On Sun, Jun 08, 2025 at 07:26:44AM +0100, Stafford Horne wrote:
> On Sat, Jun 07, 2025 at 06:53:28PM +0300, Dimitar Dimitrov wrote:
> > On Sat, Jun 07, 2025 at 11:38:46AM +0100, Stafford Horne wrote:
> > > On Fri, Jun 06, 2025 at 09:54:53PM +0100, Stafford Horne wrote:
> > > > On Fri, Jun 06, 2025 at 11:31:00PM +0300, Dimitar Dimitrov wrote:
> > > > > On Fri, Jun 06, 2025 at 06:12:13PM +0100, Stafford Horne wrote:
> > > > > > On Fri, Jun 06, 2025 at 04:41:21PM +0100, Stafford Horne wrote:
> > > > > > > On Fri, Jun 06, 2025 at 04:20:10PM +0100, Richard Sandiford wrote:
> > > > > > > > Stafford Horne <sho...@gmail.com> writes:
> > > > > > > > > Hello,
> > > > > > > > >
> > > > > > > > > This seems to be causing a build regression on the or1k port.
> > > > > > > > >
> > > > > > > > > I have not quite figured it out yet but I have bisected to
> > > > > > > > > this commit.
> > > > > > > > >
> > > > > > > > > The failure is as below, this seems to be caused by the
> > > > > > > > > cstoresi4 instruction produced by
> > > > > > > > > or1k.md. So I think its likely something we are doing funny
> > > > > > > > > in the port.
> > > > > > > > >
> > > > > > > > > Thanks to Jeff for pointing this out. If you have any idea
> > > > > > > > > let me know.
> > > > > > > > >
> > > > > > > > > Log:
> > > > > > > > >
> > > > > > > > > during RTL pass: ce1
> > > > > > > > > dump file: libgcc2.c.286r.ce1
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/libgcc/libgcc2.c: In
> > > > > > > > > function '__muldi3':
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/libgcc/libgcc2.c:538:1:
> > > > > > > > > internal compiler error: in emit_move_multi_word, at
> > > > > > > > > expr.cc:4492
> > > > > > > > > 538 | }
> > > > > > > > > | ^
> > > > > > > > > 0x1b094df internal_error(char const*, ...)
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/diagnostic-global-context.cc:517
> > > > > > > > > 0x6459c1 fancy_abort(char const*, int, char const*)
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/diagnostic.cc:1815
> > > > > > > > > 0x473a2e emit_move_multi_word
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/expr.cc:4492
> > > > > > > > > 0x93fd7d emit_move_insn(rtx_def*, rtx_def*)
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/expr.cc:4746
> > > > > > > > > 0x9175f1 copy_to_reg(rtx_def*)
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/explow.cc:630
> > > > > > > > > 0xd18977 gen_lowpart_general(machine_mode, rtx_def*)
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/rtlhooks.cc:56
> > > > > > > > > 0x9414c7 convert_mode_scalar
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/expr.cc:818
> > > > > > > > > 0x9421a1 convert_modes(machine_mode, machine_mode,
> > > > > > > > > rtx_def*, int)
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/expr.cc:961
> > > > > > > > > 0xc0afa0 prepare_operand(insn_code, rtx_def*, int,
> > > > > > > > > machine_mode, machine_mode, int)
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/optabs.cc:4637
> > > > > > > > > 0xc0b33a prepare_cmp_insn
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/optabs.cc:4538
> > > > > > > > > 0xc0f0dd emit_conditional_move(rtx_def*, rtx_comparison,
> > > > > > > > > rtx_def*, rtx_def*, machine_mode, int)
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/optabs.cc:5098
> > > > > > > > > 0x192cc7b noce_emit_cmove
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/ifcvt.cc:1777
> > > > > > > > > 0x193365b try_emit_cmove_seq
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/ifcvt.cc:3410
> > > > > > > > > 0x193365b noce_convert_multiple_sets_1
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/ifcvt.cc:3705
> > > > > > > > > 0x1933c71 noce_convert_multiple_sets
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/ifcvt.cc:3496
> > > > > > > > > 0x1938687 noce_process_if_block
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/ifcvt.cc:4045
> > > > > > > > > 0x1938687 noce_find_if_block
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/ifcvt.cc:4726
> > > > > > > > > 0x1938687 find_if_header
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/ifcvt.cc:4931
> > > > > > > > > 0x1938687 if_convert
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/ifcvt.cc:6078
> > > > > > > > > 0x1938d01 rest_of_handle_if_conversion
> > > > > > > > >
> > > > > > > > > /home/shorne/work/gnu-toolchain/gcc/gcc/ifcvt.cc:6143
> > > > > > > > >
> > > > > > > > > -Stafford
> > > > > > > >
> > > > > > > > Can you attach the .i?
> > > > > > >
> > > > > > > Sure please find attached, I can reproduce with the or1k compiler
> > > > > > > using:
> > > > > > >
> > > > > > > $HOME/work/gnu-toolchain/build-gcc/./gcc/cc1 -O2 __muldi3.i
> > > > > > >
> > > > > > > I am still debugging, but I notice when reverting the patch the
> > > > > > > compiler never
> > > > > > > goes to emit_move_multi_word(). I am trying to work backwards as
> > > > > > > to where it goes
> > > > > > > different, it's just taking a bit of time as I haven't debugged
> > > > > > > GCC ICE's for a
> > > > > > > while.
> > > > > >
> > > > > > I made some progress on this. In openRISC we generate cstoresi4
> > > > > > with special
> > > > > > register SR_F (condition flag) bit register.
> > > > > >
> > > > > > This is represented as (reg:BI 34 ?sr_f).
> > > > > >
> > > > > > The RTL produced for cstoresi4 becomes something like:
> > > > > >
> > > > > > // The compare
> > > > > > (insn 49 48 50 (set (reg:BI 34 ?sr_f)
> > > > > > (leu:BI (reg/v:SI 68 [ __x2D.6783 ])
> > > > > > (reg/v:SI 69 [ __x1D.6782 ])))
> > > > > > "/home/shorne/work/gnu-toolchain/gcc/libgcc/libgcc2.c":532:22 -1
> > > > > > (nil)
> > > > > > // The conditional branch
> > > > > > (jump_insn 50 49 0 (set (pc)
> > > > > > (if_then_else (ne (reg:BI 34 ?sr_f)
> > > > > > (const_int 0 [0]))
> > > > > > (label_ref 0)
> > > > > > (pc)))
> > > > > > "/home/shorne/work/gnu-toolchain/gcc/libgcc/libgcc2.c":532:22 -1
> > > > > > (int_list:REG_BR_PROB 536870916 (nil)))
> > > > > >
> > > > > > During ce1 the try_emit_cmove_seq would try to convert (reg:BI 34
> > > > > > ?sr_f) to SI.
> > > > > > Before this patch it was done in gen_lowpart_general by
> > > > > > simplify_context::simplify_gen_subreg.
> > > > > >
> > > > > > The simplity to subreg operation would convert:
> > > > > > (reg:BI 34 ?sr_f)
> > > > > >
> > > > > > To:
> > > > > > (subreg:SI (reg:BI 34 ?sr_f) 0)
> > > > > >
> > > > > > But now the call to validate_subreg is returning false and
> > > > > > simplify_gen_subreg
> > > > > > returns NULL_RTX, this causes gen_lowpart_general to try to do
> > > > > > copy_to_reg,
> > > > > > which fails.
> > > > > >
> > > > > > I need to step away for a bit, but I would next look into why
> > > > > > validate_subreg
> > > > > > is now failing to allow (reg:BI 34 ?sr_f) to be subreg'd.
> > > > >
> > > > > I think that the REG_CAN_CHANGE_MODE_P call in validate_subreg is
> > > > > rejecting the subreg, because or1k_can_change_mode_class rejects the
> > > > > mode change of sr_f from BI to SI.
> > > > >
> > > > Hi Dimitar,
> > > >
> > > > That's right, I noticed the same thing and I am testing out a patch to
> > > > or1k_can_change_mode_class now, it fixes the ICE. But I'll have to
> > > > spend a bit
> > > > more time testing.
> > >
> > > Hi All,
> > >
> > > The patch I have now is as below and it looks to work ok:
> > >
> > > --- a/gcc/config/or1k/or1k.cc
> > > +++ b/gcc/config/or1k/or1k.cc
> > > @@ -1408,8 +1408,9 @@ static bool
> > > or1k_can_change_mode_class (machine_mode from, machine_mode to,
> > > reg_class_t rclass)
> > > {
> > > + /* Allow cnoverting special flags to SI mode subregs. */
> > > if (rclass == FLAG_REGS)
> > > - return from == to;
> > > + return from == to || (from == BImode && to == SImode);
> > > return true;
> > > }
> >
> > Another patch I intend to merge tomorrow will tighten the checks even
> > more: https://gcc.gnu.org/pipermail/gcc-patches/2025-June/685837.html
> >
> > Seeing that or1k_hard_regno_mode_ok forbids SImode for class FLAG_REGS,
> > the validate_subreg may once again start rejecting subregs for or1k.
>
> OK, I tested with this patch and confirm it causes the same ICE for the or1k
> port again, with a slighly differnt code path.
>
> I tried to patch with the below and it causes another. I think this will not
> work as we should not allow FLAG_REGS directly as SI.
>
> --- a/gcc/config/or1k/or1k.cc
> +++ b/gcc/config/or1k/or1k.cc
> @@ -1390,9 +1390,10 @@ static bool
> or1k_hard_regno_mode_ok (unsigned int regno, machine_mode mode)
> {
> /* For OpenRISC, GENERAL_REGS can hold anything, while
> - FLAG_REGS are really single bits within SP[SR]. */
> + FLAG_REGS are really single bits within SP[SR].
> + We also allow SImode for FLAG_REGS to allow subregs. */
> if (REGNO_REG_CLASS (regno) == FLAG_REGS)
> - return mode == BImode;
> + return mode == BImode || mode == SImode;
> return true;
> }
>
> With the above patch we get the failure:
>
> In file included from
> /home/shorne/work/gnu-toolchain/gcc/libstdc++-v3/src/c++17/floating_from_chars.cc:86:
>
> /home/shorne/work/gnu-toolchain/gcc/libstdc++-v3/src/c++17/fast_float/fast_float.h:
> In function 'std::from_chars_result
> {anonymous}::fast_float::from_chars_advanced(const char*, const char*, T&,
> parse_options) [with T = double]':
>
> /home/shorne/work/gnu-toolchain/gcc/libstdc++-v3/src/c++17/fast_float/fast_float.h:3040:1:
> error: insn does not satisfy its constraints:
> 3040 | }
> | ^
> (insn 9056 5757 7408 339 (set (reg:SI 21 r27 [orig:3711+-3 ] [3711])
> (reg:SI 34 ?sr_f [+-3 ]))
> "/home/shorne/work/gnu-toolchain/gcc/libstdc++-v3/src/c++17/fast_float/fast_float.h":261:30
> 28 {*movsi_internal}
> (nil))
> during RTL pass: reload
>
> /home/shorne/work/gnu-toolchain/gcc/libstdc++-v3/src/c++17/fast_float/fast_float.h:3040:1:
> internal compiler error: in extract_constrain_insn, at recog.cc:2783
> 0x19b4af9 internal_error(char const*, ...)
>
> /home/shorne/work/gnu-toolchain/gcc/gcc/diagnostic-global-context.cc:517
> 0x1998bf0 fancy_abort(char const*, int, char const*)
> /home/shorne/work/gnu-toolchain/gcc/gcc/diagnostic.cc:1803
> 0xcfec74 _fatal_insn(char const*, rtx_def const*, char const*, int, char
> const*)
> /home/shorne/work/gnu-toolchain/gcc/gcc/rtl-error.cc:108
> 0xcfeca5 _fatal_insn_not_found(rtx_def const*, char const*, int, char
> const*)
> /home/shorne/work/gnu-toolchain/gcc/gcc/rtl-error.cc:118
> 0xcd2684 extract_constrain_insn(rtx_insn*)
> /home/shorne/work/gnu-toolchain/gcc/gcc/recog.cc:2783
> 0xb954ea check_rtl
> /home/shorne/work/gnu-toolchain/gcc/gcc/lra.cc:2202
> 0xb974cd lra(_IO_FILE*, int)
> /home/shorne/work/gnu-toolchain/gcc/gcc/lra.cc:2636
> 0xb5dc6f do_reload
> /home/shorne/work/gnu-toolchain/gcc/gcc/ira.cc:5986
> 0xb5dfb6 execute
> /home/shorne/work/gnu-toolchain/gcc/gcc/ira.cc:6174
>
> The RTL generated: (reg:SI 34 ?sr_f [+-3 ]) is incorrect, and I think it's
> being allowed due to my patch.
Hi,
I do not have a fix for this yet. I feel using or1k_hard_regno_mode_ok to
control allowing paradoxical subreging of openriscs sr_f special register is not
right. It seems we would want to have something like
or1k_regno_paradoxical_subreg_mode_ok.
OpenRISC seems pretty unique in how we use (reg:BI 34 ?sr_f) for representing
the
condition register. The SR[F] bit is a flag bit in the supervisor register that
represents if the last comparison was true (flag set) or false (flag unset).
The sr_f RTL never actually gets conveted to instructions that read the
supervisor register bits directly.
I am still reading though code to see what options there are, one option is to
try to convert openrisc to use CC_MODE instead of the sr_f special register.
-Stafford
