[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[avr-gcc-list] avr-gcc: Add --with-avrlibc configure option for AVR-Libc

From: Georg-Johann Lay
Subject: [avr-gcc-list] avr-gcc: Add --with-avrlibc configure option for AVR-Libc cooperation?
Date: Sun, 26 Aug 2012 12:09:51 +0200
User-agent: Thunderbird (Windows/20100228)


for historical reasons, AVR-Libc implements functions that
GCC expects to live in libgcc, namely float functions like

Currently, avr-gcc is not configurable to accommodate that
situation which leads to performance problems if the float
support functions from libgcc are used.

This happens at least in the following situations:

* The user does not specify -lm.  -lm should only be needed
  if function from math.h are used, not for language core
  features like int i = (int) float.

* The application is LTO compiled, i.e. linked with -flto.
  The plugin machinery passes lgcc -lc lgcc through to the
  linker by means of -plugin-opt=-pass-through=-lgcc etc.
  so that -lgcc is linked prior to -lm.

* The user uses fixed <-> float conversion routines from
  libgcc.  These routines refer float functions, and the linker
  resolves these function in libgcc if they are there.

* avr-g++ is used as linker driver.

In order to fix all of these cases, at least two changes
are needed in the compiler.

1) Remove respective functions from libgcc

2) libm.a is not a "math library", it is a "libgcc supplement".
   Thus, -lm has to be treated like libgcc.

avr.h currently has:

#define LIBGCC_SPEC "\
%{!mmcu=attiny15:%{!mmcu=attiny28: -lgcc }}}}}"

LINK_GCC_C_SEQUENCE_SPEC is defined and defaults to
"%G %L %G"

Together with 1) which is already contained in the proposed
patch for libgcc, a reasonable spec layout would be

#define LIBGCC_SPEC "\
%{!mmcu=attiny15:%{!mmcu=attiny28: -lgcc -lm }}}}}"

"--start-group %G %L --end-group"

Could a AVR-Libc maintainer test this setup?

To test the attached patch, use avr-gcc from trunk
SVN 190644 or newer and configure --with-avrlibc=yes

To test the new specs, you can

$ avr-gcc -dumpspecs > my.specs

and then

$ avr-gcc -specs my.specs ...

The specs in question are libgcc and link_gcc_c_sequence.

This means that the option makes avr-gcc assume that
AVR-Libc comes with a specific layout:  -lm is available
and comes with libgcc-like support and fills out the gaps
introduced by LIB2FUNCS_EXCLUDE from t-avrlibc.

Just like the t-avrlibc in libgcc, there would be a t-avrlibc
Makefile snip in gcc that takes care of the specs and
(re)defines them if configured for AVR-Libc cooperation.

Moreover, it must be ensured that no function causes ABI
clashes.  For example, libgcc brings ffs with optimized ABI
similar to __divmodxx4:  The compiler knows what registers
are clobbered and what are not, and/or it may pass values
in non-ABI registers.

Functions in question are, in all available modes:
ffs, ctz, clz, bswap, popcount, parity

BTW: The last point is an issue already present in the current
releases and not related to a new configure option.



Index: libgcc/config.host
--- libgcc/config.host  (revision 190644)
+++ libgcc/config.host  (working copy)
@@ -380,6 +380,9 @@ avr-*-rtems*)
        # Make HImode functions for AVR
        tmake_file="${cpu_type}/t-avr t-fpbit"
+       if test x$with_avrlibc = xyes; then
+           tmake_file="$tmake_file ${cpu_type}/t-avrlibc"
+       fi
        tm_file="$tm_file avr/avr-lib.h"
Index: libgcc/Makefile.in
--- libgcc/Makefile.in  (revision 190644)
+++ libgcc/Makefile.in  (working copy)
@@ -516,6 +516,10 @@ FPBIT_FUNCS := $(filter-out _sf_to_tf,$(
 DPBIT_FUNCS := $(filter-out _df_to_tf,$(DPBIT_FUNCS))
 fpbit-src := $(srcdir)/fp-bit.c
 # Build FPBIT.
Index: libgcc/config/avr/t-avrlibc
--- libgcc/config/avr/t-avrlibc (revision 0)
+++ libgcc/config/avr/t-avrlibc (revision 0)
@@ -0,0 +1,36 @@
+       _compare_sf \
+       _eq_sf \
+       _ne_sf \
+       _gt_sf \
+       _ge_sf \
+       _lt_sf \
+       _le_sf \
+       _unord_sf \
+       \
+       _negate_sf \
+       _addsub_sf \
+       _mul_sf \
+       _div_sf \
+       \
+       _si_to_sf  _sf_to_si \
+       _usi_to_sf _sf_to_usi
+# DF
+       _sf_to_df
+       _fixunssfsi \
+       _fixsfdi \
+       _fixunssfdi \
+       _floatdisf \
+       _floatundisf
+# DF
+       _fixdfdi \
+       _fixunsdfsi \
+       _fixunsdfdi \
+       _floatdidf \
+       _floatundidf

reply via email to

[Prev in Thread] Current Thread [Next in Thread]