r205999 - Debug info: (Bugfix) Make sure artificial functions like _GLOBAL__I_a
Eric Christopher
echristo at gmail.com
Fri Apr 11 21:37:55 PDT 2014
Cool, thanks.
-eric
On Fri, Apr 11, 2014 at 4:53 PM, Adrian Prantl <aprantl at apple.com> wrote:
>
> On Apr 11, 2014, at 3:27 PM, Eric Christopher <echristo at gmail.com> wrote:
>
>> On Fri, Apr 11, 2014 at 3:26 PM, Adrian Prantl <aprantl at apple.com> wrote:
>>>
>>> On Apr 11, 2014, at 3:24 PM, Eric Christopher <echristo at gmail.com> wrote:
>>>
>>>> So from looking at this it doesn't appear that we have any address
>>>> range or location information for _GLOBAL__I_a anymore which is
>>>> exactly what I was worried about:
>>>>
>>>> 0x000000d0: DW_TAG_subprogram [13]
>>>> DW_AT_MIPS_linkage_name [DW_FORM_strp] (
>>>> .debug_str[0x00000099] = "_GLOBAL__I_a")
>>>> DW_AT_artificial [DW_FORM_flag_present] (true)
>>>> DW_AT_accessibility [DW_FORM_data1] (0x01)
>>>>
>>>> versus before:
>>>>
>>>> 0x000000d0: DW_TAG_subprogram [13]
>>>> DW_AT_MIPS_linkage_name [DW_FORM_strp] (
>>>> .debug_str[0x00000099] = "_GLOBAL__I_a")
>>>> DW_AT_decl_file [DW_FORM_data1] (0x03)
>>>> DW_AT_decl_line [DW_FORM_data1] (0x66)
>>>> DW_AT_artificial [DW_FORM_flag_present] (true)
>>>> DW_AT_accessibility [DW_FORM_data1] (0x01)
>>>> DW_AT_low_pc [DW_FORM_addr] (0x0000000000000020)
>>>> DW_AT_high_pc [DW_FORM_data4] (0x0000000b)
>>>> DW_AT_frame_base [DW_FORM_exprloc] (<0x1> 56 )
>>>>
>>>> Now we don't have any note that this code belongs to this compile
>>>> unit, etc. I think this part is incorrect. We should have scope and
>>>> line information for this - even if it's line 0 etc. (Though I think
>>>> the location of the variable would be a better one).
>>>>
>>>
>>> Interesting! That was unintentional. I'll fix that.
>>>
>>
>> Thanks!
>
> Fixed in CFE r206088.
>
> 0x000000e5: TAG_subprogram [13]
> AT_MIPS_linkage_name( "_GLOBAL__I_a" )
> AT_artificial( 0x01 )
> AT_accessibility( DW_ACCESS_public )
> AT_low_pc( 0x00000000000003e0 )
> AT_high_pc( 0x00000000000003eb )
> AT_frame_base( rbp )
>
> -- adrian
>
More information about the cfe-commits
mailing list