PANIC: page fault, but interrupts were disabled. Touching address 0x00000001 from eip 0x80b32094

[oco]

When booting Haiku (hrev28043) through pxe on my old machine (Celeron 300 / 192 Mo of RAM), a page fault occur. If i remove the gcc2 demangler from the image, it boots fine.

Here is the end of the debug output :

...
KDiskDeviceManager::_Scan(/dev/disk/virtual)
nbd:>init_hardware()
nbd:>init_driver()
KDiskDeviceManager::_Scan(/dev/disk/virtual/remote_disk)
KDiskDeviceManager::_Scan(/dev/disk/virtual/remote_disk/0)
KDiskDeviceManager::_Scan(/dev/disk/virtual/remote_disk/0/raw)
Failed to resize /var/swap to 402653184 bytes: No space left on device
Failed to add swap file /var/swap: Invalid Argument
allocate_commpage_entry(4, 34) -> 0xffff0118
set_memory_write_back base 0 length c000000
find_nearest c000000 0
find_nearest 4000000 1
find_nearest 4000000 1
solutions: 0x8000000 0x4000000
allocate MTRR slot 0, base = 0, length = 8000000, type=0x6
allocate MTRR slot 1, base = 8000000, length = 4000000, type=0x6
kernel debugger extension "debugger/disasm/v1": loaded
kernel debugger extension "debugger/hangman/v1": loaded
kernel debugger extension "debugger/invalidate_on_exit/v1": loaded
kernel debugger extension "debugger/demangle/gcc2/v1": loaded
PANIC: page fault, but interrupts were disabled. Touching address 0x00000001 fro
m eip 0x80b32094

[axeld]

Can you also post a stack trace if that is possible? You can use "sc -d" to disable the demangler in case this is needed.

[oco]

Oups ! This bug is obviously invalid. I just figure today that i was using an old haiku-netboot.tgz which contains a kernel. And because of bug fixes during the coding sprint, some modules in an up to date image were probably incompatible with this old kernel...

Everything work fine in hrev28314.

[anevilyak]

Thanks for the update!