre-move ctf.h in GCC into the gcc/
> subtree and have it use that until it learns to emit v4 as well, so that
> future syncs of a CTFv4 ctf.h can just happen without anything special
> going on.
>
> I'll submit a suitable patch.

Thanks, please CC me when you send it.

Pietro


More information about the Binutils mailing list