summaryrefslogtreecommitdiff
path: root/sys/arch
AgeCommit message (Collapse)Author
2003-08-07dead protoMichael Shalayeff
2003-08-07make all intr_establish take the same last arg for nameMichael Shalayeff
2003-08-07first allocate space on stack then copyout signal shits. add a DEBUG check ↵Michael Shalayeff
for a whacky user pc (as impossible due to the recent trap.c change)
2003-08-07be more careful when loading trap_tmp_save and a bit better sigcodeMichael Shalayeff
2003-08-07In sendsig() and sigreturn(), use the SS_xxx flags rather than the SA_xxxMiod Vallat
flags (purely cosmetic, as they have the same value). Also some minor indentation/KNF repairs.
2003-08-07better dmesg output and do some spaces; markus@ okMichael Shalayeff
2003-08-07Allow nested interrupts in the openpic interrupt controller, tested byDale Rahn
brad@ and miod@
2003-08-07watchdog driver for Geode SC1100 (e.g. in soekris net4801); ok mickey@Markus Friedl
2003-08-07do not treat traps from the syscall gate page as user codeMichael Shalayeff
2003-08-06Remove some double semicolons (hmm, do two semis equal a maxi?).Todd C. Miller
I've skipped the GNU stuff for now. From Patrick Latifi.
2003-08-05a180c blinks too, 715/80 supposedly tooMichael Shalayeff
2003-08-05support for shared interrupts as required by isa/eisa/pci. remove a gsckbc ↵Michael Shalayeff
kludge; miod@ testing
2003-08-04make 'em all unsignedMichael Shalayeff
2003-08-03Resistance is futile, you will be KNF'ed.Miod Vallat
And then it will be easier to debug this mess (no functional change yet).
2003-08-03oopsMichael Shalayeff
2003-08-03prefer fdc to fdce for cache flushing hereMichael Shalayeff
2003-08-03clear branch and nullify bits in psw on sending a signalMichael Shalayeff
2003-08-01The pmap potpourri du jour, while hunting for evil bugs:Miod Vallat
- provide a simpler flush_atc_entry() in the NCPUS == 1 case - remove some can't happen tests in pmap_protect() - handle pool_get failure and PMAP_CANFAIL correctly in pmap_enter() - don't forget to initialize pv_flags in new pv_entry items - de-cretinize pmap_testbit() and pmap_page_protect()
2003-08-01Nuke uvm_useracc() call here too.Miod Vallat
2003-08-01The Vigra VS10-EK is recognized as a cgthree clone, but due to its utterlyMiod Vallat
ridiculous resolution, we need to plug the "clear the screen and switch font" trick here too. [Thanks to Andrey Smagin for providing a card for testing]
2003-08-01pmap_kremove was so broken I'd rather pretend I did not write it...Miod Vallat
2003-08-01Nuke calls to uvm_useracc() and instead check copyin() and copyout() forMiod Vallat
failure, since the former is not reliable; requested by art@ long ago.
2003-08-01No semicolon at the end of macros supposed to be atomic statements.Miod Vallat
2003-08-01Compensate for the assym.h lossage, which defines NBPG for now (this will beMiod Vallat
cleaned up soon...)
2003-08-01Do not use hardcoded values for constants which are in fact computed from otherMiod Vallat
constants.
2003-08-01Turn flush_pipeline() in a simple macro.Miod Vallat
2003-08-01Define a more reasonable label_t size.Miod Vallat
2003-08-01Even better varargs code, inspired by and more closer to the powerpc code,Miod Vallat
necessary for proper gcc 2.95 operation.
2003-08-01Fix pcctwo locator typosMiod Vallat
2003-08-01${LIBGCC} is useless, since libgcc.a does not live in /usr/lib; instead,Miod Vallat
use gcc to get the correct location.
2003-08-01Fix the *longjmp() behaviour - it is legal to reuse a jmp_buf several times.Miod Vallat
Gets us a working perl 5.8.
2003-08-01remove the strcpy/strcat macrosAnil Madhavapeddy
deraadt@ ok
2003-07-31remove autoconf_nzs; from netbsdJason Wright
2003-07-30an absolute cure for an inherited arrhythmia w/ uninitialized variableMichael Shalayeff
2003-07-30kill sample device mappings as not really usefulMichael Shalayeff
2003-07-30deduce the leds address from the model number and thus make some more use of ↵Michael Shalayeff
that lichtenblinkenschmutz on some more machinens
2003-07-30give those leds some hasteMichael Shalayeff
2003-07-30support for vt8231 from netbsd via pr3362 and tested by matthew.gream@pobox.comMichael Shalayeff
2003-07-29load_raw did not set all the needed fields in the dmamap.Michael Shalayeff
make sure to allocate dmamem in one segment. do not allocate extra va for dmamem allocated -- just reuse the already there equ mapping of pa and thus implement the dmamem_free as a pglist free.
2003-07-29move spl checking wrap over the syscall into trap.c:syscall();Michael Shalayeff
this gives several benefits: the arguments for a syscall printed are the exact args and not some wild guess fetched from the user stack; the check does not include the AST traps that may be checked on their own in the trap() routine should be desired; kill last use and thus itself the fuword() function and it's glue.
2003-07-29Remove a diagnostic print which was left in the code to determine howDale Rahn
bad integer division is for calculation of the tick timer. Turns out that the typical error on most machines is < .005% just a few seconds per week. Not enough to bother with at this time.
2003-07-28nuke more MATH_EMULATEJason Wright
2003-07-28remove the non-licensed i386 math emulation stuff. This only leaves theJason Wright
gnu stuff as an option until it is replaced RSN. ok deraadt.
2003-07-26on syscall exit restore old cpl if there is a need; move call for softints ↵Michael Shalayeff
processing towards after that
2003-07-25use names not numbers for msrs, do not define yet another implementation of ↵Michael Shalayeff
wrmsr/rdmsr
2003-07-25s/CPUID_SYS2/CPUID_SEP/ to avoid confusionMichael Shalayeff
2003-07-25let ALWAYS_MATH_EMULATE actually compileJason Wright
2003-07-25make sure SYSENTER registers are 0 at boot time. should fix an unusualTed Unangst
DOS reported by Michal Zalewski to bugtraq. ok mickey@ toby@
2003-07-24a whole bunch of tyop fixes from Andrey SmaginJason Wright
2003-07-23add ahc and clean things a bitJason Wright