[lld] r193662 - [ELF] Implement minimal support for .eh_frame_hdr.

[Shankar Easwaran]

Ping ?

Thanks

Shankar Easwaran

On 10/29/2013 8:11 PM, Shankar Easwaran wrote:
> On 10/29/2013 8:02 PM, Nick Kledzik wrote:
>> On Oct 29, 2013, at 5:45 PM, Shankar Easwaran 
>> <shankare at codeaurora.org> wrote:
>>> We were also discussing on IRC, that if this should be done by the 
>>> reader by having each atom refer to a FDE entry that belongs to a PC 
>>> range.
>>>
>>> This way the garbage collection would also be consistent ?
> I was mentioning this as on ELF even with -ffunction-sections (each 
> function in a seperate section), there is only one single eh_frame 
> section that contains all the FDE entries. If we intended to garbage 
> collect any of the functions, their FDE entry will be left as is.
>
> So the Reader would need to convert the eh_frame section into FDE/CIE 
> Atoms , for this to happen.
>
> I dont know where the CIE reference should be contained though in the 
> above approach ?
>
>> Well, the eh_frame_hdr is an index.  You need to build it after the 
>> Resolver is done (dead stripping (aka GC)) and after the layout pass 
>> which orders some atoms.
> Yes, I agree constructing the eh_frame_hdr to be a pass.
>
> Thanks
>
> Shankar Easwaran
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>
>


-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by the Linux Foundation