summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2008-07-18Add a macro that clears the want_resched flag that need_resched sets.Artur Grabowski
Right now when mi_switch picks up the same proc, we didn't clear the flag which would mean that every time we service an AST we would attempt a context switch. For some architectures, amd64 being probably the most extreme, that meant attempting to context switch for every trap and interrupt. Now we clear_resched explicitly after every context switch, even if it didn't do anything. Which also allows us to remove some more code in cpu_switchto (not done yet). miod@ ok
2008-07-18Add ohci(4) and ehci(4) at cardbus(4).Mark Kettenis
2008-07-18Determine the free address space by looking at the "available" property ofMark Kettenis
the PCI host bridge if we're not running on an UltraBook. Fix allocation of bus number such that it works on machines that have OpenBoot 4.x.
2008-07-18For negotation first get the devices default values (GET_DEF) insteadMarcus Glocker
asking for (GET_CUR) directly. This gets us better negotation values.
2008-07-18Protect dma operations with splvm instead of splhigh, that's faster thanMiod Vallat
splbio and won't delay clock interrupts.
2008-07-18Make IPL_VM level 5, which is guaranteed to be above bio tty and net, insteadMiod Vallat
of making it dynamic and the smallest value above the former three. Idea from NetBSD.
2008-07-18Implement VIDIOC_S_FMT and VIDIOC_TRY_FMT. Now V4L2 applications canMarcus Glocker
set their desired image size, and therefore users can manipulate the image size, too via the application. Also tested by brad@
2008-07-18Add support for the kernel to recognize, load and execute positionKurt Miller
independent executables using the uvm_map_pie() function to randomize the load address. okay miod@, kettenis@, drahn@
2008-07-18Add new uvm function called uvm_map_pie() which takes align as aKurt Miller
parameter and returns an aligned random load address for position independent executables to use. This also adds three new vmparam.h defines to specify the maximum address, minimum address and minimum allowed alignment for uvm_map_pie() to use. The PIE address range for i386 was carefully selected to work well within the i386 W^X framework. With much help and feedback from weingart@. okay weingart@, miod@, kettenis@, drahn@
2008-07-18o Use mbufs, for the RX ring, instead of malloc()'ing an MCLBYTES sized buffer.Thordur I. Bjornsson
o On non-strict alignment archs, dont copy the mbuf, every time, unload it, and send it up the stack and just get a new one for the rx ring. We still do the copy on strict alignment archs though... o create a function to handle mbuf allocation for the rx ring, vr_mbuf_alloc(), use it to allocate the mbufs and shuffle the bus dma setup around. ideas/code from vic(4) and sis(4); ok reyk@, brad@, dlg@ tested by many, been in snapshots for a while.
2008-07-18the Belkin F5D7010 version # is a RTL8185 cardbus device. thisReyk Floeter
chipset revision is not supported yet, so add it commented out.
2008-07-18syncReyk Floeter
2008-07-18Add Belkin Wireless G Notebook Card F5D7010 (rtw(4) RTL8185-based)Reyk Floeter
2008-07-18POLL causes concurrency issues with event driven batteries. Remove for nowMarco Peereboom
until I have time to implement proper fix. Noted by several folks.
2008-07-17Zero mbreg_t structure *before* initializing any members, not when we'reMark Kettenis
halfway through. Makes the ISP12160 correctly load its firmware again on sparc64. ok krw@
2008-07-17Add a flag and allow the driver to identify the BCM5788 chipsets. They'reBrad Smith
slightly different from the other chipsets using the BCM5705 ASIC. ok jsg@
2008-07-17Remove clause 3 from NetBSD license.Joel Sing
ok deraadt@
2008-07-17Increase the size of the diskinfo struct, which has grownJonathan Gray
4 bytes in the new firmware. To allow this to work on older firmwares add a "short read allowed" flag to arc_msgbuf(), which will allow short reads by getting the read length from the header of the response the firmware sends. Only set the flag on the arc_msgbuf() call that deals with the diskinfo struct for now. Discussed with and ok dlg@
2008-07-16Add a new parameter to pckbc_cnattach(): flags to put in the pckbc bowels.Miod Vallat
For now, only one such flag is defined, PCKBC_CANT_TRANSLATE. It hints pckbc that the device does not behave correctly to the ``set translation'' commands. Set this flag if we are running on a Tadpole Ultrabook machine, which needs it. This makes the built-in keyboard work correctly on this laptop (with the help of the software translation pckbd diff). tested & ok kettenis@
2008-07-16If pckbc_xt_translation() fails, do not try to force the keyboard in tableMiod Vallat
1; instead, try table 2 first, then 1: some older keyboards, connected to a castrated 8042, do not know how to talk in table 1 (or worse, will report they do while they don't). If automatic translation is not possible, remember which table the keyboard ended configured in, and convert the scancodes from table 2 to table 1 if necessary. ok kettenis@
2008-07-16Send unused string literal ``mapme'' to nirvana.Miod Vallat
2008-07-16This file is long unused.Miod Vallat
2008-07-16It is no longer necessary to regen gsckbdmap after changing this file.Miod Vallat
2008-07-16Implement pckbc_xt_translation() in gsckbc. Thus, we can attach pckbd instead ofMiod Vallat
gsckbd; the former will cause a proper translation page to be selected by the keyboard. Because of this, we no longer depend on the page the keyboard is left in by the PDC (page 2 for all machines but the PrecisionBook, which is in page 3), and there is no longer any need to use separate keyboard maps.
2008-07-16Now that uvm_pglistalloc() does not lose on large memory gaps, do notMiod Vallat
restrict the memory allocation range in _dmamem_alloc().
2008-07-16link udp pcbs to pf states, same as done for tcp alreadyHenning Brauer
ok markus, also tested david sthen
2008-07-16Enable the read DMA engine's PCI read request burst length long burstBrad Smith
mode (4KB) for PCIe chips. This resolves the poor TX performance for the PCIe chips. The result being a bit under double the TX performance on a Gig connection (roughly 495 Mb/s -> 940 Mb/s). Tested by reyk@, sthen@, brad@ and a few end users.
2008-07-15Add a timecounter based on the $sys_tick register, and use it on machinesMark Kettenis
that have it. Initial diff from art@.
2008-07-15Use -Wa,-Av9b instead of -Wa,-Av9a such that I can use %sys_tick in anMark Kettenis
upcoming change.
2008-07-15Backout rev 1.162. This change made us muck with with pci config space atMark Kettenis
address 0x1a and 0x1e, and that's not where the PCIe capability stuff lives. Potentially it was mucking with an IO BAR (super dangerous). But probably it was achieving nothing at all. ok dlg@, marco@, brad@
2008-07-15Remove some debug code that i forgot to remove in a previous commit.Owain Ainsworth
Pointed out by Chris Cappuccio, thanks!
2008-07-15Adapt some code from the FreeBSD driver required to makeJonathan Gray
newer PCI Express adapters (ie 8168C*/8102*) work. V2 Checksum offload format in RTL8102 devices not yet supported. No objections from brad@. Thanks to everyone who tested.
2008-07-15theres no need too m_adj() the mbuf by 8bytes (u_int64_t) inThordur I. Bjornsson
sis_newbuf(), so remove the call. ok and lots of prodding dlg@, brad@
2008-07-14Fix struct sysioreg layout after the struct iommureg layout change;Miod Vallat
ok kettenis@
2008-07-14Make VIDIOC_ENUM_FMT list all available formats for the attached device.Marcus Glocker
2008-07-14Don't lock timeouts in db_show_callout.Artur Grabowski
All cpus are stopped and this cpu blocks all interrupts. It doesn't make sense to grab locks that ddb can then jump past with longjmp. Noticed by Pierre Riteau. I just forgot about the bug until reminded today.
2008-07-14Use uvm_km_valloc_prefer_wait() instead of uvm_km_valloc_wait() in vmapbuf().Miod Vallat
ok art@
2008-07-14Be sure to propagate PSL_O in psw on PCXU* processors to userland processesMiod Vallat
and signal handlers. ok kettenis@
2008-07-14Be sure to propagate PSL_O in psw on PCXU* processors to userland processesMiod Vallat
and signal handlers. ok kettenis@
2008-07-14m_copy can return NULL, so check for itHenning Brauer
problem found by "Adrian M. Whatley" <amw> and "Stephan A. Rickauer" <stephan.rickauer>, both @ni.phys.ethz.ch fix by me, ok mpf
2008-07-14Zap some dead commons that are no longer used.Artur Grabowski
"Commitski!" miod@
2008-07-14Finish support for uncompressed payloads (UDESCSUB_VS_FORMAT_UNCOMPRESSED).Marcus Glocker
2008-07-13Do not specify the gateway on RTM_DELETE -- similar fix was done in the arpClaudio Jeker
code. This fixes in6_ifloop_request warnings seen on carp interfaces. OK henning@, found and tested by david@
2008-07-13Spacing.Marcus Glocker
2008-07-13Save format and frame descriptors to a format group structure so weMarcus Glocker
can use them later for stuff like VIDIOC_S_FMT (e.g. set custom resolution).
2008-07-13Add some additional hardware revisions from FreeBSD needed forJonathan Gray
upcoming changes and sort list.
2008-07-12reference vmt (openbsd) rather than vmware (xorg)Chris Kuethe
ok dlg
2008-07-12Make the console interrupt driven.Mark Kettenis
2008-07-12Make sure the label is read in rdopen(), this isn't the case if youMiod Vallat
boot bsd.rd -a and want to mount the ramdisk later on. ok otto@ krw@ millert@
2008-07-12Right now agp_generic_enable() is wrong. It has been since 2006. ItOwain Ainsworth
assumes that the display device and the agp bridge are the same device. In almost all cases this is incorrect. In order to correctly enable the agp device with the correct mode, we need to set the AGP_COMMAND register on both the display and the bridge with the right bits. Since agp is currently attaching at vga(4), due to the problems with the intel integrated graphics (a problem that I still need to solve, to be honest), for now just provide both pci_attach_args to the agp init, and get what we need from there to do the enable. This fixes the "agp bug" which i have been known to rant about, a lot. So agp radeons now work with dri without needing to be forced to pci mode. This wasn't detected before, since the only non-drm consumer of agp is the intel X driver, the i810 agp driver has its own enable function. tested by many. ok kettenis@.