]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoRename the local variable 'person' to 'lpr_username', to avoid compile-time
gad [Tue, 9 Oct 2001 00:27:14 +0000 (00:27 +0000)]
Rename the local variable 'person' to 'lpr_username', to avoid compile-time
warnings about conflict with a global variable used by 'lprm'.

MFC after: 4 days

22 years agoNote that this driver is soon to be deprecated and removed from FreeBSD.
mjacob [Tue, 9 Oct 2001 00:14:41 +0000 (00:14 +0000)]
Note that this driver is soon to be deprecated and removed from FreeBSD.

22 years agoGet rid of some minor compile-time errors by changing copyright/rcsid
gad [Tue, 9 Oct 2001 00:09:46 +0000 (00:09 +0000)]
Get rid of some minor compile-time errors by changing copyright/rcsid
definitions to the same format used in other lpr source files, and by
adding parenthesis to the right spot in one 'if' statement.

MFC after: 4 days

22 years agoFix minor compile-time error in definition of 'copyright'.
gad [Tue, 9 Oct 2001 00:06:52 +0000 (00:06 +0000)]
Fix minor compile-time error in definition of 'copyright'.

MFC after: 4 days

22 years agoFix minor compile-time warning that snunk in with changes to sortq() rtn.
gad [Tue, 9 Oct 2001 00:04:59 +0000 (00:04 +0000)]
Fix minor compile-time warning that snunk in with changes to sortq() rtn.

MFC after: 4 days

22 years agoDon't dereference td if it's NULL.
julian [Mon, 8 Oct 2001 23:47:44 +0000 (23:47 +0000)]
Don't dereference td if it's NULL.

Submitted by: Alexander N. Kabaev <ak03@gte.com>

22 years agoDon't attempt to initialize the pcaudio device on open. The attach routine
jhb [Mon, 8 Oct 2001 23:35:40 +0000 (23:35 +0000)]
Don't attempt to initialize the pcaudio device on open.  The attach routine
already does the initialization (though it didn't set pca_initialized, so
we always initialized twice) and since attach calls make_dev(), there's no
way that pcaopen() can be called before pcaattach().

22 years agoUpdate for FILE v3.37.
obrien [Mon, 8 Oct 2001 22:57:03 +0000 (22:57 +0000)]
Update for FILE v3.37.
FreeBSD no longer needs the daylight hack.
(we really should sync our timezone code up ...)

22 years agoThis commit was generated by cvs2svn to compensate for changes in r84685,
obrien [Mon, 8 Oct 2001 22:50:54 +0000 (22:50 +0000)]
This commit was generated by cvs2svn to compensate for changes in r84685,
which included commits to RCS files with non-trunk default branches.

22 years agoVirgin import of Christos Zoulas's FILE 3.37.
obrien [Mon, 8 Oct 2001 22:50:54 +0000 (22:50 +0000)]
Virgin import of Christos Zoulas's FILE 3.37.

22 years agoDo not clobber users hostname.mc file if freebsd.mc changes (likely after an
gshapiro [Mon, 8 Oct 2001 22:35:24 +0000 (22:35 +0000)]
Do not clobber users hostname.mc file if freebsd.mc changes (likely after an
installworld).

Submitted by: Steve Watt <steve@Watt.COM>
MFC after: 2 days

22 years agoUpdate relevant release note item with SA-01:62.
bmah [Mon, 8 Oct 2001 21:33:56 +0000 (21:33 +0000)]
Update relevant release note item with SA-01:62.

22 years agoReplace 'curproc' with 'td->td_proc'.
jhb [Mon, 8 Oct 2001 21:05:46 +0000 (21:05 +0000)]
Replace 'curproc' with 'td->td_proc'.

22 years agoAllow atomic ops to be somewhat safely used in userland. We always use
jhb [Mon, 8 Oct 2001 20:58:24 +0000 (20:58 +0000)]
Allow atomic ops to be somewhat safely used in userland.  We always use
lock prefixes in the userland case so that the binaries will work on both
SMP and UP systems.

22 years agoUpdate references to 'device' entries for the kernel config
wilko [Mon, 8 Oct 2001 20:33:46 +0000 (20:33 +0000)]
Update references to 'device' entries for the kernel config
files to use 5.x syntax (I hope..)

Note: I suppose one might need to hack the hints file as well?

Submitted by: Mike Barcroft <mike@FreeBSD.ORG>, Bruce Mah <bmah@FreeBSD.ORG>

22 years agoMake printtrap() more informative.
dfr [Mon, 8 Oct 2001 20:21:04 +0000 (20:21 +0000)]
Make printtrap() more informative.

22 years agoWS Cleanup
dillon [Mon, 8 Oct 2001 19:51:13 +0000 (19:51 +0000)]
WS Cleanup

22 years agoOOPS, delete some commented out debugging code committed along with r.1.25.
sobomax [Mon, 8 Oct 2001 17:09:08 +0000 (17:09 +0000)]
OOPS, delete some commented out debugging code committed along with r.1.25.

MFC after: 2 weeks

22 years ago(forced)
sobomax [Mon, 8 Oct 2001 17:05:34 +0000 (17:05 +0000)]
(forced)

Forgot to mention: creation of a full-fleged package archive requires some
changes in bsd.ports.mk, because currently it doesn't install some files
that are usually packed into the package archive (+MTREE_DIRS, +DISPLAY
and so on). Support for this will be committed later.

22 years ago- Allow package archive to be created from a locally installed package. This
sobomax [Mon, 8 Oct 2001 17:01:35 +0000 (17:01 +0000)]
- Allow package archive to be created from a locally installed package. This
  allows for an easy way to backup old version of port prior to installing
  a new one;
- silence compiler warnings by killing some unused variables and adding
  all includes necessary.

MFC after: 2 weeks

22 years agoCatch up with the SMPng reduced interrupt frame size. The corresponding
iedowse [Mon, 8 Oct 2001 12:46:01 +0000 (12:46 +0000)]
Catch up with the SMPng reduced interrupt frame size. The corresponding
change was made to DDB months ago (i386/i386/db_trace.c revision 1.37).

Reviewed by: bde

22 years agoAdd $FreeBSD$.
asmodai [Mon, 8 Oct 2001 12:28:31 +0000 (12:28 +0000)]
Add $FreeBSD$.
Describe the other two directories in here.

22 years agoPrint interface name for each ARP entry.
ru [Mon, 8 Oct 2001 11:52:23 +0000 (11:52 +0000)]
Print interface name for each ARP entry.

Prodded by: Igor Podlesny <poige@morning.ru>
MFC after: 1 week

22 years agoPage align the data segment.
mp [Mon, 8 Oct 2001 10:43:34 +0000 (10:43 +0000)]
Page align the data segment.

22 years agomdoc(7) police: fix markup.
ru [Mon, 8 Oct 2001 10:12:41 +0000 (10:12 +0000)]
mdoc(7) police: fix markup.

22 years agoPort to ia64. Actually, just do like the alpha.
dfr [Mon, 8 Oct 2001 08:29:52 +0000 (08:29 +0000)]
Port to ia64. Actually, just do like the alpha.

22 years agouse pcm_getbuffersize()
cg [Mon, 8 Oct 2001 06:10:39 +0000 (06:10 +0000)]
use pcm_getbuffersize()
only creat a mic recording channel if the codec supports it

22 years agouse pcm_getbuffersize()
cg [Mon, 8 Oct 2001 06:01:13 +0000 (06:01 +0000)]
use pcm_getbuffersize()
use seperate dma tags for buffers and control info

22 years agouse pcm_getbuffersize()
cg [Mon, 8 Oct 2001 05:59:54 +0000 (05:59 +0000)]
use pcm_getbuffersize()

22 years agoAdd man page for device.hint(5).
yokota [Mon, 8 Oct 2001 05:37:56 +0000 (05:37 +0000)]
Add man page for device.hint(5).

22 years agoRemove references to non-existant pvcsif(8) and pvctxctl(8).
murray [Mon, 8 Oct 2001 05:05:31 +0000 (05:05 +0000)]
Remove references to non-existant pvcsif(8) and pvctxctl(8).

PR: docs/30446
Submitted by: Giorgos Keramidas <charon@labs.gr>

22 years agoRemove reference to non-existant crd(4) man page.
murray [Mon, 8 Oct 2001 04:46:07 +0000 (04:46 +0000)]
Remove reference to non-existant crd(4) man page.

PR: docs/30445
Submitted by: Giorgos Keramidas <charon@labs.gr>

22 years agos/alpha/${MACHINE_ARCH}/g
marcel [Mon, 8 Oct 2001 01:41:45 +0000 (01:41 +0000)]
s/alpha/${MACHINE_ARCH}/g

22 years agoAdd a call to init_param() to initialize some necessary variables.
mp [Mon, 8 Oct 2001 00:44:21 +0000 (00:44 +0000)]
Add a call to init_param() to initialize some necessary variables.

22 years agoRemove panics for rename() race conditions. The panics are inappropriate
dillon [Mon, 8 Oct 2001 00:37:54 +0000 (00:37 +0000)]
Remove panics for rename() race conditions.  The panics are inappropriate
because the IN_RENAME flag only fixes a few of the huge number of race
conditions that can result in the source path becoming invalid even
prior to the VOP_RENAME() call.  The panics created a serious security
issue whereby an attacker could fairly easily cause the panic to
occur, crashing the machine.

The correct solution requires a great deal of work in the namei
path cache code.

MFC after: 0 days

22 years ago* register the correct number of channels
cg [Sun, 7 Oct 2001 23:31:52 +0000 (23:31 +0000)]
* register the correct number of channels
* include hw buffer size in sndstat info

22 years agouse mixer channel numbers, not channel masks- this should fix sb/sbpro mixer
cg [Sun, 7 Oct 2001 23:28:34 +0000 (23:28 +0000)]
use mixer channel numbers, not channel masks- this should fix sb/sbpro mixer
problems.

22 years agoForced commit to note that the previous entry referred to p_candebug().
des [Sun, 7 Oct 2001 22:52:07 +0000 (22:52 +0000)]
Forced commit to note that the previous entry referred to p_candebug().

22 years agoImplement inline versions of ntohl etc.
dfr [Sun, 7 Oct 2001 21:09:35 +0000 (21:09 +0000)]
Implement inline versions of ntohl etc.

22 years agoDissociate ptrace from procfs.
des [Sun, 7 Oct 2001 20:08:42 +0000 (20:08 +0000)]
Dissociate ptrace from procfs.

Until now, the ptrace syscall was implemented as a wrapper that called
various functions in procfs depending on which ptrace operation was
requested.  Most of these functions were themselves wrappers around
procfs_{read,write}_{,db,fp}regs(), with only some extra error checks,
which weren't necessary in the ptrace case anyway.

This commit moves procfs_rwmem() from procfs_mem.c into sys_process.c
(renaming it to proc_rwmem() in the process), and implements ptrace()
directly in terms of procfs_{read,write}_{,db,fp}regs() instead of
having it fake up a struct uio and then call procfs_do{,db,fp}regs().

It also moves the prototypes for procfs_{read,write}_{,db,fp}regs()
and proc_rwmem() from proc.h to ptrace.h, and marks all procfs files
except procfs_machdep.c as "optional procfs" instead of "standard".

22 years agoAlways succeed if the target process is the same as the requesting process.
des [Sun, 7 Oct 2001 20:06:03 +0000 (20:06 +0000)]
Always succeed if the target process is the same as the requesting process.

22 years agoThese flags aren't just for procfs - in fact, these days theye are primarily
des [Sun, 7 Oct 2001 19:47:37 +0000 (19:47 +0000)]
These flags aren't just for procfs - in fact, these days theye are primarily
used by ptrace(2) - so tweak the accompanying comments a little.

22 years agoRemove some useless preprocesor paranoia.
des [Sun, 7 Oct 2001 19:41:19 +0000 (19:41 +0000)]
Remove some useless preprocesor paranoia.

22 years agoIn procfs_readdir(), when the directory being read was a process directory,
des [Sun, 7 Oct 2001 19:37:13 +0000 (19:37 +0000)]
In procfs_readdir(), when the directory being read was a process directory,
the target process was being held locked during the uiomove() call.  If the
process calling readdir() was the same as the target process (for instance
'ls /proc/curproc/'), and uiomove() caused a page fault, the result would
be a proc lock recursion.  I have no idea how long this has been broken -
possibly ever since pfind() was changed to lock the process it returns.

Also replace the one and only call to procfs_findtextvp() with a direct
test of td->td_proc->p_textvp.

22 years ago* Use srlz.i to serialise changes to psr.ic
dfr [Sun, 7 Oct 2001 18:40:01 +0000 (18:40 +0000)]
* Use srlz.i to serialise changes to psr.ic
* Don't enable psr.i at the same time as psr.dt and psr.ic

These changes improve stability considerably.

22 years agoAdd some somewhat vague documentation for this driver and a list
mjacob [Sun, 7 Oct 2001 18:26:47 +0000 (18:26 +0000)]
Add some somewhat vague documentation for this driver and a list
of Hardware that might, in fact, work.

22 years agoMake the interrupt routing a bit less chatty unless bootverbose is set.
dfr [Sun, 7 Oct 2001 18:25:43 +0000 (18:25 +0000)]
Make the interrupt routing a bit less chatty unless bootverbose is set.

22 years agoSome patches from Doug for ia64 support- the principle one being the
mjacob [Sun, 7 Oct 2001 18:18:50 +0000 (18:18 +0000)]
Some patches from Doug for ia64 support- the principle one being the
appropriate cache flush that provides MEMORY_BARRIER in between handoffs
between host && RISC processor for the shared memory request/response
queues.

Submitted by: dfr@nlsystems.com

22 years agoUse host keys from the config directory if they exist, instead of
luigi [Sun, 7 Oct 2001 17:22:46 +0000 (17:22 +0000)]
Use host keys from the config directory if they exist, instead of
regenerating them.

Make the script handle relative paths for the source directory.

22 years agoMake fetch_mask return ${default_mask} if entry not found.
luigi [Sun, 7 Oct 2001 17:17:26 +0000 (17:17 +0000)]
Make fetch_mask return ${default_mask} if entry not found.
Add a couple of diagnostic messages.

22 years agoExport interrupt statistics via sysctl.
luigi [Sun, 7 Oct 2001 17:03:56 +0000 (17:03 +0000)]
Export interrupt statistics via sysctl.

MFC-after: 3 days

22 years agoMention that ed requires miibus.
iedowse [Sun, 7 Oct 2001 16:39:40 +0000 (16:39 +0000)]
Mention that ed requires miibus.

Suggested by: Eugene Grosbein <eugen@grosbein.pp.ru>

22 years agoFix a typo in do_sigaction() where sa_sigaction and sa_handler were
iedowse [Sun, 7 Oct 2001 16:11:37 +0000 (16:11 +0000)]
Fix a typo in do_sigaction() where sa_sigaction and sa_handler were
confused. Since sa_sigaction and sa_handler alias each other in a
union, the bug was completely harmless. This had been fixed as part
of the SIGCHLD changes in revision 1.125, but it was reverted when
they were backed out in revision 1.126.

22 years agoRemove bogus include.
dfr [Sun, 7 Oct 2001 14:20:01 +0000 (14:20 +0000)]
Remove bogus include.

22 years agoWhitespace fixes.
benno [Sun, 7 Oct 2001 13:27:27 +0000 (13:27 +0000)]
Whitespace fixes.

22 years agoAdd new files needed by previous commit.
benno [Sun, 7 Oct 2001 13:24:56 +0000 (13:24 +0000)]
Add new files needed by previous commit.

22 years agoMega-patch for OpenFirmware loader support.
benno [Sun, 7 Oct 2001 13:22:25 +0000 (13:22 +0000)]
Mega-patch for OpenFirmware loader support.

- Flesh out ofw_readin routine.
- Add OpenFirmware load and exec routines.
- Make sure memory allocation for the kernel is done correctly.
- Change the way the heap is allocated so as to make it easier to deallocate
  when we hand over.
- Add a command to print memory maps similar to the one for ia64.

With this patch, I can now load and hand over to a kernel on my iMac.  There
are some problems with OpenFirmware routines failing after the hand over that
still need to be addressed.

22 years agoRewrite the pc98 bus_space stuff.
nyan [Sun, 7 Oct 2001 10:04:18 +0000 (10:04 +0000)]
Rewrite the pc98 bus_space stuff.

The type of bus_space_tag_t is now a pointer to bus_space_tag structure,
and the bus_space_tag structure saves pointers to functions for direct
access and relocate access.

Added bsh_bam member to the bus_space_handle structure, it saves access
method either direct access or relocate access which is called by
bus_space_* functions.

Added the mecia device support. If the bs_da and bs_ra in bus tag are set
NEPC_io_space_tag and NEPC_mem_space_tag respectively, new bus_space stuff
changes the register of mecia automatically for 16bit access.

Obtained from: NetBSD/pc98

22 years agoUpdate to reflect the change in ata.h
sos [Sun, 7 Oct 2001 07:48:12 +0000 (07:48 +0000)]
Update to reflect the change in ata.h

Forgotten by: sos

22 years agoo Introduce an 'options REGRESSION'-dependant sysctl namespaces,
rwatson [Sun, 7 Oct 2001 03:51:22 +0000 (03:51 +0000)]
o Introduce an 'options REGRESSION'-dependant sysctl namespaces,
  'regression.*'.
o Add 'regression.securelevel_nonmonotonic', conditional on 'options
  REGRESSION', which allows the securelevel to be lowered for the purposes
  of efficient regression testing of securelevel policy decisions.
  Regression tests for securelevels will be committed shortly.

NOTE: 'options REGRESSION' should never be used on production machines, as
it permits violation of system invariants so as to improve the ability to
effectively test edge cases, and improve testing efficiency.

22 years agoLimit maximum poll interval to 60 seconds. This prevents an overflow
deischen [Sun, 7 Oct 2001 02:34:43 +0000 (02:34 +0000)]
Limit maximum poll interval to 60 seconds.  This prevents an overflow
from occurring when converting from a timeval/timespec to a timeout in
milliseconds.

Submitted by: dwmalone

22 years agoAdd code to print interrupt statistics when they will be
luigi [Sun, 7 Oct 2001 02:13:23 +0000 (02:13 +0000)]
Add code to print interrupt statistics when they will be
exported via sysctl.

22 years agoClarify foundation of GDR/DDR.
grog [Sun, 7 Oct 2001 01:22:44 +0000 (01:22 +0000)]
Clarify foundation of GDR/DDR.

22 years agoFix a typo (missing "; at the end of a string)
luigi [Sun, 7 Oct 2001 00:29:59 +0000 (00:29 +0000)]
Fix a typo (missing "; at the end of a string)

22 years agoRegen.
marcel [Sun, 7 Oct 2001 00:21:18 +0000 (00:21 +0000)]
Regen.

22 years agoFix a typo and sync a comment with the version in RELENG_4
luigi [Sun, 7 Oct 2001 00:21:13 +0000 (00:21 +0000)]
Fix a typo and sync a comment with the version in RELENG_4

22 years agoFix the NODEF entry for setpgid() so that it matches the NODEF
marcel [Sun, 7 Oct 2001 00:20:08 +0000 (00:20 +0000)]
Fix the NODEF entry for setpgid() so that it matches the NODEF
entries in sys/kern/syscalls.master.

22 years agoMake this compile again after the changes to make bridge kld'able
luigi [Sun, 7 Oct 2001 00:18:48 +0000 (00:18 +0000)]
Make this compile again after the changes to make bridge kld'able

22 years agoFix breakage caused by previous commit. The lkmnosys and lkmressys
marcel [Sun, 7 Oct 2001 00:16:31 +0000 (00:16 +0000)]
Fix breakage caused by previous commit. The lkmnosys and lkmressys
syscalls are of type NODEF but not in a way that fits the given
definition of that type. The exact difference of lkmressys and
lkmnosys is unclear, which makes it all the more confusing. A
reevaluation of what we have and what we really need is in order.

Spotted by: Maxime Henrion <mux@qualys.com>
Pointy hat: marcel

22 years agoMisunderstanding documentation caused me to try and set 1Gbps/2Gps/Auto
mjacob [Sat, 6 Oct 2001 20:41:18 +0000 (20:41 +0000)]
Misunderstanding documentation caused me to try and set 1Gbps/2Gps/Auto
connection speed for the 2300 in the wrong offset in the ICB. Oops.

Respect some QLogic errat wrt PCI errors on certain shared host/RISC registers.

22 years agoWhups- remember to zero the isr pointer arg.
mjacob [Sat, 6 Oct 2001 19:34:43 +0000 (19:34 +0000)]
Whups- remember to zero the isr pointer arg.

22 years agoRespect QLogic's errata- read BIU_ISR even on the 2300
mjacob [Sat, 6 Oct 2001 19:19:24 +0000 (19:19 +0000)]
Respect QLogic's errata- read BIU_ISR even on the 2300
to see if there's an interrupt (avoids PCI parity errors
which can occur on the 2312 if you access some registers
from the host at the same time the RISC on the 2312 is
C accessing them).

MFC after: 1 day

22 years agoDocument -depth
obrien [Sat, 6 Oct 2001 18:07:32 +0000 (18:07 +0000)]
Document -depth

22 years ago- Moved the bus_dma declarations from bus_{at386,pc98}.h into bus_dma.h.
nyan [Sat, 6 Oct 2001 16:27:21 +0000 (16:27 +0000)]
- Moved the bus_dma declarations from bus_{at386,pc98}.h into bus_dma.h.
  (bus_dma.h is repo-copied from bus_at386.h)
- Added '#include <machine/bus_dma.h>' into bus.h for backward compatibility.

22 years agoMove console probes until after we set boothowto so that 'boot -h' works.
dfr [Sat, 6 Oct 2001 16:06:48 +0000 (16:06 +0000)]
Move console probes until after we set boothowto so that 'boot -h' works.

22 years agoMake this work on ia64. I have no idea why it works on alpha - it
dfr [Sat, 6 Oct 2001 15:58:54 +0000 (15:58 +0000)]
Make this work on ia64. I have no idea why it works on alpha - it
shouldn't.

22 years agoAssume round-to-nearest mode for floating point.
dfr [Sat, 6 Oct 2001 15:57:22 +0000 (15:57 +0000)]
Assume round-to-nearest mode for floating point.

22 years agoEnable console. There are far too many 'ifdef __i386__' in this file.
dfr [Sat, 6 Oct 2001 13:34:30 +0000 (13:34 +0000)]
Enable console. There are far too many 'ifdef __i386__' in this file.

22 years agoAdd /etc files for ia64.
dfr [Sat, 6 Oct 2001 12:55:50 +0000 (12:55 +0000)]
Add /etc files for ia64.

22 years agoPut the label at the same place as i386 for compatibility.
dfr [Sat, 6 Oct 2001 12:35:22 +0000 (12:35 +0000)]
Put the label at the same place as i386 for compatibility.

22 years agoAdd an entry with my birthday.
keramida [Sat, 6 Oct 2001 11:40:36 +0000 (11:40 +0000)]
Add an entry with my birthday.

Approved by: nik

22 years agoUpdates for changes made by KSE import.
sheldonh [Sat, 6 Oct 2001 11:19:41 +0000 (11:19 +0000)]
Updates for changes made by KSE import.

Submitted by: Chad David <davidc@acns.ab.ca>

22 years agoUpdate with latest ATA/ATAPI ver 6 rev 2 items.
sos [Sat, 6 Oct 2001 11:07:04 +0000 (11:07 +0000)]
Update with latest ATA/ATAPI ver 6 rev 2 items.

22 years agoDelete legacy pcib code - we can't possibly work without acpi on ia64.
dfr [Sat, 6 Oct 2001 10:09:57 +0000 (10:09 +0000)]
Delete legacy pcib code - we can't possibly work without acpi on ia64.

22 years agoRemove nexus_pcib.c.
dfr [Sat, 6 Oct 2001 10:09:14 +0000 (10:09 +0000)]
Remove nexus_pcib.c.

22 years agoo Change ia64_memory_address to explicitly take a u_int64_t
marcel [Sat, 6 Oct 2001 09:31:43 +0000 (09:31 +0000)]
o  Change ia64_memory_address to explicitly take a u_int64_t
o  Add memcpy_fromio, memcpy_io, memcpy_toio, memset_io,
   memsetw and memsetw_io. I'm not sure this is the right
   place for it, though.

22 years agoMake this compile on ia64.
marcel [Sat, 6 Oct 2001 09:27:43 +0000 (09:27 +0000)]
Make this compile on ia64.

22 years agoImplement these using mux1 and extr.u. I'll update the userland versions
dfr [Sat, 6 Oct 2001 08:35:05 +0000 (08:35 +0000)]
Implement these using mux1 and extr.u. I'll update the userland versions
similarly. Actually, they should be inline on gcc.

22 years agoNew release notes: eaccess(2), vfs.vmiodirenable=1.
bmah [Sat, 6 Oct 2001 05:16:27 +0000 (05:16 +0000)]
New release notes:  eaccess(2), vfs.vmiodirenable=1.

MFCs noted:  wall(1) -g, route(8) indirect routes, route(8) net/bits
syntax, route(8) proxy only published ARP entries.

22 years ago- Fix typo in "didn't find tag in list" code -- != should have been ==.
fenner [Sat, 6 Oct 2001 05:02:11 +0000 (05:02 +0000)]
- Fix typo in "didn't find tag in list" code -- != should have been ==.
  This fixes the panic when receiving a packet with an unknown tag, and
  also allows reception of packets with known tags.
- Allow overlapping tag number spaces when using multiple hardware-assisted
  VLAN parent devices (by comparing the parent interface in
  vlan_input_tag() just as in vlan_input() ).
- fix typo in comment

MFC after: 1 week

22 years agoFix some 64-bit uncleanliness.
peter [Sat, 6 Oct 2001 04:09:26 +0000 (04:09 +0000)]
Fix some 64-bit uncleanliness.

22 years agoFix some style bugs before fixing some real bugs.
peter [Sat, 6 Oct 2001 04:03:20 +0000 (04:03 +0000)]
Fix some style bugs before fixing some real bugs.

22 years agoFix a stack trashing bug when int != sizeof(pointer)
peter [Sat, 6 Oct 2001 03:49:38 +0000 (03:49 +0000)]
Fix a stack trashing bug when int != sizeof(pointer)
This fixes the ia64 boot!  We have scsi disks!

22 years agoFix a warning. (unused p if not INVARIANTS)
peter [Sat, 6 Oct 2001 02:13:58 +0000 (02:13 +0000)]
Fix a warning. (unused p if not INVARIANTS)

22 years agoAdd ia64/ia64/sapic.c
peter [Sat, 6 Oct 2001 02:03:00 +0000 (02:03 +0000)]
Add ia64/ia64/sapic.c

22 years agoUpdate list of supported ata(4) controllers (from ata.4 manpage).
bmah [Fri, 5 Oct 2001 23:52:15 +0000 (23:52 +0000)]
Update list of supported ata(4) controllers (from ata.4 manpage).

22 years agoNew release note: MAXMEM/hw.physmem.
bmah [Fri, 5 Oct 2001 22:52:43 +0000 (22:52 +0000)]
New release note:  MAXMEM/hw.physmem.

Use entity for agp(4) driver.

22 years agoAdd an abstract for this document, flesh out the introduction a
bmah [Fri, 5 Oct 2001 22:03:35 +0000 (22:03 +0000)]
Add an abstract for this document, flesh out the introduction a
bit, and fix up the upgrading section.  While I'm here, delete some
of the SGML comments that were left-over from merging the source
files.

MFC after: 1 day

22 years agoAdd a flag TF_LASTIDLE, that forces a previously idle connection
jayanth [Fri, 5 Oct 2001 21:33:38 +0000 (21:33 +0000)]
Add a flag TF_LASTIDLE, that forces a previously idle connection
to send all its data, especially when the data is less than one MSS.
This fixes an issue where the stack was delaying the sending
of data, eventhough there was enough window to send all the data and
the sending of data was emptying the socket buffer.

Problem found by Yoshihiro Tsuchiya (tsuchiya@flab.fujitsu.co.jp)

Submitted by: Jayanth Vijayaraghavan

22 years agoDon't core dump if we can't get the index automatically.
imp [Fri, 5 Oct 2001 20:38:32 +0000 (20:38 +0000)]
Don't core dump if we can't get the index automatically.

Debug info from: Ben Hockenhull <benh@jpj.net>