Bug 5871 - eglibc build failure on several ppc machines
Summary: eglibc build failure on several ppc machines
Status: CLOSED FIXED
Alias: None
Product: BSPs
Classification: Build System, Metadata & Runtime
Component: bsps-meta-fsl-ppc (show other bugs)
Version: 1.6
Hardware: Other ppc
: Medium+ critical
Target Milestone: 1.6 M4
Assignee: Zhenhua Luo
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2014-02-24 16:03 UTC by Tudor Florea
Modified: 2014-04-08 02:35 UTC (History)
5 users (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: No (bug/feature does not impact docs)

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Tudor Florea 2014-02-24 16:03:13 UTC 
Below is one example of the build fail:
bitbake eglibc
Loading cache: 100% |############################################################################################################| ETA:  00:00:00
Loaded 1255 entries from dependency cache.
Parsing recipes: 100% |##########################################################################################################| Time: 00:00:01
Parsing of 877 .bb files complete (789 cached, 88 parsed). 1250 targets, 63 skipped, 0 masked, 0 errors.
NOTE: Resolving any missing task queue dependencies

Build Configuration:
BB_VERSION        = "1.21.1"
BUILD_SYS         = "x86_64-linux"
NATIVELSBSTRING   = "Ubuntu-12.10"
TARGET_SYS        = "powerpc-poky-linux"
MACHINE           = "p2041rdb"
DISTRO            = "poky"
DISTRO_VERSION    = "1.5+snapshot-20140224"
TUNE_FEATURES     = "m32 fpu-hard ppce500mc"
TARGET_FPU        = "hard"
meta              = "master:8d7dd9ae471ad6154cd4e59152560ff5f24b3c2a"
meta-fsl-ppc      = ...[custom]...
meta-yocto        
meta-yocto-bsp    = "master:8d7dd9ae471ad6154cd4e59152560ff5f24b3c2a"
...
powerpc-enea-linux-gcc  -m32 -mhard-float -mcpu=e500mc --sysroot/path/tobuild_p2041rdb/tmp/sysroots/p2041rdb-tcbootstrap ../sysdeps/ieee754/dbl-64/k_tan.c -c -std=gnu99 -fgnu89-inline  -O2 -Wall -Winline -Wwrite-strings -feliminate-unused-debug-types -fmerge-all-constants -frounding-math -g -mcpu=e500mc -pipe -mhard-float -Wstrict-prototypes -mlong-double-128     -Wno-uninitialized   -D__NO_MATH_INLINES -D__LIBC_INTERNAL_MATH_INLINES -I../include -I/media/sdd1/fb/tufl/linux/master-enea15/build_p2041rdb/tmp/work/ppce500mc-enea-linux/eglibc/2.19-r0/build-powerpc-enea-linux/math  -I/media/sdd1/fb/tufl/linux/master-enea15/build_p2041rdb/tmp/work/ppce500mc-enea-linux/eglibc/2.19-r0/build-powerpc-enea-linux  -I../sysdeps/unix/sysv/linux/powerpc/powerpc32/e500mc/fpu  -I../sysdeps/powerpc/powerpc32/e500mc/fpu  -I../ports/sysdeps/powerpc/powerpc32/e500mc/fpu  -I../sysdeps/unix/sysv/linux/powerpc/powerpc32/e500mc  -I../sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/nptl  -I../sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu  -I../nptl/sysdeps/unix/sysv/linux/powerpc/powerpc32  -I../sysdeps/unix/sysv/linux/powerpc/powerpc32  -I../nptl/sysdeps/unix/sysv/linux/powerpc  -I../sysdeps/unix/sysv/linux/powerpc  -I../nptl/sysdeps/unix/sysv/linux  -I../nptl/sysdeps/pthread  -I../sysdeps/pthread  -I../ports/sysdeps/unix/sysv/linux  -I../sysdeps/unix/sysv/linux  -I../sysdeps/gnu  -I../sysdeps/unix/inet  -I../nptl/sysdeps/unix/sysv  -I../ports/sysdeps/unix/sysv  -I../sysdeps/unix/sysv  -I../sysdeps/unix/powerpc  -I../nptl/sysdeps/unix  -I../ports/sysdeps/unix  -I../sysdeps/unix  -I../sysdeps/posix  -I../ports/sysdeps/powerpc/powerpc32/e500mc  -I../sysdeps/powerpc/powerpc32/e500mc  -I../sysdeps/powerpc/powerpc32/fpu  -I../ports/sysdeps/powerpc/powerpc32  -I../sysdeps/powerpc/powerpc32  -I../sysdeps/wordsize-32  -I../sysdeps/powerpc/fpu  -I../nptl/sysdeps/powerpc  -I../ports/sysdeps/powerpc  -I../sysdeps/powerpc  -I../sysdeps/ieee754/ldbl-128ibm  -I../sysdeps/ieee754/ldbl-opt  -I../sysdeps/ieee754/dbl-64  -I../sysdeps/ieee754/flt-32  -I../sysdeps/ieee754  -I../sysdeps/generic  -I../nptl  -I../ports  -I.. -I../libio -I. -nostdinc -isystem /media/sdd1/fb/tufl/linux/master-enea15/build_p2041rdb/tmp/sysroots/x86_64-linux/usr/lib/ppce500mc-enea-linux.gcc-cross-initial/gcc/powerpc-enea-linux/4.8.2/include -isystem /media/sdd1/fb/tufl/linux/master-enea15/build_p2041rdb/tmp/sysroots/x86_64-linux/usr/lib/ppce500mc-enea-linux.gcc-cross-initial/gcc/powerpc-enea-linux/4.8.2/include-fixed -isystem /media/sdd1/fb/tufl/linux/master-enea15/build_p2041rdb/tmp/sysroots/p2041rdb/usr/include  -D_LIBC_REENTRANT -include ../include/libc-symbols.h   -DNOT_IN_libc=1 -DIS_IN_libm=1 -DIN_LIB=libm    -o /media/sdd1/fb/tufl/linux/master-enea15/build_p2041rdb/tmp/work/ppce500mc-enea-linux/eglibc/2.19-r0/build-powerpc-enea-linux/math/k_tan.o -MD -MP -MF /media/sdd1/fb/tufl/linux/master-enea15/build_p2041rdb/tmp/work/ppce500mc-enea-linux/eglibc/2.19-r0/build-powerpc-enea-linux/math/k_tan.o.dt -MT /media/sdd1/fb/tufl/linux/master-enea15/build_p2041rdb/tmp/work/ppce500mc-enea-linux/eglibc/2.19-r0/build-powerpc-enea-linux/math/k_tan.o
../sysdeps/powerpc/powerpc32/e500mc/fpu/e_sqrt.c: In function '__slow_ieee754_sqrt':
../sysdeps/powerpc/powerpc32/e500mc/fpu/e_sqrt.c:128:15: error: subscripted value is neither array nor pointer nor vector
       if ((u.l[1] & FE_INVALID) == 0)
               ^
../sysdeps/powerpc/powerpc32/e500mc/fpu/e_sqrt.c:127:20: warning: variable 'u' set but not used [-Wunused-but-set-variable]
       fenv_union_t u = { .fenv = fegetenv_register () };
                    ^
Comment 1 Saul Wold 2014-02-27 09:12:12 UTC 
This is more of a FSL issue since you are including the meta-fsl-ppc layer and may set additional flags for compiler/linker.
Comment 2 Tudor Florea 2014-02-27 16:10:41 UTC 
This is a ppc specific problem that comes from this patch: pmeta/recipes-core/eglibc/eglibc-2.19/ppc_slow_ieee754_sqrt.patch
Comment 3 Khem Raj 2014-02-28 03:17:49 UTC 
Please test out the patch I posted to contrib tree

http://git.openembedded.org/openembedded-core-contrib/commit/?h=kraj/misc&id=c5a2242079722ab266f4c9c9cddcd75caa3f1505
Comment 4 Zhenhua Luo 2014-03-03 11:07:19 UTC 
(In reply to comment #3)
> Please test out the patch I posted to contrib tree
> 
> http://git.openembedded.org/openembedded-core-contrib/commit/?h=kraj/
> misc&id=c5a2242079722ab266f4c9c9cddcd75caa3f1505

The build issue goes away after applying the patch.
Comment 5 Khem Raj 2014-03-03 16:30:01 UTC 
(In reply to comment #4)
> (In reply to comment #3)
> > Please test out the patch I posted to contrib tree
> > 
> > http://git.openembedded.org/openembedded-core-contrib/commit/?h=kraj/
> > misc&id=c5a2242079722ab266f4c9c9cddcd75caa3f1505
> 
> The build issue goes away after applying the patch.

How about boot and run
Comment 6 Chunrong Guo 2014-03-11 04:09:43 UTC 
apply the patch and test system in p4080ds boards.
the system can boot successfully and gcc work well.


1. Test step:
  (1) tftp 1000000 p4080ds/uImagetest.bin;tftp c00000 p4080ds/uImage.dtb
  (2) setenv bootargs root=/dev/nfs rw nfsroot=10.193.20.117:/home/guo/tool/nfsroot ip=10.193.20.169:10.193.20.117:10.193.20.254:255.255.254.0:p4080ds:eth0:off console=ttyS0,115200 $othbootargs;bootm 1000000 - c00000

the system can boot successfully

2. Test gcc in rootfs:
gcc -lpthread -o main main.c
./main

gcc work well.
Comment 7 Tudor Florea 2014-03-12 01:31:12 UTC 
Build and boot tested.
There are any testcases available to verify the correctness of __slow_ieee754_sqrt implementation?
Comment 8 Chunrong Guo 2014-03-12 03:10:40 UTC 
I will verify "__slow_ieee754_sqrt " by testfloat case.
Comment 9 Chunrong Guo 2014-03-13 02:40:53 UTC 
testfloat case work well.

the following information:
root@p1020rdb:~# testfloat -all
Testing int32_to_float32, rounding nearest_even.
372 tests total.
95 tests performed.
No errors found in int32_to_float32, rounding nearest_even.
Testing int32_to_float32, rounding to_zero.
372 tests total.
77 tests performed.
No errors found in int32_to_float32, rounding to_zero.
Testing int32_to_float32, rounding up.
372 tests total.
92 tests performed.
No errors found in int32_to_float32, rounding up.
Testing int32_to_float32, rounding down.
372 tests total.
92 tests performed.
No errors found in int32_to_float32, rounding down.
Testing int32_to_float64.
372 tests total.
0 tests performed.
No errors found in int32_to_float64.
Testing float32_to_int32_round_to_zero.
Comment 10 Richard Purdie 2014-04-04 15:00:11 UTC 
This was resolved in master: http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=b130d75c037c3d458817711221a86aa74903b509
Comment 11 Chunrong Guo 2014-04-08 02:35:26 UTC 
the patch is verified.