This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Successful install of egcs-1.1.2


I just completed installation of egcs-1.1.2 on my machine. Running
'config.guess' produces nothing, but I guess my setup is similar to the
"i686-unknown-linux-gnu" entry in your "build and test status" page.

Running "uname -a" on my system (a Pionex system with an AMD-K6-2-300
CPU, 100 MHz bus, 256 MB of RAM, an ATI 8 MB Rage Pro AGP video card and
dual-boot Linux/Win98) produces the following:

    Linux hercules.olympus 2.0.36 #3 Thu Dec 24 09:59:02 PST 1998 i586
unknown

I installed Linux from the "RedHat Linux 5.2 Deluxe" package from
McMillan publishing Co. A couple of months ago I rebuilt the kernel to
remove sound support (a preliminary step before installing a custom
driver for my sound card). I used the gcc compiler provided as part of
the 5.2 Red Hat distribution. While building the kernel, I ran into the
"signal 11" problem discussed below; I ran 'make' again (which worked
without interruptions) and did not worry about it at the time.

I had a fair amount of trouble getting rid of the nasty "SIG11" problem
while installing egcs-1.1.2. Every time I ran "make bootstrap", the make
would abort sooned or later (usually in the middle of stage 1) due to a
signal 11 fault. Consulting your "FAQ" page, I found the reference to
item 4.12 (Signal 11 on GNU/Linux) most helpful. I followed the link to
www.bitwizard.nl site, which convinced me that I DID have a hardware
problem. I followed a suggestion therein, namely to add WAIT states
through the BIOS. This took a couple of tries (my ATX board, model M5ALA
from ???, disables manual chipset settings by default and it took me a
while to figure that out). Telling the BIOS to treat the DRAM as "Slow"
instead of "Normal" seems to have done the trick.

After setting the BIOS and rebooting, I was able to do a clean 'make'
without signal 11 interruptions. This proves nothing, but was definitely
a big improvement...

Thanks for the excellent documentation and for an outstanding product.

Pierre F. Maldague



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]