When comparing differences between your output and the output of my K7-GCC4 binary, I notice the following differences:
1. Your kernel is 2.6.13.2, while mine is the default 10.1 kernel of 2.6.16.13-4
2. After the "Loading Module..." messages and before "Bootstrapping...", mine shows : Copying Module 0x039be000-0x039e0c99 [hello]
3. Your running a Pentium-M (I'm running a K-7), so my TLBs and Caches are different.
4. My init code/data freed is 7 pages.
5. I don't get a SIGSEGV after MIKEE: Exiting Thread::initialize(), it appears to go into a tight loop that ignores CTRL+C.
Not sure what all that means but thought I'd point it out in case someone has an "Aha!" they might offer up.
Thanks, MikeE
On Tuesday 27 June 2006 05:28, you wrote:
Hi,
I get the following...
ch12@krishna:/tmp/fiasco-ux> ./fiasco
Fiasco-UX on Linux 2.6.13.2-krishna (i686) Mapped 64 MB Memory + 0 KB Framebuffer + 0 MB Input Area on FD 3
Loading Module 0x00090000-0x00096924 [sigma0-ux] Loading Module 0x00120000-0x002bc9c4 [roottask]
Bootstrapping...
Welcome to Fiasco(ux)! DD-L4(v2)/x86 microkernel (C) 1998-2006 TU Dresden Rev: Mon Jun 26 22:50:26 2006 compiled with gcc 3.4.6 for Intel Pentium
CPU: GenuineIntel (6:9:5:16) Model: Pentium-M (Banias) at 598 MHz
128 Entry I TLB (4K pages) 2 Entry I TLB (4M pages) 128 Entry D TLB (4K pages) 8 Entry D TLB (4M pages) 32 KB L1 I Cache (8-way associative, 64 bytes per line) 32 KB L1 D Cache (8-way associative, 64 bytes per line) 1024 KB L2 U Cache (8-way associative, 64 bytes per line)
Freeing init code/data: 24576 bytes (6 pages)
MIKEE: Initializing Architecture... MIKEE: Starting System... MIKEE: Creating Kernel Thread... MIKEE: Bootstrapping Kernel Thread... MIKEE: Activating Helping Lock... MIKEE: Initializing System Clock... MIKEE: Initializing Scheduler... MIKEE: Enabling Timers... MIKEE: Bootstrapping Architecture... MIKEE: Initialization Delay Loop... MIKEE: Running Kernel Thread... MIKEE: Freeing InitCall Section... MIKEE: Initializing Idle Thread... MIKEE: Initializing Workload... MIKEE: Creating Sigma0 Task... MIKEE: Creating Sigma0 Thread... MIKEE: Initializng Sigma0 Workload... MIKEE: Initializng Sigma0 Thread... MIKEE: Entering Thread::initialize()... MIKEE: Initializing Entry Frame... MIKEE: Check for Dead Thread... MIKEE: Resurrect Thread... MIKEE: Check for Exception... MIKEE: Examine Instruction Pointer... MIKEE: Initializing Instruction Pointer... MIKEE: Changing State... MIKEE: Initializing Stack Pointer... MIKEE: Check for Preempter... MIKEE: Exiting Thread::initialize()... Segmentation fault
So Fiasco-UX stops on SIGSEGV. Maybe you should also send me your globalconfig.out from your Fiasco build directory.
Up to now I have no glue what the problem could be...
On Tue Jun 27, 2006 at 16:26:33 -0400, Michael J. Emswiler wrote:
When comparing differences between your output and the output of my K7-GCC4 binary, I notice the following differences:
- Your kernel is 2.6.13.2, while mine is the default 10.1 kernel of
2.6.16.13-4
- After the "Loading Module..." messages and before "Bootstrapping...", mine
shows : Copying Module 0x039be000-0x039e0c99 [hello]
- Your running a Pentium-M (I'm running a K-7), so my TLBs and Caches are
different.
My init code/data freed is 7 pages.
I don't get a SIGSEGV after MIKEE: Exiting Thread::initialize(), it
appears to go into a tight loop that ignores CTRL+C.
Not sure what all that means but thought I'd point it out in case someone has an "Aha!" they might offer up.
I am currently suspecting an issue caused by the new libc on opensuse 10.1 (2.4 vs. 2.3.5 in opensuse 10.0). Unfortunately we got distracted by some heavy thunderstorm while discussing this...
Adam
l4-hackers@os.inf.tu-dresden.de