what's the usage of -flimit-inline

[Wu Mingqiao <mingqiao dot wu at gmail dot com>]

I have built a toolchain with crosstool-0.28-rc34 for arm-iwmmxt. I
used gcc-3.4.0, glibc-2.3.2, binutils-2.14.90 and linux-2.6.8.1 kernel
head files. I find it is needed to add
GLIBC_EXTRA_CONFIG="--without-fp" in arm-iwmmxt.dat, or there is an
error report: selected processor does not support "rfs r2". Then it
progressed well.

The toolchain can be used to compiled linux-2.6.8.1 kernel. But there
is a weird feature about this toolchain. I built a linux-2.6.8.1
kernel image and busybox with it, then I run it in a xscale board, it
always report: illegal instructions when the busybox is loaded. I
overwrite the libc-2.3.2.so with an old one built with an old
toolchain, then it works fine. I checked the two toolchains and found
there is a different option in configuring GLIBC-2.3.2. The old
toolchain uses CC="arm-linux-gcc -finline-limit=10000". After adding
the option GLIBC_EXTRA_CC_ARGS="-finline-limit=10000" to
arm-iwmmxt.dat, I got a new toolchain and run the kernel and busybox
successfully.

I could not find clear explanation about -finline-limit option with
materials at hand. Does anybody know the usage of -flimit-inline? Any
comment is appreciated.

--Bridge

------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sources.redhat.com