public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH 0/2] RISC-V: Define not broken prefetch builtins
@ 2023-09-22  7:11 Tsukasa OI
  2023-09-22  7:11 ` [PATCH 1/2] " Tsukasa OI
                   ` (2 more replies)
  0 siblings, 3 replies; 6+ messages in thread
From: Tsukasa OI @ 2023-09-22  7:11 UTC (permalink / raw)
  To: Tsukasa OI, Kito Cheng, Palmer Dabbelt, Andrew Waterman,
	Jim Wilson, Jeff Law
  Cc: gcc-patches

Hello,

As I explained earlier:
<https://gcc.gnu.org/pipermail/gcc-patches/2023-August/626916.html>,
the builtin function for RISC-V "__builtin_riscv_zicbop_cbo_prefetchi" is
completely broken.  Instead, this patch set (in PATCH 1/2) creates three
new, working builtin intrinsics.

void __builtin_riscv_prefetch_i(void *addr, [intptr_t offset,] ...);
void __builtin_riscv_prefetch_r(void *addr, [intptr_t offset,] ...);
void __builtin_riscv_prefetch_w(void *addr, [intptr_t offset,] ...);


For consistency with "prefetch.i" and the reason I describe later (which
requires native instructions for "prefetch.r" and "prefetch.w"), I decided
to make builtin functions for "prefetch.[rw]" as well.

Optional second argument (named "offset" here) defaults to zero and must be
a compile-time integral constant.  Also, it must be a valid offset for a
"prefetch.[irw]" HINT instruction (x % 32 == 0 && x >= -2048 && x < 2048).

They are defined if the 'Zicbop' extension is supported and expands to:

> prefetch.i offset(addr_reg)  ; __builtin_riscv_prefetch_i
> prefetch.r offset(addr_reg)  ; __builtin_riscv_prefetch_r
> prefetch.w offset(addr_reg)  ; __builtin_riscv_prefetch_w


The hardest part of this patch set was to support builtin function with
variable argument (making "offset" optional).  It required:

1.  Support for variable argument function prototype for RISC-V builtins
    (corresponding "..." on C-based languages)
2.  Support for (non-vector) RISC-V builtins with custom expansion
    (on RVV intrinsics, custom expansion is already implemented)


... and PATCH 2/2 fixes an ICE while I'm investigating regular prefetch
builtin (__builtin_prefetch).  If the 'Zicbop' extension is enabled,
__builtin_prefetch with the first argument NULL or (not all but) some
fixed addresses (like ((void*)0x20)) can cause an ICE.  This is because
the "r" constraint is not checked and a constant can be a first argument
of target-specific "prefetch" RTL instruction.

PATCH 2/2 fixes this issue by:

1.  Making "prefetch" not an instruction but instead an expansion
    (this is not rare; e.g. on i386) and
2.  Coercing the address argument into a register in the expansion

It requires separate instructions for "prefetch.[rw]" and I decided to make
those prefetch instructions very similar to "prefetch.i".  That's one of the
reasons I created builtins corresponding those.


Sincerely,
Tsukasa




Tsukasa OI (2):
  RISC-V: Define not broken prefetch builtins
  RISC-V: Fix ICE by expansion and register coercion

 gcc/config/riscv/riscv-builtins.cc            | 112 +++++++++++++++++-
 gcc/config/riscv/riscv-cmo.def                |   8 +-
 gcc/config/riscv/riscv-ftypes.def             |   1 +
 gcc/config/riscv/riscv.md                     |  67 ++++++++---
 gcc/testsuite/gcc.target/riscv/cmo-zicbop-1.c |  41 ++++---
 gcc/testsuite/gcc.target/riscv/cmo-zicbop-2.c |  33 ++----
 gcc/testsuite/gcc.target/riscv/cmo-zicbop-3.c |  29 +++++
 gcc/testsuite/gcc.target/riscv/cmo-zicbop-4.c |  14 +++
 gcc/testsuite/gcc.target/riscv/cmo-zicbop-5.c |  14 +++
 gcc/testsuite/gcc.target/riscv/cmo-zicbop-6.c |  38 ++++++
 .../gcc.target/riscv/cmo-zicbop-by-common-1.c |  17 +++
 .../gcc.target/riscv/cmo-zicbop-by-common-2.c |   7 ++
 .../gcc.target/riscv/cmo-zicbop-by-common-3.c |  13 ++
 .../riscv/cmo-zicbop-by-common-ice-1.c        |  13 ++
 .../riscv/cmo-zicbop-by-common-ice-2.c        |   7 ++
 15 files changed, 350 insertions(+), 64 deletions(-)
 create mode 100644 gcc/testsuite/gcc.target/riscv/cmo-zicbop-3.c
 create mode 100644 gcc/testsuite/gcc.target/riscv/cmo-zicbop-4.c
 create mode 100644 gcc/testsuite/gcc.target/riscv/cmo-zicbop-5.c
 create mode 100644 gcc/testsuite/gcc.target/riscv/cmo-zicbop-6.c
 create mode 100644 gcc/testsuite/gcc.target/riscv/cmo-zicbop-by-common-1.c
 create mode 100644 gcc/testsuite/gcc.target/riscv/cmo-zicbop-by-common-2.c
 create mode 100644 gcc/testsuite/gcc.target/riscv/cmo-zicbop-by-common-3.c
 create mode 100644 gcc/testsuite/gcc.target/riscv/cmo-zicbop-by-common-ice-1.c
 create mode 100644 gcc/testsuite/gcc.target/riscv/cmo-zicbop-by-common-ice-2.c


base-commit: 40ac613627205dd4d24ae136917e48b357fee758
-- 
2.42.0


^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2023-10-30 21:38 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-09-22  7:11 [PATCH 0/2] RISC-V: Define not broken prefetch builtins Tsukasa OI
2023-09-22  7:11 ` [PATCH 1/2] " Tsukasa OI
2023-09-22  7:11 ` [PATCH 2/2] RISC-V: Fix ICE by expansion and register coercion Tsukasa OI
2023-09-26 21:38 ` [PATCH 0/2] RISC-V: Define not broken prefetch builtins Jeff Law
2023-10-23  3:55   ` Tsukasa OI
2023-10-30 21:38     ` Jeff Law

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).