ia32 needs to be able to detect page misses caused by executing incorrect pseudocode

[Jakub Jermář]

The ia32 kernel needs to be able to handle page faults caused by processing IRQ pseudocode. Note that if the pseudocode is correct, it is not expected to cause a page fault on ia32 so the only case we need to handle is when the pseudocode is not correct. Thus getting to a page fault handler on ia32 while processing IRQ pseudocode is a reason to panic the system or, if running in the context of the pseudocode owner task, kill the current task.

Also note that a failure to detect this situation will result in hitting an assertion and panicking anyway as soon as as_page_fault() attempts to lock a mutex.

[Jakub Jermář]

Not applicable in current kernel which maps the physical addresses used by the pseudocode.