summaryrefslogtreecommitdiff
path: root/sys/arch
AgeCommit message (Expand)Author
2009-10-24Match on `esm' as the device name, not `anything starting with esm'; ok dlg@Miod Vallat
2009-10-23Fix logic in ip27_hub_intr_makemasks() to correctly {un,}mask interruptMiod Vallat
2009-10-22Completely overhaul interrupt handling on sgi. Cpu state now only stores aMiod Vallat
2009-10-22With the splx() changes, it is no longer necessary to remember which interruptMiod Vallat
2009-10-22Implement bus_space_vaddr() for macepcibr.Miod Vallat
2009-10-22Make macebus_intr_disestablish() signature sane, and update its caller.Miod Vallat
2009-10-22The recent cleanups make blatantly visible that the pending_int handlerMiod Vallat
2009-10-22unifdef -DIMASK_EXTERNAL to the mips code. Support for interrupt masking atMiod Vallat
2009-10-22Replace intrmask_t with uint32_t. This types only describes interrupt masksMiod Vallat
2009-10-22Introduce a logical xbpci(4) device between xbridge and pci, since more thanMiod Vallat
2009-10-22Change the #define controlling use of RM7k/RM9k coprocessor 0 ICR toMiod Vallat
2009-10-22Correctly initialize the second HUB PI interrupt and calias registers onMiod Vallat
2009-10-22Remove a never hit debug panic I commited by accident sometime ago.Miod Vallat
2009-10-22Only play with RM7k coprocessor 0 ICR if IMASK_EXTERNAL is not defined.Miod Vallat
2009-10-22Do not bother invoking hw_setintrmask() in splinit(), spl0() will do it forMiod Vallat
2009-10-22At the end of a context switch and in proc_trampoline(), instead of doing theMiod Vallat
2009-10-22Crank VM_MIN_ADDRESS to prevent userland from being able to mmap zero,Miod Vallat
2009-10-21Replace IP32 hw_setintrmask() .S routine with a two line C routine.Miod Vallat
2009-10-21In atoi(), only check for a base indication iff the string starts with `0'Miod Vallat
2009-10-19Correct a target name so that we don't rebuild vers.o (and thenPhilip Guenthe
2009-10-16Honour serial console speed on O2 too.Miod Vallat
2009-10-16Make Octane kernels compile again after recent changes. My bad.Miod Vallat
2009-10-16Get serial console speed from prom, and use it instead of hardcoding 9600 bps,Miod Vallat
2009-10-15regenMiod Vallat
2009-10-15Sort widgets by type instead of manufacturer code, add some TIO widgets,Miod Vallat
2009-10-15The Octane boot PROM is accessible through the PCI space of the on-board i/oMiod Vallat
2009-10-14On coherent systems, all bus_dmamap_sync() needs to do is writebacks, noMiod Vallat
2009-10-14On IP30 and IP35 systems, try harder to figure out exactly what model we areMiod Vallat
2009-10-14Report the crossbow widget part of bedrock as PXBow for consistency withMiod Vallat
2009-10-14Add some comments to explain why the DMA32 physseg is really 2**31 bytesMiod Vallat
2009-10-13Make iof(4) pass its bus speed to children, this in turn allows com(4) toMiod Vallat
2009-10-13Get rid of devact enum, substitute it with an int and coresponding defines.Paul Irofti
2009-10-11Make sure com@io[cf] attachment stanzas with locators win over generic com*Miod Vallat
2009-10-10Simplify interrupt address programming to avoid the need to act differentlyMiod Vallat
2009-10-10Add extra com@iof stanzas to have the first IOC4 card attach com0-3, insteadMiod Vallat
2009-10-10Locators for extra IOC3 cards.Miod Vallat
2009-10-09initialize the disk size instead of getting a random value (always 0) offTheo de Raadt
2009-10-08Program the widget interrupt address register as a whole 64 bit registerMiod Vallat
2009-10-08regenMiod Vallat
2009-10-08Recognize the Bedrock as an XBow instead of reporting it as ``unknownMiod Vallat
2009-10-08Recognize more brick types and probe fooX bricks in the same order asMiod Vallat
2009-10-08PIC actually comes with two sets of widget registers, with different IDs,Miod Vallat
2009-10-07Attach DS1742 style dsrtc to iof (IOC4) too.Miod Vallat
2009-10-07Do not truncate bridge register values to 32 bits, allows the few 64 bit PICMiod Vallat
2009-10-07ipending, cpl moved into cpu_infoTakuya ASADA
2009-10-07This file is supposed to compile.Theo de Raadt
2009-10-07Try to keep existing IOC4 mappings, the same way we do on IOC3, in caseMiod Vallat
2009-10-07Missing semicolon in dmesgMiod Vallat
2009-10-07Small steps towards getting the serial clock correct on IOC4 serial ports.Miod Vallat
2009-10-07Add an IOC4 section there tooMiod Vallat