summaryrefslogtreecommitdiff
path: root/sys/arch
AgeCommit message (Collapse)Author
2001-11-28more typedef zapping vm_page_t -> struct vm_page *Artur Grabowski
2001-11-28zap some typedefs.Artur Grabowski
vm_map_t -> struct vm_map * vm_map_entry_t -> struct vm_map_entry * simple_lock_data_t -> struct simplelock (uvm not done yet, coming in the next commit)
2001-11-28Make pmap_update functions into nops so that we can have a consistentArtur Grabowski
pmap_update API (right now it's nop).
2001-11-28Don't use pmap_update when we mean tlbflush. make pmap_update into a noop.Artur Grabowski
2001-11-28make pmap_virtual_space madatory in all pmaps.Artur Grabowski
2001-11-28pmap_kenter_pgs is not used and not really useful. remove.Artur Grabowski
2001-11-28pmap_kenter_pgs(.., 1) is pointless. pmap_kenter_pa instead.Artur Grabowski
2001-11-28Sync in more uvm changes from NetBSD.Artur Grabowski
This time we're getting rid of KERN_* and VM_PAGER_* error codes and use errnos instead.
2001-11-28rewrite large chunks of intr_establish to make it more clear what's going on.Jason Wright
Also, while here, fix a work around (hack) for the problem of more than 3 devices sharing a particular level (make a copy of the interrupt handler rather than modifying the one added to the pil table). This allows qec+qe (and probably qfe) to get through autoconf without hanging.
2001-11-27Various pmap_k* optimizations, as well as uvm interface updates,Miod Vallat
from NetBSD. Soon to be found in other m68k pmap, this one is just a teaser to please art@.
2001-11-27Use pools for pmap and pv_entry structures.Miod Vallat
2001-11-27Adapt to include files changes.Miod Vallat
2001-11-27A few cleanups to have less include files include each other.Miod Vallat
2001-11-27Use symbolic constants to identify ddb traps.Miod Vallat
Fix a comment as well.
2001-11-27Declare symbolic names for the various interrupts used by ddb.Miod Vallat
Make this file #include-able from locore.
2001-11-27Merge in the unified buffer cache code as found in NetBSD 2001/03/10. TheArtur Grabowski
code is written mostly by Chuck Silvers <chuq@chuq.com>/<chs@netbsd.org>. Tested for the past few weeks by many developers, should be in a pretty stable state, but will require optimizations and additional cleanups.
2001-11-27Move the declaration of EH_DEBUG and ERRATA__XXX_USR to the kernelMiod Vallat
configuration files.
2001-11-27garbage collect sc_nticksJason Wright
2001-11-26Trident 4DWAVE-DX/NX, SiS 7018, ALi M5451 Sound Driver; from SOMEYA ↵Michael Shalayeff
Yoshihiko and KUROSAWA Takahiro; tested by Matt Behrens <matt@zigg.com>
2001-11-25Extend the magic formulae to choose option FPSP only if 68040 processorMiod Vallat
support needed.
2001-11-25Typos.Miod Vallat
2001-11-25Need <machine/cpu.h> to be sure to bring userret() prototype for all ports.Miod Vallat
2001-11-25Be cross-compilation friendly.Miod Vallat
2001-11-25Remove comments which only apply to the HP MMU found on HP300 modelsMiod Vallat
318, 319, 320 and 350, which bear no sense in the mvme68k world.
2001-11-25Oops, commited this file from the wrong tree.Miod Vallat
2001-11-25Harmonize userret() prototypes across m68k arches.Miod Vallat
XXX The userret() code is not factorized out yet, as this will require XXX insane amiga cleaning work.
2001-11-24Harmonize boot() logic across arches:Miod Vallat
- ensure RB_DUMP | RB_HALT will cause a dump - or RB_HALT if (cold) While there, honor RB_TIMEBAD on sparc64.
2001-11-23Remove splsched() declaration and related stuff.Miod Vallat
2001-11-23dev{cons,cninit.c} are mandatory.Miod Vallat
2001-11-22Use pmap_kenter_pa in pmap_map..Artur Grabowski
Remove the (commented out) body of pmap_copy, it's very unlikely that it will be ever used, and right now it's just confusing.
2001-11-22more pmap_enter vs. pmap_kenter.Artur Grabowski
2001-11-22simplify the iommu page table mapping. use pmap_kenter.Artur Grabowski
2001-11-22Map the interrupt enable register with pmap_kenterArtur Grabowski
2001-11-22Use pmap_map to map the msg buffer.Artur Grabowski
2001-11-22Now that pmap_enter is allowed to fail in some cases, allowArtur Grabowski
pmap_enk* and pmap_enu* to fail and return failure to pmap_enter. pmap_enter will panic when PMAP_CANFAIL is not set, but the alternative is worse. Besides, it shouldn't happen.
2001-11-22Use pool to allocate pv list entries.Artur Grabowski
2001-11-22Don't check for PMAP_NC in pmap_zero_page.Artur Grabowski
Also make the code in there slightly more efficient.
2001-11-22Simplify pmap_is_* and pmap_clear_*.Artur Grabowski
This should also allow sun4/sun4c to work a bit better.
2001-11-21Lower splvm() level, after discussing with art.Miod Vallat
2001-11-21Provide COMPAT_SUNOS files via the common m68k configuration files.Miod Vallat
2001-11-20Oops.Miod Vallat
2001-11-20Prevent more symbol clashes when compiling userland.Miod Vallat
2001-11-20All sensitive pmap operations should operate at splvm, NOT splimp.Miod Vallat
2001-11-19more than typoTheo de Raadt
2001-11-19Remove pointless additional 'syncing' instructions in the powerpc spl*()Dale Rahn
functions. Seems to have no effect on system run time (it should have...) reduces GENERIC kernel size by 52k.
2001-11-19Use netisr_dispatch instead of handcoded list.Dale Rahn
In this instance it actually is completely identical, but had ATM been added, or something else added to netisr_dispatch, this would have been incorrect. Currently generated asm code is identical.
2001-11-18Fix RB_NOSYM clash here, too.Miod Vallat
2001-11-18Revert niklas's broken ICU interrupt handling changes; deraadt@ okAaron Campbell
2001-11-18Remove \n from more panic messagesMiod Vallat
2001-11-17put more common stuff into m68k/genassym.cf; miod@ okMichael Shalayeff