From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 1256) id 5C3023858C50; Wed, 6 Apr 2022 22:46:00 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 5C3023858C50 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable From: Maciej W. Rozycki To: gdb-cvs@sourceware.org Subject: [binutils-gdb] gdb: mips: Fix the handling of complex type of function return value X-Act-Checkin: binutils-gdb X-Git-Author: Youling Tang X-Git-Refname: refs/heads/master X-Git-Oldrev: 8fb105930851c7a722b4794f8348f72f29721cd0 X-Git-Newrev: 089169c0037bc2ae153580de56df6acceab738fc Message-Id: <20220406224600.5C3023858C50@sourceware.org> Date: Wed, 6 Apr 2022 22:46:00 +0000 (GMT) X-BeenThere: gdb-cvs@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-cvs mailing list List-Unsubscribe: , List-Archive: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 06 Apr 2022 22:46:00 -0000 https://sourceware.org/git/gitweb.cgi?p=3Dbinutils-gdb.git;h=3D089169c0037b= c2ae153580de56df6acceab738fc commit 089169c0037bc2ae153580de56df6acceab738fc Author: Youling Tang Date: Wed Apr 6 23:38:21 2022 +0100 gdb: mips: Fix the handling of complex type of function return value =20 $ objdump -d outputs/gdb.base/varargs/varargs 00000001200012e8 : ... 1200013b8: c7c10000 lwc1 $f1,0(s8) 1200013bc: c7c00004 lwc1 $f0,4(s8) 1200013c0: 46000886 mov.s $f2,$f1 1200013c4: 46000046 mov.s $f1,$f0 1200013c8: 46001006 mov.s $f0,$f2 1200013cc: 46000886 mov.s $f2,$f1 1200013d0: 03c0e825 move sp,s8 1200013d4: dfbe0038 ld s8,56(sp) 1200013d8: 67bd0080 daddiu sp,sp,128 1200013dc: 03e00008 jr ra 1200013e0: 00000000 nop =20 From the above disassembly, we can see that when the return value of the function is a complex type and len <=3D 2 * MIPS64_REGSIZE, the return = value will be passed through $f0 and $f2, so fix the corresponding processing in mips_n32n64_return_value(). =20 $ make check RUNTESTFLAGS=3D'GDB=3D../gdb gdb.base/varargs.exp --outdir= =3Dtest' =20 Before applying the patch: FAIL: gdb.base/varargs.exp: print find_max_float_real(4, fc1, fc2, fc3= , fc4) FAIL: gdb.base/varargs.exp: print find_max_double_real(4, dc1, dc2, dc= 3, dc4) =20 # of expected passes 9 # of unexpected failures 2 =20 After applying the patch: # of expected passes 11 =20 This also fixes: FAIL: gdb.base/callfuncs.exp: call inferior func with struct - returns= float _Complex =20 Signed-off-by: Youling Tang Co-Authored-By: Maciej W. Rozycki Diff: --- gdb/mips-tdep.c | 34 ++++++++++++++++++++++++---------- 1 file changed, 24 insertions(+), 10 deletions(-) diff --git a/gdb/mips-tdep.c b/gdb/mips-tdep.c index 93945891407..4bdbf0fcdb7 100644 --- a/gdb/mips-tdep.c +++ b/gdb/mips-tdep.c @@ -5217,30 +5217,44 @@ mips_n32n64_return_value (struct gdbarch *gdbarch, = struct value *function, that all composite results be handled by conversion to implicit first parameters. The MIPS/SGI Fortran implementation has always made a specific exception to return COMPLEX results in the floating point - registers.] */ + registers.] + + From MIPSpro Assembly Language Programmer's Guide, Document Number: + 007-2418-004 + + Software + Register Name(from + Name fgregdef.h) Use and Linkage + ----------------------------------------------------------------- + $f0, $f2 fv0, fv1 Hold results of floating-point type function + ($f0) and complex type function ($f0 has the + real part, $f2 has the imaginary part.) */ =20 if (TYPE_LENGTH (type) > 2 * MIPS64_REGSIZE) return RETURN_VALUE_STRUCT_CONVENTION; - else if (type->code () =3D=3D TYPE_CODE_FLT - && TYPE_LENGTH (type) =3D=3D 16 + else if ((type->code () =3D=3D TYPE_CODE_COMPLEX + || (type->code () =3D=3D TYPE_CODE_FLT && TYPE_LENGTH (type) =3D=3D 1= 6)) && tdep->mips_fpu_type !=3D MIPS_FPU_NONE) { - /* A 128-bit floating-point value fills both $f0 and $f2. The - two registers are used in the same as memory order, so the - eight bytes with the lower memory address are in $f0. */ + /* A complex value of up to 128 bits in width as well as a 128-bit + floating-point value goes in both $f0 and $f2. A single complex + value is held in the lower halves only of the respective registers. + The two registers are used in the same as memory order, so the + bytes with the lower memory address are in $f0. */ if (mips_debug) gdb_printf (gdb_stderr, "Return float in $f0 and $f2\n"); mips_xfer_register (gdbarch, regcache, (gdbarch_num_regs (gdbarch) + mips_regnum (gdbarch)->fp0), - 8, gdbarch_byte_order (gdbarch), + TYPE_LENGTH (type) / 2, gdbarch_byte_order (gdbarch), readbuf, writebuf, 0); mips_xfer_register (gdbarch, regcache, (gdbarch_num_regs (gdbarch) + mips_regnum (gdbarch)->fp0 + 2), - 8, gdbarch_byte_order (gdbarch), - readbuf ? readbuf + 8 : readbuf, - writebuf ? writebuf + 8 : writebuf, 0); + TYPE_LENGTH (type) / 2, gdbarch_byte_order (gdbarch), + readbuf ? readbuf + TYPE_LENGTH (type) / 2 : readbuf, + (writebuf + ? writebuf + TYPE_LENGTH (type) / 2 : writebuf), 0); return RETURN_VALUE_REGISTER_CONVENTION; } else if (type->code () =3D=3D TYPE_CODE_FLT