[Bug c++/66239] New: Unoptimized sqrt(float or double) returns wrong values for ARM Cortex-A8 -mfloat-abi=[soft,softfp]
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66239
Bug ID: 66239
Summary: Unoptimized sqrt(float or double) returns wrong values
for ARM Cortex-A8 -mfloat-abi=[soft,softfp]
Product: gcc
Version: unknown
Status: UNCONFIRMED
Severity: major
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: maciej.andrzejewski at data dot pl
Target Milestone: ---
Tested both on GCC 4.9.1 (compiled toolchain using crosstool-ng) and 4.9.2
(compiled toolchain using buildroot).
Simple code:
-- CODE --
#include
#include
#include
int main( void )
{
double sq3 = 9.1;
double ret3 = sqrtf(sq3);
printf("%f\n", ret3);
float sq4 = 9.1;
double ret4 = sqrtf(sq4);
printf("%f\n", ret4);
double sq1 = 9.1;
double ret1 = sqrt(sq1);
printf("%f\n", ret1);
float sq2 = 9.1;
double ret2 = sqrt(sq2);
printf("%f\n", ret2);
return 0;
}
-- CODE --
compiled with command:
g++ -mcpu=cortex-a8 -march=armv7-a -mtune=cortex-a8 -mfpu=neon
-mthumb-interwork -mfpu=neon -Wall -Wextra -mfloat-abi=softfp
or
g++ -mcpu=cortex-a8 -march=armv7-a -mtune=cortex-a8 -mfpu=neon
-mthumb-interwork -mfpu=neon -Wall -Wextra -mfloat-abi=soft
prints output when run:
-- OUTPUT --
0.00
0.00
89884613882771507935772421602449274280826426490922860415370742828850803088708436568909338933268615382725731836148779976703521876921396883553861971381426763394584730974161643341227168116324626810241973964225774272233175406843466371141886318608237834273423597057507238373804952327520531541920130815989439791104.00
89884613882771507935772421602449274280826426490922860415370742828850803088708436568909338933268615382725731836148779976703521876921396883553861971381426763394584730974161643341227168116324626810241973964225774272233175406843466371141886318608237834273423597057507238373804952327520531541920130815989439791104.00
-- OUTPUT --
when compiled with added ANY optimization flag (O, O1, O2, O3, Og) prints
proper output:
-- OUTPUT 2 --
3.016621
3.016621
3.016621
3.016621
-- OUTPUT 2 --
*** Additional info ***
g++ --version:
arm-none-linux-gnueabi-g++ (Buildroot 2015.02) 4.9.2
readelf -a a.out:
-- READELF --
macieja@ubuntu:~/Projects/sqrtf$ readelf -a a.out
ELF Header:
Magic: 7f 45 4c 46 01 01 01 00 00 00 00 00 00 00 00 00
Class: ELF32
Data: 2's complement, little endian
Version: 1 (current)
OS/ABI:UNIX - System V
ABI Version: 0
Type: EXEC (Executable file)
Machine: ARM
Version: 0x1
Entry point address: 0x103d0
Start of program headers: 52 (bytes into file)
Start of section headers: 5012 (bytes into file)
Flags: 0x5000202, has entry point, Version5 EABI,
soft-float ABI
Size of this header: 52 (bytes)
Size of program headers: 32 (bytes)
Number of program headers: 8
Size of section headers: 40 (bytes)
Number of section headers: 29
Section header string table index: 26
Section Headers:
[Nr] Name TypeAddr OffSize ES Flg Lk Inf
Al
[ 0] NULL 00 00 00 0 0
0
[ 1] .interp PROGBITS00010134 000134 13 00 A 0 0
1
[ 2] .note.ABI-tag NOTE00010148 000148 20 00 A 0 0
4
[ 3] .hash HASH00010168 000168 38 04 A 4 0
4
[ 4] .dynsym DYNSYM 000101a0 0001a0 90 10 A 5 1
4
[ 5] .dynstr STRTAB 00010230 000230 d3 00 A 0 0
1
[ 6] .gnu.version VERSYM 00010304 000304 12 02 A 4 0
2
[ 7] .gnu.version_rVERNEED 00010318 000318 40 00 A 5 2
4
[ 8] .rel.dyn REL 00010358 000358 08 08 A 4 0
4
[ 9] .rel.plt REL 00010360 000360 20 08 AI 4 11
4
[10] .init PROGBITS00010380 000380 0c 00 AX 0 0
4
[11] .plt PROGBITS0001038c 00038c 44 04 AX 0 0
4
[12] .text PROGBITS000103d0 0003d0 0001f8 00 AX 0 0
4
[13] .fini PROGBITS000105c8 0005c8 08 00 AX 0 0
4
[14] .rodata PROGBITS000105d0 0005d0 08 00 A 0 0
4
[15] .ARM.exidxARM_EXIDX 000105d8 0005d8 18 00 AL 12 0
4
[16] .eh_frame PROGBITS000105f0 0005f0 04 00 A 0 0
4
[17] .init_array INIT_ARRAY 000205f4 0005f4 04 00 WA 0 0
4
[18] .fini_array FINI_ARRAY 000205
[Bug c++/66239] Unoptimized sqrt(float or double) returns wrong values for ARM Cortex-A8 -mfloat-abi=[soft,softfp]
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66239
--- Comment #1 from Maciej Andrzejewski ---
It is getting even more interesting.
I have disassabled 4 binaries compiled with options:
1) -mfloat-abi=softfp
2) -mfloat-abi=softfp -O
3) -mfloat-abi=hard
4) -mfloat-abi=hard -O
and from what I understand if we turn ON the optimization the FPU is turned
OFF!
I dont see in assembler that FPU s** registers are used in those two cases
where optimization is turned on:
-- DISASSAMBLE OPTION 1 --
00010570 :
10570: e92d4800push{fp, lr}
10574: e28db004add fp, sp, #4
10578: e24dd040sub sp, sp, #64 ; 0x40
1057c: e3032333movwr2, #13107 ; 0x
10580: e3432333movtr2, #13107 ; 0x
10584: e303movwr3, #13107 ; 0x
10588: e3443022movtr3, #16418 ; 0x4022
1058c: e14b20fcstrdr2, [fp, #-12]
10590: ed5b0b03vldrd16, [fp, #-12]
10594: eef77be0vcvt.f32.f64s15, d16
10598: ee170a90vmovr0, s15
1059c: eba1bl 10428
105a0: ee070a90vmovs15, r0
105a4: eef70ae7vcvt.f64.f32d16, s15
105a8: ed4b0b05vstrd16, [fp, #-20] ; 0xffec
105ac: e30006ccmovwr0, #1740 ; 0x6cc
105b0: e341movtr0, #1
105b4: e14b21d4ldrdr2, [fp, #-20] ; 0xffec
105b8: eba0bl 10440
105bc: e309399amovwr3, #39322 ; 0x999a
105c0: e3443111movtr3, #16657 ; 0x4111
105c4: e50b3018str r3, [fp, #-24] ; 0xffe8
105c8: e51b0018ldr r0, [fp, #-24] ; 0xffe8
105cc: eb95bl 10428
105d0: ee070a90vmovs15, r0
105d4: eef70ae7vcvt.f64.f32d16, s15
105d8: ed4b0b09vstrd16, [fp, #-36] ; 0xffdc
105dc: e30006ccmovwr0, #1740 ; 0x6cc
105e0: e341movtr0, #1
105e4: e14b22d4ldrdr2, [fp, #-36] ; 0xffdc
105e8: eb94bl 10440
105ec: e3032333movwr2, #13107 ; 0x
105f0: e3432333movtr2, #13107 ; 0x
105f4: e303movwr3, #13107 ; 0x
105f8: e3443022movtr3, #16418 ; 0x4022
105fc: e14b22fcstrdr2, [fp, #-44] ; 0xffd4
10600: e14b02dcldrdr0, [fp, #-44] ; 0xffd4
10604: eb8abl 10434
10608: e14b03f4strdr0, [fp, #-52] ; 0xffcc
1060c: e30006ccmovwr0, #1740 ; 0x6cc
10610: e341movtr0, #1
10614: e14b23d4ldrdr2, [fp, #-52] ; 0xffcc
10618: eb88bl 10440
1061c: e309399amovwr3, #39322 ; 0x999a
10620: e3443111movtr3, #16657 ; 0x4111
10624: e50b3038str r3, [fp, #-56] ; 0xffc8
10628: ed5b7a0evldrs15, [fp, #-56] ; 0xffc8
1062c: eef70ae7vcvt.f64.f32d16, s15
10630: ec510b30vmovr0, r1, d16
10634: eb7ebl 10434
10638: e14b04f4strdr0, [fp, #-68] ; 0xffbc
1063c: e30006ccmovwr0, #1740 ; 0x6cc
10640: e341movtr0, #1
10644: e14b24d4ldrdr2, [fp, #-68] ; 0xffbc
10648: eb7cbl 10440
1064c: e3a03000mov r3, #0
10650: e1a3mov r0, r3
10654: e24bd004sub sp, fp, #4
10658: e8bd8800pop {fp, pc}
-- DISASSAMBLE OPTION 1 --
-- DISASSAMBLE OPTION 2 --
000104f4 :
104f4: e92d40d0push{r4, r6, r7, lr}
104f8: e30045d4movwr4, #1492 ; 0x5d4
104fc: e3404001movtr4, #1
10500: e3a06000mov r6, #0
10504: e302720amovwr7, #8714 ; 0x220a
10508: e3447008movtr7, #16392 ; 0x4008
1050c: e1a4mov r0, r4
10510: e1a02006mov r2, r6
10514: e1a03007mov r3, r7
10518: eba9bl 103c4
1051c: e1a4mov r0, r4
10520: e1a02006mov r2, r6
10524: e1a03007mov r3, r7
10528: eba5bl 103c4
1052c: e1a4mov r0, r4
10530: e30f2d38movwr2, #64824 ; 0xfd38
10534: e34f2ea1movtr2, #65185 ; 0xfea1
10538: e3023209movwr3, #8713 ; 0x2209
[Bug c++/66239] Unoptimized sqrt(float or double) returns wrong values for ARM Cortex-A8 -mfloat-abi=[soft,softfp]
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66239 --- Comment #6 from Maciej Andrzejewski --- Libm and other libs have been mixed. Properly compiled toolchain with libs solved the problem. Thank you all for assistance!
[Bug c++/66239] Unoptimized sqrt(float or double) returns wrong values for ARM Cortex-A8 -mfloat-abi=[soft,softfp]
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66239 Maciej Andrzejewski changed: What|Removed |Added Status|WAITING |RESOLVED Resolution|--- |INVALID --- Comment #5 from Maciej Andrzejewski --- Libm and other libs have been mixed. Properly compiled toolchain with libs solved the problem. Thank you all for assistance!
