Handling of local variables in the backend

Handling of local variables in the backend

[Dibyendu Majumdar]

Hi,

One issue I am facing at the moment is that there is no explicit
alloca for each symbol that is used in a function - so that the
backend has to detect when a symbol is accessed and then do alloca to
associate stack space. In the LLVM backend this is not so much an
issue as LLVM allows instructions to be inserted at a particular point
- so my code just ensures that all the allocas go at the start of the
function as required by LLVM. But I am having trouble with the other
backend that doesn't have this capability.

I think that there is no easy way to identify all the symbols used for
which stack space is needed and do the allocas at the start - is
there? I mean the only way is to go through all the instructions to
find these references as far as I know.

Is there a better way?

Regards
Dibyendu

Re: Handling of local variables in the backend

[Christopher Li]

On Tue, Sep 12, 2017 at 5:54 PM, Dibyendu Majumdar
<mobile@majumdar.org.uk> wrote:
>
> I think that there is no easy way to identify all the symbols used for
> which stack space is needed and do the allocas at the start - is
> there? I mean the only way is to go through all the instructions to
> find these references as far as I know.


Have you try the sym->symbol_list for the function node?
It should contain all the symbol used in that function.

I can even see that symbol_list under test-inspect.
under
[SYM_NODE:
       [ ctype.base_type: SYM_FN]
       [ symbol_list: symbol_list,
            ... <============== here.
       ]
]

It has other label symbol as well. You should be able to filter
out those pretty easily.

Chris

Re: Handling of local variables in the backend

[Dibyendu Majumdar]

Hi Chris,

On 12 September 2017 at 23:44, Christopher Li <sparse@chrisli.org> wrote:
> On Tue, Sep 12, 2017 at 5:54 PM, Dibyendu Majumdar
> <mobile@majumdar.org.uk> wrote:
>>
>> I think that there is no easy way to identify all the symbols used for
>> which stack space is needed and do the allocas at the start - is
>> there? I mean the only way is to go through all the instructions to
>> find these references as far as I know.
>
>
> Have you try the sym->symbol_list for the function node?
> It should contain all the symbol used in that function.
>

Did you mean struct entrypoint -> syms instead?
I see that this contains a list of symbols ... I don't know if this is
guaranteed to have all symbols defined in the function.

Regards
Dibyendu

Re: Handling of local variables in the backend

[Luc Van Oostenryck]

On Wed, Sep 13, 2017 at 10:56:16AM +0100, Dibyendu Majumdar wrote:
> Hi Chris,
> 
> On 12 September 2017 at 23:44, Christopher Li <sparse@chrisli.org> wrote:
> > On Tue, Sep 12, 2017 at 5:54 PM, Dibyendu Majumdar
> > <mobile@majumdar.org.uk> wrote:
> >>
> >> I think that there is no easy way to identify all the symbols used for
> >> which stack space is needed and do the allocas at the start - is
> >> there? I mean the only way is to go through all the instructions to
> >> find these references as far as I know.
> >
> >
> > Have you try the sym->symbol_list for the function node?
> > It should contain all the symbol used in that function.
> >
> 
> Did you mean struct entrypoint -> syms instead?
> I see that this contains a list of symbols ... I don't know if this is
> guaranteed to have all symbols defined in the function.

ep->accesses contains all the pseudos corresponding to the symbols
used for doing load & store accesses.

So, you probably should doing something like: 

	pseudo_t pseudo;

	FOR_EACH_PTR(ep->accesses, pseudo) {
		struct symbol *sym = pseudo->sym;
		unsigned log mods = sym->ctype.modifiers;

		if (mods & MOD_VOLATILE)
			...
		if (mods & (MOD_NONLOCAL | MOD_STATIC | MOD_ADDRESSABLE))
			...


-- Luc

Re: Handling of local variables in the backend

[Dibyendu Majumdar]

Hi Luc,

On 13 September 2017 at 11:18, Luc Van Oostenryck
<luc.vanoostenryck@gmail.com> wrote:
> On Wed, Sep 13, 2017 at 10:56:16AM +0100, Dibyendu Majumdar wrote:
>> Hi Chris,
>>
>> On 12 September 2017 at 23:44, Christopher Li <sparse@chrisli.org> wrote:
>> > On Tue, Sep 12, 2017 at 5:54 PM, Dibyendu Majumdar
>> > <mobile@majumdar.org.uk> wrote:
>> >>
>> >> I think that there is no easy way to identify all the symbols used for
>> >> which stack space is needed and do the allocas at the start - is
>> >> there? I mean the only way is to go through all the instructions to
>> >> find these references as far as I know.
>> >
>> >
>> > Have you try the sym->symbol_list for the function node?
>> > It should contain all the symbol used in that function.
>> >
>>
>> Did you mean struct entrypoint -> syms instead?
>> I see that this contains a list of symbols ... I don't know if this is
>> guaranteed to have all symbols defined in the function.
>
> ep->accesses contains all the pseudos corresponding to the symbols
> used for doing load & store accesses.
>
> So, you probably should doing something like:
>
>         pseudo_t pseudo;
>
>         FOR_EACH_PTR(ep->accesses, pseudo) {
>                 struct symbol *sym = pseudo->sym;
>                 unsigned log mods = sym->ctype.modifiers;
>
>                 if (mods & MOD_VOLATILE)
>                         ...
>                 if (mods & (MOD_NONLOCAL | MOD_STATIC | MOD_ADDRESSABLE))
>                         ...
>
>

Thanks. I will look at this.

Regards
Dibyendu