Check whether the fxp(4) actually attached before calling fxp_stop().

(dyoung)

Add a self-suspensor.  It's not used, yet.

In ex_detach(), call ex_stop(), first.  ex_stop() stops the callout.
Destroy the callout in ex_detach().

(dyoung)

Ensure cd9660_read() gives accurate EOF indication.
Should fix PR/42983.

(jakllsch)

Add a DIOCTUR (test unit ready) ioctl, to allow userland to easily
(and silently) check for presence of media. While this would be easy
to implement via SCIOCCOMMAND, our kernel has quirks for this
command, which otherwise would not be readily available to userland
applications.

(martin)

Ignore ioapic write failures.
They happen when writing to ioapic pins Xen already disabled.
This makes netbsd dom0 booting.

(cegger)

Add a detach routine.  Enable detachment at shutdown.

psc_pwrmgmt_csr_reg was never initialized to the actual offset of the
PCI PMCSR!  Just delete psc_pwrmgmt_csr_reg and all uses of it.  In
this way, fxp_pci_disable() becomes a no-op, so delete it, too.

(dyoung)

In fxp_detach(), fxp_stop(), first.  fxp_stop() stops the callout.
Destroy the callout in fxp_detach().

(dyoung)

Fix the previous commit properly: we need to set all sensor-flags during the
initialization of sysmon_envsys(9), otherwise, if _STA does not return true
during the attachment of the driver, it is possible that a flag is never
set. This flaw has always been present, but it only got exposed after the
conversion to use the sysmon-limits.

(jruoho)

Include bthub support.

As mario knows, it should be in another component, but just cram
it here for now until someone solves kernel module dependencies
for us.

(pooka)

deal with ioconf.h

(pooka)

Some drivers want to include ioconf.h, so generate it if doing
"ioconf" config.

(pooka)

Break the (too) big lead paragraph into three smaller ones.

(jruoho)

Add a missing word introduced in the previous commit. Thanks, snj@.

(jruoho)

Add rumpdev_ubt and rumpnet_netbt components.

(pooka)

Add an example/test program for rump bluetooth.  This does discovery
and prints the address and name of the first peer encountered (if any):

ubt0 at uhub0 port 1
ubt0: Cambridge Silicon Radio Bluetooth USB Adapter, rev 2.00/19.15, addr 2
device ubt0, addr xx:xx:xx:xx:xx:xx
my peer: yy:yy:yy:yy:yy:yy, major class: 2 (phone)
peer name: Mokia maailmalta

(pooka)

USB bluetooth device rump component.

This is a curious driver in the sense that unlike all other current
device drivers, it does not require vfs.  This is because the driver
is controlled via bluetooth, which is controlled via PF_BLUETOOTH
sockets (as opposed to a /dev node).

(pooka)

bluetooth stack component for rump

(pooka)

* support isochronous transfers (for as much as they can be supported
  considering our ugen driver doesn't support isochronous write.
  kern/25960 seems to contain a patch and could be investigated for that)
* add a dirty rotten hack which makes interrupt transfers of >16
  work for me

(pooka)

fix build.sh -N 4 distribution failure:

+ echo 'MKSOLARIS=yes'
xargs: unterminated quote

(cegger)

Regen.

(jruoho)

Add IFX0102 and PNP0C15. Fix INT0800.

Observed from the dmesg of nojoly@'s Sony VAIO VGN-BZ12VN.

(jruoho)

Document the questionable ISSET/SET/CLR macros.

(jruoho)

Mark up NULL with Dv, as usual.

(wiz)

Note that some definitions, such as NULL, were already in v4 UNIX.

(jruoho)

Wrap DOCALL and DOCALL_KLOCK in a new block (for easy adding of
debug prints before them.

XXX: can't use the standard do-while0 because lint doesn't like it.

(pooka)

Set the capacity flags already in the acpibat_get_info(),
instead of acpibat_get_status().

(jruoho)

Small changes to wording and markup.

(jruoho)

- in cpu_setup_trampoline_callback, add some DIAGNOSTIC checks,
and make sure we get needed sign extension when storing 32 bit pointers
into trampoline args structure.

(cliff)

- in rmixlfw_wakeup_cpu, properly nuke the upper half of sp
before OR-ing in KSEG0_START.  This is only needed in _LP64 case.
- in rmixl_cpu_trampoline:
trampoline args addr needs reconstructing the upper half only in _LP64 case.
delete set of MIPS_SR_INT_IE bit.
use REG_L instead of PTR_L to load trampoline args
so we get 64 bit loads in both 32 and 64 bit kernels.

(cliff)

add + _SYMQUIRK='|\.(rump|RUMP)' for powerpc64, as recommended by pooka.
it gets rid of many of the warnings trying to link rump_cd9960, but
there are a bunch remaining.  atomic ops seem missing, and some ktrace
points.

(mrg)

for mips64*, use mips64 subdir, otherwise use $MACHINE_ARCH if exists,
or $MACHINE_CPU.  idea from dennis.c.ferguson@gmail.com in PR#43021.

(mrg)

rename text() to wrtext(), aligning it with most other output routines,
and avoiding a compile error on powerpc64.

XXX: this tries to make a ".text" symbol, which already exists...

(mrg)

fix some powerpc64 build issues.

(mrg)

in cpunode_rmixl_attach use the right macro to obtain number of cores
provided by this chip.

(cliff)

set MACHINE_CPU to powerpc for MACHINE_ARCH of powerpc64.

from pooka via PR#43021.

(mrg)

define _NOREGNAMES before including <machine/asm.h>, like powerpc does.

from dennis.c.ferguson@gmail.com in PR#43022.

(mrg)

add a nop to provide enough space for the linker and long calls.
fixes link errors with 'cerror' and libposix.

from dennis.c.ferguson@gmail.com in PR#43010.

(mrg)

default powerpc64 mbufs to 512 bytes, like other 64 bit platforms.
from dennis.c.ferguson@gmail.com in PR#43028.

(mrg)

add powerpc64 support.  from dennis.c.ferguson@gmail.com in PR#43029.

(mrg)

replace kernel-only 'stmd' and 'lmd' operations in userland with equivalent
sequences of code, as done in eg, powerpc64/gen/_setjmp.S rev 1.2.

from dennis.c.ferguson@gmail.com in PR#43009.

(mrg)

convert -idirafter to -isystem, as recommended by cgd in PR 11843.
most of the list in that PR were already done.

(mrg)

fix the dependancy list for this.

(mrg)

since pmap_activate/pmap_deactivate need to not block, but the rest of
pmap is OK, add a new IPL_VM pmap_ctx_lock that is held while allocating
or freeing a context, and revert pmap_lock back to IPL_NONE.

i considered a per-cpu context lock, but that would require locking
each active CPUs lock in turn in pmap_remove_all() and pmap_destroy().

thanks to ad for explaining how these parts interact.

(mrg)

declaration of rmixl_cpu_trampoline_args now #ifdef MULTIPROCESSOR

(cliff)

moved a number of RMI interrupt items to rmixl_intr.h

(cliff)

- add defines for Coprocessor 2 (FMN) registers
- add defines for non-CPU-core FMN bucket size and credit counter regs
arch/mips/rmi/rmixlvar.h

(cliff)

- obtain interrupt routing mask from obio_attach_args,
pass along to rmixl_intr_establish

(cliff)

- add rmixlfw_wakeup_cpu, performs callback to RMI firmware wakeup function
- add rmixl_cpu_trampoline, entry point for CPU wakeup following
RMI firmware wakeup callback.

(cliff)

- include rmixl_intr.h
- obtain interrupt routing mask from obio_attach_args,
pass along to rmixl_intr_establish
- cut out some dead (#if 0) code
- in rmixl_pcie_intr_establish(), construct interrupt event name
in the interrupt dispatch structure, not on the stack.  also
improve the name.

(cliff)

- add sc_tmsk interrupt routing mask to the softc
- add count_name for naming the evcnt count in
each struct rmixl_pcie_link_dispatch

(cliff)

- add obio_tmsk interrupt routing mask to obio_attach_args

(cliff)

- add use of the 'tmsk' locator
- adapt to new rmixl_intr_establish() args
- establishing the system bridge controller
address error interrupt is now explicitly XLS specific

(cliff)

- rework to make full use of RMI extended interrupt management
provided by EIRR/EIMR registers
- depends on rmixl_spl.S
- add support for IRT based interrupt routing; for now we are still routing
all IRT interrupts to CPU#0.
- note that count/compare clock, IPI and FMN are handled by each CPU since
these are local interrupt sources.
- further changes are still needed for XLR and XLP support

(cliff)

- include rmixl_intr.h for externs

(cliff)

- except for core#0, pmap tlb0info lives in the cpucore softc;
cpucore_rmixl_attach() calls pmap_tlb_info_init() with a KSEG0 or
XKPHYS_CACHED conversion of the address of that structure in the softc

(cliff)

- use different wakeup trampolines depending on firmware type
so far only RMI formware with callback wakeup method is tested.
- add a (chip-specific) softc structure, and cross link it
with struct cpu_info
- remove exception vectors mem allocation, it is now provided
by cpu_attach_common
- remove common cpu_info initialization, now provided
by cpu_attach_common
- use per-CPU bit in 'cpus_hatched' as handshake with the CPU
we just hatched
- call rmixl_intr_init_cpu for per-CPU interrupt initialization
- call cpu_attach_common() for common mips stuff
- cpu_rmixl_attach_once() groups together chip-specific initializations
that are only to be one once, i.e. by CPU#0
- cpu_rmixl_hatch() provides chip-specific hatch code called from
(mips-common) cpu_hatch() via lsw_cpu_init
- cpu_setup_trampoline_callback() provides method to wake up
subordinate CPUs using RMI firmware callback.
- rmixl_cpuinfo_print() can be called for debugging e.g. from DDB
should maybe evolve into DDB 'mach cpu <n>' ?

(cliff)

Document the weird <iso646.h>.

(jruoho)

- pass obio_tmsk to rmixl_intr_establish

(cliff)

- specify rmixl_splsw to mips_vector_init
  note that __INTR_PRIVATE must be defined to do that
- print MEMLIMIT if it is defined, after printing memsize
- improve some comments

(cliff)

- add files rmixl_spl.S, rmixl_fmn.c
- add 'tmsk' locator to obio, allows specifying a thread (vCPU) mask
for interrupt routing; default is -1 which means route to any

(cliff)

add include for RMI specific cpu stuff: softc and wakeup trampoline args structures

(cliff)

add splswitch variant using RMI chip-specific EIRR/EIMR interrupt extensions

(cliff)

Reorganizing Chelsio 10 gig files into separate directory.

(jklos)

Reorganizing all Chelsio 10 gig files into separate directory.

(jklos)

add device driver for RMI Fast Message Network

(cliff)

Bump date for previous.

(jruoho)

Note inttypes(3) also here.

(jruoho)

Update version to 20100320 for Christos' commit to popen.c 1.37 for:
PR/43023: Bruce Cran: FTPD bug remote crash

(lukem)

send the SET_SCALE11 command before the 'sliced' command as per the
Linux driver (and elantech(4)), via PR kern/42853

(plunky)

Extended capability probes can confuse the passthrough device,
reset the touchpad on enable to cure that.

(plunky)

Mention the recommended usage of inttypes(3) when applicable.

(jruoho)

some definitions from the Linux driver (via PR kern/42853)

(plunky)

Add inttypes(3).

(jruoho)

Document <inttypes.h>.

(jruoho)

- added this file of RMI interrupt stuff

(cliff)

- in _splsw_splintr, add 'nop' in delay slot to avoid incorrect return
of IPL_VM when nothing is pending

(cliff)

- mips_int5_evcnt, mips_int5_missed_evcnt and next_cp0_clk_intr
were moved to struct cpu_info to allow per-CPU MIPS count/compare
clock programming

(cliff)

- in cpu_info_alloc, move calling pmap_tlb_info_attach() to
after called mi_cpu_attach(), so we dont trigger assert on
(ci->ci_data.cpu_idlelwp != NULL) in pmap_tlb_info_attach()

(cliff)

- if __INTR_PRIVATE is not defined, declare (but do not define) struct splsw

(cliff)

- define MIPS_SR_COP_2_BIT to control enable/disable of coprocessor 2

(cliff)

- add ci_next_cp0_clk_intr, ci_count_compare_evcnt, ci_count_compare_missed_evcnt
to struct cpu_info, for per-CPU count/compare clock programming
- add ci_request_ipis to struct cpu_info to allow passing IPI tags
on systems where hardware does not provide such feature; use atomic ops for this.
- declaration of mips_vector_init was moved from here to mips/include/locore.h

(cliff)

mips_vector_init now takes an argument to specify splsw.
NULL specifies use the default 'std_splsw'

(cliff)

Remove superfluous Pp.

(wiz)

sync with OpenBSD:r1.3
> bus_dmamap_sync fixes to cope with real life.

(nonaka)

revert for now.

(christos)

revert previous for now.

(christos)

Emphasize imaxint_t.

(jruoho)

Use \e

(joerg)

more fix.

(nonaka)

don't create obsoleted directories

(pooka)

fix wrong op priority.

(nonaka)

lock vnode when calling VOP_GETATTR.

(yamt)

fix merge botches

(yamt)

yamt-nfs-mp: update base tag

(yamt)

Note that some of these were already in ANSI C.

XXX: Maybe someone with insight adds a word or two about the history?

(jruoho)

Reference stddef(3).

(jruoho)

Add stddef(3).

(jruoho)

While at it, add also stddef(3).

(jruoho)

gdb-7.1 out.

(wiz)

Add stdbool(3).

(jruoho)

Reference stdbool(3).

(jruoho)

As we have added stdint(3) and tgmath(3), we may as well add stdbool(3).

(jruoho)

popen.c security fix

(lukem)

Apply rev 1.37 from NetBSD:
PR/43023: Bruce Cran: FTPD bug remote crash
Since we specify NOCHECK, in the NOMATCH case gl_pathv can be NULL.
(From FreeBSD)

(lukem)

examples are installed. not bothering with separate email address

(lukem)

Convert acpibat to use ioconf instead of homerolled config glue.

Note, config whines about things like:
files.acpi:30: attaching undefined device `apm'

I intentionally was a lazy boy and didn't include a ton of dirt to
keep the hot napalm and riot blasts in the config thread rolling ;)

tested by jruoho (thanks!)

(pooka)

include bsd.ioconf.mk when necessary

(pooka)

Move make snippets for use with config(1) ioconf functionality from
rump into share/mk.  This is to make it useful for all kernel
builders.

Note: we have waaay too many weird and wonderful ways of making
kernel code (monolithic kernel, modules, rump).  There should be
only one way to build kernel code instead of a maze of twisty little
.mk files, all not quite alike.  When that is fixed, this snippet can go
into the more generic .mk file.

(pooka)

Ticket 1344.

(snj)

Pull up following revision(s) (requested by joerg in ticket #1344):
sys/sys/null.h: revision 1.8
Enclose (void *)0 in an extra set of parenthese to make the result usable
in arbitrary expressions.
Fixes PR/41890, I can't think of any downsides!

(snj)

in nfs_bootparam(), set the corresponding flag for each field that we fill in.

(chs)

assert that pcb_onfault is NULL in places where it should be.

(chs)

remove unnecessary cast now that lint is more intelligent.

(christos)

right shift of unsigned quantities with constant.

(christos)

make lint recognize constant shift values and not warn if the number of bits
shifted results in a narrower type.

(christos)

fix copy{in,out}{,str}() to return the error returned by uvm_fault().
fixes PR 41813.

(chs)

Drop redundant .TP

(joerg)

convert pmap_lock into an IPL_VM lock, and avoid calling
KASSERT(!mutex_owned(&pmap_lock)) since that's no longer valid.

this fixes hangs on 4-core machines like the U80.

tested on U80, SB2500 and SB2000.

XXX: my initial attempts to replace pseg_lock with this new equivalent
XXX: lock hang in pmap_bootstrap it seems..  but this should be able
XXX: to go away now.

(mrg)

Quote lines that start with . and should appear literally.

(joerg)

Drop redundant TPs

(joerg)

Use correct format directive in aprint_debug_dev()

Thanks to Kurt Schreiner for reporting this on current-users@

(pgoyette)

PR/43023: Bruce Cran: FTPD bug remote crash
Since we specify NOCHECK, in the NOMATCH case gl_pathv can be NULL.
(From FreeBSD)

(christos)

fix byte order mark related bug introduced by previous commit,
reported by Sverre Froyen via current-user, thanks!

(tnozaki)

Remove quotes around Nd argument, they are not needed.

(wiz)

Add stdint(3).

(jruoho)

Only apply -Wno-stack-protector where alloca is used.

(skrll)

Add a manual page for <stdint.h>. From FreeBSD, by Mike Barcroft.

(jruoho)

Fix build failed.
  Add forgot zmiireg.h since support 405EX.

(kiyohara)

Do not use .pp with %P, as noted by joerg@.

(jruoho)

/*LINTED*/ is too wide a brush.

(christos)

Shut up the broken piece of junk called lint. Use __inline.

(joerg)

Bump date for previous.

(wiz)

New sentence, new line.

(wiz)

Sort SEE ALSO. New sentence, new line.

(wiz)

Serial comma; prepare for HTML-output.

(wiz)

Add the (ACM) journal where this appeared. Fix URL.

(jruoho)

Note also here that this is a legacy daemon that is more or less deprecated.

(jruoho)

Add a note that APM was deprecated by the PC manufacturers in about 1996.
Mention the APM emulation layer of ACPI. In addition, some small fixes.

(jruoho)

don't pass "maxaddr" as a boundary in bus_dmamap_create() like rev 1.8
did for _alloc().

(mrg)

Fix the dbcool_{get,set}_limits() routines to properly deal with min/max
values on temp limits.  The actual ranges for temp limits varies if the
chip is running in "extended temperature" mode, so we need to handle this
difference, too.

Verified on my home machine which runs in extended-temp mode and by
njoly@ on a machine which does runs in the older 2s-complement mode.

(pgoyette)

Move openssl to the front of SUBDIR list and add .WAIT after it since
both netpgp and openssh depend on it.  This should fix intermittent
failures of parallel builds (as seen e.g. on the autobuild cluster).

(uwe)

more debugging compilation fixes.

(christos)

fix debugging code.

(christos)

minimize ifdefs and avoid duplicated code.

(christos)

Add PCI_BAR(n) for the nth Base Address Register.

(dyoung)

Add a prototype for pci_mmio_range_infer() that will infer the
range of memory forwarded by the host chipset to PCI.

(dyoung)

Cosmetic: abbreviate: use `pc' instead of `pci_chipset_tag'.

(dyoung)

When booting the secondary processors, try to make the cpu_counter is
close to cpu0's counter.

(matt)

- Make maximum memory limits for various things #define constants and use the
  consistently across the code.
- Re-do note parsing code to read the section headers instead of the program
  headers because the new binutils merge all the note sections in one program
  header. This fixes all the pax note parsing which has been broken for all
  binaries built with the new binutils.
- Add diagnostics to the note parsing code to detect malformed binaries.
- Allocate and free note scratch space only once, not once per note.

(christos)

Use the gcc directive to disable SSP warnings.

(roy)

Fix set for tparm.html.

(joerg)

Add man page for fast_divide32.

(joerg)

Bump date for the previous commit.

(jruoho)

Some cosmetic modifications, including removal of white space.

(jruoho)

Document some missing options. Fixes my own PR # 41913.

(jruoho)

Apply the equivalent of r426 and r427 from openpam repository:
Fix doc: pam_strerror() ignores pamh.
Portability tip.

(joerg)

Apply r432 from openpam repository:
pam_end() already contains a NULL check, and it is not unreasonable to
call it with a NULL pamh in a cleanup / error-handling situation. Remove
OPENPAM_NONNULL, which may cause gcc to optimize away the NULL check.

(joerg)

Namespace protection for mi_vector_hash.

(joerg)

First draft of a manual page for Matt Thomas' red-black trees.  Please
review for correctness.

(dyoung)

Add functions for replacing runtime invariant 32bit unsigned divisions
with simpler full width multiplications and shifts + adds. The main
operations are generally at least 50% faster when serialised and often
better for parallelism as well.

(joerg)

Avoid known unknown if the kernel presents us with an unknown
unknown for module class and/or source.

related to PR kern/43014

(pooka)

Replace some magic numbers with macro. Tested on SMC_TT on atari.

(tsutsui)

Misc KNF.

(tsutsui)

No need to override mem_ring value in ne2000_ipkdb_attach()
since a wrong calculation in dp8390_ipkdb_attach() was fixed.

(tsutsui)

ANSI'fy, misc KNF.

(tsutsui)

(re-)run gcc mknative for powerpc64.

Fixes PR/43008.

(skrll)

u_intN_t -> uintN_t

(tsutsui)

bus_space_write_region_N() should also take a const buffer.

(tsutsui)

Remove trailing whitespace. Use Nm instead of Xr to itself.

(wiz)

Constify the buffer pointer argument of the _bus_space_write_multi_<n>
inline functions, so that this builds again.  (Otherwise, ne2000.c driver
would now fail to build.)

(he)

Fix PMAPDEBUG build for previous.

(skrll)

Not all PA CPUs have the U-bit (uncacheable) for non-IO memory.  In fact
most don't.  Deal with non-equivalent aliases by removing and flushing the
managed mappings, and flushing the unmanaged mappings.

When flushing caches/TLB flush the cache before purging the TLB just in
case the flush enters the mapping into the TLB.

(skrll)

document aiboost(4) removal and replacement with aibs(4) in doc/CHANGES

(cnst)

Remove aiboost(4) as obsolete and redundant since the introduction of aibs(4).

      http://mail-index.netbsd.org/tech-kern/2010/03/06/msg007458.html

Reviewed by <pgoyette>, <cegger>, <jruoho>, <tech-kern>.

(cnst)

Document the new sysmon_envsys_sensor_event() function.

(pgoyette)

Modify previous commit.  The routine to deliver events is intended to be
exported to device drivers, so rename it and move the declaration to
<dev/sysmon/sysmonvar.h>

(pgoyette)

support W83667HG (0xa5) in wbsio(4); its hw mon is already supported by lm(4) as W83627DHG (0xc1)

(cnst)

Remove superfluous cast (device_t to device_t).

(dyoung)

Use Cardbus_intr_disestablish() instead of cardbus_intr_disestablish().

Get rid of a (short) staircase in ex_cardbus_detach().

(dyoung)