LFS-6.6, Stage2, glibc, nscd.c:442

linux fan linuxscratch at gmail.com
Fri Jun 4 12:44:44 PDT 2010

On 6/4/10, Paul Rogers <paulgrogers at fastmail.fm> wrote:

>> So the problem was the Chapter 5 gcc?
> I'm not sure that follows.

Seems to.

Intending to attempt to duplicate __stack_chk_guard error in chapter 6
while building lfs-6.6 from lfs-6.1.1 ...
The theory being that the host gcc/glibc "ideal" gets passed to the
toolchain build and then passed to the finished build as if it were an
olympic torch.

I'm using the exemplary jhalfs.
lfs-6.2 original host (hostname lfs)
lfs-6.1.1 target partition /dev/sdb8 (hostname lfs-1)
lfs-6.6 target partition /dev/sdb9

Stage 1 - build lfs-6.1.1
me at lfs:~/jhalfs-0.2$
su # stay in root for this version
mount /dev/sdb8 /mnt/lfs
./jhalfs -L 6.1.1
cd /mnt/lfs/jhalfs
... it goes on for hours and then is done.
... chroot into it and build a requisite kernel.
... shutdown and boot to it (entertaining grub in the process).

Stage 2 - build lfs-6-6
me at lfs-1:~/jhalfs-2.3.2$
mount /dev/sdb9 /mnt/lfs
exit # to regular user for this version
.... select the options and approve
cd /mnt/lfs/jhalfs
make BREAKPOINT=038-gcc-pass2

... while that is building, I see in gcc-build/config.log
"Reading specs from ...  Configured with: ... --disable-libssp"
... I decide to break out of it, and try something
... This diff in commands
--- /mnt/lfs/jhalfs/lfs-commands/chapter05/038-gcc-pass2.orig
+++ /mnt/lfs/jhalfs/lfs-commands/chapter05/038-gcc-pass2
@@ -43,7 +43,8 @@
     --enable-shared --enable-threads=posix \
     --enable-__cxa_atexit --enable-languages=c,c++ \
     --disable-libstdcxx-pch --disable-multilib \
-    --disable-bootstrap
+    --disable-bootstrap \
+    --enable-libssp
 make install
 ln -vs gcc /tools/bin/cc

... remove gcc-4.4.3 and gcc-build dirs
... restart ...
make BREAKPOINT=038-gcc-pass2
... libssp is produced in tools/lib
... continue building

I imagine this will be ok.



More information about the lfs-support mailing list