--Apple-Mail-24-609105046
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII; format=flowed
okays. here the journey ends. i got it :)
problem is that gcc compiled with -mrtd disabled but sio.S/sio_putc()
still used that calling convention. so fix is easy:
Index: boot/i386/boot2/Makefile
===================================================================
RCS file: /home/dcvs/src/sys/boot/i386/boot2/Makefile,v
retrieving revision 1.11
diff -u -r1.11 Makefile
--- boot/i386/boot2/Makefile 6 Aug 2004 20:50:35 -0000 1.11
+++ boot/i386/boot2/Makefile 11 Sep 2004 05:19:20 -0000
@@ -60,13 +60,7 @@
-Wmissing-declarations -Wmissing-prototypes -Wnested-externs \
-Wpointer-arith -Wshadow -Wstrict-prototypes -Wwrite-strings
-# -mrtd will produce bad code (missing stack pops) when combined with
-# gcc-3.4's (default on) unit-at-a-time optimization. See the code
-# generated for 'xputc' as an example.
-#
-.if ${CCVER} == "gcc2"
CFLAGS+= -mrtd
-.endif
# Unfortunately, unit-at-a-time creates other issues as well, as yet
# not tracked down, in boot2, so we have to turn it off. Note that
done. gcc3 boots again. /me bows
simon
ps: oh yea, please commit ASAP :)
On 11.09.2004, at 06:41, Simon 'corecode' Schubert wrote:
> HELL! I think I got it! It's got to do with serial console!
> I'm on track. Fault here easily reproducable by typing something and
> then hitting backspace. If entering -h first this won't happen.
>
> cheers
> simon
>
> On 11.09.2004, at 04:35, Simon 'corecode' Schubert wrote:
>
>> okay... more information:
>>
>> i've been playing around with qemu to test that boot issue.
>> unfortunately gdb hangs at some point and it's really hard to debug
>> without symbols and with the CPU happily switching between protected
>> and real mode...
>>
>> anyways:
>>
>> - latest sys/boot sources
>> - gcc2 bootblocks work
>> - gcc3 bootblocks don't. failure: BTX error stuff
>>
>> while debugging it became obvious that it must be something within
>> boot2.c: sometimes it showed "No /boot/loader" and then just 2 BTX
>> halted stuff.
>> Another fact is that it takes 3 seconds for the BTX to fail, whereas
>> pressing any key instantly crashes it. must be somewhere between that
>> printf() in load() and the printf() in main() that outputs the
>> prompt. or this one gets overwritten by the error message. oh well.
>>
>> so long for this report, cheers
>> simon
>>
>> On 10.09.2004, at 20:37, Scott Ullrich wrote:
>>
>>> Jeroen Ruigrok/asmodai wrote:
>>>> Now do that from a current snapshot.
>>>
>>> I just updated all of the boot code from Sep 10th snapshot ISO and
>>> now everything works as exepcted.
>>>
>>> Cheers,
>>>
>>> Scott
>>>
>> --
>> /"\
>> \ /
>> \ ASCII Ribbon Campaign
>> / \ Against HTML Mail and News
>>
>>
> --
> /"\
> \ /
> \ ASCII Ribbon Campaign
> / \ Against HTML Mail and News
>
>
--
/"\
\ /
\ ASCII Ribbon Campaign
/ \ Against HTML Mail and News
--Apple-Mail-24-609105046
content-type: application/pgp-signature; x-mac-type=70674453;
name=PGP.sig
content-description: This is a digitally signed message part
content-disposition: inline; filename=PGP.sig
content-transfer-encoding: 7bit
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)
iD8DBQFBQoyor5S+dk6z85oRAtcWAKC/kQUmyuEHdu+aCOsD0sgMy4E2XgCfZ0M+
3SHb8jMFm8iLRrqUlJIxru4=
=bkD9
-----END PGP SIGNATURE-----
--Apple-Mail-24-609105046--