llvm dev - Jan 2020 - aarch64 does not emit DW_AT_Location

Hi Devs,
debug info emitted by llvm does not contain DW_AT_Location for Formal
parameter
if it is an aggregate like below case
1) aggregate contain more than 4 homogeneous and size more than 128 bits
i.e.
typedef struct{
int a,b,c,d,e;
}mystruct;
void foo(mystruct ms){
}

2) aggregate contain hetrogeneous type and size more than 128 bits.
i.e.
typedef struct{
int a,b;
float c,d,e;
}mystruct;
void foo(mystruct ms){

}

This occur only when aarch64 is used, not with arm and x86.
like to know community thoughts wether this can be considered as bug or it's
an abi restriction.


./kamlesh
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200121/67ca32d5/attachment.html>

Hi Kamlesh,

Does it also happen when the parameters are used? From what I quickly
tried, if the parameter is used, clang emits
the location, as expected, e.g. `(DW_OP_breg0 (x0): 0)`

On Tue, Jan 21, 2020 at 11:55 AM kamlesh kumar via llvm-dev <
llvm-dev at lists.llvm.org> wrote:
>
> Hi Devs,
> debug info emitted by llvm does not contain DW_AT_Location for Formal
> parameter
> if it is an aggregate like below case
> 1) aggregate contain more than 4 homogeneous and size more than 128 bits
> i.e.
> typedef struct{
> int a,b,c,d,e;
> }mystruct;
> void foo(mystruct ms){
> }
>
> 2) aggregate contain hetrogeneous type and size more than 128 bits.
> i.e.
> typedef struct{
> int a,b;
> float c,d,e;
> }mystruct;
> void foo(mystruct ms){
>
> }
>
> This occur only when aarch64 is used, not with arm and x86.
> like to know community thoughts wether this can be considered as bug or it
> 's an abi restriction.
>
>
> ./kamlesh
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>

-- 
Compiler scrub, Arm
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200121/9710a1d8/attachment-0001.html>

Hi Momchil,
I am using something like this and we do not get Location information.

#include<stdio.h>
typedef struct {
double a,b,c,d,e;
}mystruct;

void foo(mystruct ms){
printf("%llu\n",(unsigned long long) &ms);
}




On Tue, Jan 21, 2020 at 5:33 PM Momchil Velikov <momchil.velikov at
gmail.com>
wrote:
> Hi Kamlesh,
>
> Does it also happen when the parameters are used? From what I quickly
> tried, if the parameter is used, clang emits
> the location, as expected, e.g. `(DW_OP_breg0 (x0): 0)`
>
> On Tue, Jan 21, 2020 at 11:55 AM kamlesh kumar via llvm-dev <
> llvm-dev at lists.llvm.org> wrote:
>
>>
>> Hi Devs,
>> debug info emitted by llvm does not contain DW_AT_Location for Formal
>> parameter
>> if it is an aggregate like below case
>> 1) aggregate contain more than 4 homogeneous and size more than 128
bits
>> i.e.
>> typedef struct{
>> int a,b,c,d,e;
>> }mystruct;
>> void foo(mystruct ms){
>> }
>>
>> 2) aggregate contain hetrogeneous type and size more than 128 bits.
>> i.e.
>> typedef struct{
>> int a,b;
>> float c,d,e;
>> }mystruct;
>> void foo(mystruct ms){
>>
>> }
>>
>> This occur only when aarch64 is used, not with arm and x86.
>> like to know community thoughts wether this can be considered as bug or
it
>> 's an abi restriction.
>>
>>
>> ./kamlesh
>> _______________________________________________
>> LLVM Developers mailing list
>> llvm-dev at lists.llvm.org
>> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>>
>
>
> --
> Compiler scrub, Arm
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200121/6fb8cf53/attachment.html>

If this is unoptimized code this is definitely a bug, with optimizations it may
not be possible to do any better. In both cases it's worth filing a bug and
attaching a clang reproducer (by running clang through env
FORCE_CLANG_DIAGNOSTICS_CRASH=1).

-- adrian
> On Jan 21, 2020, at 3:54 AM, kamlesh kumar via llvm-dev <llvm-dev at
lists.llvm.org> wrote:
> 
> 
> Hi Devs,
> debug info emitted by llvm does not contain DW_AT_Location for Formal
parameter
> if it is an aggregate like below case
> 1) aggregate contain more than 4 homogeneous and size more than 128 bits
> i.e.
> typedef struct{
> int a,b,c,d,e;
> }mystruct;
> void foo(mystruct ms){
> }
> 
> 2) aggregate contain hetrogeneous type and size more than 128 bits.
> i.e.
> typedef struct{
> int a,b;
> float c,d,e;
> }mystruct;
> void foo(mystruct ms){
> 
> }
> 
> This occur only when aarch64 is used, not with arm and x86.
> like to know community thoughts wether this can be considered as bug or
it's an abi restriction.
> 
> 
> ./kamlesh
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200121/85fafb26/attachment.html>

Adrian,
Yes it is unoptimized.I will file a bug for this.

Thanks


On Wed, Jan 22, 2020, 1:40 AM Adrian Prantl <aprantl at apple.com> wrote:
> If this is unoptimized code this is definitely a bug, with optimizations
> it may not be possible to do any better. In both cases it's worth
filing a
> bug and attaching a clang reproducer (by running clang through env
> FORCE_CLANG_DIAGNOSTICS_CRASH=1).
>
> -- adrian
>
> On Jan 21, 2020, at 3:54 AM, kamlesh kumar via llvm-dev <
> llvm-dev at lists.llvm.org> wrote:
>
>
> Hi Devs,
> debug info emitted by llvm does not contain DW_AT_Location for Formal
> parameter
> if it is an aggregate like below case
> 1) aggregate contain more than 4 homogeneous and size more than 128 bits
> i.e.
> typedef struct{
> int a,b,c,d,e;
> }mystruct;
> void foo(mystruct ms){
> }
>
> 2) aggregate contain hetrogeneous type and size more than 128 bits.
> i.e.
> typedef struct{
> int a,b;
> float c,d,e;
> }mystruct;
> void foo(mystruct ms){
>
> }
>
> This occur only when aarch64 is used, not with arm and x86.
> like to know community thoughts wether this can be considered as bug or it
> 's an abi restriction.
>
>
> ./kamlesh
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200122/477d5c66/attachment.html>