Hi,
Still having to use t2-minimal-glibc-gcc-x86-64-r49128.iso as the basis
for my chroot, the newer iso:
t2-minimal-xorg-glibc-gcc-x86-64-r49198.iso was not able to configure a
t2 target, lua was having errors, I will try that again,when I have some
time, and post the errors.
meanwhile, using r49128 as the chroot base :
every day I try a svn update, to see if any new fixes show up, but still
mostly hitting the same issues.
1. mpc is still not building unless I Downgrade mpfr to the previous
version.
2. glib stage 1 was failing with errors like:
/opt/t2-trunk/build/default-9.0-svn-generic-x86-64-linux/TOOLCHAIN/tools.cross/crosscc/../lib/gcc/x86_64-t2-linux-gnu/9.2.0/../../../../x86_64-t2-linux-gnu/bin/ld:
/lib64/librt.so.1: undefined reference to `__clock_settime@GLIBC_PRIVATE'
/opt/t2-trunk/build/default-9.0-svn-generic-x86-64-linux/TOOLCHAIN/tools.cross/crosscc/../lib/gcc/x86_64-t2-linux-gnu/9.2.0/../../../../x86_64-t2-linux-gnu/bin/ld:
/lib64/librt.so.1: undefined reference to
`__clock_getcpuclockid@GLIBC_PRIVATE'
/opt/t2-trunk/build/default-9.0-svn-generic-x86-64-linux/TOOLCHAIN/tools.cross/crosscc/../lib/gcc/x86_64-t2-linux-gnu/9.2.0/../../../../x86_64-t2-linux-gnu/bin/ld:
/lib64/librt.so.1: undefined reference to `__clock_nanosleep@GLIBC_PRIVATE'
/opt/t2-trunk/build/default-9.0-svn-generic-x86-64-linux/TOOLCHAIN/tools.cross/crosscc/../lib/gcc/x86_64-t2-linux-gnu/9.2.0/../../../../x86_64-t2-linux-gnu/bin/ld:
/lib64/librt.so.1: undefined reference to `__clock_getres@GLIBC_PRIVATE'
The thing that got rid of these errors was upgrading my rootfs's glibc
from 2.30 to 2.31, to match the current package version in the SDE. and
then it built OK, so there is a system dependency there. I just checked
my config, and "cross build" was not selected, hopefully if cross build
is hapening, it would be insulated from something like this, (i am going
to do the rest of my testing with crossbuild enabled)
For this initial testing I picked a very simple generic minimal, and did
not change many values from the defaults.
3. libaio is installing things too deeply, it no longer needs the
line:var_append makeinstopt ' ' 'prefix=$root/$prefix libdir=$root$libdir'
Index: package/filesystem/libaio/libaio.conf
===================================================================
--- package/filesystem/libaio/libaio.conf (revision 49361)
+++ package/filesystem/libaio/libaio.conf (working copy)
@@ -17,4 +17,5 @@
var_append makeopt ' ' "ARCH=$LIBAIO_ARCH"
# no DESTDIR
-var_append makeinstopt ' ' 'prefix=$root/$prefix libdir=$root$libdir'
+# JLR: No longer needed:
+#var_append makeinstopt ' ' 'prefix=$root/$prefix libdir=$root$libdir'
4. The above caused lvm2 to fail since it could not find libaio.h
5. grub2 was failing: First with:
from symlist.h:2:
../include/grub/symbol.h:36:20: error: #if with no expression
36 | #if HAVE_ASM_USCORE
I found a patch on the web for that error: (attached)
but there is a 2nd error after that, which I have not solved yet:
make[3]: *** [Makefile:31011: usbserial_pl2303.module] Error 1
/opt/t2-trunk/build/default-9.0-svn-generic-x86-64-linux/TOOLCHAIN/tools.cross/crosscc/../lib/gcc/x86_64-t2-linux-gnu/9.2.0/../../../../x86_64-t2-linux-gnu/bin/ld:
relocatable linking with relocations from format elf64-x86-64
(bus/usb/serial/usbserial_ftdi_module-ftdi.o) to format elf32-i386
(usbserial_ftdi.module) is not supported
collect2: error: ld returned 1 exit status
/opt/t2-trunk/build/default-9.0-svn-generic-x86-64-linux/TOOLCHAIN/tools.cross/crosscc/../lib/make[3]:
*** [Makefile:31004: usbserial_ftdi.module] Error 1
gcc/x86_64-t2-linux-gnu/9.2.0/../../../../x86_64-t2-linux-gnu/bin/ld:
relocatable linking with relocations from format elf64-x86-64
(bus/usb/uhci_module-uhci.o) to format elf32-i386 (uhci.module) is not
supported
collect2: error: ld returned 1 exit status
Seems to be building for i386, and does not like the lib64 libs?? not
sure what is going on here.
Will be keeping you updated.
Jan
>From 7fe03fc44bd342995756c2c3710dcc7340bfa3eb Mon Sep 17 00:00:00 2001
From: Mark Wutzke <[email protected]>
Date: Wed, 10 Oct 2018 07:52:15 +1100
Subject: [PATCH 1/1] grub: Allow 'configure' be re-run with cache
The grub_ASM_USCORE macro breaks one of the
tenants of AC_CACHE_VAL [1]:
The commands-to-set-it must have no side effects
except for setting the variable cache-id
This patch reworks grub_ASM_USCORE so that the
side effect of setting HAVE_ASM_USCORE is now
outside the scope of AC_CACHE_VAL.
This removes the following error from Grub
bulids that re-use the cache:
../include/grub/symbol.h:36:20: error: #if with no expression
[1] https://www.gnu.org/savannah-checkouts/gnu/autoconf/manual/autoconf-2.69/html_node/Caching-Results.html#Caching-Results
---
acinclude.m4 | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/acinclude.m4 b/acinclude.m4
index 32d5477d1..7467b9711 100644
--- a/acinclude.m4
+++ b/acinclude.m4
@@ -60,16 +60,20 @@ else
fi
if $EGREP '(^|[^_[:alnum]])_func' conftest.s >/dev/null 2>&1; then
- HAVE_ASM_USCORE=1
grub_cv_asm_uscore=yes
else
- HAVE_ASM_USCORE=0
grub_cv_asm_uscore=no
fi
rm -f conftest*])
AC_MSG_RESULT([$grub_cv_asm_uscore])
+
+if test "x$grub_cv_asm_uscore" = xyes; then
+ HAVE_ASM_USCORE=1
+else
+ HAVE_ASM_USCORE=0
+fi
])
--
2.18.0
-----------------------------------------------------------
If you wish to unsubscribe from this mailing, send mail to
[email protected] with a subject of: unsubscribe t2
