How to write pattern for addition with carry operation

How to write pattern for addition with carry operation

[Mohamed Shafi]

Hello all,

The 16bit target that i am porting to gcc4.1.2 doesn't have any
instructions for 32bit operations. But for addition and subtraction
there is
addc
subc
instructions that consider carry bit also. Presently i have patterns
for SImode addition and subtraction such that the template will have

add %0, %1\naddc %N0, %N1
sub %0, %1\nsubc %N0, %N1

Will it be possible for me to write separate patterns for the
instructions add and addc?

Regards,
Shafi

Re: How to write pattern for addition with carry operation

[Ian Lance Taylor]

"Mohamed Shafi" <shafitvm@gmail.com> writes:

>> You can look into config/i386.md, how i.e. adddi3 is expanded and split in
>> case of !TARGET_64BIT.
>
> But is it scheduling safe?
> I mean you can't have addc executed before add. If i am right there
> will be no dependency between the two instructions. So there can be a
> case where addc gets scheduled before add. Am i right on both counts?

You are right in the sense that you must express the dependency
between the insns in the RTL.

The i386 backend is safe.  It expresses this dependency via (reg:CC
FLAGS_REG).

Ian

Re: How to write pattern for addition with carry operation

[Uros Bizjak]

On Sat, Jun 7, 2008 at 11:18 AM, Mohamed Shafi <shafitvm@gmail.com> wrote:

>>> The 16bit target that i am porting to gcc4.1.2 doesn't have any
>>> instructions for 32bit operations. But for addition and subtraction
>>> there is
>>> addc
>>> subc
>>> instructions that consider carry bit also. Presently i have patterns
>>> for SImode addition and subtraction such that the template will have
>>>
>>> add %0, %1\naddc %N0, %N1
>>> sub %0, %1\nsubc %N0, %N1
>>>
>>> Will it be possible for me to write separate patterns for the
>>> instructions add and addc?
>>>
>>
>> You can look into config/i386.md, how i.e. adddi3 is expanded and split in
>> case of !TARGET_64BIT.
>
> But is it scheduling safe?
> I mean you can't have addc executed before add. If i am right there
> will be no dependency between the two instructions. So there can be a
> case where addc gets scheduled before add. Am i right on both counts?

No, they are connected through carry flag in the sense that "add" is
the producer "and addc" is the consumer of CC reg, At least this is
the way i386 pattern works for a couple of decades without problems.

Uros.

Re: How to write pattern for addition with carry operation

[Mohamed Shafi]

2008/6/7 Uros Bizjak <ubizjak@gmail.com>:
> Hello!
>
>> The 16bit target that i am porting to gcc4.1.2 doesn't have any
>> instructions for 32bit operations. But for addition and subtraction
>> there is
>> addc
>> subc
>> instructions that consider carry bit also. Presently i have patterns
>> for SImode addition and subtraction such that the template will have
>>
>> add %0, %1\naddc %N0, %N1
>> sub %0, %1\nsubc %N0, %N1
>>
>> Will it be possible for me to write separate patterns for the
>> instructions add and addc?
>>
>
> You can look into config/i386.md, how i.e. adddi3 is expanded and split in
> case of !TARGET_64BIT.

But is it scheduling safe?
I mean you can't have addc executed before add. If i am right there
will be no dependency between the two instructions. So there can be a
case where addc gets scheduled before add. Am i right on both counts?

Regards,
Shafi

Re: How to write pattern for addition with carry operation

[Uros Bizjak]

Hello!

> The 16bit target that i am porting to gcc4.1.2 doesn't have any
> instructions for 32bit operations. But for addition and subtraction
> there is
> addc
> subc
> instructions that consider carry bit also. Presently i have patterns
> for SImode addition and subtraction such that the template will have
>
> add %0, %1\naddc %N0, %N1
> sub %0, %1\nsubc %N0, %N1
>
> Will it be possible for me to write separate patterns for the
> instructions add and addc?
>   

You can look into config/i386.md, how i.e. adddi3 is expanded and split 
in case of !TARGET_64BIT.

Uros.