fix wrong integer promotion rule(removed U suffix from UINT{8,16}_C).
see ISO/IEC 9899:1999 7.18.4.3.

(tnozaki)

Nvidia EHCI controllers do not ignore one or more of the "Port Number",
"Hub Address", "Split Completion Mask" fields in the Queue Head marked
"This field is ignored by the host controller unless the EPS field
indicates a full- or low-speed device.".

Therefore, only populate these fields for full- and low-speed devices.

Fixes PR#37884.

(jakllsch)

Move "compiled" files from dist directory to reachover directory (suggested
by mrg).

(ghen)

Move "compiled" files from dist directory to reachover directory (suggested
by mrg).  Add a note in the Makefile for future upgrades.

(ghen)

The rule of W-ZERO3 was added again.

(nonaka)

sharp_vndr was added to the build object.

(nonaka)

unload usbverbose, not pciverbose

(cegger)

Rename ci_node_id to ci_package_id, as some claim that the former might
be confused with NUMA node.

(rmind)

avoid vmware module crashing with "X -configure".

(mrg)

Extract USBVERBOSE into a kernel module.  The module can be builtin
by defining 'options USBVERBOSE' in the kernel config file (no change
from current behavior), or it can be loaded at boot time on those
architectures that support the boot loader's "load" command.

The module is built for all architectures, whether or not USB support
exists.

(pgoyette)

uvm_fault_{upper,lower}_done: move drop-swap outside the page-queues lock.
Assert for object lock being held (or ref count 0) in uao_set_swslot().

(rmind)

rmixl_spl.S:
- where possible, stop using CP0 STATUS to disable all interrupts,zero EIMR instead.  more efficient since less meddling with CP0.
assume STATUS[IE] is normally set.
- add rmixl_spl_init_cpu(), to initialize cp0 interrupt control for this cpu

rmixl_intr.c:
- rmixl_intr_init_cpu() calls rmixl_spl_init_cpu()
to set up CP0 interrupt controls for this cpu

(cliff)

When use N32 and we are on a MIPS64 chip, enable 64-bit instructions via
PX so that we can still use the TLB miss exception instead of the XTLB miss
exception path.

(matt)

Make sure that user stack starts 32KB below maximum so that accesses with
displacements will never cross the VM_MAXUSER_ADDRESS boundary.

(matt)

Added optimization reduction for gdb/dwarf2-frame.c for Amigas. We don't
see the problem on other m68k platforms, but that could be because other
platforms don't have m68060s (except Atari, perhaps).

(jklos)

Deal with the "we get a portably arbitrary set of headers on
different archs" problem.

(pooka)

Update to xkeyboard-config 1.9.

(ghen)

Improve the CPU scheduler for a host MP system with multithreaded
access.  The old scheduler had a global freelist which caused a
cache crisis with multiple host threads trying to schedule a virtual
CPU simultaneously.

The rump scheduler is different from a normal thread scheduler, so
it has different requirements.  First, we schedule a CPU for a
thread (which we get from the host scheduler) instead of scheduling
a thread onto a CPU.  Second, scheduling points are at every
entry/exit to/from the rump kernel, including (but not limited to)
syscall entry points and hypercalls.  This means scheduling happens
a lot more frequently than in a normal kernel.

For every lwp, cache the previously used CPU.  When scheduling,
attempt to reuse the same CPU.  If we get it, we can use it directly
without any memory barriers or expensive locks.  If the CPU is
taken, migrate.  Use a lock/wait only in the slowpath.  Be very
wary of walking the entire CPU array because that does not lead to
a happy cacher.

The migration algorithm could probably benefit from improved
heuristics and tuning.  Even as such, with the new scheduler an
application which has two threads making rlimit syscalls in a tight
loop experiences almost 400% speedup.  The exact speedup is difficult
to pinpoint, though, since the old scheduler caused very jittery
results due to cache contention.  Also, the rump version is now
70% faster than the counterpart which calls the host kernel.

(pooka)

QNAP support (untested): PCI fixup and corrected PCI address lines (+1).
Patch submitted by Toru Nishimura.

(phx)

Redo previous.

(uebayasi)

pmap_enter:
This pmap treats all pmap_enter()'ed entries as "managed", and always
allocate PV entries.  Put back the original behavior.  There is no need
to check the new PMAP_UNMANAGED flag.

(uebayasi)

Comment.

(uebayasi)

Re-add compiled rules files (removed from distribution).

(ghen)

Bump date for previous.
Remove trailing space.

(wiz)

Install pkgconfig file for xkeyboard-config.

(ghen)

Update to xkeyboard-config 1.4.

(ghen)

+libgeode_drv.so.2.debug for i386.

(njoly)

un-obsolete ws.0 for i386.

(njoly)

+libopenchrome_drv.so.0.debug for amd64 and i386.

(njoly)

Resolve conflicts.

(ghen)

Initial import of xkeyboard-config-1.9

Status:

Vendor Tag: xorg
Release Tags: xkeyboard-config-1-9

(ghen)

Restore original (1.4 distfile) order to ease merges.

(ghen)

Remove the "xip" option from mount_ffs(8) for simplicity.

(uebayasi)

Reduce changes against HEAD.

(uebayasi)

Fix a typo; pointed out by Chuck Silvers.

(uebayasi)

Fix build.

(uebayasi)

Change getnum() to use intmax_t instead of int, so it doesn't overflow
between 2^31 and 2^32. Adjust call sites accordingly. PR 43354.

(dholland)

specdev::v_phys_addr is now specdev::v_physseg.

(uebayasi)

Remove an old XXX comment.  Fix a conditional build.

(uebayasi)

build libchromeXvMC on x86.

(mrg)

add / retire some xf86-video drivers.

(mrg)

we're now at mesa 7.7.1.

XXX: really really really need to make this in one place...

(mrg)

fix MKDEBUG=yes for X11R7 libs

(lukem)

geode_drv is a .so.2.

(mrg)

fix compile error in previous.

(mrg)

X11: Added support for xf86-video-geode and xf86-video-openchrome.

(mrg)

enable xf86-video-openchrome for x86 and xf86-video-geode for i386.

(mrg)

Remove opt_pci.h - it's not needed for anything.

(pgoyette)

obsolete xf86-video-cyrix sources.

(mrg)

obsolete xf86-video-cyrix

(mrg)

add build glue for xf86-video-geode and xf86-video-openchrome and
obsolete xf86-video-via.

(mrg)

obsolete these old sources.

(mrg)

geode wants a config.h.  give it one.

(mrg)

Fix symbol name in multiple-inclusion protection. (missing trailing '_')

(pgoyette)

initial import of xf86-video-geode-2.11.8

Status:

Vendor Tag: xorg
Release Tags: xf86-video-geode-2-11-8

(mrg)

.. except we need to check the flag while still clinging to iLock.

(pooka)

Assert that a VI_CLEAN vnode is not succesfully vget'd.

(pooka)

Mark files removed in the in-memory structure.  This allows us
to do two things:

1) properly set "recycle?" in inactive
2) easily check if we are renaming a removed vnode.  without the
    check, it was possible to enter a dirent in the file system for
    a removed (and hence scheduled to be vcleaned) vnode.  this would
    lead to the succesful vget() of a clean vnode.  the use of the
    cleaned vnode was, however, less succesful, except for purposes
    of crashing.

(pooka)

Remove a left-over debugging call to aprint_normal()

Thanks njoly@

(pgoyette)

help the next person tripping over PR kern/43203 :
list the error messages, the reason for the error and the location where
the firmware ought to be

(spz)

Added some more precise LinkStation/TeraStation and Synology product names.

(phx)

Disable EPIC_DEBUGIRQ. Reenable when needed.

(phx)

Fix build.

(uebayasi)

Release pathname buffers after use.

(pooka)

Rerun valid_format() both before and after unescaping the format
string, in case the format string contains printf conversions
assembled from escape sequences.

A better approach might be to adjust the unescape logic to avoid
generating printf conversions (e.g. convert \x25 to %% instead of %)
but that's somewhat problematic and it's not really worth taking the
trouble.

Running valid_format() only after unescaping would also be somewhat
tidier but makes printing the invalid format string problematic,
because the unescape logic runs in place.

None of these cases are really worth worrying about in detail, but now
at least they don't result in SIGSEGV.

Related to PR 43355.

(dholland)

Fix valid_format() to be more careful about allowing only valid printf
formats.

Also, accept %a and %A, which are new since this logic was last updated,
and also allow %F even though it's not functionally different from %f.
Document these additions and bump date of man page.

Fixes PR 43355.

(dholland)

initial import of xf86-video-openchrome-0.2.904

Status:

Vendor Tag: xorg
Release Tags: xf86-video-openchrome-0-2-904

(mrg)

add r600_blit.c and radeon_cs.c.

(mrg)

Introduce .WAIT into the cleandir rule for blasting the libsa/libkern/etc.
trees, because it can race with the libsa/libkern/etc. makefiles' own
cleandir rules. I think I've found all of the uses of the offending idiom...
Closes PR 43360.

(dholland)

Extract pci_verbose interface definitions into their own header file so
we don't need to include pcivar.h in the module.  pcivar.h tries to pull
in machine/pci_machdep.h which doesn't exist on all machine/architecture
combos.  Keeping track and building the module only for those that work
would have been a maintenance headache;  this change allows us to build
the module on all systems, regardless of whether the system has pci
support or not.

(pgoyette)

Don't recognize "+3" after -- or after the first non-option argument.
This prevents converting "+3" into "-k4.1" in places where getopt
won't recognize it, which in turn prevents silly error messages and
lossage trying to sort files whose names begin with +. PR 43358.

(dholland)

Minor grammar improvement in status message.

(pgoyette)

VM_PHYSSEG_NOADD is no more.

(uebayasi)

Include uvm/uvm.h, because this touches uvm internal.

(uebayasi)

new Mesa r300_dri driver has a "compiler" subdir of objects needed.
mostly dieter roelants in PR#43367.

(mrg)

Feed dust to a few linkset uses and explicitly call the constructor.

(pooka)

Change sc_wrports from an int to a bool and "test truth" instead of
comparing with 0.

Add 'volatile' to several other state variables that need it.

(dyoung)

regen: rump_vfs_mount_print

(pooka)

Add public namespace helper routine for dumping info on mountpoints.

(pooka)

print vm object refcount

(pooka)

fix refcounting

(pooka)

fix empty suffix case.

(christos)

Initialize *vpp to NULL: relookup() requires it without initializing
the value before the call (yea, changing relookup would probably
be smart, but other file systems already initialize vpp, so I'm
letting someone else experiment with tylenol od).

(pooka)

Make sure SYN_CACHE_TIMER_ARM() has been run before calling syn_cache_put()
as it will reschedule the timer.  Fixes PR kern/43318.

(bouyer)

Make sure tty00 and tty01 are linked to com(4).
The default for the unified PPC devices was to set a hard link to ttyZ*,
which uses the zstty(4) driver (not used on any Sandpoint board so far).

(phx)

PR/43364: Nicolas Joly: mkdep(1) wrong target names when suffix list miss .o

(christos)

Small typo.

(njoly)

Regen.

(jruoho)

Split the IrDA devices to their own section. Add HWPC224 and IBM0071,
both for the PC8738x-family of SuperIO chips from NSC (or Winbond).

(jruoho)

Do not try to unload the pciverbose module if it had never been loaded.

(martin)

pmap_map_ptes: handle emap on TLB flush.

(rmind)

Split x86 TLB shootdown code into a separate file.
Code part is under TNF license, as per pmap.c 1.105.2.4 revision.

(rmind)

bsd.subdir.mk - don't mention "make subdir" where subdir is any
directory listed in ${SUBDIR} as ${dir}: all-${dir} alias was removed
from 11 years ago (revision 1.31)

PR bin/43356

(uwe)

Fix typo in comment.

(uwe)

Added keyboard config files for Sharp WS011SH/WS020SH.

(nonaka)

Added keyboard config files for Sharp WS011SH/WS020SH.

(nonaka)

Add a comment describing an observed boom-crash-burn problem in
the code.  Fixing it will require a full tank of gas, half a pack
of cigarettes, sunglasses, darkness, and most importantly:
someone else.

(pooka)

Don't double unlock fvp if source file disappears during rename.
Problem found by njoly's awesome stresstester.

(pooka)

build fix: revert rev. 1.51

(cegger)

Revert earlier changes to callers of pci_find{vendor,product} since
the original routine entrypoint name has been restored.

Requested by dyoung@ and mrg@

(pgoyette)

Update libpci for change in pci_find{vendor,product} entrypoint names.

(pgoyette)

Rework the pciverbose module dispatch vectors to avoid renaming the
externally-visible entrypoint name.  Also this avoids a potential
need to bump kernel version.

Requested by dyoung@ and mrg@

(pgoyette)

For some reason the 6502 clock detection routine takes 1553 iterations
of the delay loop to finish on a normally clocked A2232.

(jklos)

Fix the names of 82577L[MC] LAN controllers (for mobile).
Fix typo.

(msaitoh)

WARNS=5 now

(agc)

Simplify and shorten the internals of packet processing by getting rid of
the intermediate pseudo-abstraction layer, which detracted from understanding
and had no benefit whatsoever. Rename some enums and some definitions.

(agc)

Don't missinterpret targets that start with .info

(sjg)

Adapt libpci build to include new pci_verbose.c

(pgoyette)

Update all callers of the pci_find{vendor,product} routines to now call
these routines through their global pointers.

(pgoyette)

Extract the vendor/product tables and related access routines into a
separate kernel module.  Update pci bus attach routine to load the
module (if available) when we're about to start scanning the bus, and
unload the module after the scan is finished.

On architectures which support loading of modules by the boot loader,
the 'pciverbose' module can be loaded and executed without needing to
rebuild the kernel.  On all architectures, using 'options PCIVERBOSE'
in the kernel configuration file will create a 'builtin' module which
is functionally equivalent to previous behavior.

XXX Although not nearly as large as the vendor and product tables,
XXX the PCI class and subclass tables might also be offloaded into
XXX the module at a future time.

XXX Cardbus (and possibly other) drivers should also be modified to
XXX load the module before scanning/attaching devices.

(pgoyette)

Note xsrc Xorg package updates, and radeondrm(4) updates.

(mrg)

Move definition of module_load_vfs_vec from kern_module_vfs.c to
kern_module.c - keeps rump happy.

Thanks, pooka!

(pgoyette)

Clean up previous, thanks to pooka@ for detailed review and suggestions.

(pgoyette)

fix MKDEBUG=yes for X11R7 libs

(lukem)

Make C declarations match assembler reality: ipi functions get two arguments
(which most of them (partly) ignore, but anyway).

(martin)

use round_page() for PAGE_ALIGN().  from Onno van der Linden
<o.vd.linden@quicknet.nl> in PR#38700, though apparently i didn't
commit it way back when i closed that bug.  oops?

(mrg)

Replace __appmansuffix__ in fsconfig.cpp

(veego)

Protect against attempting to load modules from the filesystem until we
have mounted the root file-system.  This allows us to load built-in and
boot-loader-provided modules much earlier during startup.

(pgoyette)

this needs <machine/psl.h> for PSTATE_AM on 32 bit kernels.

(mrg)

add r600_blit.c and radeon_cs.c.

(mrg)

implement DRM_RADEON_CS ioctl, necessary for r600_dri to function.

this is mostly from freebsd (though it also exists in other versions
of the drm code) svn commits:

--
SVN rev 196470 on 2009-08-23 14:55:57Z by rnoland

Add kernel support for Radeon R6/7xx 3D.

You will still need Mesa from git and possibly an updated DDX driver,
but this is working fairly well now.
--
SVN rev 196142 on 2009-08-12 12:57:02Z by rnoland

Add support for radeon RS880 IGP chips to drm.
--
SVN rev 195501 on 2009-07-09 16:39:28Z by rnoland

Add support for Radeon HD 4770 (RV740) chips.
--
SVN rev 196471 on 2009-08-23 15:02:58Z by rnoland

Add GET_PARAM support for Z pipes.

This is needed for occulsion queries on rv530 chips.
--
SVN rev 198695 on 2009-10-30 18:07:22Z by rnoland

A bit of cleanup work on radeon_freelist_get()

  * Fix the main loop to search all buffers before sleeping.
  * Remove dead code
--
SVN rev 197606 on 2009-09-28 22:41:28Z by rnoland

Fix offset handling
--
SVN rev 197605 on 2009-09-28 22:40:29Z by rnoland

radeon_family is an enum, so ordering can be important.

sync up with what amd is shipping.
--
SVN rev 197603 on 2009-09-28 22:37:07Z by rnoland

R600 doesn't support IRQs yet, so don't try to use them.
--

special thanks to robert noland @ freebsd for making this an
easy port!

(mrg)

Explicitly quote | and : meant as separate operator. Kill trailing
whitespace. Don't use \Z' with obscure character entity. Reorder sections
to canonical order.

(joerg)

PRIxPADDR instead of %lx to print bus_addr_t/paddr_t variables.

(jym)

Cast the sum of iobase and iosize to long, not just iobase. Avoids size
warnings during build when bus_addr_t is 64 bits.

(jym)

these depend upon libdrm_radeon now.

(mrg)

re-add the final new-line.

(mrg)

simplify the debugging code and make it more informative.

(christos)

no need for double the newlines!

(christos)

- centralize the snprintf code.
- use err where appropriate.
- add machclass which should be x86 when i386 and amd64 and can be specified
  in man.conf as:
_i386 x86
_amd64 x86
  so that we can support merged pages. Nothing uses this yet.

(christos)

make these build with Mesa 7.7.1.

(mrg)

Replace UTMP_FILE and WTMP_FILE for Xstartup here and not in the cpp file
to make a xdm update easier.
xdm was updated but the local change in Xstartup.cpp was lost.

(veego)

updates for Mesa 7.7.1.

(mrg)

Use ${XDMPIDDIR} and ${XDMLOGDIR} from ../xdm/Makefile.xdm instead of
hard coding the path here.

(veego)

build fixes for Mesa 7.7.1.  introduce r600 driver!

(mrg)

merge MesaLib 7.7.1

(mrg)

merge MesaGLUT 7.7.1

(mrg)

merge MesaDemos 7.7.1

(mrg)

Remove a / between a ${DESTDIR}${X11USRLIBDIR} which would result in a
double /

(veego)

initial import of MesaLib-7.7.1

Status:

Vendor Tag: xorg
Release Tags: MesaLib-7-7-1

(mrg)

initial import of MesaGLUT-7.7.1

Status:

Vendor Tag: xorg
Release Tags: MesaGLUT-7-7-1

(mrg)