Help required re ARM EABI ZCX Ada (4.6.1)

Luke A. Guest laguest@archeia.com
Sun Oct 9 19:43:00 GMT 2011


Hi,

I got back to this yesterday and have tried to port over the ARM code
from the C++/Java exception files. I have managed to build a native
compiler (using make bootstrap) and run the test suite (logs not
included due to their size), but the general summary is as follows:

                === gcc Summary ===

# of expected passes            78536
# of unexpected failures        40
# of unexpected successes       38
# of expected failures          222
# of unsupported tests          1145

                === gnat Summary ===

# of expected passes            871
# of expected failures          10
# of unsupported tests          2

                === acats Summary ===
# of expected passes            2321
# of unexpected failures        0

My host machine is:

Linux rogue 2.6.32-5-amd64 #1 SMP Tue Jun 14 09:42:28 UTC 2011 x86_64
GNU/Linux


I have then built for arm-linux-gnueabi and have an installed toolchain
(no compile errors) and a running QEMU which can be logged into
automatically by me via SSH and which has a TAP set up, I run this with:

qemu-system-arm -k en-gb -m 256 -M versatilepb -kernel
~/opt/arm/vmlinuz-2.6.32-5-versatile -initrd
~/opt/arm/initrd.img-2.6.32-5-versatile -hda ~/opt/arm/arm-debian.img
-append "root=/dev/sda1" -net nic -net
tap,ifname=tap0,script=no,downscript=no

I have set up 2 env vars for DejaGNU:

DEJAGNU=/home/laguest/src/others/gcc-zxc-arm/dejagnu/site.exp
DEJAGNU_TIMEOUT=-1

I have the set up a board called angel.exp and these can be seen in the
attachments.

When I initially had the results, I had DEJAGNU_TIMEOUT=5400, I didn't
realise there was a separate timeout for the compiler. Due to the speed
I've not rerun these tests.

There is a problem with the Ada test suite for cross targets, on
starting I get the following error:

make[1]: Entering directory
`/home/laguest/src/others/gcc-zxc-arm/arm-build-gcc-4.6.1/gcc'
gnattools missing, exiting.
make[1]: *** [check-acats] Error 1

It then continues with the gnat tests, but fails to run the acats tests.
That message comes from the run_acats script.

I have compiled a test program, exc.adb which does not catch the
exceptions, so I know this ARM code does not work properly yet. The
error on the arm side is:

raised EXC.T : exc.adb:6

I'm obviously missing something here, can someone knowledgeable look
over the patch and see what it is?

Can anyone confirm that the Ada test suite should work for cross
targets?

If anyone has a target this can be tested for and has a working test
suite, can they apply the patch and see what results they get?

Thanks,
Luke.

-------------- next part --------------
append boards_dir "/home/laguest/src/others/gcc-zxc-arm/dejagnu/boards"
case "$target_triplet" in {
      { "arm*-linux-gnueabi" } { set target_list { angel } }
}
-------------- next part --------------
load_generic_config "unix"
set_board_info hostname "192.168.0.5"
set_board_info username laguest
set_board_info rsh_prog /usr/bin/ssh
set_board_info rcp_prog /usr/bin/scp
set_board_info gcc,timeout -1
-------------- next part --------------
A non-text attachment was scrubbed...
Name: arm-linux-gnat.log.bz2
Type: application/x-bzip
Size: 62067 bytes
Desc: not available
URL: <https://gcc.gnu.org/pipermail/gcc/attachments/20111009/3c2e55d0/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: arm-linux-gnat.sum.bz2
Type: application/x-bzip
Size: 4642 bytes
Desc: not available
URL: <https://gcc.gnu.org/pipermail/gcc/attachments/20111009/3c2e55d0/attachment-0001.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: arm-linux-zcx.diff
Type: text/x-patch
Size: 23546 bytes
Desc: not available
URL: <https://gcc.gnu.org/pipermail/gcc/attachments/20111009/3c2e55d0/attachment-0002.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: exc.adb
Type: text/x-adasrc
Size: 211 bytes
Desc: not available
URL: <https://gcc.gnu.org/pipermail/gcc/attachments/20111009/3c2e55d0/attachment-0003.bin>


More information about the Gcc mailing list