summaryrefslogtreecommitdiff
path: root/sys/arch
AgeCommit message (Collapse)Author
2007-11-25Enable et/etphy now it can pass packets both ways.Jonathan Gray
2007-11-25It helps if you initialise timeouts before you try using them...Joel Sing
ok miod@
2007-11-25Get rid of the kernel 'libcompat' framework, and instead use conf/files toTheo de Raadt
decide which files must be pulled into the kernel. Also conditionalize the pulling of those files based on the COMPAT_* options.
2007-11-25Add entries for the Transmeta Efficeon CPU to the CPUID table.Tom Cosgrove
Without this we think it's an 80386-compatible, which is no longer supported. Problem found, and solution tested, by Anders Langworthy (anders (at) psilanthropy (dot) org); thanks. Initial suggestion for fix by Travers Buda (traversbuda (at) gmail (dot) com). ok deraadt@
2007-11-25When available and ACPI is enabled use the ACPI reset methodJonathan Gray
to reboot. This is required for some quirky sis machines and likely others to reboot. Tested by many. 'commit early commit often' deraadt@
2007-11-25Give the CarBus side of CardBus-PCI bridges a bus number, since OpenFirmwareMark Kettenis
doesn't do it for us.
2007-11-24make ALTENTRY() use _C_LABEL() for the 2nd argument, and adjust callersTheo de Raadt
ok miod
2007-11-24Recognize the wait r5k instruction in ddb and use it in cpu_idle_cycle.Miod Vallat
ok jsing@ a while ago (and this diff is much much older)
2007-11-24Fix a stupid underscore bug in memory area computations in the cache functions,Miod Vallat
it's amazing things didn't break.
2007-11-24Some macppc models (e.g. the 1.5 GHz G4 MacMini) have a large clock driftMarc Balmer
which is due to the timecounter running at a different frequency than what OpenFirmware tells us. On such machines, measure the actual timecounter frequency (triggered by a table of affected models) and use this measured value instead. This brings down the drift to values where adjtime()/adjfreq() triggered by ntpd(8) can tame the clock. feedback many, ok miod, kettenis
2007-11-24Be sure to program the busswitch interrupt selection register on 197SPMiod Vallat
and 197DP to route interrupts to the processor we're booting on. This allows a 197DP to run when booting from the second cpu.
2007-11-24some spelling fixes from Martynas VenckusJason McIntyre
2007-11-24enable PCIAGP on amd64, tested for some time now by various peopleReyk Floeter
ok tedu@ oga@
2007-11-24Do not mark any page as global on monoprocessor kernels if running on aMiod Vallat
88110, to avoid snooping notification bus cycles, as recommended by the manual.
2007-11-24Slightly faster cache flushing operations on MP systems.Miod Vallat
2007-11-22Split the cmmu code routines into single 88110 (MVME197LE) and 88110+88410Miod Vallat
combos (MVME197SP/DP), and implement supposedly smarter cache routines. There is still room for improvement, however, cache flush operation errata permissing. Tested on 197LE and 197DP.
2007-11-22Introduce an inline function to skip an instruction on 88110 and use itMiod Vallat
whenever necessary, instead of duplicating the same code 10+ times.
2007-11-22A decent workaround for errata #16. I have verified it gets triggered, theMiod Vallat
hard way.
2007-11-22Do not reenable interrupts for netsted interrupts if interrupts wereMiod Vallat
previously disabled on 88110, similar to a recent 88100 fix.
2007-11-22Do not bother saving the fault address registers if the fault statusMiod Vallat
registers are empty, to speed up a bit non-fault traps.
2007-11-22Do not redirect exception occuring when enabling the fpu to the error handler,Miod Vallat
for we are supposed to be able to recover from them.
2007-11-22Do not save a copy of the previous exception frame in a fixed memory location.Miod Vallat
2007-11-22disable i386 ipmi as well, because marco has an acpi interaction bug to fix ↵Theo de Raadt
still
2007-11-22On 88110, skip the faulting instruction in double_reg_fixup() if it canMiod Vallat
recover, otherwise we'll fault in a loop.
2007-11-22Quote a few errata to explain why odd things are done in oddly ways on 88110.Miod Vallat
2007-11-22Remove the cpu parameter from cmmu_set_sapr(), since it is only invokedMiod Vallat
for the current processor. And remove now unused cmmu_flush_data_page().
2007-11-22Remove explicit tlb flush from vmapbuf(), since pmap_enter() will takeMiod Vallat
care of this.
2007-11-22Add an extra parameter to pmap_remove_pte() to control whether it willMiod Vallat
flush tlbs or not. This is used by pmap_enter() to avoid flushing the same tlb entry twice.
2007-11-22Move the cmmu lock to 88200-specific code. 88110 MP code will use ipisMiod Vallat
and will not require such a lock.
2007-11-21Use bcopy instead of assignment to avoid a possible misaligned access.Miod Vallat
2007-11-21TypoMiod Vallat
2007-11-21Better cache invalidation functions (still horribly unoptimal, but lessMiod Vallat
broken now).
2007-11-21Don't bother flushing caches in pmap_{copy,zero}_page(), since our cachesMiod Vallat
are physically addressed. Might be revisited for 88110 SMP, but we're not there yet.
2007-11-21Move pmap_set_modify() prototype to a header fileMiod Vallat
2007-11-21Move external interrupt and ast handling from m881[01]0_trap() to theirMiod Vallat
own functions, which do not need to be processor-specific. This speeds up the exception return sequence a bit.
2007-11-21A really horrible fix for sigreturn on 88110. This will probably be replacedMiod Vallat
by something better in the future, but for now this will do. Just don't tell anyone I did this.
2007-11-20A line got lost while hand-editing diffs...Miod Vallat
2007-11-20On 88110, use a shorter path in pmap_is_modified(), since we performMiod Vallat
accurate mod tracking through the write fault traps; there is no need to walk the pv list since, unlike on 88200, hardware does not set PG_M or PG_U in page table entries.
2007-11-20Move 88110 trap-on-write processing from two duplicated sections in trap.c toMiod Vallat
a single function in pmap.c, which will do the right thing and also perform mod/ref vm_page flags accounting.
2007-11-20Fix MP locking issues when processing data access faults on 88110, similarMiod Vallat
to the fix already applied to 88100.
2007-11-20Overhaul of the pmap debug messages. They now provide slightly betterMiod Vallat
information, and are easier to parse. While there, a few never-happening checks move from option DIAGNOSTIC to option PMAPDEBUG.
2007-11-20Fix setregs() on 88110, so that binaries do not skip their first instruction.Miod Vallat
Makes binaries beyond init(8) running on MVME197, to some extent.
2007-11-20In data_access_emulation(), make sure that if we want the pipeline replayMiod Vallat
to be disabled, this effectively disables the three slots, instead of only the first. This will only make (rare) things faster.
2007-11-18Add missing locking around ktrsyscall() call.Mark Kettenis
2007-11-17- move the re/bwi CardBus entries in the GENERIC kernel config intoBrad Smith
the CardBus section. - add forgotten re CardBus entry to the RAMDISK kernel config. ok dlg@
2007-11-17AMD Geode CS5536 GPIO can invert input and output pins.Marc Balmer
2007-11-17Allow for more than one flag per GPIO pin.Marc Balmer
ok grange.
2007-11-17Support GPIO_PIN_PULLDOWN (pulldown-enable) as GPIO pin flag.Marc Balmer
2007-11-17The initial ipl on luna/mvme188 like interrupt arbiters is IPL_HIGH, notMiod Vallat
IPL_NONE; fixes a false splassert warning on boot.
2007-11-17Replace many ``unsigned'' variables with ``unsigned int'', ``u_int'' or otherMiod Vallat
appropriate types. No functional change.
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-23 08:24:09 +0000