[Barrelfish-users] Kernel.0: Waiting for GDB connection!!

[jiban sarma]

Hi All,

We are trying to boot Barrelfish on few HP-Proliant servers over tftp/pxe. Initially we are getting the below exception for all the machines:

Faulting instruction pointer (or next instruction): 0xffffff800757805d
=>i.e unrelocated kernel address 0x10f05d

And finally, the boot process stuck with the below message:
kernel.0 waiting for GDB connection...

To resolve the issue we have commented out few assert statements in the "startup_arch.c",  after that  two machines are able to bypass that error. But still we are getting the same error for  one machine.

===================================================================
assert(vbase + size - X86_64_INIT_VBASE < X86_64_INIT_SPACE_LIMIT);
//assert(err_is_ok(err));
assert(glbl_core_data->mmap_length < BOOTINFO_SIZE);
assert(curmmap->size == 20); // FIXME: The multiboot specification does not require this size
assert(nextmmap->size == 20);
//assert((clean_mmap_length + tmpmmap.length + 4) < BOOTINFO_SIZE);
//assert(err_is_ok(err));
//assert(err_is_ok(err));
assert(mmap->base_addr > local_phys_to_gen_phys(init_alloc_addr));
// assert(err_is_ok(err));
assert(last_end_addr != 0);
// assert(err_is_ok(err));
// assert(err_is_ok(err));
assert(apic_is_bsp());
//assert(err_is_ok(err));

=========================================================

Can anyone give us a pointer why we are getting this ? and how can we deal with the  unrelocated kernel address error.
Also, kernel.0 waiting for GDB connection implies for remote debugging(remote gdb stub) option in barrelfish?
We have checked remote debugging in QEMU, but not sure how can we do it in a tftp/pxe setup. Please provide some pointers.

Thanks,
Jiban
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.inf.ethz.ch/pipermail/barrelfish-users/attachments/20140218/2845a0a0/attachment.html