[OE-core] [PATCH 1/3] glibc: Upgrade to 2.23

Fri Apr 1 14:11:11 UTC 2016

On Fri, Apr 1, 2016 at 1:44 AM, ChenQi <Qi.Chen at windriver.com> wrote:
> On 03/31/2016 03:01 PM, ChenQi wrote:
>>
>> On 03/29/2016 10:16 AM, Khem Raj wrote:
>>>>
>>>> On Mar 28, 2016, at 6:38 PM, ChenQi <Qi.Chen at windriver.com> wrote:
>>>>
>>>> On 02/23/2016 06:31 AM, Andre McCurdy wrote:
>>>>>
>>>>> On Sat, Feb 20, 2016 at 12:07 AM, Khem Raj <raj.khem at gmail.com> wrote:
>>>>>>
>>>>>> Drop kconfig and options-group support
>>>>>> Forward port cross-localedef support
>>>>>> Assume ssp support in libc when building gcc-initial
>>>>>
>>>>> Breaks for me when building for x86 with -Os (building with
>>>>> -O1/-O2/-O3 are all OK):
>>>>>
>>>>> i686-rdk-linux-gcc  -m32 -march=core2 -mtune=core2 -msse3 -mfpmath=sse
>>>>>
>>>>> --sysroot=/home/andre/rdk/rdk-master/build-vbox32/tmp/sysroots/vbox32-tcbootstrap
>>>>> ../sysdeps/unix/sysv/linux/posix_fallocate.c -c -std=gnu11
>>>>> -fgnu89-inline  -Os -Wall -Werror -Wno-error -Wundef -Wwrite-strings
>>>>>
>>>>> -fdebug-prefix-map=/home/andre/rdk/rdk-master/build-vbox32/tmp/sysroots/vbox32=
>>>>>
>>>>> -fdebug-prefix-map=/home/andre/rdk/rdk-master/build-vbox32/tmp/sysroots/x86_64-linux=
>>>>>
>>>>> -fdebug-prefix-map=/home/andre/rdk/rdk-master/build-vbox32/tmp/work/core2-32-rdk-linux/glibc/2.23-r0/build-i686-rdk-linux=/usr/src/glibc
>>>>>
>>>>> -fdebug-prefix-map=/home/andre/rdk/rdk-master/build-vbox32/tmp/work/core2-32-rdk-linux/glibc/2.23-r0/git=/usr/src/glibc
>>>>> -feliminate-unused-debug-types -fmerge-all-constants -frounding-math
>>>>> -g -pipe -Wstrict-prototypes -Wold-style-definition -Wa,-mtune=i686
>>>>> -fexceptions  -ftls-model=initial-exec      -I../include
>>>>>
>>>>> -I/home/andre/rdk/rdk-master/build-vbox32/tmp/work/core2-32-rdk-linux/glibc/2.23-r0/build-i686-rdk-linux/io
>>>>>
>>>>> -I/home/andre/rdk/rdk-master/build-vbox32/tmp/work/core2-32-rdk-linux/glibc/2.23-r0/build-i686-rdk-linux
>>>>>   -I../sysdeps/unix/sysv/linux/i386/i686 -I../sysdeps/i386/i686/nptl
>>>>> -I../sysdeps/unix/sysv/linux/i386 -I../sysdeps/unix/sysv/linux/x86
>>>>> -I../sysdeps/i386/nptl  -I../sysdeps/unix/sysv/linux/include
>>>>> -I../sysdeps/unix/sysv/linux  -I../sysdeps/nptl -I../sysdeps/pthread
>>>>> -I../sysdeps/gnu  -I../sysdeps/unix/inet -I../sysdeps/unix/sysv
>>>>> -I../sysdeps/unix/i386  -I../sysdeps/unix -I../sysdeps/posix
>>>>> -I../sysdeps/i386/i686/fpu/multiarch -I../sysdeps/i386/i686/fpu
>>>>> -I../sysdeps/i386/i686/multiarch  -I../sysdeps/i386/i686
>>>>> -I../sysdeps/i386/fpu  -I../sysdeps/x86/fpu/include
>>>>> -I../sysdeps/x86/fpu  -I../sysdeps/i386  -I../sysdeps/x86
>>>>> -I../sysdeps/wordsize-32  -I../sysdeps/ieee754/ldbl-96
>>>>> -I../sysdeps/ieee754/dbl-64  -I../sysdeps/ieee754/flt-32
>>>>> -I../sysdeps/ieee754  -I../sysdeps/generic  -I.. -I../libio -I.
>>>>> -nostdinc -isystem
>>>>>
>>>>> /home/andre/rdk/rdk-master/build-vbox32/tmp/sysroots/x86_64-linux/usr/bin/i686-rdk-linux.gcc-cross-initial-i686/../../lib/i686-rdk-linux.gcc-cross-initial-i686/gcc/i686-rdk-linux/5.3.0/include
>>>>> -isystem
>>>>> /home/andre/rdk/rdk-master/build-vbox32/tmp/sysroots/x86_64-linux/usr/bin/i686-rdk-linux.gcc-cross-initial-i686/../../lib/i686-rdk-linux.gcc-cross-initial-i686/gcc/i686-rdk-linux/5.3.0/include-fixed
>>>>> -isystem
>>>>> /home/andre/rdk/rdk-master/build-vbox32/tmp/sysroots/vbox32/usr/include
>>>>>   -D_LIBC_REENTRANT -include
>>>>>
>>>>> /home/andre/rdk/rdk-master/build-vbox32/tmp/work/core2-32-rdk-linux/glibc/2.23-r0/build-i686-rdk-linux/libc-modules.h
>>>>> -DMODULE_NAME=libc -include ../include/libc-symbols.h -o
>>>>>
>>>>> /home/andre/rdk/rdk-master/build-vbox32/tmp/work/core2-32-rdk-linux/glibc/2.23-r0/build-i686-rdk-linux/io/posix_fallocate.o
>>>>> -MD -MP -MF
>>>>> /home/andre/rdk/rdk-master/build-vbox32/tmp/work/core2-32-rdk-linux/glibc/2.23-r0/build-i686-rdk-linux/io/posix_fallocate.o.dt
>>>>> -MT
>>>>> /home/andre/rdk/rdk-master/build-vbox32/tmp/work/core2-32-rdk-linux/glibc/2.23-r0/build-i686-rdk-linux/io/posix_fallocate.o
>>>>> ../sysdeps/unix/sysv/linux/posix_fallocate.c: In function
>>>>> 'posix_fallocate':
>>>>> ../sysdeps/unix/sysv/linux/posix_fallocate.c:39:1: error: bp cannot be
>>>>> used in asm here
>>>>>   }
>>>>>   ^
>>>>> ../sysdeps/unix/sysv/linux/posix_fallocate.c:39:1: error: bp cannot be
>>>>> used in asm here
>>>>
>>>>
>>>> Hi All,
>>>>
>>>> For the latest master, if 'DEBUG_BUILD' is enabled, the above error
>>>> appers.
>>>>
>>>> | ../sysdeps/unix/sysv/linux/posix_fallocate.c: In function
>>>> 'posix_fallocate':
>>>> | ../sysdeps/unix/sysv/linux/posix_fallocate.c:39:1: error: bp cannot be
>>>> used in asm here
>>>> |  }
>>>> |  ^
>>>>
>>>> The problem could be reproduced if we add in local.conf the following
>>>> line.
>>>> DEBUG_BUILD = "1"
>>>>
>>>> Do you have any idea what's going on there?
>>>
>>> posix_fallocate is calling a syscall helper with 6 argument, and there
>>> has been i386 specific optimization added to glibc
>>> which checks for gcc version and if its >= 5.0 then relies upon gcc to do
>>> right thing when spilling ebp, and switched to using
>>> inline syscall handler. But this spill is working/tested with -O2 and
>>> everything else is just a bonus if it works. Can you try this
>>> patch to glibc  ?
>>>
>>>
>>> --- a/sysdeps/unix/sysv/linux/i386/sysdep.h
>>> +++ b/sysdeps/unix/sysv/linux/i386/sysdep.h
>>> @@ -46,7 +46,7 @@
>>>      to compile glibc.  Disable GCC 5 optimization when compiling for
>>>      profiling since asm ("ebp") can't be used to put the 6th argument
>>>      in %ebp for syscall.  */
>>> -#if __GNUC_PREREQ (5,0) && !defined PROF
>>> +#if __GNUC_PREREQ (5,0) && !defined PROF && !defined __OPTIMIZE_SIZE__
>>>   # define OPTIMIZE_FOR_GCC_5
>>>   #endif
>>>
>>>
>>>
>>>
>>
>> Hi Khem,
>>
>> Thanks for your reply.
>> I've tried the above patch, but the problem is still there :(
>>
>> Best Regards,
>> Chen Qi
>>
>
> And if we remove " -fno-omit-frame-pointer" from CFLAGS, things work.
> CFLAGS_remove_pn-glibc_x86 = " -fno-omit-frame-pointer"

this option did not appear in the cmdline you pasted in your mails. But this
would let gcc not need frames and ebp will be free to clobber and it would work
but check how stack traces are looking then.

>
> Regards,
> Chen Qi