table rather than the old linear list search.
On my "hardlink detection torture test", this reduced
user time from 4700 seconds down to 4.2 seconds
and wallclock time from 1:24:48 down to 1:08.
(Yes, that's over one THOUSAND times reduction in user time. ;-)
In the worst case, the new code doubles peak memory usage,
though it could actually reduce memory usage in many cases.
MFC after: 1 week
PR: misc/42167, bin/51151
table for the hardlink cache. This dramatically improves
performance when archiving millions of hardlinked files.
While I'm here, clean up some style bugs (per Bruce Evans)
and clarify some comments.
<sys/user.h> for the definition of TDF_SINTR. Fixed anachronous
spelling of TDF_SINTR in a comment
Demangled VCS ids. There were 2 misplaced copies of $FreeBSD$ and of
the include before it. The vendor id infrastructure was edited.
Fixed the only other remaining style bug since rev.1.1 (expansion of
struct member names made a line too long).
While I'm here, add a couple of extra sanity-checks to
the argument parsing (reject -j -z, for instance) and
update the docs a bit.
Requested by: most everyone ;-)
ACLs. This is similar to what ls(1) can do. It is handy to
have it so that it can be used in conjunction with
"-exec setfacl {} \;" (to find(1)), among others.
This is the submitter's patch, but slightly modified.
PR: bin/65016
Submitted by: Christian S.J. Peron <maneo@bsdpro.com>
as a result of mis-sorting.
PR Submitter wanted to see a sorted list of non-national holidays followed
by a sorted list of national holidays; I'm not going to do this, because I
don't think we want to get into the "what is a nation?" debate.
PR: docs/64848
Submitted by: Ulrich Spoerlein <q@uni.de> (approximately)
MFC after: 3 days
"system processes" to always ignore. Based on my testing with `-D',
I am pretty sure this is what we want for 5.x-current. If my thinking
is wrong, this also makes it easier to switch to a different check.
what are supported in `ps':
-M Extract values associated with the name list from the
specified core instead of the default /dev/kmem.
-N Extract the name list from the specified system instead
of the default /kernel.
Written by: Mario Sergio Fujikawa Ferreira <lioux@FreeBSD.org>
Obtained from: the sysutils/pkill port
straight from NetBSD (except to add the RCS-ID lines for FreeBSD).
These will probably require a few updates before they are added to
the FreeBSD buildworld. I might MFC these to 4.x-stable after 4.10.
Discussed on: freebsd-arch
Obtained from: NetBSD (and OpenBSD also has these)
Fd_set and Sigaction structures. Use these for printing the arguments
to sigaction(), nanosleep(), select(), poll(), gettimeofday(),
clock_gettime(), recvfrom(), getitimer() and setitimer().
This is based on Dan's patch from the PR but I've hacked it for
style and some other issues. While Dan has checked this patch, any
goofs are probably my fault.
(The PR also contains support for the dual return values of pipe().
These will follow once I've ported that support to platforms other
than i386.)
PR: 52190
Submitted by: Dan Nelson <dnelson@allantgroup.com>
- 0 should have been -1 in previous commit (just to stay consistent),
- Spell null pointers as NULL, not 0,
- Fixed the comment about pr_usesysctl to not confuse it with boolean.
Pointed by: bde
included twice by lprint.c, which included both finger.h and extern.h.
finger.h, in turn, includes extern.h. The redundant include of extern.h
was removed from lprint.c, as part of this change, but the include guards
were added anyway out of spite.
Fixed a nearby bug. The "play it safe" code in dosysctl() was unsafe
because it overran the buffer by 1 if sysctl() filled all of the buffer.
Fixed a nearby style bug in output. Not just 1, but 2 extra newlines
were printed at the end by "vmstat -m" and "vmstat -z". Don't print
any newlines explicitly. This depends on 2 of the many formatting
bugs in the corresponding sysctls. First, the sysctls return an extra
newline at the end of the strings. This also messes up output from
sysctl(8). Second, the sysctls return an extra newline at the beginning
of the strings. This is good for separating the 2 tables output by
"vmstat -mz" and for starting the header on a new line in plain sysctl
output, but gives a bogus extra newline at the beginning for "vm -[m | z]"
and "sysctl -n [kern.malloc | vm.zone]".
Fixed some nearby style bugs in the source code:
- the same line that misspelled 0 as NULL also spelled NULL as 0.
- the size was doubled twice in the realloc loop.
- the "play it safe" comment was misleading. Terminating the buffer
is bogus because dosysctl() is only meant to work with sysctls that
return strings and the terminator is part of a string. However, the
kern.malloc sysctl has more than style bugs. It also doesn't return
a string. Termination is needed to work around this bug.
the .for loop:
- Replaced four global variables in parse.c with one.
- Made Parse_FromString() accept the "lineno" as an argument.
- Fixed line numbering when there are escaped newlines in the
body of the .for loop.
Adopted from: NetBSD
1) Constness, unusedness and size_tness.
2) Don't clobber the value returned by hid_report_size in the daemon case.
3) Don't misspell "sizeof buf" as 100.
4) Don't run off the end of a buffer if the pid is about a google.
5) Avoid shadowing the usage function.
being reported by /usr/bin/printf.
This bug has been around for 22 months... either nobody uses printf
with floating-point values, or people are forgetting to check their
return codes.
Approved by: rwatson (mentor)
(message save as first recipient) options for standards
conformance.
Submitted by: Wartan Hachaturow <wart@tepkom.ru> (with some changes)
PR: standards/61934
1: Document -follow under COMPATIBILITY.
2: Update an example to be a little more 'safe'.
3: Use '/' in place of '.' for an example; similar to other manual pages.
PR: 40196 (1), 39532 (2, 3)
Submitted by: Marc Silver <marcs@draenor.org> (2 and 3)
Discussed with: des (1)
<netinet/tcp_var.h>'s prerequisites. Prerequistes should not grow for
userland headers, and <netinet/tcp_var.h> is unfortunately still needed
in userland.
Use %ju and an (uintmax_t) cast to print size_t values.
Use %jd and an (intmax_t) cast to print off_t values.
Use off_t variables to hold the difference between two off_t values.
Don't bother with unsigned char where char will do nicely.
1) Include string.h for strcpy.
2) Don't make duplicate declaration of dump_file, we now include extern.h.
3) Help out with some constness.
4) Cast to slightly better types in some comparisons.
read EOF from STDIN_FILENO, write zero bytes into the pseudo-terminal;
this is interpreted as an EOF by the program being scripted.
I've tested this with two non-interactive scripts:
# echo 5 | script foo sh -c 'read x; sleep $x; echo bar'
# echo bar | xargs script foo echo
and one interactive program:
# script foo more /etc/passwd
and everything seems to work properly...
PR: bin/56166, bin/57414, ports/57415, ports/60534
Approved by: rwatson (mentor)
MFC after: 2 weeks
parentheses if the continuation indent is exactly half of the main
indent. Indenting one contination indent for every level of
parentheses gives bad results in most cases and is not what is done
in about 90% of properly hand-formatted KNF code (sys/kern/*.c,
nvi/common/*.c). The main advantage of the non-default KNF options
-nlp -ci4 is that continuation lines don't accidentally line up with
the next main indentation level or march to the right, and increasing
their indentation defeats this.
This behaviour change is limited to when the continuation indent is
exactly half of the main indent to avoid adding yet another option.
The default is to be backwards compatible and non-KNF (use the same
indentation for locals as for globals; -ldi0 gives KNF indentation
for locals (none)). The indentation for globals also applies to struct
member names in local declatations. The indentation of variable names
in multi-line declarations is broken in various ways and this commit
gives some new variations.
indent.1:
Also clarified the description of -di<N>.
nonzero (so that the 1-char indentation given by -di0 is never rendered
by a tab).
Removed garbage commented out code for setting the indentation of variable
names.
worse than useless because it caused things like mangling of
"^int\tfoo" to "int foo" (this for N = 8). This quick fix breaks
the invariant that characters between s_code and e_code have width 1,
but nothing seems to depend on this.
was mangled to "struct foo * bar". There should be an option to control
this, but no space is normal. This finishes fixing the bugs in rev.1.4.
indent(1) still doesn't really understand types in parameter lists. It
thinks keywords inside parentheses are for casts or sizeofs. This works
accidentally for scalar types and this quick fix makes it work similarly
but not so accidentally for struct/union/enum types.
the number of typedef-names is not so limited. Same as in rev.1.4.
Added the "const" and "volatile" to the keyword table. Rev.1.4 added
these but they were misclassified so they were not formatted as types.
indent still doesn't really understand them. E.g., it mangles
"char * const *foo" and "char *const *foo". This change mainly stops
it mangling "char const foo" to "char<declaration-indent>const foo".
properly. Of the 3 changes mentioned in the log message for rev.1.4,
the first (implementing -[n]fcb) was correct but didn't touch this
file, the second (no-space-after-sizeof) was not actually done (it is
the default and is controlled by the undcoumented -[n]bs options), and
the third (no-space-after 'struct foo *') was very buggy and was reduced
to wrong comments and other style bugs by backing out the main part
of it in rev.1.6. Rev.1.4 had 2 changes which were not mentioned in
its commit log: expand specials[] so that more than -83 typedef-names
can be specified (this was the one working change in rev.1.4), and add
"const" and "volatile" to specials[] (this was buggy).
re-breaks non-interactive portupgrade (or at least old versions of
portupgrade); I'll see if I can put together a solution which avoids
breaking anything later.
Approved by: rwatson (mentor)
Noticed by: Stefan Farfeleder, Joshua Goodall
the double quotes ("" and '') as a separate argument.
Reported by: ache
The fix in this and previous revisions combined is functionally
equivalent to the below patch against rev. 1.27 but the code is
now much easier to follow:
%%%
Index: str.c
===================================================================
RCS file: /home/ncvs/src/usr.bin/make/str.c,v
retrieving revision 1.27
diff -u -r1.27 str.c
--- str.c 28 Oct 2002 23:33:57 -0000 1.27
+++ str.c 25 Jan 2004 12:09:21 -0000
@@ -168,7 +168,7 @@
inquote = (char) ch;
/* Don't miss "" or '' */
if (start == NULL && p[1] == inquote) {
- start = t + 1;
+ start = t;
break;
}
}
%%%
1. Don't do tty stuff to stdin if stdin isn't a tty.
2. When running in non-interactive mode, don't select(2)
on the standard input.
This un-breaks non-interactive portupgrade.
PR: bin/59036 [1]
PR: bin/56166, bin/57414, ports/57415, ports/60534 [2]
MFC after: 7 days
Approved by: rwatson (mentor)
return for getopt() and comparing to -1, ditto with fgetc() and EOF,
and using the kg_nice value from <sys/user.h>
Submitted by: Stefan Farfeleder <stefan@fafoe.narf.at>
Reviewed by: obrien, bde (a while back)
Tested lightly on: ppc, i386, make universe
1. If fgets fails, don't go into an infinite cpu-intensive loop. Instead,
check to see if the terminal still exists, and sleep(1) otherwise.
2. When we check to see if the terminal still exists, make sure we're not
mislead by EINTR. This could have been a security issue, but fortunately
the current implementation of tcgetattr doesn't EINTR.
PR: bin/60758
Approved by: rwatson (mentor)
a new function dup_shell() to replace ok_shell() and make it unconditionnally
strdup() its result to make the caller's code simplier. Change ok_shell() to
just return an integer value suitable for tests (it was used mainly for that
purpose). Do not use strdup() in the caller's code but rely on dup_shell()
that will do the job for us.
PR: bin/2442
- Unify the conditional assignments section so that architectural
exclusions come first, then options and !options, sorted by the
option name, also in directory order, then architecture specific
sections, sorted by the architecture name, with i386 being a
traditional exception.
Prodded by: bde
normally succeed (because root can su to anyone), but pam_acct_mgmt()
will most likely fail, causing su to log a confusing "pam_acct_mgmt:
error in service module" message. To avoid this, call getpwnam()
before pam_acct_mgmt().
Sponsored by: registrar.no
calculations. Long longs should never be used, since they break compiling
with C90 compilers and don't necessarily work any better than longs for
avoiding overflow.
Print percentages with another digit of precision since they can be small
and this is easy to do now that the format is floating point.
Restored some more of the old -m output:
Print the percentage of allocated memory that is in use. This is the
amount of memory in active mbufs and mbuf clusters relative to the
total amount of memory soft-allocated for mbufs and mbuf clusters.
Print the percentage of allocated memory that is wired (cached). The
old mbuf allocator never freed memory so printing this value wasn't
useful. A previous version of netstat for the new allocator printed
the in-use amount as a percentage of the wired amount.
Fixed some nearby style bugs (excessive parenthesization and a redundant
return).
Reviewed by: alfred
printf format warnings for inet6.c (pluralies() was implicit int, but
the context requires a "char *").
Added WARNS?=2 to the Makefile so that such errors don't come back.
Added NO_WERROR?= to the Makefile because I haven't checked that setting
WARNS doesn't uncover more bugs except on i386's.
In my last change I made sure that the signal as reported from a truss
exit is the same as if truss wasn't between parent and trussed
program. I was smart enough to not have it coredump on SIGQUIT but it
didn't ocur to me SIGSEGV might cause a coredump, too :-)
So get rid of SIGQUIT extra hack and limit coredumpsize to zero
instead.
Tested: still works, correct signal reported. No more codedumps from
SIGSEGV in the trussed proces. This file compiles cleanly on AMD64
(sledge).
PR:
Submitted by:
Reviewed by:
Approved by:
Obtained from:
MFC after:
overflow was breaking a bunch of the stats, specifically the
percentage displayed for wired memory.
Fix the output for current/peak/max lines, I forgot to output the types.
161/320/51200 (current/peak/max):
-to-
639/25696/51200 mbufs in use (current/peak/max):
using the old 'cached' value but reporting it as 'cached'.
I've decided to report the 'cached' as 'peak', why? Well because
it is the peak, the peak of what is actually allocated. 'cached'
doesn't make sense to me as a user.
terminates. Without this patch, 'make -j1 buildworld' takes about 30%
longer than 'make -B buildworld' on my 2.4 GHz P4; the difference is
probably even larger on faster systems. With this patch, there is no
perceptible difference in wall time between the two.
Submitted by: bde
MFC after: 3 days
- Replace overly-complicated (and buggy) -a logic with a much simpler
version: -a causes all interrupts to be displayed, otherwise only
those that have occurred are displayed. This removes the need for
any MD code.
- Instead of just making sure intrcnt is large enough, figure out the
exact size it needs to be. We derive nintr from this number, and we
don't want to risk printing garbage. Note that on sparc64, we end up
printing garbage anyway because the names of non-existent interrupts
are left uninitialized by the kernel.
Tested on: alpha, i386, sparc64
- sort the -E switch into the right place.
- add previously missing -p pid in usage (from the last few commits).
- add -E to usage.
- consistently use trfile in the man page.
I knew I shouldn't have touched the man page. If I commit to a man page,
it just makes people suspicious. :-)
o nintr and inamlen must by of type size_t, not int,
o Remove now unnecessary casts,
o Handle the aflag differently, because the intr. names have a
fixed width and almost always have trailing spaces.
The use of libkvm for post-mortem analysis is still supported (though it
could use more testing). We can now remove vmstat's setgid bit.
While I'm here, hack the interrupt listing code to not display interrupts
that haven't occurred unless the -a option was given on the command line,
and document this change.
