Subversion Repositories HelenOS-historic

Rev

Go to most recent revision | Show changed files | Details | Compare with Previous | Blame | RSS feed

Filtering Options

Rev Age Author Path Log message Diff
758 6850 d 22 h jermar /kernel/trunk/arch/ sparc64 bugfix.
When disabling IMMU and DMMU the kernel has to perform synchronization operation
(e.g flush %r or membar #Sync instruction). There is no guarantee that the address
contained in %r is in DTLB and therefore the flush instruction can fault. Normally
this would be recognized and fixed by the OpenFirmware Fast Data MMU fault handler.
However, this handler lives in virtually mapped memory and an attempt to execute
there while the MMUs are disabled would result in a nested trap leading to error state.
Replacing flush %r instruction with membar #Sync, wich is sufficient in this case,
fixes this problem.		  
757 6851 d 1 h jermar /kernel/trunk/ Blacklist addresses between 0xa0000 and 0xfffff on ia64 for frame allocator.
This area contains VGA text frame buffer and should be avoided.
falloc2 test now passes on ia64.		  
756 6851 d 21 h jermar /kernel/trunk/ Memory management work.
Remove the last (i.e. 'root') argument from page_mapping_insert() and page_mapping_find().
Page table address is now extracted from the first (i.e. 'as') argument.
Add a lot of infrastructure to make the above possible.
sparc64 is now broken, most likely because of insufficient identity mapping of physical memory.		  
753 6853 d 5 h jermar /kernel/trunk/ Convert ASID management of ia64 to ASID FIFO mechanism.
18-bit RIDs are supported.		  
746 6857 d 7 h jermar /kernel/trunk/ Page hash table modifications.  
743 6859 d 0 h jermar /kernel/trunk/arch/ sparc64 work.
13 bits wide means the max value is 8191 and not 0x8191.		  
742 6859 d 0 h jermar /kernel/trunk/arch/ sparc64 work.
Memory context (ASID) is, in fact, 13 bits (and not 12 bits) wide.		  
741 6859 d 1 h jermar /kernel/trunk/ Unlock address space prior TLB shootdown in get_asid() to unify
the locking order among mips32, sparc64 and ia64.

Add ASID_STEALING_ENABLED macro to disable the stealing part on ia64
in a clean way.		  
727 6863 d 23 h jermar /kernel/trunk/ New ASID management subsystem (initial work, more is required).
Some TLB invalidation changes.		  
703 6870 d 8 h jermar /kernel/trunk/ Memory management work.
- vm.* -> as.* (as like address space is, imho, more fitting)
- Don't do TLB shootdown on vm_install(). Some architectures only need to call tlb_invalidate_asid().
- Don't allocate all frames for as_area in as_area_create(), but let them be allocated on-demand by as_page_fault().
- Add high-level page fault handler as_page_fault().
- Add as_area_load_mapping().		  
700 6873 d 7 h jermar /kernel/trunk/arch/sparc64/include/ Remove fmath.h  
699 6873 d 8 h jermar /kernel/trunk/ Memory management work.
Proto-interface and dummy implementation of generic page hash table subsytem.		  
691 6875 d 5 h jermar /kernel/trunk/arch/ Cleanup.  
669 6883 d 23 h jermar /kernel/trunk/ sparc64 work.
kconsole support.
Add non-blocking ofw_getchar().		  
667 6885 d 2 h jermar /kernel/trunk/arch/sparc64/ sparc64 work.
Rename saving_handler() to preemptible_handler()
and fix it to make sparc64 kernel preemptive.
Add two handlers for two fatal exceptions (i.e.
instruction_access_exception and mem_address_not_aligned.
Fix panic_printf() to not allocate its own register window.		  
666 6887 d 0 h jermar /kernel/trunk/ sparc64 work.
Fix clean_window trap handler so that it clears output registers instead of input registers!
Fix interrupt handlers to save global registers.		  
665 6887 d 9 h jermar /kernel/trunk/arch/sparc64/ sparc64 work.
Tick interrupt support.		  
664 6888 d 20 h jermar /kernel/trunk/ sparc64 work.
Interrupt Levels 1 - 15 serviced.
Minor changes in the exc_* functions.		  
658 6892 d 10 h jermar /kernel/trunk/arch/sparc64/include/ sparc64 work.
Add functions to read and write TICK and TICK_compare registers.
Add types describing TICK and TICK_compare registers.		  
657 6892 d 22 h jermar /kernel/trunk/ sparc64 work.
Context again. It turns out, according to SCD 2.4, that registers that didn't have to be saved were saved (%o1 - %o5) and
registers that had to be saved were not saved (%i0 - %i5, %l0 - %l7).		  

←Prev12Next→

Show All