[Bug inline-asm/97667] New: a bug in asm_operand_ok() recog.c:1801

[Bug inline-asm/97667] New: a bug in asm_operand_ok() recog.c:1801

[huntazhang at tencent dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97667

            Bug ID: 97667
           Summary: a bug in asm_operand_ok() recog.c:1801
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: inline-asm
          Assignee: unassigned at gcc dot gnu.org
          Reporter: huntazhang at tencent dot com
  Target Milestone: ---

when i used gcc-11.0 compile linux kernel in
arch/x86/include/asm/checksum_64.h:171 trigger a bug 

here is compile source code argurment 
gcc -Wp,-MMD,net/core/.utils.o.d  -nostdinc -isystem
/usr/lib/gcc/x86_64-linux-gnu/11.0.0/include -I../arch/x86/include
-I./arch/x86/include/generated -I../include -I./include
-I../arch/x86/include/uapi -I./arch/x86/include/generated/uapi
-I../include/uapi -I./include/generated/uapi -include
../include/linux/kconfig.h -include ../include/linux/compiler_types.h
-D__KERNEL__ -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs
-fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE
-Werror=implicit-function-declaration -Werror=implicit-int -Werror=return-type
-Wno-format-security -std=gnu89 -mno-sse -mno-mmx -mno-sse2 -mno-3dnow -mno-avx
-m64 -falign-jumps=1 -falign-loops=1 -mno-80387 -mno-fp-ret-in-387
-mpreferred-stack-boundary=3 -mskip-rax-setup -mtune=generic -mno-red-zone
-mcmodel=kernel -Wno-sign-compare -fno-asynchronous-unwind-tables
-mindirect-branch=thunk-extern -mindirect-branch-register -fno-jump-tables
-fno-delete-null-pointer-checks -Wno-frame-address -Wno-format-truncation
-Wno-format-overflow -Wno-address-of-packed-member -O2
-fno-allow-store-data-races -Wframe-larger-than=2048 -fstack-protector-strong
-Wno-unused-but-set-variable -Wimplicit-fallthrough -Wno-unused-const-variable
-fomit-frame-pointer -Wdeclaration-after-statement -Wvla -Wno-pointer-sign
-Wno-stringop-truncation -Wno-zero-length-bounds -Wno-array-bounds
-Wno-stringop-overflow -Wno-restrict -Wno-maybe-uninitialized
-fno-strict-overflow -fno-stack-check -fconserve-stack -Werror=date-time
-Werror=incompatible-pointer-types -Werror=designated-init
-fmacro-prefix-map=../= -fcf-protection=none -Wno-packed-not-aligned -I
../net/core -I ./net/core    -DKBUILD_MODFILE='"net/core/utils"'
-DKBUILD_BASENAME='"utils"' -DKBUILD_MODNAME='"utils"' -c -o net/core/utils.o
../net/core/utils.c

here is gcc configure:

# gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-linux-gnu/11.0.0/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../configure --enable-languages=c,c++ --program-suffix=-11
--program-prefix=x86_64-linux-gnu- --build=x86_64-linux-gnu --enable-multilib
--enable-pugin --prefix=/usr
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 11.0.0 20201027 (experimental) (GCC) 

here is trigger  bug source code :
static inline unsigned add32_with_carry(unsigned a, unsigned b)
{
        asm("addl %2,%0\n\t"
            "adcl $0,%0"
            : "=r" (a)
            : "0" (a), "rm" (b));
        return a;
}

gcc-11.0 combine inline-asm 

(insn 43 84 44 5 (parallel [
            (set (reg:SI 126 [ a ])
                (asm_operands:SI ("addl %2,%0
        adcl $0,%0") ("=r") 0 [
                        (reg:SI 126 [ a ])
                        (not:SI (mem/j:SI (plus:DI (reg/v/f:DI 108 [ skb ])
                                    (const_int 136 [0x88])) [169
skb_7(D)->D.41008.csum+0 S4 A64]))
                    ]
                     [
                        (asm_input:SI ("0")
../arch/x86/include/asm/checksum_64.h:171)
                        (asm_input:SI ("rm")
../arch/x86/include/asm/checksum_64.h:171)
                    ]
                     [] ../arch/x86/include/asm/checksum_64.h:171))
            (clobber (reg:CC 17 flags))
        ]) "../arch/x86/include/asm/checksum_64.h":171:2 -1
     (expr_list:REG_UNUSED (reg:CC 17 flags)
        (nil)))

gcc-10 combine inline-asm 

(insn 43 84 44 5 (parallel [
            (set (reg:SI 126 [ a ])
                (asm_operands:SI ("addl %2,%0
        adcl $0,%0") ("=r") 0 [
                        (reg:SI 126 [ a ])
                        (reg:SI 127)
                    ]
                     [
                        (asm_input:SI ("0")
../arch/x86/include/asm/checksum_64.h:171)
                        (asm_input:SI ("rm")
../arch/x86/include/asm/checksum_64.h:171)
                    ]
                     [] ../arch/x86/include/asm/checksum_64.h:171))
            (clobber (reg:CC 17 flags))
        ]) "../arch/x86/include/asm/checksum_64.h":171:2 -1
     (expr_list:REG_DEAD (reg:SI 127)
        (expr_list:REG_UNUSED (reg:CC 17 flags)
            (nil))))


the difference between gcc10 and gcc 11 in asm_operand_ok() recog.c 

gcc-11 is 
case CT_MEMORY:
            case CT_SPECIAL_MEMORY:
              /* Every memory operand can be reloaded to fit.  */
              result = result || memory_operand (extract_mem_from_operand (op),
                                                 VOIDmode);
              break;

and gcc-10 is 

    case CT_MEMORY:
            case CT_SPECIAL_MEMORY:
              /* Every memory operand can be reloaded to fit.  */
              result = result || memory_operand (op, VOIDmode);
              break;

so in gcc-11 the insn  (not:SI (mem/j:SI (plus:DI (reg/v/f:DI 108 [ skb ])
                                    (const_int 136 [0x88])) [169
skb_7(D)->D.41008.csum+0 S4 A64])   
can not reocg  in lra_set_insn_recog_data()

here is the bactrace :
during RTL pass: reload
../net/core/utils.c: In function 'inet_proto_csum_replace_by_diff':
../net/core/utils.c:485:1: internal compiler error: in lra_set_insn_recog_data,
at lra.c:1004
  485 | }
      | ^
0x6a0287 lra_set_insn_recog_data(rtx_insn*)
        ../../gcc/lra.c:1004
0xc23bcf lra_get_insn_recog_data
        ../../gcc/lra-int.h:488
0xc23bcf remove_scratches_1
        ../../gcc/lra.c:2064
0xc23d17 lra_emit_move(rtx_def*, rtx_def*)
        ../../gcc/lra.c:506
0xc361d6 curr_insn_transform
        ../../gcc/lra-constraints.c:4523
0xc37666 lra_constraints(bool)
        ../../gcc/lra-constraints.c:5123
0xc2426c lra(_IO_FILE*)
        ../../gcc/lra.c:2415
0xbdfde1 do_reload
        ../../gcc/ira.c:5529
0xbdfde1 execute
        ../../gcc/ira.c:5715
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

reported by TencentOS team

[Bug inline-asm/97667] [11 Regression] a bug in asm_operand_ok() recog.c:1801

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97667

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code
             Target|                            |x86_64-*-*
   Target Milestone|---                         |11.0
            Summary|a bug in asm_operand_ok()   |[11 Regression] a bug in
                   |recog.c:1801                |asm_operand_ok()
                   |                            |recog.c:1801

[Bug inline-asm/97667] [11 Regression] a bug in asm_operand_ok() recog.c:1801

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97667

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |crazylht at gmail dot com,
                   |                            |jakub at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
I think we have a dup of this already.

[Bug inline-asm/97667] [11 Regression] a bug in asm_operand_ok() recog.c:1801

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97667

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |marxin at gcc dot gnu.org
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2020-11-02

--- Comment #2 from Martin Liška <marxin at gcc dot gnu.org> ---
Likely dup of PR97540.

[Bug inline-asm/97667] [11 Regression] a bug in asm_operand_ok() recog.c:1801

[crazylht at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97667

--- Comment #3 from Hongtao.liu <crazylht at gmail dot com> ---
(In reply to Martin Liška from comment #2)
> Likely dup of PR97540.

Yes, it should be.

[Bug inline-asm/97667] [11 Regression] a bug in asm_operand_ok() recog.c:1801

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97667

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |DUPLICATE
             Status|NEW                         |RESOLVED

--- Comment #4 from Martin Liška <marxin at gcc dot gnu.org> ---
Dup.

*** This bug has been marked as a duplicate of bug 97540 ***

[Bug inline-asm/97667] [11 Regression] a bug in asm_operand_ok() recog.c:1801

[huntazhang at tencent dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97667

--- Comment #5 from huntazhang <huntazhang at tencent dot com> ---
YES DUP.