[llvm-dev] arm tailcall with many parameters?

Jay K via llvm-dev llvm-dev at lists.llvm.org
Wed Mar 28 02:40:45 PDT 2018


Oh nevermind, that works.


 - Jay


________________________________
From: Jay K
Sent: Wednesday, March 28, 2018 9:39 AM
To: llvm-dev at lists.llvm.org
Subject: Re: arm tailcall with many parameters?


Sorry that also suffered from signature mismatch.


So how about more like:


typedef struct vtable_t {
 int (*pf4)(int a, int b, int c, int d);
 int (*pf5)(int a, int b, int c, int d, int e);
} vtable_t;

int f3(int (*pf5)(int a, int b, int c, int d, int e), int a, int b, int c, int d)
{
 return pf5(a, b, c, d, d + d);
}
int f4(vtable_t *vt, int a, int b, int c, int d)
{
 return vt->pf5(a, b, c, d, d + d);
}




 - Jay




________________________________
From: Jay K
Sent: Wednesday, March 28, 2018 9:29 AM
To: llvm-dev at lists.llvm.org
Subject: arm tailcall with many parameters?


typedef struct vtable_t {
 int (*pf4)(int a, int b, int c, int d);
 int (*pf5)(int a, int b, int c, int d, int e);
} vtable_t;


int f1(int (*pf4)(int a, int b, int c, int d))
{
 return pf4(1, 2, 3, 4);
}


int f2(vtable_t *vt)
{
 return vt->pf4(1, 2, 3, 4);

}

gcc and clang will not tailcall these on arm with -O3.


int f3(int (*pf5)(int a, int b, int c, int d, int e))
{
 return pf5(1, 2, 3, 4, 5);
}

int f4(vtable_t *vt)
{
 return vt->pf5(1, 2, 3, 4, 5);
}


Could you not allocate an extra word on the stack,
in the first position, store the return address there, and pop
it to pc (or ip and then move to pc) after restoring nonvolatiles and lr?
Is that not describable in unwind data?

I only tested Debian's 3.3 so far but I'll try to try 6.0.



I ask because I'm working on tailcalls in another project.

I figured I'd look at gcc/clang output for hints here.



I understand there can be multiple exits of a function -- tailcalls of varying signatures and non-tailcalls.

I don't think that breaks the idea.



Thank you,

 - Jay

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20180328/06095bb9/attachment-0001.html>


More information about the llvm-dev mailing list