0000047: ARM BigEndian does not allow Soft Float
When ARMEB is selected for the arch I am unable to select soft float support.
The toolchain/Config.in does not have a depend for armeb, it just has arm.
The attached patch should cover it.
0000040: Adds options to jffs2 filesystem creation
Brings jffs2 up to the 20041007 build level so it will compile with both 2.4
and 2.6 kernels.
Adds ability to specify eraseblock and pad sizes, permissions squash,
big/little endian and device table in the configuration.
Also adds capability to copy the resulting image to a secondary location like
a tftp server's root directory.
mtd.patch adds capabilities of the previously attached patch plus allows you to
compile a version of the utilities to run on the target system.
So now there's a set of config options under package to compile the utils for
distribution to the target and a set of config options under target to create a
jffs2 filesystem.
0000050: gdb_server and gdb_client fixes and split
Here is a patch with 2 issues.
gdbserver does not depend on ncurses (just the full gdb needs it)
Split out gdb client config option, to be separate from gdbserver. If you
build gdb client for the host once it does not need to be rebuilt when the rest
of libs change. So I build it and save it off.
0000033: [PATCH] libpng: New download site
This patch changes the URL to download libpng, since the actual seems to be
offline or lack that particular file
0000039: Adds options to ext2 filesystem creation
Adds ability to specify block and inode counts, reserved blocks, permissions
squash and device table in the configuration.
Also adds capability to gzip output and to copy the resulting image to a
secondary location like a tftp server's root directory.
This patch changes the path used as dependency for liblzo, which appears to be
pointing to $(STAGING_DIR)/usr prefix while the package is actually installed
with prefix ${STAGING_DIR}
Hi Eric,
I writing privately since I have not heard from Mike for 2 weeks about
a one line patch to the gcc Makefile.in to allow the extra gcc options
to be passed in. The patch is below and without it -- gcc extra's like
--with-cpu=xscale
do not work.
The patch is as follows:
ello,
Enclosed is a small documentation update. It adds a section that answers
a very often asked question : "How to use the uClibc toolchain generated
by Buildroot".
Erik, could you apply it if you agree with the modifications ?
Thanks,
Thomas
