dragonfly.git
19 years agoThis is an ugly hack, but for now I am spliting ParseModifiers() out
Max Okumoto [Fri, 18 Feb 2005 01:21:52 +0000 (01:21 +0000)]
This is an ugly hack, but for now I am spliting ParseModifiers() out
from VarParseLong().  I am passing many the variables, but most of them
will go away once I fix the lengthPtr calculations.

19 years agoVarParseLong():
Max Okumoto [Fri, 18 Feb 2005 01:20:57 +0000 (01:20 +0000)]
VarParseLong():
Create new else block for if (haveModifier) {}
Move VarExpand() down into both branches of if () {}
Move cleanup section up into both branches.
Eliminate dead code.

19 years agoVarParseShort(): Change else blocks to fall through.
Max Okumoto [Fri, 18 Feb 2005 01:20:09 +0000 (01:20 +0000)]
VarParseShort(): Change else blocks to fall through.

19 years agoVarParseLong():
Max Okumoto [Fri, 18 Feb 2005 01:19:47 +0000 (01:19 +0000)]
VarParseLong():
Move setting of dynamic back before the if statement.
Moved lengthPtr down.

19 years agoChanged code to avoid modifying the input string, its not really
Max Okumoto [Fri, 18 Feb 2005 01:18:27 +0000 (01:18 +0000)]
Changed code to avoid modifying the input string, its not really
nessisary to null terminate substrings, when you have the length
of the substring you are interested in.

19 years ago- Add battd to rc.conf(5)
Liam J. Foy [Thu, 17 Feb 2005 22:44:39 +0000 (22:44 +0000)]
- Add battd to rc.conf(5)

Submitted by: Devon H. O'Dell <dodell@sitetronics.com>

19 years ago- Add battd rcng script
Liam J. Foy [Thu, 17 Feb 2005 22:35:38 +0000 (22:35 +0000)]
- Add battd rcng script

Submitted by: Our good friend Devon H. O'Dell <dodell@sitetronics.com>

19 years agoVarParseLong() - Removed space at end of line.
Max Okumoto [Thu, 17 Feb 2005 20:11:53 +0000 (20:11 +0000)]
VarParseLong() - Removed space at end of line.

19 years agoGCC supports two pseudo variables to get the function name, __FUNCTION__
Joerg Sonnenberger [Thu, 17 Feb 2005 14:00:10 +0000 (14:00 +0000)]
GCC supports two pseudo variables to get the function name, __FUNCTION__
and __func__. The latter is C99, prefer that.

19 years agoUse -pthread instead of adding -lc_r for linking the thread library.
Joerg Sonnenberger [Thu, 17 Feb 2005 13:59:00 +0000 (13:59 +0000)]
Use -pthread instead of adding -lc_r for linking the thread library.

19 years agoRename PACKAGES to REL_PACKAGES. PACKAGES is used by ports already.
Joerg Sonnenberger [Thu, 17 Feb 2005 12:22:24 +0000 (12:22 +0000)]
Rename PACKAGES to REL_PACKAGES. PACKAGES is used by ports already.
Change the default for PACKAGES_LOC to ${PACKAGES}/All if it is defined.

19 years ago- Better handling of a given(-p) invalid port number.
Liam J. Foy [Thu, 17 Feb 2005 11:53:33 +0000 (11:53 +0000)]
- Better handling of a given(-p) invalid port number.

From: OpenBSD

19 years agoThe old USB ethernet code utilized a netisr to hand packets over
Joerg Sonnenberger [Wed, 16 Feb 2005 22:50:28 +0000 (22:50 +0000)]
The old USB ethernet code utilized a netisr to hand packets over
the ether_input and cleanup after txeof. This was broken some
time ago and is a workaround for the different IPL of USB and
the network stack.

Reduce the amount of hacks by removing usb_tx_done, it can be
done mostly in-place with a slight race-condition in the enqueue
code. usb_ether_input now directly calls ether_input, the caller
is responsible for rearming the receiver. That's doesn't open
a race since the receiver doesn't have to handle the interface
queues, for the driver state is the driver responsible.

The IPL protection of the USB ethernet drivers has to be reviewed,
it's mostly not existing.

19 years agoAdd better comments.
Max Okumoto [Wed, 16 Feb 2005 22:38:03 +0000 (22:38 +0000)]
Add better comments.

19 years agoAdd __pure as attribute. A __pure function can only depend on the
Joerg Sonnenberger [Wed, 16 Feb 2005 21:50:00 +0000 (21:50 +0000)]
Add __pure as attribute. A __pure function can only depend on the
arguments, not on global memory. A __pure2 function in contrast
just has to return the same value for the same arguments. It can
use a hash table or similiar means.

For GCC before 3.0 __pure is mapped to __pure2.

Inspired-by: FreeBSD
19 years ago- isupper(x) can be true for x in [128..255], but since tolower(x) is
Liam J. Foy [Wed, 16 Feb 2005 20:35:09 +0000 (20:35 +0000)]
- isupper(x) can be true for x in [128..255], but since tolower(x) is
  required to return x if it cannot be translated, we don't need the
  isupper() test. - NetBSD 1.16
- Kill registers
- Constify
- Style(9)

19 years agoImport of FreeBSD rev. 1.2 by mlaier:
Simon Schubert [Wed, 16 Feb 2005 19:42:35 +0000 (19:42 +0000)]
Import of FreeBSD rev. 1.2 by mlaier:
  Don't rely on properly setup linker.hints to figure out that pflog is now
  part of the pf module.
  While here fix a comment that was c'n'ped from rc.d/pf

  PR:             bin/71096 (partly)
  Submitted by:   Ville-Pertti Keinonen
  MFC after:      2 days

19 years agoTest for my existance
Simon Schubert [Wed, 16 Feb 2005 19:37:39 +0000 (19:37 +0000)]
Test for my existance

19 years agoMake pgrep(1) print a newline at EOL, as expected, instead of at the
Chris Pressey [Tue, 15 Feb 2005 20:31:29 +0000 (20:31 +0000)]
Make pgrep(1) print a newline at EOL, as expected, instead of at the
delimiter specified with the -d switch.

Submitted-by: Robert Nagy <robert_at_bsd_dot_hu>
Same-as-in: OpenBSD: src/usr.bin/pkill.c:1.8

19 years agoALTQ support.
Joerg Sonnenberger [Tue, 15 Feb 2005 20:28:38 +0000 (20:28 +0000)]
ALTQ support.

19 years agoAdd "exit" command as a synonym for "quit".
Chris Pressey [Tue, 15 Feb 2005 20:23:24 +0000 (20:23 +0000)]
Add "exit" command as a synonym for "quit".

Submitted-by: Robert Nagy <robert_at_bsd_dot_hu>
19 years agoALTQ support.
Joerg Sonnenberger [Tue, 15 Feb 2005 20:23:10 +0000 (20:23 +0000)]
ALTQ support.

19 years agoALTQ support.
Joerg Sonnenberger [Tue, 15 Feb 2005 20:09:13 +0000 (20:09 +0000)]
ALTQ support.

19 years agoForgotten in last commit.
Joerg Sonnenberger [Tue, 15 Feb 2005 20:06:50 +0000 (20:06 +0000)]
Forgotten in last commit.

19 years agoRaise WARNS to 6.
Chris Pressey [Tue, 15 Feb 2005 19:49:06 +0000 (19:49 +0000)]
Raise WARNS to 6.

19 years agoALTQ support.
Joerg Sonnenberger [Tue, 15 Feb 2005 19:39:40 +0000 (19:39 +0000)]
ALTQ support.

19 years agoALTQ support.
Joerg Sonnenberger [Tue, 15 Feb 2005 19:19:11 +0000 (19:19 +0000)]
ALTQ support.

19 years agoFix lvalue cast.
Joerg Sonnenberger [Tue, 15 Feb 2005 19:08:06 +0000 (19:08 +0000)]
Fix lvalue cast.

Obtained-from: FreeBSD

19 years agoReplace list of checks with loop. Fix lvalue casts.
Joerg Sonnenberger [Tue, 15 Feb 2005 18:40:38 +0000 (18:40 +0000)]
Replace list of checks with loop. Fix lvalue casts.

19 years agoFix lvalue cast.
Joerg Sonnenberger [Tue, 15 Feb 2005 18:31:48 +0000 (18:31 +0000)]
Fix lvalue cast.

19 years agoChange second argument of dc_crc_le to c_caddr_t to fix warning.
Joerg Sonnenberger [Tue, 15 Feb 2005 18:25:42 +0000 (18:25 +0000)]
Change second argument of dc_crc_le to c_caddr_t to fix warning.

19 years agoALTQ support.
Joerg Sonnenberger [Tue, 15 Feb 2005 18:23:41 +0000 (18:23 +0000)]
ALTQ support.

19 years agoMake the second argument to ncp_conn_asssert a const char *.
Joerg Sonnenberger [Tue, 15 Feb 2005 18:20:50 +0000 (18:20 +0000)]
Make the second argument to ncp_conn_asssert a const char *.
Use __func__ while being here insteed of __FUNCTION__.

19 years agoReplace lvalue cast.
Joerg Sonnenberger [Tue, 15 Feb 2005 18:11:59 +0000 (18:11 +0000)]
Replace lvalue cast.

19 years agoReplace lvalue cast with explicit cast from u_short via int to void *.
Joerg Sonnenberger [Tue, 15 Feb 2005 18:08:52 +0000 (18:08 +0000)]
Replace lvalue cast with explicit cast from u_short via int to void *.

19 years agoAvoid casts as lvalues.
Joerg Sonnenberger [Tue, 15 Feb 2005 17:00:19 +0000 (17:00 +0000)]
Avoid casts as lvalues.

Obtained-from: FreeBSD

19 years agoDon't cast lvalues.
Joerg Sonnenberger [Tue, 15 Feb 2005 16:47:41 +0000 (16:47 +0000)]
Don't cast lvalues.

19 years agoALTQ support.
Joerg Sonnenberger [Tue, 15 Feb 2005 16:44:23 +0000 (16:44 +0000)]
ALTQ support.

19 years agoBe more careful when doing el_parse() - only do it when el is
Joerg Sonnenberger [Tue, 15 Feb 2005 16:32:46 +0000 (16:32 +0000)]
Be more careful when doing el_parse() - only do it when el is
properly initialized, that happens when lpc is called from a tty.
Without this change, it's possible to get SIGSEGV simply doing:
        echo "..:" | lpc

Reported by:    Wojciech A. Koszek <dunstan at freebsd czest pl>
PR:             77462 (patch rewritten by myself)

lpc.c, 1.29
Obtained-from: FreeBSD
Submitted-by: Xin Li (delphij@freebsd)

19 years ago- Use strcasecmp() for upper and lower cases.
Liam J. Foy [Tue, 15 Feb 2005 14:15:16 +0000 (14:15 +0000)]
- Use strcasecmp() for upper and lower cases.
- Style(9)

19 years agoCreate two new functions VarParseShort() and VarParseLong() from the
Max Okumoto [Tue, 15 Feb 2005 11:15:11 +0000 (11:15 +0000)]
Create two new functions VarParseShort() and VarParseLong() from the
contents of the if-else-if-else blocks.   VarParseShort() handles
variables that are only one letter, while VarParseLong() handles
variables that have multiple characters, and might possibly have
modifiers.

19 years ago- Remove more dead code, since VarExpand() is never going to be the same
Max Okumoto [Tue, 15 Feb 2005 11:12:44 +0000 (11:12 +0000)]
- Remove more dead code, since VarExpand() is never going to be the same
  as v->var.

19 years ago- Push more variables into local scope of the if else blocks. This
Max Okumoto [Tue, 15 Feb 2005 11:11:29 +0000 (11:11 +0000)]
- Push more variables into local scope of the if else blocks.  This
  is in preperation for making those blocks into functions.

19 years ago- Killing some dead code, since the conditions can never be true.
Max Okumoto [Tue, 15 Feb 2005 11:09:33 +0000 (11:09 +0000)]
- Killing some dead code, since the conditions can never be true.

19 years ago- Copy destroy code at end of Var_Parse() up into else-if blocks.
Max Okumoto [Tue, 15 Feb 2005 11:05:22 +0000 (11:05 +0000)]
- Copy destroy code at end of Var_Parse() up into else-if blocks.
  This temporarily increases the code size.  Much of that goes
  away, because the certain branches become unreachable, and thus
  can be deleted.

19 years ago- Cleanup Destroy code at end of Var_Parse()
Max Okumoto [Tue, 15 Feb 2005 11:00:39 +0000 (11:00 +0000)]
- Cleanup Destroy code at end of Var_Parse()

19 years ago- Clean up return statements at end of Var_Parse().
Max Okumoto [Tue, 15 Feb 2005 10:59:46 +0000 (10:59 +0000)]
- Clean up return statements at end of Var_Parse().

19 years ago- Moved haveModifier handing up into else-if block. This make that
Max Okumoto [Tue, 15 Feb 2005 10:58:32 +0000 (10:58 +0000)]
- Moved haveModifier handing up into else-if block.  This make that
  block huge, but we are going to turn that into a function.

19 years ago- Factor out code into a new function VarExpand().
Max Okumoto [Tue, 15 Feb 2005 10:54:26 +0000 (10:54 +0000)]
- Factor out code into a new function VarExpand().

19 years agoUse ifq_is_empty to allow USB drivers to support ALTQ.
Joerg Sonnenberger [Tue, 15 Feb 2005 10:30:11 +0000 (10:30 +0000)]
Use ifq_is_empty to allow USB drivers to support ALTQ.

19 years agoIntroduce vnodepv_entry_t as type for the vnodeopv_entry functions.
Joerg Sonnenberger [Tue, 15 Feb 2005 08:32:19 +0000 (08:32 +0000)]
Introduce vnodepv_entry_t as type for the vnodeopv_entry functions.
This is slightly better than casting all the functions to void *,
which is a data pointer.

19 years agoALTQ support.
Joerg Sonnenberger [Tue, 15 Feb 2005 08:29:02 +0000 (08:29 +0000)]
ALTQ support.

19 years agoHook ru.koi8-r.win.kbd up to Makefile and INDEX.keymaps.
Chris Pressey [Tue, 15 Feb 2005 02:27:03 +0000 (02:27 +0000)]
Hook ru.koi8-r.win.kbd up to Makefile and INDEX.keymaps.

19 years agoImport (slightly modified) ru.koi8-r.win.kbd:1.1 from FreeBSD (fjoe):
Chris Pressey [Tue, 15 Feb 2005 02:16:24 +0000 (02:16 +0000)]
Import (slightly modified) ru.koi8-r.win.kbd:1.1 from FreeBSD (fjoe):

"Add ru.koi8-r.win.kbd based on ru.koi8-r.shift.kbd which closely mimics
 ru(winkeys) XFree86 layout.

 Differences with ru.koi8-r.shift.kbd are:
 - ctrl+shift = alock
 - left/right win = meta, menu = saver
 - alock'ed shift+[0-9] = )!@#$%^&*("

Submitted-by: Vyacheslav Bocharov <list_at_smz_dot_com_dot_ua>
Reviewed-by: Alexander 'alxl' Lobachov <alxl_at_alxl_dot_info>
19 years agoFurther cleanup:
Chris Pressey [Tue, 15 Feb 2005 01:53:05 +0000 (01:53 +0000)]
Further cleanup:
- Remove `register' keyword.
- Declare local functions as `static'.
- Remove unneeded unions, and their associated #define'd aliases,
  around the struct fs and struct cg that are being used.

19 years agoMerge with FreeBSD, revision 1.30 (markm):
Chris Pressey [Tue, 15 Feb 2005 01:24:49 +0000 (01:24 +0000)]
Merge with FreeBSD, revision 1.30 (markm):

"Easy warns fixes; use const, mark unused function parameters. Remove
 unused includes."

This allows us to raise WARNS to 6.

19 years agoMerge with FreeBSD, revision 1.29 (imp):
Chris Pressey [Tue, 15 Feb 2005 01:12:07 +0000 (01:12 +0000)]
Merge with FreeBSD, revision 1.29 (imp):

"o unifdef -D__STDC__
 o use ansi function definitions
 o remove OLDSUN"

19 years agoMake sure the $FreeBSD$ tags are correct.
Max Okumoto [Tue, 15 Feb 2005 01:01:18 +0000 (01:01 +0000)]
Make sure the $FreeBSD$ tags are correct.

19 years agoMerge with FreeBSD, revision 1.27 (fenner):
Chris Pressey [Tue, 15 Feb 2005 01:00:39 +0000 (01:00 +0000)]
Merge with FreeBSD, revision 1.27 (fenner):

"Use SIGUSR1 to propogate SIGURG to the child; security measures prevent
  the parent from sending SIGURG itself to the child.  This fixes the
  problem of occasionally failing to pass the window size."

19 years agoRaise WARNS to 6:
Chris Pressey [Tue, 15 Feb 2005 00:26:00 +0000 (00:26 +0000)]
Raise WARNS to 6:
- Constify strings as needed;
- Use correct number of initializer elements;
- Use ssize_t instead of size_t where needed;
- Correct usage of signal().

19 years agoPartial merge with FreeBSD: revisions 1.23 through 1.34.
Chris Pressey [Mon, 14 Feb 2005 23:15:51 +0000 (23:15 +0000)]
Partial merge with FreeBSD: revisions 1.23 through 1.34.

Highlights:

- Support C99 format specifiers %z, %t, %j, %hh.
- Support locale-sensitive "thousands" seperators.
- Some internal cleanup.

19 years agoMake VGA palette registers available.
Sascha Wildner [Mon, 14 Feb 2005 23:05:30 +0000 (23:05 +0000)]
Make VGA palette registers available.

19 years agoChange type of find_compare() so that it doesn't need to be casted
Chris Pressey [Mon, 14 Feb 2005 20:31:59 +0000 (20:31 +0000)]
Change type of find_compare() so that it doesn't need to be casted
when being passed to fts_open().

Requested-by: joerg
19 years agoFix up a bug that crept in during the strcpy->strlcpy conversion;
Chris Pressey [Mon, 14 Feb 2005 20:09:43 +0000 (20:09 +0000)]
Fix up a bug that crept in during the strcpy->strlcpy conversion;
prefix the device name with _PATH_DEV each time through the loop,
instead of just once before the loop.

19 years agoApply less dodgy range checks.
Chris Pressey [Mon, 14 Feb 2005 19:18:53 +0000 (19:18 +0000)]
Apply less dodgy range checks.

Requested-by: joerg
19 years ago- Remove unnecessary headers (string.h, sys/uio.h, utmp.h)
Liam J. Foy [Mon, 14 Feb 2005 18:06:28 +0000 (18:06 +0000)]
- Remove unnecessary headers (string.h, sys/uio.h, utmp.h)
- State any unlinkable file
- Style(9)
- Err magic
- Constify

19 years agoALTQ support. Remove bogus check of ifq max len.
Joerg Sonnenberger [Mon, 14 Feb 2005 17:45:08 +0000 (17:45 +0000)]
ALTQ support. Remove bogus check of ifq max len.

19 years agoALTQ support.
Joerg Sonnenberger [Mon, 14 Feb 2005 17:38:30 +0000 (17:38 +0000)]
ALTQ support.

19 years agoSet ALTQ ready.
Joerg Sonnenberger [Mon, 14 Feb 2005 17:11:12 +0000 (17:11 +0000)]
Set ALTQ ready.

19 years agoALTQ support.
Joerg Sonnenberger [Mon, 14 Feb 2005 17:09:58 +0000 (17:09 +0000)]
ALTQ support.

19 years agoRename IFM_homePNA to IFM_HPNA_1 and IFM_1000_TX to IFM_1000_T.
Joerg Sonnenberger [Mon, 14 Feb 2005 16:21:34 +0000 (16:21 +0000)]
Rename IFM_homePNA to IFM_HPNA_1 and IFM_1000_TX to IFM_1000_T.

Obtained-from: FreeBSD
Submitted-by: Andreas Hauser
19 years agoGreatly reduce the size of ISOFS's inode hash table. CDs and DVDs are small
Matthew Dillon [Mon, 14 Feb 2005 16:11:42 +0000 (16:11 +0000)]
Greatly reduce the size of ISOFS's inode hash table.  CDs and DVDs are small
and slow compared to hard disks, an ultra-efficient inode hash table is not
necessary.

Suggested-by: Joerg Sonnenberger <joerg@britannica.bec.de>
19 years agoAdd support for ICH6 and some nForce AC97 chips.
Joerg Sonnenberger [Mon, 14 Feb 2005 14:48:49 +0000 (14:48 +0000)]
Add support for ICH6 and some nForce AC97 chips.

Obtained-from: FreeBSD
Submitted-by: Andreas Hauser
19 years agoIncrease size of limit column by one character to better fit
Joerg Sonnenberger [Mon, 14 Feb 2005 13:18:39 +0000 (13:18 +0000)]
Increase size of limit column by one character to better fit
on screen on machines with >= 1 GB RAM.

19 years agoExpand *n't contractions.
Max Okumoto [Mon, 14 Feb 2005 11:12:03 +0000 (11:12 +0000)]
Expand *n't contractions.

Author: ru
Taken-from: FreeBSD

19 years ago- Changed type of vlen from int to size_t.
Max Okumoto [Mon, 14 Feb 2005 09:58:28 +0000 (09:58 +0000)]
- Changed type of vlen from int to size_t.
- Replaced usage of function scope variable 'rw_str' with vname
  from a local scope.  This makes it clear that the value in rw_str
  is dead.

19 years ago- Remove adding of '\0' to a Buffer object, since that is already
Max Okumoto [Mon, 14 Feb 2005 09:55:58 +0000 (09:55 +0000)]
- Remove adding of '\0' to a Buffer object, since that is already
  done by the Buffer object itself.
- Renamed vname to name, I want to use that variable name in a
  different scope.

19 years agoPartial merge with recent revisions from FreeBSD:
Chris Pressey [Mon, 14 Feb 2005 02:09:12 +0000 (02:09 +0000)]
Partial merge with recent revisions from FreeBSD:

- pass -1 to setfile() (cp.c:1.49)
- When -R is not specified, fail to copy the contents of dangling
  symlinks instead of making a copy of the link itself, as required
  by SUSv3 (cp.c:1.36)
- Fix the case of a trailing slash following the destination
  mistakenly creating a file instead of a directory (cp.c:1.34)
- Mechanically kill hard sentence breaks in manual page (cp.1:1.29)
- Clarify what -p option does and spell "set-{user,group}-ID" with
  hyphens (cp.1:1.27)

Submitted-by: Sarunas Vancevicius <svan_at_redbrick_dot_dcu_dot_ie>
19 years agoClean up:
Chris Pressey [Mon, 14 Feb 2005 00:39:04 +0000 (00:39 +0000)]
Clean up:
- Remove casts to void when ignoring return values.
- Use explicit cast when calling fts_open().
- Raise WARNS to 6.

Submitted-by: Alexander 'alxl' Lobachov <alxl_at_alxl_dot_info>
19 years agoPartial merge with recent revisions[1] from FreeBSD.
Chris Pressey [Sun, 13 Feb 2005 23:49:53 +0000 (23:49 +0000)]
Partial merge with recent revisions[1] from FreeBSD.

Highlights:
- Use a larger field for the size in blocks; the current width
  (4 digits) is only good for 5 MB.
- Introduce a new primary `-depth n', which tests whether
  the depth of the current file relative to the starting
  point of the traversal is n. The usual +/- modifiers
  to the argument apply.
- Support SysV-style -exec args {} + function, required by SUSv3.
- Remove partial support for building on NetBSD.
- Lots of style(9) work.
- Update manual page.

Features not merged:
- The -acl option for locating files with extended ACLs.

Submitted-by: Alexander 'alxl' Lobachov <alxl_at_alxl_dot_info>
[1] Makefile:1.17 extern.h:1.22 find.1:1.71 find.c:1.17 find.h:1.17
    function.c:1.52 getdate.y:1.3 ls.c:1.17 main.c:1.15 misc.c:1.7
    operator.c:1.14 option.c:1.23

19 years agoRaise WARNS to 6:
Chris Pressey [Sun, 13 Feb 2005 19:22:42 +0000 (19:22 +0000)]
Raise WARNS to 6:
- Double-cast values when testing their ranges for validity.

19 years agoRaise WARNS to 6:
Chris Pressey [Sun, 13 Feb 2005 19:12:26 +0000 (19:12 +0000)]
Raise WARNS to 6:
- Use ssize_t instead of size_t for storing size of inode buffer.

19 years agoClean up:
Chris Pressey [Sun, 13 Feb 2005 18:57:30 +0000 (18:57 +0000)]
Clean up:
- ANSIfy function definitions.
- Constify where appropriate and remove `auto' declarations.
- Normalize whitespace and placement of braces.
- Parentheses around arguments to sizeof.
- Explicitly test against NUL character.
- Use curses functions instead of accessing internal members of the
  WINDOW structure.
- Eliminate a convoluted and unnecessary goto.
- Raise WARNS to 6.

Brought-to-my-attention-by: my CPSC 213 lab assignment
19 years agoCopy input parameter 'char *str' pointer into two variables. One which
Max Okumoto [Sun, 13 Feb 2005 10:12:04 +0000 (10:12 +0000)]
Copy input parameter 'char *str' pointer into two variables.  One which
is a 'const char *' pointer and the other a 'char *'.  This way we can
seperate out the code that actually modifies the input string.

19 years agoMove statement into both branches of a if-statement so that
Max Okumoto [Sun, 13 Feb 2005 10:08:36 +0000 (10:08 +0000)]
Move statement into both branches of a if-statement so that
common grouping of the statements become more apperent.

19 years agoMove statement into both branches of a if-statement so that
Max Okumoto [Sun, 13 Feb 2005 10:05:07 +0000 (10:05 +0000)]
Move statement into both branches of a if-statement so that
common grouping of the statements become more apperent.

19 years agoUse strchr() instead of a switch statement to determine if a char
Max Okumoto [Sun, 13 Feb 2005 10:03:11 +0000 (10:03 +0000)]
Use strchr() instead of a switch statement to determine if a char
is part of a set.

19 years agoRe-order conditions in if-statement a group of so common conditions
Max Okumoto [Sun, 13 Feb 2005 10:01:04 +0000 (10:01 +0000)]
Re-order conditions in if-statement a group of so common conditions
are visible.

19 years agoRemove unneeded conditionals.
Sascha Wildner [Sun, 13 Feb 2005 03:29:17 +0000 (03:29 +0000)]
Remove unneeded conditionals.

19 years agoScrap PC98 support.
Sascha Wildner [Sun, 13 Feb 2005 03:02:26 +0000 (03:02 +0000)]
Scrap PC98 support.

19 years ago* Remove Alpha support and redundant flags from the Makefile.
Sascha Wildner [Sun, 13 Feb 2005 02:11:26 +0000 (02:11 +0000)]
* Remove Alpha support and redundant flags from the Makefile.
* Rename 2 local variables.
* Cast 2 sizeof's to int for comparison.
* Add WARNS?= 6.

Submitted by: Sepherosa Ziehau <sepherosa@softhome.net>
              (with minor modifications by me)

19 years agoFix a little redundancy from my last commit. The border cannot be outside
Sascha Wildner [Sat, 12 Feb 2005 23:44:54 +0000 (23:44 +0000)]
Fix a little redundancy from my last commit. The border cannot be outside
of the drawing area.

19 years agoIn VESA_800x600, never draw the mouse pointer in the border area
Sascha Wildner [Sat, 12 Feb 2005 23:25:41 +0000 (23:25 +0000)]
In VESA_800x600, never draw the mouse pointer in the border area
which means the border does no longer have to be redrawn in
remove_pxlmouse_planar().

19 years agoI have added a per cpu buffer for ktr, so this note can be scraped.
Eirik Nygaard [Sat, 12 Feb 2005 21:12:01 +0000 (21:12 +0000)]
I have added a per cpu buffer for ktr, so this note can be scraped.

19 years agoAdd per cpu buffer for storing KTR information.
Eirik Nygaard [Sat, 12 Feb 2005 21:09:46 +0000 (21:09 +0000)]
Add per cpu buffer for storing KTR information.
Dynamically allocate entries to each cpu buffer on boot.

19 years agoAllow numbers to be used after a letter in the modifier section after a
Eirik Nygaard [Sat, 12 Feb 2005 21:04:41 +0000 (21:04 +0000)]
Allow numbers to be used after a letter in the modifier section after a
command.

19 years agoFix a cache_resolve() vs cache_inval() race which can result in a livelock.
Matthew Dillon [Sat, 12 Feb 2005 18:56:47 +0000 (18:56 +0000)]
Fix a cache_resolve() vs cache_inval() race which can result in a livelock.
The namecache invalidation code was being a bit overzealous when asked to
invalidate a subhierarchy.  It was retrying until the subhierarchy was
completely invalidated, re-invalidating new entries created after the initial
call to cache_inval().  This can occur if the filesystem is heavily loaded
and a high level directory is being recursively invalidated.

It is unnecessary to retry in this case... the purpose is to invalidate
as-of the call to cache_inval(), so it is acceptable to allow new entries
to be resolved within the subhierarchy undergoing the invalidation.

Certain higher level entities... rename, and vnode reclamation, require
complete invalidation.  The retry has been moved to a higher level for
these entities.  The basic cache_inval() code is now single-pass.

Reported-by: Richard Nyberg <rnyberg@it.su.se>
19 years agoMark it as ALTQ ready too.
Joerg Sonnenberger [Sat, 12 Feb 2005 04:18:05 +0000 (04:18 +0000)]
Mark it as ALTQ ready too.

19 years agoALTQ support.
Joerg Sonnenberger [Sat, 12 Feb 2005 04:15:52 +0000 (04:15 +0000)]
ALTQ support.

19 years agoALTQ support.
Joerg Sonnenberger [Sat, 12 Feb 2005 04:07:34 +0000 (04:07 +0000)]
ALTQ support.