public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug target/97032] New: [8/9/10/11 Regression] ICE output_operand: invalid use of register 'frame' since r8-1911
@ 2020-09-12 11:16 jakub at gcc dot gnu.org
  2020-09-12 11:16 ` [Bug target/97032] " jakub at gcc dot gnu.org
                   ` (8 more replies)
  0 siblings, 9 replies; 10+ messages in thread
From: jakub at gcc dot gnu.org @ 2020-09-12 11:16 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97032

            Bug ID: 97032
           Summary: [8/9/10/11 Regression] ICE output_operand: invalid use
                    of register 'frame' since r8-1911
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: jakub at gcc dot gnu.org
  Target Milestone: ---

Since r8-1911-g35c95658180e67503f9de96567ab4d6e71b10a6e the following testcase
distilled from firefox:
#include <stdarg.h>

extern int *__errno_location (void);

long
sys_socketcall (int op, ...)
{
  long int res;
  va_list ap;
  va_start (ap, op);
  asm volatile ("push %%ebx; movl %2, %%ebx; int $0x80; pop %%ebx"
                : "=a" (res) : "0" (102), "ri" (16), "c" (ap) : "memory",
"esp");
  if (__builtin_expect (res > 4294963200UL, 0))
    *__errno_location () = -res;
  va_end (ap);
  return res;
}
ICEs on x86_64-linux or i686-linux with
-O2 -m32 -mincoming-stack-boundary=2 -fstack-protector-all

test.c: In function ‘sys_socketcall’:
test.c:11:3: warning: listing the stack pointer register ‘esp’ in a clobber
list is deprecated [-Wdeprecated]
   11 |   asm volatile ("push %%ebx; movl %2, %%ebx; int $0x80; pop %%ebx"
      |   ^~~
test.c:11:3: note: the value of the stack pointer after an ‘asm’ statement must
be the same as it was before the statement
during RTL pass: final
test.c:17:1: internal compiler error: output_operand: invalid use of register
'frame'
   17 | }
      | ^
0xd3b872 output_operand_lossage(char const*, ...)
        ../../gcc/final.c:3609
0x17e0869 print_reg(rtx_def*, int, _IO_FILE*)
        ../../gcc/config/i386/i386.c:12274
0x17e2fa1 ix86_print_operand_address_as
        ../../gcc/config/i386/i386.c:13311
0x17e22d2 ix86_print_operand(_IO_FILE*, rtx_def*, int)
        ../../gcc/config/i386/i386.c:13060
0xd3c92a output_operand(rtx_def*, int)
        ../../gcc/final.c:4051
0xd3c550 output_asm_insn(char const*, rtx_def**)
        ../../gcc/final.c:3963
0x1f325e1 output_1129
        ../../gcc/config/i386/i386.md:20656
0xd38b91 get_insn_template(int, rtx_insn*)
        ../../gcc/final.c:2070
0xd3a8b6 final_scan_insn_1
        ../../gcc/final.c:3039
0xd3ad3e final_scan_insn(rtx_insn*, _IO_FILE*, int, int, int*)
        ../../gcc/final.c:3152
0xd389c0 final_1
        ../../gcc/final.c:2020
0xd3dd0d rest_of_handle_final
        ../../gcc/final.c:4658
0xd3e06c execute
        ../../gcc/final.c:4736
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.

Seems we set crtl->stack_realign_needed, which prevents elimination of frame
register to hard frame pointer, but have the frame register used in the IL and
therefore not eliminated.

^ permalink raw reply	[flat|nested] 10+ messages in thread
end of thread, other threads:[~2020-09-24 17:44 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-09-12 11:16 [Bug target/97032] New: [8/9/10/11 Regression] ICE output_operand: invalid use of register 'frame' since r8-1911 jakub at gcc dot gnu.org
2020-09-12 11:16 ` [Bug target/97032] " jakub at gcc dot gnu.org
2020-09-12 11:18 ` jakub at gcc dot gnu.org
2020-09-12 14:16 ` hjl.tools at gmail dot com
2020-09-12 17:12 ` hjl.tools at gmail dot com
2020-09-16 12:31 ` cvs-commit at gcc dot gnu.org
2020-09-24 16:05 ` cvs-commit at gcc dot gnu.org
2020-09-24 16:16 ` cvs-commit at gcc dot gnu.org
2020-09-24 17:40 ` cvs-commit at gcc dot gnu.org
2020-09-24 17:44 ` hjl.tools at gmail dot com

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).