summaryrefslogtreecommitdiff
path: root/sys/arch/sgi
AgeCommit message (Expand)Author
2009-05-22Drop almost unused <machine/psl.h> on sgi; move USERMODE() definition fromMiod Vallat
2009-05-21Overhaul root device detection; the code will now follow the ARCS bootpathMiod Vallat
2009-05-21Do not attempt the interrupt deadlock workaround on IP30 yet, since theMiod Vallat
2009-05-21If we had interrupts to service and interrupts to mark as pending, afterMiod Vallat
2009-05-21Make sure splx() reenables hardware interrupt sources, even there aren'tMiod Vallat
2009-05-18Cleanup the ioctl and receive filter code for mec(4). This alsoJoel Sing
2009-05-16Add option ISP_COMPILE_FW to kernels with isp(4).Miod Vallat
2009-05-16need that include file...Theo de Raadt
2009-05-15Oops, forgot to add this file, part of the dsrtc DS174x changes.Miod Vallat
2009-05-15TOD clock support for IOC3 flavours with DS174x chips (which are rebadgedMiod Vallat
2009-05-15Don't forget to make sure IMR writes hit the hardware (by reading backMiod Vallat
2009-05-15A few stanzas from GENERIC-IP32 I forgot to uncomment or carry out.Miod Vallat
2009-05-15Forgot to remove that now unused function in previous commit.Miod Vallat
2009-05-15Update to match new kernel directories.Miod Vallat
2009-05-15Do not explicitely clear interrupt sources in the interrupt handler, butMiod Vallat
2009-05-14Add configuration files for IP27 (includes some IP35 support) and IP30Miod Vallat
2009-05-14Rename sgi GENERIC and RAMDISK with -IP32 suffixes, and have the buildMiod Vallat
2009-05-14More interrupt deadlock tomfoolery; turns out the non-XBridge workaroundMiod Vallat
2009-05-14When walking cpu configuration, report fpu type as the cpu type, sinceMiod Vallat
2009-05-14In addition to the existing IP32 boot block, build a real 64 bit boot blockMiod Vallat
2009-05-14Remove unused local.Miod Vallat
2009-05-13Set D_VENDOR when spoofing a label on archs that use D_VENDOR whenKenneth R Westerback
2009-05-09Promote types in ARCBios function prototypes from int to long wheneverMiod Vallat
2009-05-08Remove xhub boilerplate; Hub doesn't appear as a widget itself but as theMiod Vallat
2009-05-08In _dmamem_alloc_range() DIAGNOSTIC code, check for the address being withinMiod Vallat
2009-05-08Add a new page freelist, to which memory suitable for 32-bit dma onMiod Vallat
2009-05-08Fix devio byteswap, this makes isp(4) happy on IP27 and IP35.Miod Vallat
2009-05-08Make interrupt handling code closer to IP30 XHeart code.Miod Vallat
2009-05-08Attach children with config_found_sm() instead of directly invokingMiod Vallat
2009-05-08IP30 Dallas clock runs in non-BCD mode, so make sure to check which modeMiod Vallat
2009-05-06Workaround a bridge deadlock, as advised by comments found in the linux sn1Miod Vallat
2009-05-06IP27 and IP35 do not layout physical memory the same way, IP35 is muchMiod Vallat
2009-05-03Complete overhaul of the PCI bridge initialization. It will now allocateMiod Vallat
2009-05-03Pass 0 as base offset, not -1, for child devices which ignore the value.Miod Vallat
2009-05-02Let ioc(4) decide which child devices to attach depending on its identificationMiod Vallat
2009-05-02More progress taming the xbow and the pci bridge; still needs code to write,Miod Vallat
2009-05-02Explain why the short window of widget #0 needs to be accessed through aMiod Vallat
2009-05-02More KL configuration structures.Miod Vallat
2009-04-27Revert mtx_enter_try. It didn't compile on hppa, it doesn't compile onMark Kettenis
2009-04-25Handle unknown processor types as r5k family on O2, and r10k family otherwise.Miod Vallat
2009-04-25More IOC3 flavour partnumbers. And apparently there are some flavours whichMiod Vallat
2009-04-25RegenMiod Vallat
2009-04-25More widgets found on high end hardware we don't run on (yet (-: )Miod Vallat
2009-04-25Enter mtx_enter_try. In part for completeness, things may startTobias Weingartner
2009-04-25Extend PCI mmio address space by using a bigger window that requires 64-bitMark Kettenis
2009-04-24Fix the superio irq discovery heuristic to work on systems where there isMiod Vallat
2009-04-22Only report the IP number in sysctl hw.model; vendor and machine name areMiod Vallat
2009-04-20Some recent IOC3 do not have Dallas timekeepers, so perform a few more testsMiod Vallat
2009-04-20Add a BUS_DMA_ZERO flag for bus_dmamem_alloc() to return zeroed memory.Owain Ainsworth
2009-04-19On Octane, force a nonzero _dma_mask value so that we do not risk trying toMiod Vallat