summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2009-05-24regenMarcus Glocker
2009-05-24List product name instead chip type.Marcus Glocker
2009-05-24ansi func declsJonathan Gray
2009-05-24regenStuart Henderson
2009-05-24- rename Audiovox vendor to HyundaiStuart Henderson
- add Pantech UM175 to umsm(4) from Dan Colish on tech@; ok jsg.
2009-05-24attach to more devices.Damien Bergamini
2009-05-24regenDamien Bergamini
2009-05-2410 new RT3070/RT3072 product ids.Damien Bergamini
one new vendor id (Encore).
2009-05-24fix prev commit (s/ic_flags/ic_caps/).Damien Bergamini
from dhill@
2009-05-24Improve the ARMv7 support, still work in progress.Dale Rahn
2009-05-24If flags to bind_page include BUS_DMA_COHERENT, set the snooped bit onOwain Ainsworth
the pte. This is currently unused, but will be used soon.
2009-05-24Instead of using magic numbers, define the bits that we set in gtt ptesOwain Ainsworth
and use those defines.
2009-05-24Rework for 3530.Dale Rahn
2009-05-24remove debug print.Dale Rahn
2009-05-23match on D-Link DWA-130 rev D1Jonathan Gray
2009-05-23regenJonathan Gray
2009-05-23D-Link DWA-130 rev D1.Jonathan Gray
2009-05-23Remove redefinition of ETHER_ADDR_LEN.Jonathan Gray
2009-05-23remove PCI vendor/product defines which duplicate those in pcidevsJonathan Gray
2009-05-23More PG_RELEASED cleaning.Owain Ainsworth
similar to the aobj.c changes, this one does vnodes. Vnodes are more complex because they actaully have to sync to backing store. So firstly, convert sync to sleep instead of setting released. Now, for backing store, in the PGO_FREE case, if we set PG_RELEASED *before* an async io, (nothing else will see it, the page is busy), then we can ignore the page after the io is done. We could do something similar for PGO_DEACTIVATE too, but that is another change. On error we just clear the released flag, nothing else sets it for uobj pages other than aiodoned. ok thib@, beck@, ariane@
2009-05-23FALL THROUGH -> FALLTHROUGH to appease lintJonathan Gray
2009-05-23Seems I confused Mb and MB and was in too much of a rush when IJonathan Gray
added the DDR3 bits. Take into account more of the spd data and compute DDR3 module size in a way similiar to NetBSD so it comes out correctly. Tested by Olivier Cherrier and Christian Stuermer.
2009-05-22According to pics on ebay, CADDuo boards indeed have Ethernet connectors.Miod Vallat
2009-05-22Drop almost unused <machine/psl.h> on sgi; move USERMODE() definition fromMiod Vallat
there to trap.c which is its only user. This also cleans up multiple inclusion of <machine/cpu.h> (because <machine/psl.h> includes it) in many places.
2009-05-22Add Unitek Y-2240. -mojMats O Jansson
2009-05-22Regen. -mojMats O Jansson
2009-05-22Yet another udl device. -mojMats O Jansson
2009-05-22retire the nfs rtt log code, this isnt really all thatThordur I. Bjornsson
usefull, and has been disabled for a long long time. Cleans house a bit. OK blambert@
2009-05-21The only value that d_npartitions should have is MAXPARTITIONS.Kenneth R Westerback
2009-05-21add support for BCM576x, 577xx, and some new 572x/578x devices.Stuart Henderson
from Brad; testing by phessler, naddy, myself and others on a range of older and newer devices. ok dlg@.
2009-05-21regenStuart Henderson
2009-05-21add BCM5784, from Brad.Stuart Henderson
2009-05-21Overhaul root device detection; the code will now follow the ARCS bootpathMiod Vallat
component by component, from device_register(). IP27 and IP35 systems using the dksc() syntax get their dksc() syntax converted to an ARCS bootpath, so the same logic will apply.
2009-05-21Add König CMP-USBVGA10. -mojMats O Jansson
2009-05-21Do not attempt the interrupt deadlock workaround on IP30 yet, since theMiod Vallat
IP27 logic will obviously not work there.
2009-05-21If we had interrupts to service and interrupts to mark as pending, afterMiod Vallat
servicing interrupts, correctly mask the other (marked as pending) interrupts in imr. Otherwise we get another interrupt immediately for nothing (which updates imr correctly the second time).
2009-05-21Regen. -mojMats O Jansson
2009-05-21yet another udl device. -mojMats O Jansson
2009-05-21Make sure cpu_switchto() returns to the new proc with interrupts enabled (thisMiod Vallat
is bandaid until interrupt handling is made more sane.
2009-05-21In splinit(), adjust proc0 pcb values, for kthread to start with correctMiod Vallat
interrupt masks.
2009-05-21Make sure splx() reenables hardware interrupt sources, even there aren'tMiod Vallat
any such interrupts marked as pending.
2009-05-21always set VR_RXCFG_RX_MULTI when setting the chip into promiscuousStuart Henderson
mode; VR_RXCFG_RX_PROMISC isn't enough to allow reception of multicast frames. Discussed with Brad, ok dlg@.
2009-05-21Add ForwardVideo EasyCAP008. -mojMats O Jansson
2009-05-21Regen. -mojMats O Jansson
2009-05-21Another udl device. -mojMats O Jansson
2009-05-21IO_METASYNC has been a noop since around 4.4BSD-Lite, theThordur I. Bjornsson
idea never really caught on anywhere so retire it. "gank this shizzle", blambert@
2009-05-20Add code to attach sun mice on sparc and sparc64 as wsmouse devices. SinceMiod Vallat
not all Sun mice run at the canonical 1200 bps, hop between 1200, 4800 and 9600 bps by paying attention to breaks on the line. Attachement and engine code written 7 years ago for OpenBSD/sparc, except for sparc64 com(4) attachment. Speed hop idea borrowed from Opensolaris. This allows sparc and sparc64 users to run X11 without needing a configuration file anymore, as it was in the XFree86 3.x days. Multihead configurations will still need a minimal configuration file, though.
2009-05-20add code to read from OTPROM (1000 and 6000 series only).Damien Bergamini
2009-05-19Processing AST can trigger more AST. Loop on astpending instead of checkingMiod Vallat
only once.
2009-05-19prevent injection of raw 802.11 control frames through bpf on driversDamien Bergamini
that do not explicitly state this capability (using flag IEEE80211_C_RAWCTL). also, perform all the sanity checks on injected raw 802.11 frames earlier (in ieee80211_output instead of ieee80211_encap). prevent kernel panics with many drivers when running aircrack-ng. when/if all drivers are capable, we can remove this C_RAWCTL flag.