Subject: Re: fixing gcc RTL template library
To: Jonathan Stone <jonathan@DSG.Stanford.EDU>
From: Todd Vierling <>
List: tech-toolchain
Date: 06/09/1999 23:42:28
On Wed, 9 Jun 1999, Jonathan Stone wrote:

: My take is that if our GCC ever emits a call to a libgcc2 routine
: which the native FSF/EGCS build never called, then that's an egregious
: bug in gcc:), and one where busting the application is probably a
: /good/ thing :).

No.  We can remove anything which has *never* been used by an in-tree
compiler.  That is, all calls which were not used by a particular port's
compiler after that port attained shlibs.  Binary compatibility.

: More seriously, it'll lose in /very/ ugly ways if egcs ever starts
: using nonstandard calling sequences for libgcc2 routines.

We'll address it if/when that happens.  This would be a relatively simple
change to the gcc internals for naming the libgcc2 functions, while we can
keep the existing code in libc for binaries.

: Last time I tried this (which was yers ago) only the insn routines
: from libgcc2 that werent already in the .md got compiled into libgcc2.

This varies wildly between architectures.

But as it is, the way we load support code for LKMs is aggravatingly broken.
We need an *installed* -lkern that will pull in not-linked-into-kernel-yet
support stuff at LKM load time, not at kernel link time.

-- Todd Vierling (Personal; Bus.