Commit Graph

94533 Commits

Author SHA1 Message Date
Robert Drehmel
4cc9f52f78 Move some tracing related code into its own function as it will
be needed for system call related ptrace functionality I plan
to commit soon.
2003-09-26 15:09:46 +00:00
Ralf S. Engelschall
ffd1bc0626 fix typo: s/Instaed/Instead/ 2003-09-26 12:24:16 +00:00
Ralf S. Engelschall
d1f602f79e fix typo: s/sytem/system/ 2003-09-26 12:22:28 +00:00
Poul-Henning Kamp
405f776c83 Typo in last commit: missing ')' 2003-09-26 12:13:35 +00:00
Tim J. Robbins
aa808a7fa6 Allow the [, ], and = characters in non-8.3 filenames since they
are allowed by Windows (ref: MS KB article 120138).

XXX From my reading of the CIFS specification, it's not clear that
clients need to validate filenames at all.

PR:		57123
Submitted by:	Paul Coucher
MFC after:	1 month
2003-09-26 12:11:08 +00:00
David Xu
90c6d14eb1 Add __volatile keyword. 2003-09-26 11:59:04 +00:00
Bruce Evans
3a6a55a60b Use the correct speed in the delay for the transmission of a character
in the loopback test in the probe.  The delay was too short for consoles
at speeds lower than about 3200 bps.  This shouldn't have caused many
problems, since such low speeds are rare and the probe is forced to
succeed for consoles.
2003-09-26 11:36:09 +00:00
Poul-Henning Kamp
2f06c3e7f2 Initialize the cn_name instead of the cn_dev 2003-09-26 10:55:07 +00:00
Poul-Henning Kamp
d31593c1a5 Initialize cn_name instead of cn_dev 2003-09-26 10:53:29 +00:00
Poul-Henning Kamp
4866f95d76 Change fb_attach() and fb_detach() to take a integer unit number rather
than a dev_t.

All of the dev_t's passed were bogusly created with makedev()
2003-09-26 10:41:44 +00:00
Poul-Henning Kamp
26b0e90ca2 Set cn_name, not cn_dev 2003-09-26 10:37:16 +00:00
Poul-Henning Kamp
5e27a46ce9 As far as we know, there is no reason to not expose /dev/crypto in
jails so code in there can take advantage of hardware assisted
crypto.
2003-09-26 10:32:21 +00:00
Poul-Henning Kamp
3d4274a52b Update the list of CDROM device names to try for booting with RB_CDROM
flag set.
2003-09-26 09:07:27 +00:00
Poul-Henning Kamp
ff2cf8a51d Eliminate bogus use of makedev(): rather than synthesize the dev_t, use
our already cached copy.
2003-09-26 09:05:57 +00:00
Peter Grehan
02b63ceaff DELAY must be a routine, not a macro definition. 2003-09-26 09:02:24 +00:00
Poul-Henning Kamp
3623186cbc Initialize cn_name, ignore cn_dev. 2003-09-26 08:51:54 +00:00
Poul-Henning Kamp
c5c8f9652b Adhere to the new console API:
Initialize cn_name
   Use cn_unit for internal housekeeping.
   Forget about cn_dev.
2003-09-26 07:29:34 +00:00
Poul-Henning Kamp
0d44087987 Remove wrongly sized cnd_name field, we now store the name in the
consdev structure.

If the consdev name is not set and we have a cn_dev, set the name
from there.  Try to issue a printf about this, even though it may
not have a place to go.

Modify the sysctl related code to pick up the name from the consdev
instead.
2003-09-26 07:26:54 +00:00
Poul-Henning Kamp
d6cfdd73bb Add a cn_unit and cn_name field to the consdev structure.
Most of the actual use of the cn_dev field is merely to get the name,
and most of the actual initializations are bogusly using makedev()
because the probe/attach has not been completed.

Instead we will migrate console drivers to fill in the name and if
the driver needs it: the unit number, thereby avoiding the bogus
calls to makedev().
2003-09-26 07:22:37 +00:00
Marcel Moolenaar
2dd4819fe6 s/ia64/alpha/g 2003-09-26 06:53:54 +00:00
Nate Lawson
656b9dd5c3 Consistently print attach messages. 2003-09-26 05:24:55 +00:00
Marcel Moolenaar
875f70dba4 Revert the introduction of iobase in struct uart_bas. Both the SAB82532
and the Z8530 drivers used the I/O address as a quick and dirty way to
determine which channel they operated on, but formalizing this by
introducing iobase is not a solution. How for example would a driver
know which channel it controls for a multi-channel UART that only has a
single I/O range?

Instead, add an explicit field, called chan, to struct uart_bas that
holds the channel within a device, or 0 otherwise. The chan field is
initialized both by the system device probing (i.e. a system console)
or it is passed down to uart_bus_probe() by any of the bus front-ends.
As such, it impacts all platforms and bus drivers and makes it a rather
large commit.

Remove the use of iobase in uart_cpu_eqres() for pc98. It is expected
that platforms have the capability to compare tag and handle pairs for
equality; as to determine whether two pairs access the same device or
not. The use of iobase for pc98 makes it impossible to formalize this
and turn it into a real newbus function later. This commit reverts
uart_cpu_eqres() for pc98 to an unimplemented function. It has to be
reimplemented using only the tag and handle fields in struct uart_bas.

Rewrite the SAB82532 and Z8530 drivers to use the chan field in struct
uart_bas. Remove the IS_CHANNEL_A and IS_CHANNEL_B macros. We don't
need to abstract anything anymore.

Discussed with: nyan
Tested on: i386, ia64, sparc64
2003-09-26 05:14:56 +00:00
Marcel Moolenaar
153bbe3e6a Make the port number available to the sub-device with PUC_IVAR_PORT. 2003-09-26 04:44:55 +00:00
Nate Lawson
297835bcd5 Sort debugging levels and update the man page to match reality. Also
update man page to reflect iasl(8) import.
2003-09-26 04:32:40 +00:00
Alan Cox
d2a81cdbed MFi386
Allocate the page table directory page as "no object" pages.
2003-09-26 04:12:41 +00:00
Peter Wemm
ee9003796a Fix fabs(). This commit brought to you by the letter 'l'.
(fstp stores a mem32 value, fstpl stores a mem64 value)

This fixes ghostscript for 'make release' on amd64.  Ghostscript for some
reason thinks it is a good idea to use -fno-builtin, which means it is
vulnerable to bugs in libc that are normally hidden by the builtin gcc
functions.  Oops.
2003-09-26 01:49:48 +00:00
Sam Leffler
bffc912d5d o add information about the debugging tools
o give proper attribution to Atsushi Onoe
2003-09-25 23:56:20 +00:00
Bruce M Simpson
2951193672 Fix MLINKS.
Noticed by:	bde
Pointy hat to:	bms
2003-09-25 23:44:15 +00:00
Scott Long
ba079c0d9f aic79xx_pci.c:
aic7xxx_pci.c:
	When performing our register test, be careful
	to avoid resetting the chip when pausing the
	controller.  The test reads the HCNTRL register
	and then writes it back with the PAUSE bit
	explicitly set.  If the last write to the controller
	before our probe is to reset it, the CHIPRST
	bit will still be set, so we must mask it off
	before the PAUSE operation.  On some chip versions,
	we cannot access registers for a few 100us after
	a reset, so this inadvertant reset was causing PCI
	errors to occur on the read to check for paused
	status.

Submitted by:	gibbs
2003-09-25 23:36:41 +00:00
Wilko Bulte
f018fde49c Clarify SRM bootability on Alpha for DAC960. 2003-09-25 21:15:55 +00:00
Sam Leffler
89a688272a update to reflect new api
Submitted by:	Max Laier <max@love2party.net>
Obtained from:	NetBSD (with changes)
2003-09-25 20:49:28 +00:00
Bruce M Simpson
8019c643f1 Add manual page documentation for the machine-independent PMAP subsystem.
Include documentation of alc's new pmap_extract_and_hold() function.

Reviewed by:	hmp
Approved by:	jake (mentor)
2003-09-25 19:14:40 +00:00
Sam Leffler
17dcd0265b note PFIL_HOOKS must be explicitly configured when using IPFILTER 2003-09-25 16:12:12 +00:00
Sam Leffler
28cfb8fc84 indicate PFIL_HOOKS is now required by IPFILTER; it used to automagically
be defined in net/pfil.h
2003-09-25 16:06:17 +00:00
David Xu
58effe49ae pthread API should return error code in return value not in errno.
Reviewed by: deischen
2003-09-25 13:53:49 +00:00
Hajimu UMEMOTO
5c6ebad8f6 add /*CONSTCOND*/ to reduce diffs against latest KAME.
Obtained from:	KAME
2003-09-25 13:40:06 +00:00
Hajimu UMEMOTO
4bcf9f8e6f panic() doesn't need `\n'.
Obtained from:	KAME
2003-09-25 13:36:51 +00:00
Hidetoshi Shimokawa
a8587980d8 Add -p option to dump phy registers. 2003-09-25 09:33:16 +00:00
Hidetoshi Shimokawa
aed9eb791a Phy register map. 2003-09-25 09:27:05 +00:00
Hidetoshi Shimokawa
77ceb8f469 Add -s option to strip subdirectory from module path.
e.g. moudle-path/firewire/firewire.ko -> module-path/firewire.ko

Reviewed by: grog
2003-09-25 07:56:48 +00:00
David Xu
a551fe8dc1 If syscall failed, restore old sigaction and return error to thread. 2003-09-25 06:23:40 +00:00
Hidetoshi Shimokawa
c4b3637b44 * scsi_cmds.c
- Fill in autosense data.
- Add compatibility for RELENG_4.

* scsi_target.c
- Raw device support
- Set correct value in c_descr->offset for CAM_DIR_NONE case.
- Support for CTIO abort.
2003-09-25 05:43:26 +00:00
Alan Cox
d91440cd46 MFi386
Reimplement pmap_release() such that it uses the page table rather than
 the pte object to locate the page table directory pages.  (Temporarily,
 retain an assertion on the emptiness of the pte object.)
2003-09-25 05:38:18 +00:00
Hidetoshi Shimokawa
388f522d79 Detach black hole device on AC_PATH_DEREGISTERED. 2003-09-25 05:02:19 +00:00
Alan Cox
f3fd831cdd - Eliminate the pte object.
- Use kmem_alloc_nofault() rather than kmem_alloc_pageable() to allocate
   KVA space for the page directory page(s).  Submitted by: tegge
2003-09-25 02:51:06 +00:00
Peter Wemm
cc3112f108 Re-raise the default datasize and stacksize now that the 32 bit exec
support can clip it to sensible values.
2003-09-25 01:11:17 +00:00
Peter Wemm
c460ac3a00 Add sysentvec->sv_fixlimits() hook so that we can catch cases on 64 bit
systems where the data/stack/etc limits are too big for a 32 bit process.

Move the 5 or so identical instances of ELF_RTLD_ADDR() into imgact_elf.c.

Supply an ia32_fixlimits function.  Export the clip/default values to
sysctl under the compat.ia32 heirarchy.

Have mmap(0, ...) respect the current p->p_limits[RLIMIT_DATA].rlim_max
value rather than the sysctl tweakable variable.  This allows mmap to
place mappings at sensible locations when limits have been reduced.

Have the imgact_elf.c ld-elf.so.1 placement algorithm use the same
method as mmap(0, ...) now does.

Note that we cannot remove all references to the sysctl tweakable
maxdsiz etc variables because /etc/login.conf specifies a datasize
of 'unlimited'.  And that causes exec etc to fail since it can no
longer find space to mmap things.
2003-09-25 01:10:26 +00:00
Warner Losh
26824d8c47 CARDMEM is the offset of the address ON THE CARD (eg which page).
Instead, use EXCA_MEMREG_WIN_SHIFT which is the amount we shift the
bus address by to write into upper memory (eg above 24MB).  Use the
latter in this case.
2003-09-24 22:13:25 +00:00
Bruce M Simpson
85cc199400 Fix a logic error in the check to see if arplookup() should free the route.
Noticed by:	Mike Hogsett
Reviewed by:	ru
2003-09-24 20:52:25 +00:00
Dag-Erling Smørgrav
44172b702c Update version string. 2003-09-24 19:20:23 +00:00