]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years agoIt turns out that we do not need to add a new ioctl to unbreak a
dillon [Sun, 12 Jan 2003 03:31:10 +0000 (03:31 +0000)]
It turns out that we do not need to add a new ioctl to unbreak a
default-to-deny firewall.  Simply turning off IPFW via a preexisting
sysctl does the job.  To make it more apparent (since nobody picked up
on this in a week's worth of flames), the boolean sysctl's have been
integrated into the /sbin/ipfw command set in an obvious and straightforward
manner.  For example, you can now do 'ipfw disable firewall' or
'ipfw enable firewall'.  This is far easier to remember then the
net.inet.ip.fw.enable sysctl.

Reviewed by: imp
MFC after: 3 days

21 years agoMerged from sys/isa/fd.c revision 1.244.
nyan [Sun, 12 Jan 2003 03:11:21 +0000 (03:11 +0000)]
Merged from sys/isa/fd.c revision 1.244.

21 years agoChange struct file f_data to un_data, a union of the correct struct
dillon [Sun, 12 Jan 2003 01:37:13 +0000 (01:37 +0000)]
Change struct file f_data to un_data, a union of the correct struct
pointer types, and remove a huge number of casts from code using it.

Change struct xfile xf_data to xun_data (ABI is still compatible).

If we need to add a #define for f_data and xf_data we can, but I don't
think it will be necessary.  There are no operational changes in this
commit.

21 years agoNote that the 1E is the 1 Enterprise.
obrien [Sat, 11 Jan 2003 20:46:55 +0000 (20:46 +0000)]
Note that the 1E is the 1 Enterprise.

21 years agoMake iso.1 more readable.
obrien [Sat, 11 Jan 2003 20:34:34 +0000 (20:34 +0000)]
Make iso.1 more readable.

21 years agoBail out of fd_clone() if the parsed unit number doesn't match our
joerg [Sat, 11 Jan 2003 20:10:41 +0000 (20:10 +0000)]
Bail out of fd_clone() if the parsed unit number doesn't match our
expectation.

This solves the problem, where in a constellation with two (or more)
drives, an attempt is made to access a device name for that device
using a historic partition letter, like /dev/fd1c.  This is supposed
to create a symlink to the master device, but previously, the link was
always created to /dev/fd0, even if the request was for fd1*.

21 years agoAdd support for the Intel 82820 UP-only AGP bridge.
anholt [Sat, 11 Jan 2003 20:08:28 +0000 (20:08 +0000)]
Add support for the Intel 82820 UP-only AGP bridge.

PR: 41466
Submitted by: NIIMI Satoshi <sa2c@sa2c.net>
MFC after: 1 week

21 years agoIn vm_page_alloc(), fuse two if statements that are conditioned on the same
alc [Sat, 11 Jan 2003 20:07:17 +0000 (20:07 +0000)]
In vm_page_alloc(), fuse two if statements that are conditioned on the same
expression.

21 years agoCleanup the formatting from the last commit, convert everything to
dillon [Sat, 11 Jan 2003 20:03:43 +0000 (20:03 +0000)]
Cleanup the formatting from the last commit, convert everything to
a more human-readable 'kilobytes' instead of pages.

21 years agoFBSDIDize and clean up whitespace nits.
des [Sat, 11 Jan 2003 16:11:21 +0000 (16:11 +0000)]
FBSDIDize and clean up whitespace nits.

21 years agoRead sanpei's mind, and unbreak the build.
des [Sat, 11 Jan 2003 16:08:59 +0000 (16:08 +0000)]
Read sanpei's mind, and unbreak the build.

21 years agoChange to correct card entry. RATOC REX-R280 is 10Base-T Ether Card,
shiba [Sat, 11 Jan 2003 13:04:59 +0000 (13:04 +0000)]
Change to correct card entry. RATOC REX-R280 is 10Base-T Ether Card,
RATOC REX-9530 is SCSI2 Card.

21 years agoRegen.
shiba [Sat, 11 Jan 2003 13:01:40 +0000 (13:01 +0000)]
Regen.

21 years agoRATOC REX-R280(10Base-T Ether Card) and REX-9530(SCSI2 Card) have
shiba [Sat, 11 Jan 2003 13:00:56 +0000 (13:00 +0000)]
RATOC REX-R280(10Base-T Ether Card) and REX-9530(SCSI2 Card) have
same product id. So use CIS info(PCCARD_CISTPL_VERS_1).

21 years agoFix kernel build.
mux [Sat, 11 Jan 2003 12:39:45 +0000 (12:39 +0000)]
Fix kernel build.

Pointy hats to: dillon, Hiten Pandya <hiten@unixdaemons.com>

21 years agoDon't restrict MBR sectorsize to 512 bytes.
phk [Sat, 11 Jan 2003 12:30:51 +0000 (12:30 +0000)]
Don't restrict MBR sectorsize to 512 bytes.

Test data provided by: Andrey Koklin <aka@veco.ru>

21 years agoMFbed: translation update
ue [Sat, 11 Jan 2003 09:56:31 +0000 (09:56 +0000)]
MFbed: translation update
hardware/alpha/proc-alpha.sgml: 1.51  -> 1.52
hardware/common/dev.sgml: 1.119 -> 1.124
hardware/pc98/proc-pc98.sgml: 1.1   -> 1.2
hardware/sparc64/Makefile: 1.1   -> 1.3
hardware/sparc64/article.sgml: 1.2   -> 1.3
hardware/sparc64/proc-sparc64.sgml: 1.4   -> 1.6
installation/common/trouble.sgml: 1.10  -> 1.11
relnotes/common/new.sgml: 1.467 -> 1.473

21 years agoDon't count mbufs with m_type == MT_HEADER or MT_OOBDATA as control data
tjr [Sat, 11 Jan 2003 07:51:52 +0000 (07:51 +0000)]
Don't count mbufs with m_type == MT_HEADER or MT_OOBDATA as control data
in sballoc(), sbcompress(), sbdrop() and sbfree(). Fixes fstat() st_size
reporting and kevent() EVFILT_READ on TCP sockets.

21 years agoMake 'sysctl vm.vmtotal' work properly using updated patch from Hiten.
dillon [Sat, 11 Jan 2003 07:29:47 +0000 (07:29 +0000)]
Make 'sysctl vm.vmtotal' work properly using updated patch from Hiten.
(the patch in the PR was stale).

PR:             kern/5689
Submitted by:   Hiten Pandya <hiten@unixdaemons.com>

21 years agoImplement missing fpgetround() and fpsetround().
marcel [Sat, 11 Jan 2003 07:24:54 +0000 (07:24 +0000)]
Implement missing fpgetround() and fpsetround().

21 years agoAdd support for Corega FEtherII CB-TXD (CardBus 100M/10M).
sanpei [Sat, 11 Jan 2003 07:10:35 +0000 (07:10 +0000)]
Add support for Corega FEtherII CB-TXD (CardBus 100M/10M).
Fix typo (Coreaga->Corega)

21 years agocorrect printf format
sam [Sat, 11 Jan 2003 04:55:52 +0000 (04:55 +0000)]
correct printf format

Noticed by: alpha tinderbox

21 years agoReset the channel in attach if its not the console.
jake [Sat, 11 Jan 2003 04:15:46 +0000 (04:15 +0000)]
Reset the channel in attach if its not the console.

21 years agoAdd symlink support to devfs_rule_matchpath(). This allows the user
dd [Sat, 11 Jan 2003 02:36:20 +0000 (02:36 +0000)]
Add symlink support to devfs_rule_matchpath().  This allows the user
to unhide symlinks as well as hide them.

21 years agoMajor bugfixes for large memory and fast systems.
scottl [Sat, 11 Jan 2003 01:59:21 +0000 (01:59 +0000)]
Major bugfixes for large memory and fast systems.

aac.c:
        Re-arrange the interrupt handler to optimize the common case of
        the adapter interrupting us because one or more commands are complete,
        and do a read across the pci bus to ensure that all posted status
        writes are flushed.  This should close a race that could cause command
        completion interrupts to be lost.
        Follow the spec a bit closer when filling out command structures.
        Enable the Fast Response feature to eliminate the need for the card
        to DMA successfull command completions back into host memory.
        Tell the controller how much physical memory we have.  Without this
        there was a chance that our DMA regions would collide with the
        memory window used by the cache on the controller.  The result would
        be massive data corruption.  This seemed to mainly affect systems with
        >2GB of memory.
        Fix a few whitespace problems.

aac_debug.c:
        Add an extra diagnostic when printing out commands.

aac_disk.c:
        Add extra sanity checks.

aacreg.h:
        Prepare for making this 64-bit clean by reducing the use of enumeration
        types in structures.

Many thanks to Justin Gibbs for helping track these down.

21 years agoRemove all use of the LOG2() macro/inline, undoing some non-optimal cruft
dillon [Sat, 11 Jan 2003 01:09:51 +0000 (01:09 +0000)]
Remove all use of the LOG2() macro/inline, undoing some non-optimal cruft
that crept in recently.  GCC will optimize the divides and multiplies for us.

Submitted by: David Schultz <dschultz@uclink.Berkeley.EDU>
MFC after: 1 day

21 years agoIncrease the scheduler stack to 4 pages. This should prevent a stack
deischen [Sat, 11 Jan 2003 00:43:20 +0000 (00:43 +0000)]
Increase the scheduler stack to 4 pages.  This should prevent a stack
overflow when dumping thread info (generated by receipt of SIGINFO).

Reported by: jmallet

21 years agomake sem_leave return a usable errno instead of -1.
alfred [Fri, 10 Jan 2003 23:13:16 +0000 (23:13 +0000)]
make sem_leave return a usable errno instead of -1.
make ksem_close return that usable errno instead of -1 (ERESTART).

PR: 46957

21 years agoUse the fstype obtained from the GEOM dumpconf output to set the fstype
jhb [Fri, 10 Jan 2003 19:45:10 +0000 (19:45 +0000)]
Use the fstype obtained from the GEOM dumpconf output to set the fstype
of BSD part chunks when opening a disk.

Reviewed by: phk
MFC after: 2 days

21 years agoOutput the fstype of each partition in a disklabel in the configuration
jhb [Fri, 10 Jan 2003 19:44:14 +0000 (19:44 +0000)]
Output the fstype of each partition in a disklabel in the configuration
text similar to the way that the MBR module dumps its slice types.

21 years ago- Make New_Disk() non-static so it can be used in Create_Chunk_DWIM().
jhb [Fri, 10 Jan 2003 19:25:38 +0000 (19:25 +0000)]
- Make New_Disk() non-static so it can be used in Create_Chunk_DWIM().
- In Create_Chunk_DWIM(), if there is a freebsd chunk that has no
  children chunks, then trying to add a child part chunk will fail even
  though there is free space.  Handle this special case by adding an
  unused chunk the full size of the freebsd chunk as a child of the
  freebsd chunk before adding the new part chunk.  This situation can
  happen when changing the type of an existing slice to be a FreeBSD
  slice type or when installing onto a blank disk on Alpha (which has
  no slices.)

Reviewed by: phk
MFC after: 2 days

21 years agoEither compaq has two variants of the WL200, or the old entry is
imp [Fri, 10 Jan 2003 18:35:20 +0000 (18:35 +0000)]
Either compaq has two variants of the WL200, or the old entry is
wrong.  Go ahead and include a second entry for the WL200 until I can
sort this out.

Submitted by: Matt Pearce

21 years agoSort. Move /usr/heimdal to /var/heimdal where it really is.
ceri [Fri, 10 Jan 2003 18:32:39 +0000 (18:32 +0000)]
Sort.  Move /usr/heimdal to /var/heimdal where it really is.

PR: docs/46637
Submitted by: Igor Sobrado <sobrado@acm.org>
Approved by: ru

21 years agoI don't know where we got this ccd creation example from... or if it was
obrien [Fri, 10 Jan 2003 17:46:22 +0000 (17:46 +0000)]
I don't know where we got this ccd creation example from... or if it was
ever correct.  It certainly hasn't been correct for *years*, and it
conflicts with the FAQ.  So update this example to match the FAQ.

21 years agoMake the "tarcp" example more direct.
obrien [Fri, 10 Jan 2003 17:37:23 +0000 (17:37 +0000)]
Make the "tarcp" example more direct.

21 years agoAbout September 2001, I consulted with all the previous authors of
nectar [Fri, 10 Jan 2003 13:38:44 +0000 (13:38 +0000)]
About September 2001, I consulted with all the previous authors of
pam_krb5 to consolidate the copyright texts.  The semi-official
pam_krb5 module has been distributed with this new license text ever
since, but I'm just now getting around to updating the text here.

21 years agoMFi386: revision 1.552.
nyan [Fri, 10 Jan 2003 13:31:29 +0000 (13:31 +0000)]
MFi386: revision 1.552.

21 years agoAdd TAPE to the ENVIRONMENT section and standardize the section, using a
sheldonh [Fri, 10 Jan 2003 10:59:11 +0000 (10:59 +0000)]
Add TAPE to the ENVIRONMENT section and standardize the section, using a
list instead of prose.

MFC after: 1 month

21 years agoAdd TAPE to the ENVIRONMENT section.
sheldonh [Fri, 10 Jan 2003 10:55:13 +0000 (10:55 +0000)]
Add TAPE to the ENVIRONMENT section.

MFC after: 1 month

21 years agoAdd an ENVIRONMENT section so that the reader knows that the TMPDIR envar
sheldonh [Fri, 10 Jan 2003 10:45:48 +0000 (10:45 +0000)]
Add an ENVIRONMENT section so that the reader knows that the TMPDIR envar
is honoured.

Reported by: des
MFC after: 1 month

21 years agoDon't record thread pointer, it's not permanent in process life cycle,
davidxu [Fri, 10 Jan 2003 09:54:51 +0000 (09:54 +0000)]
Don't record thread pointer, it's not permanent in process life cycle,
use process pointer instead.

21 years agoMake newer integrated SiS900 cards working.
mbr [Fri, 10 Jan 2003 08:16:03 +0000 (08:16 +0000)]
Make newer integrated SiS900 cards working.

Allow to read EEPROM from LAN. It is shared
between a 1394 controller and the NIC and each
time we access it, we need to set SIS_EECMD_REQ.

Idea from: linux driver source

Reviewed by: luoqi
Obtained from: linux driver source (idea)

21 years agoSis no longer implements the enhanced phy control
mbr [Fri, 10 Jan 2003 08:14:07 +0000 (08:14 +0000)]
Sis no longer implements the enhanced phy control
register, and phy has to be directly accessed via mdio.

Patch converted to CURRENT from STABLE.

Submitted by: luoqi
Reviewed by: luoqi (again)
MFC after: 2 weeks

21 years agoSIS_CFG_EDB_MASTER_EN indicates the EDB bus is used instead of
mbr [Fri, 10 Jan 2003 08:12:20 +0000 (08:12 +0000)]
SIS_CFG_EDB_MASTER_EN indicates the EDB bus is used instead of
the PCI bus. When this bit is set, the Max DMA Burst Size
for TX/RX DMA should be no larger than 64 bytes.

Reviewed by: luoqi
Obtained from: (idea from linux driver source)
MFC after: 2 weeks

21 years agoWhen reading PHY regs over the i2c bus, the turnaround ACK bit
mbr [Fri, 10 Jan 2003 08:09:58 +0000 (08:09 +0000)]
When reading PHY regs over the i2c bus, the turnaround ACK bit
is read one clock edge too late. This bit is driven low by
slave (as any other input data bits from slave) when the clock
is LOW. The current code did read the bit after the clock was
driven high again.

Reviewed by: luoqi
MFC after: 2 weeks

21 years agoThe macro versions of putc() and putchar() are gone.
tjr [Fri, 10 Jan 2003 07:47:12 +0000 (07:47 +0000)]
The macro versions of putc() and putchar() are gone.

21 years agoThere is no macro version of getc() anymore.
tjr [Fri, 10 Jan 2003 07:45:24 +0000 (07:45 +0000)]
There is no macro version of getc() anymore.

21 years agoBring the list of function-like macros up to date with reality.
tjr [Fri, 10 Jan 2003 07:38:38 +0000 (07:38 +0000)]
Bring the list of function-like macros up to date with reality.

21 years agoo Improve wording of the comment that accompanies fs_pad. The
marcel [Fri, 10 Jan 2003 06:59:34 +0000 (06:59 +0000)]
o  Improve wording of the comment that accompanies fs_pad. The
   padding is not specific to non-i386 architectures. It is
   caused by non-i386 specific alignment requirements of
   fs_swuid,
o  Add a CTASSERT to catch a change in the size of struct fs
   at compile-time rather than run-time.

Ok'd: gordon
Tested on: i386 ia64

21 years agospell multiple correctly.
alfred [Fri, 10 Jan 2003 06:52:19 +0000 (06:52 +0000)]
spell multiple correctly.

21 years agoDocument clearerr_unlocked(), feof_unlocked(), ferror_unlocked()
tjr [Fri, 10 Jan 2003 06:22:28 +0000 (06:22 +0000)]
Document clearerr_unlocked(), feof_unlocked(), ferror_unlocked()
and fileno_unlocked().

21 years agoAdd preliminary support for the Hawking PN672TX CardBus cards.
imp [Fri, 10 Jan 2003 05:37:15 +0000 (05:37 +0000)]
Add preliminary support for the Hawking PN672TX CardBus cards.

# Preliminary because there are some subtle things the NetBSD driver does
# that we don't do yet.  My card works for me w/o them.

21 years agoAdd a manual page for flockfile(), ftrylockfile(), and funlockfile().
tjr [Fri, 10 Jan 2003 05:34:11 +0000 (05:34 +0000)]
Add a manual page for flockfile(), ftrylockfile(), and funlockfile().

21 years agoDocument getc_unlocked(), getchar_unlocked(), putc_unlocked(),
tjr [Fri, 10 Jan 2003 04:56:20 +0000 (04:56 +0000)]
Document getc_unlocked(), getchar_unlocked(), putc_unlocked(),
and putchar_unlocked().

21 years agoAdd function versions of getchar_unlocked(), getc_unlocked(),
tjr [Fri, 10 Jan 2003 04:35:08 +0000 (04:35 +0000)]
Add function versions of getchar_unlocked(), getc_unlocked(),
putchar_unlocked(), putc_unlocked(), feof_unlocked(), ferror_unlocked(),
clearerr_unlocked(), and fileno_unlocked(). The first four are required
by POSIX. The rest are provided for consistency.

21 years agoNew release notes: UFS1/UFS2 boot loaders, systat(1) -ifstat.
bmah [Fri, 10 Jan 2003 04:10:12 +0000 (04:10 +0000)]
New release notes:  UFS1/UFS2 boot loaders, systat(1) -ifstat.

Modified release note:  Rewrite matcd(4) removal to be more specific.

21 years agoAvoid a memory leak by using reallocf() instead of realloc().
tjr [Fri, 10 Jan 2003 02:58:25 +0000 (02:58 +0000)]
Avoid a memory leak by using reallocf() instead of realloc().

21 years agoPrototype __getcwd() to avoid a warning.
tjr [Fri, 10 Jan 2003 02:54:37 +0000 (02:54 +0000)]
Prototype __getcwd() to avoid a warning.

21 years ago#include <ctype.h> for prototypes (or macro versions) of
tjr [Fri, 10 Jan 2003 02:46:32 +0000 (02:46 +0000)]
#include <ctype.h> for prototypes (or macro versions) of
tolower() and isdigit().

21 years agoDon't allow user process to set an invalid window state through sigreturn.
jake [Fri, 10 Jan 2003 00:04:56 +0000 (00:04 +0000)]
Don't allow user process to set an invalid window state through sigreturn.

Spotted by: tmm

21 years agoClear the target hardware address field when generating an ARP request.
tmm [Fri, 10 Jan 2003 00:04:53 +0000 (00:04 +0000)]
Clear the target hardware address field when generating an ARP request.

Reviewed by: nectar
MFC after: 1 week

21 years agoFix superblock alignment problems on non-i386 platforms. Also change fs_uuid
gordon [Thu, 9 Jan 2003 23:53:30 +0000 (23:53 +0000)]
Fix superblock alignment problems on non-i386 platforms. Also change fs_uuid
to fs_swuid, making it more descriptive.

Submitted by: marcel
Reviewed by: peter
Pointy hat to: gordon

21 years agoFix a typo: s/save_free/safe_free/.
jhb [Thu, 9 Jan 2003 22:07:24 +0000 (22:07 +0000)]
Fix a typo: s/save_free/safe_free/.

21 years agoFix for DELL PERC firmware. Driver was hanging
emoore [Thu, 9 Jan 2003 21:02:28 +0000 (21:02 +0000)]
Fix for DELL PERC firmware. Driver was hanging
during load time; this attributed to the mailbox
busy byte not being set prior to issuing a
polling command.

Approved by: ps
MFC: 7 days

21 years agoAdd support for the Davicom DM9009 chipset.
trhodes [Thu, 9 Jan 2003 20:51:39 +0000 (20:51 +0000)]
Add support for the Davicom DM9009 chipset.

PR: 46859
Submitted by: Boaz Haberman <boaz@ool-182f8b09.dyn.optonline.net>
Approved by: rwatson

21 years agoRemove earlysetcpuclass() as it has been OBE.
jhb [Thu, 9 Jan 2003 19:59:28 +0000 (19:59 +0000)]
Remove earlysetcpuclass() as it has been OBE.

Suggested by: bde

21 years agoRework part of the previous processor name changes so that we read
jhb [Thu, 9 Jan 2003 19:54:49 +0000 (19:54 +0000)]
Rework part of the previous processor name changes so that we read
cpu_exthigh and cpu_brand in printcpuinfo() instead of in identify_cpu().
We also only do it for known-good values of cpu_vendor which is a bit more
conservative.

Reviewed by: bde (mostly)

21 years agoTop Gun Telnet has been deprecated, and the home page recomends Top Gun SSH.
trhodes [Thu, 9 Jan 2003 19:52:01 +0000 (19:52 +0000)]
Top Gun Telnet has been deprecated, and the home page recomends Top Gun SSH.
Update the URL to reflect this.

PR: 45010
Submitted by: Kim Scarborough <sluggo@unknown.nu> (Original patch)

21 years agoFix a useless use of cat. Save a process today.
schweikh [Thu, 9 Jan 2003 19:16:09 +0000 (19:16 +0000)]
Fix a useless use of cat. Save a process today.

21 years agosave_free the buf.
obrien [Thu, 9 Jan 2003 19:11:16 +0000 (19:11 +0000)]
save_free the buf.

21 years agoSeparate the uses of a buffer and make it dynamically sized.
obrien [Thu, 9 Jan 2003 17:44:29 +0000 (17:44 +0000)]
Separate the uses of a buffer and make it dynamically sized.

Approved by: scottl(re), jhb(re)
Reviewed by: kuriyama

21 years agoBuild gem(4) on powerpc also.
obrien [Thu, 9 Jan 2003 16:37:37 +0000 (16:37 +0000)]
Build gem(4) on powerpc also.

21 years agoMoved if_hme to hme.
obrien [Thu, 9 Jan 2003 16:32:12 +0000 (16:32 +0000)]
Moved if_hme to hme.
(I need a score card to keep track of the /sys layout fights)

21 years agoAdd an hme(4) module.
obrien [Thu, 9 Jan 2003 16:29:03 +0000 (16:29 +0000)]
Add an hme(4) module.

PR: 46871
Submitted by: Hartmut Brandt <brandt@fokus.gmd.de>

21 years agoOops, add zstty to the witness order list.
jake [Thu, 9 Jan 2003 15:45:28 +0000 (15:45 +0000)]
Oops, add zstty to the witness order list.

Noticed by: benno

21 years agoRegen:
sheldonh [Thu, 9 Jan 2003 15:38:06 +0000 (15:38 +0000)]
Regen:

* Boemler: vendors.txt (2002-12-27)

21 years agoDont probe PnP devices.
sos [Thu, 9 Jan 2003 14:00:07 +0000 (14:00 +0000)]
Dont probe PnP devices.

Submitted by: Takahashi Yoshihiro <nyan@FreeBSD.org>

21 years agoFix typo
sos [Thu, 9 Jan 2003 13:54:07 +0000 (13:54 +0000)]
Fix typo

PR: 45375

21 years agoSGML cleanup: Remove useless <para> </para> around lists
ue [Thu, 9 Jan 2003 13:02:09 +0000 (13:02 +0000)]
SGML cleanup: Remove useless <para> </para> around lists

21 years agoMake ofw_pci_find_node() use the reg property instead of the
benno [Thu, 9 Jan 2003 12:24:19 +0000 (12:24 +0000)]
Make ofw_pci_find_node() use the reg property instead of the
assigned-addresses property. This works a lot better.

21 years agoDocument tinderbox and whereintheworld, and move diffburst in its
des [Thu, 9 Jan 2003 12:24:17 +0000 (12:24 +0000)]
Document tinderbox and whereintheworld, and move diffburst in its
alphabetical place.

21 years agoMy version of fenner's "make world" log summarizer.
des [Thu, 9 Jan 2003 12:23:29 +0000 (12:23 +0000)]
My version of fenner's "make world" log summarizer.

21 years agoSample tinderbox script & build settings.
des [Thu, 9 Jan 2003 12:22:59 +0000 (12:22 +0000)]
Sample tinderbox script & build settings.

21 years agoMFen: Happy new year 2003
ue [Thu, 9 Jan 2003 12:11:20 +0000 (12:11 +0000)]
MFen: Happy new year 2003

21 years agoActually, the canonical way would be syntactically correct.
des [Thu, 9 Jan 2003 12:04:21 +0000 (12:04 +0000)]
Actually, the canonical way would be syntactically correct.

21 years agoAdd a pcib variant to allow us to fix up interrupt assignments.
benno [Thu, 9 Jan 2003 12:04:18 +0000 (12:04 +0000)]
Add a pcib variant to allow us to fix up interrupt assignments.

We probably want to do something wrt bus enumeration as well at some point.

21 years agoCanonical way of determining whether or not to use crypto.
des [Thu, 9 Jan 2003 11:57:58 +0000 (11:57 +0000)]
Canonical way of determining whether or not to use crypto.

21 years ago- Add error messages for bulkxfer.
simokawa [Thu, 9 Jan 2003 10:25:07 +0000 (10:25 +0000)]
- Add error messages for bulkxfer.
- Reduce register access.

21 years agoAllocate interrupts from the resource list.
benno [Thu, 9 Jan 2003 09:15:42 +0000 (09:15 +0000)]
Allocate interrupts from the resource list.

21 years agoia64 and sparc64 both have libc_r now.
des [Thu, 9 Jan 2003 08:36:05 +0000 (08:36 +0000)]
ia64 and sparc64 both have libc_r now.

21 years ago- Remove the ignore list and replace it with a quirk list of sorts.
benno [Thu, 9 Jan 2003 06:21:42 +0000 (06:21 +0000)]
- Remove the ignore list and replace it with a quirk list of sorts.
- Add a quirk type for devices whose interrupt properties are actually
  attached to their children.
- Flag the "escc" (zs-alike serial controller) device as having this quirk.
- Rework the interrupt discovery code to deal with devices that have more than
  one interrupt.

21 years agoInvoke the driver directly when a request is submitted via crypto_dispatch
sam [Thu, 9 Jan 2003 05:39:04 +0000 (05:39 +0000)]
Invoke the driver directly when a request is submitted via crypto_dispatch
or crypto_kdispatch unless the driver is currently blocked.  This eliminates
the context switch to the dispatch thread for virtually all requests.

Note that this change means that for software crypto drivers the caller
will now block until the request is completed and the callback is dispatched
to the callback thread (h/w drivers will typically just dispatch the op to
the device and return quickly).  If this is an issue we can either implement
a non-blocking interface in the s/w crypto driver or use either the
"no delay" flag in the crypto request or the "software driver" capability
flag to control what to do.

Sponsored by: Vernier Networks

21 years agoo handle jumbograms in m_clone
sam [Thu, 9 Jan 2003 05:30:25 +0000 (05:30 +0000)]
o handle jumbograms in m_clone
o fixup some comments

Submitted by: Doug Ambrisko <ambrisko@verniernetworks.com>

21 years agoThe return-path is optional in a headline, therefore don't skip a
mikeh [Thu, 9 Jan 2003 05:08:37 +0000 (05:08 +0000)]
The return-path is optional in a headline, therefore don't skip a
message if it is missing the return-path.

PR: bin/40314 (slightly different patch)
MFC after: 2 weeks

21 years agoFix up some grammar.
trhodes [Thu, 9 Jan 2003 04:35:38 +0000 (04:35 +0000)]
Fix up some grammar.

Noticed by: sheldonh

21 years agoMerge from NetBSD and arrange for FreeBSD's slight differences in
imp [Thu, 9 Jan 2003 04:24:28 +0000 (04:24 +0000)]
Merge from NetBSD and arrange for FreeBSD's slight differences in
ucom.

This gets my Sanyo SCP-4900 working.

Approved by: joe

21 years agoThese are OBE. devd.conf now lives in /etc.
imp [Thu, 9 Jan 2003 04:11:19 +0000 (04:11 +0000)]
These are OBE.  devd.conf now lives in /etc.

21 years agoMove devd.conf here and install it by default (didn't do a repo copy
imp [Thu, 9 Jan 2003 04:05:06 +0000 (04:05 +0000)]
Move devd.conf here and install it by default (didn't do a repo copy
because there's so little history to save).

MFC After: 10 millifortnights

21 years agoCorrect a few more definitions, capitalize header fields consistently.
mikeh [Thu, 9 Jan 2003 03:36:47 +0000 (03:36 +0000)]
Correct a few more definitions, capitalize header fields consistently.

21 years ago- remove unneeded includes
grehan [Thu, 9 Jan 2003 01:22:30 +0000 (01:22 +0000)]
- remove unneeded includes
- fix big in use of rid for SYS_RES_IRQ
- catch up with ATA common code by adding lock function