eglibc: [armel] make syscalls out-of-line to improve debugging

[Dave Martin]

Hi,

If possible, we should merge this patch, which changes syscalls to be out-of-line. This should improve gdb debug visibility significantly when stepping syscalls or attaching to running processes:

http://sourceware.org/ml/glibc-cvs/2010-q2/msg00038.html

Additional discussion here:

http://sourceware.org/ml/libc-ports/2010-04/msg00001.html

[Launchpad Janitor (janitor)]

This bug was fixed in the package eglibc - 2.11.1-0ubuntu8

---------------
eglibc (2.11.1-0ubuntu8) maverick; urgency=low

  * Apply from the 2.11-x86 branch:
    - x86-64 SSE4 optimized memcmp.
  * [armel] make syscalls out-of-line to improve debugging. LP: #571647 .
 -- Matthias Klose <email address hidden> Sat, 01 May 2010 14:30:23 +0200

[Matthias Klose (doko)]

fails to build:

gcc-4.4 -fno-stack-protector -U_FORTIFY_SOURCE ../ports/sysdeps/unix/sysv/linux/arm/eabi/libc-do-syscall.S -c -I../include -I/build/buildd/eglibc-2.11.1/build-tree/armel-libc/csu -I/build/buildd/eglibc-2.11.1/build-tree/armel-libc -I../ports/sysdeps/arm/elf -I../ports/sysdeps/unix/sysv/linux/arm/eabi/nptl -I../ports/sysdeps/unix/sysv/linux/arm/eabi -I../ports/sysdeps/unix/sysv/linux/arm/nptl -I../ports/sysdeps/unix/sysv/linux/arm -I../ports/sysdeps/unix/sysv/linux -I../nptl/sysdeps/unix/sysv/linux -I../nptl/sysdeps/pthread -I../sysdeps/pthread -I../sysdeps/unix/sysv/linux -I../sysdeps/gnu -I../sysdeps/unix/common -I../sysdeps/unix/mman -I../sysdeps/unix/inet -I../ports/sysdeps/unix/sysv -I../nptl/sysdeps/unix/sysv -I../sysdeps/unix/sysv -I../ports/sysdeps/unix/arm -I../ports/sysdeps/unix -I../nptl/sysdeps/unix -I../sysdeps/unix -I../sysdeps/posix -I../ports/sysdeps/arm/eabi -I../ports/sysdeps/arm/fpu -I../ports/sysdeps/arm/nptl -I../ports/sysdeps/arm -I../sysdeps/wordsize-32 -I../sysdeps/ieee754/flt-32 -I../sysdeps/ieee754/dbl-64 -I../sysdeps/ieee754 -I../sysdeps/generic/elf -I../sysdeps/generic -I../ports -I../nptl -I.. -I../libio -I. -nostdinc -isystem /usr/lib/gcc/arm-linux-gnueabi/4.4.4/include -isystem /usr/lib/gcc/arm-linux-gnueabi/4.4.4/include-fixed -isystem /build/buildd/eglibc-2.11.1/debian/include -D_LIBC_REENTRANT -include ../include/libc-symbols.h -DHAVE_INITFINI -DASSEMBLER -I/build/buildd/eglibc-2.11.1/build-tree/armel-libc/csu/. -pipe -O2 -fstrict-aliasing -g -Wa,--noexecstack -o /build/buildd/eglibc-2.11.1/build-tree/armel-libc/csu/libc-do-syscall.o -MD -MP -MF /build/buildd/eglibc-2.11.1/build-tree/armel-libc/csu/libc-do-syscall.o.dt -MT /build/buildd/eglibc-2.11.1/build-tree/armel-libc/csu/libc-do-syscall.o
../ports/sysdeps/unix/sysv/linux/arm/eabi/libc-do-syscall.S: Assembler messages:
../ports/sysdeps/unix/sysv/linux/arm/eabi/libc-do-syscall.S:36: Error: CFI instruction used without previous .cfi_startproc
../ports/sysdeps/unix/sysv/linux/arm/eabi/libc-do-syscall.S:37: Error: CFI instruction used without previous .cfi_startproc
../ports/sysdeps/unix/sysv/linux/arm/eabi/libc-do-syscall.S:38: Error: CFI instruction used without previous .cfi_startproc
make[3]: *** [/build/buildd/eglibc-2.11.1/build-tree/armel-libc/csu/libc-do-syscall.o] Error 1

[Dave Martin (dave-martin-arm)]

It looks like an additional fix is needed to allow this to build with the (stricter) current trunk binutils - see below.

I'll post more info when we have it available.

> -----Original Message-----
> From: Matthew Gretton-Dann
> Sent: 04 May 2010 11:17
> To: Dave P Martin
> Cc: 'Matthias Klose'
> Subject: RE: [Bug 571647] Re: eglibc: [armel] make syscalls
> out-of-line to improvedebugging
>
> Dave,
>
> This was trunk glibc, but I don't know what binutils Dan
> Jacobowitz used to develop against - and it is binutils that
> is producing the error.
>
> Looking at the patch - the required fix is to add
> cfi_startproc, and cfi_endproc directives in
> __libc_do_syscall
> (sysdeps/unix/sysv/linux/arm/eabi/libc-do-syscall.S):
>
> ENTRY (__libc_do_syscall)
> .fnstart
> cfi_startproc @@@@ Add this line
> push {r7, lr}
> .save {r7, lr}
> cfi_adjust_cfa_offset (8)
> cfi_rel_offset (r7, 0)
> cfi_rel_offset (lr, 4)
> mov r7, ip
> swi 0x0
> pop {r7, pc}
> cfi_endproc @@@@ Add this line
> .fnend
>
> Thanks,
>
> Matt

[Matthias Klose (doko)]

fixed in maverick and 2.12, leaving the lucid task open

[Rolf Leggewie (r0lf)]

lucid has seen the end of its life and is no longer receiving any updates. Marking the lucid task for this ticket as "Won't Fix".