Commit Graph

1922 Commits

Author SHA1 Message Date
Olof Sivertsson
586bfb7655 kill: Fix free() with bad pointer on SIG-prefixed signal-name
kill -l SIGHUP (or any other signal-name prefixed with "SIG")
would cause free() to be called with a bad pointer instead of
a pointer to what was allocated. Fix this and add test-case.
2016-09-08 22:32:18 +02:00
Jan Rybar
4a7f9fca14 pgrep - adds warning that pattern exceeds 15 chars without '-f' option 2016-09-06 18:41:56 +02:00
Jon Boden
8abd0c92ab library: Don't use SIGPWR on FreeBSD
FreeBSD doesn't have SIGPWR so makes no sense in warning and assuming
its 29.

References:
 https://bugs.debian/org/832148

Signed-off-by: Craig Small <csmall@enc.com.au>
2016-08-28 19:14:32 +10:00
Jim Warner
5313852596 top: avoid yet more overhead of accessing /proc/status
After discovering those terrible costs associated with
/proc/status vs. /proc/stat, the build_header function
changed to favor the latter for a field found in both.

Well, low-and-behold, this top program still used some
flags that needlessly caused 'status' to still be read
when 'statm' could have served. And, while top's needs
require conversion from pages to KiB, that's still far
less costly than reading that gosh darn 'status' file.

[ this patch parallels similar changes to newlib top ]

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-08-23 21:12:32 +10:00
Jim Warner
7df87ea7cf top: when setting library flags favor stat over status
Long ago, in a galaxy far away, top was convinced that
/proc/stat was to be favored over /proc/status if some
field could be satisfied with either. This was done to
avoid extra costs of 64-bit math for 32-bit platforms.

Well, its time to acknowledge the prevalence of 64-bit
platforms. And in such an environment there is a large
hidden cost currently if using status instead of stat.
In fact, that cost difference can be as high as 1400%.

So, this commit will coax top into favoring that least
costly route while also fixing an EU_TGD library flag.

[ this patch parallels similar changes to newlib top ]

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-08-23 21:12:32 +10:00
Jim Warner
67a634dc4c library: revert one ancient 'escape_str_utf8' deletion
Profiling revealed a large amount of time spent in the
'escape_str_utf8' function (escape.c) with both of our
NLS branches (newlib and master). That same result was
not seen under an ancient top-3.2.8 program & library.

Well, the 3.2.8 result was ultimately explained by the
absence of a 'setlocale', necessary under NLS support.
Thus, when that ancient library tested for locale, all
it got was 'ANSI_...' & assumed 'UTF-8' wasn't active.

But after a hack to that ancient code to place it on a
par with newlib/master, I still found cost differences
that led me to revisit an old change referenced below.

It turns out that 'iswprint' costs far more than would
a call of 'isprint', even with the extra support code.
So this commit just reverts that five year old change.

[ this patch parallels a similar change under newlib ]

Reference(s):
commit 7b0fc19e9d

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-08-23 21:12:32 +10:00
Craig Small
57619dcd81 Merge branch 'free_french_alignment' into 'master'
free: french translation alignment

address issue #24

See merge request !13
2016-08-23 10:44:53 +00:00
Craig Small
3db39cd9ed Merge branch 'pkg-config-ncursesw' into 'master'
configure.ac: use pkg-config to get ncursesw header location

Don't assume ncursesw headers are in ../usr/include/ncursesw/..
On a pure build/system without legacy ncurses that may not be true.
Since we're using pkg-config let it provide the correct include path.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

See merge request !23
2016-08-15 11:16:21 +00:00
Craig Small
45b14318cf Merge branch 'use-strchr' into 'master'
sysctl.c: use strchr() instead of index()

index() is a legacy function, which is no longer implemented by all C
libraries (example: uClibc). Instead, use the POSIX defined strchr()
function.

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

See merge request !22
2016-08-15 11:13:16 +00:00
Craig Small
41fee5ea66 Merge branch 'selinux-dlfcn-fix' into 'master'
ps/output.c: include <dlfcn.h> only when necessary

dlopen() functionality is only used when SELinux support is enabled, so
<dlfcn.h> only needs to be included when ENABLE_LIBSELINUX is
defined. This fixes the build in configurations where <dlfcn.h> is not
available.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

See merge request !21
2016-08-15 11:09:13 +00:00
Craig Small
1955657461 Merge branch 'find_elf_note_fix' into 'master'
Fixing invalid reads in find_elf_note due to setenv invocation

Resolves issue on mailing list from 12 Nov 2014 and Red Hat Bugzilla 1163404.

See merge request !20
2016-08-15 11:08:15 +00:00
Craig Small
0a247fdb28 Merge branch 'vmstat_isdev_fix' into 'master'
Fixing is_disk() which mistakes devices with name longer than 20 chars for partitions.

Associated with previous vmstat -d enhancement commit.

See merge request !19
2016-08-15 11:06:06 +00:00
Gustavo Zacarias
58559a5b64 configure.ac: use pkg-config to get ncursesw header location
Don't assume ncursesw headers are in ../usr/include/ncursesw/..
On a pure build/system without legacy ncurses that may not be true.
Since we're using pkg-config let it provide the correct include path.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
2016-08-07 11:36:10 +02:00
Yann E. MORIN
60b4208dd6 sysctl.c: use strchr() instead of index()
index() is a legacy function, which is no longer implemented by all C
libraries (example: uClibc). Instead, use the POSIX defined strchr()
function.

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
2016-08-07 11:32:52 +02:00
Thomas Petazzoni
4fc9a34802 ps/output.c: include <dlfcn.h> only when necessary
dlopen() functionality is only used when SELinux support is enabled, so
<dlfcn.h> only needs to be included when ENABLE_LIBSELINUX is
defined. This fixes the build in configurations where <dlfcn.h> is not
available.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
2016-08-07 11:08:26 +02:00
Jan Rybar
ecabdaca41 Fixing invalid reads in find_elf_note due to setenv invocation
Resolves issue on mailing list and Red Hat Bugzilla 1163404.
2016-08-05 15:21:49 +02:00
Jan Rybar
c63a5a4af0 Fixing is_disk() which mistakes devices with name longer than 20 chars for partitions.
Associated with previous vmstat -d enhancement commit.
2016-08-04 13:55:01 +02:00
Jim Warner
b4fdb27a05 top: provide for expanded potential displayable fields
This commit provides for raising the total displayable
fields from its current 70 to 86. It also bumps the id
in an rcfile representing the version from 'i' to 'j'.

The increase in number of fields will make sharing the
rcfile with an older top, once it's saved, impossible.

These changes are being done via a #define rather than
hard coded so any such sharing will still be possible.

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-08-01 20:11:11 +10:00
Craig Small
2e4959bab8 library: dont use vm_min_free on non-Linux systems
For a specific slice of kernel versions we can get a better
estimate of the available memory before the "real" available
figure appears around kernel 3.2

However, that middle method requires a /proc/sys/vm directory.
FreeBSD 9.x used to emulate a < 2.6.27 kernel procfs meaning
it never tried the middle method. FreeBSD 10.x emulates something
more modern, but without the available figure and without a
/proc/sys/vm, so the library falls into a hole.

Hurd may to one day have this bug so we'll exclude him as well
as its triggered by whatever number appears in
/proc/sys/kernel/osrelease

References:
 commit 3f3b1a59ad
 https://bugs.debian.org/831396
2016-07-17 09:06:24 +10:00
Jim Warner
63f8c16fa4 top: make compilation possible under netbsd-curses too
Whoa, I had never considered an alternative to ncurses
until the issue referenced below was raised. Thus, I'm
surprised to find that 'tparm' was the only impediment
to ultimately utilizing this alternate curses library.

And, while we could have substituted that non-standard
'tiparm' with only 2 arguments, we'll utilize the full
parms compliment in the spirit of that original patch.

Frankly, the task of developing an alternative library
to that ncurses implementation really boggles my mind.

Congratulations to rofl0r, whoever that masked man is.

Reference(s):
. issue raised
https://gitlab.com/procps-ng/procps/issues/38
. netbsd-curses home
https://github.com/sabotage-linux/netbsd-curses

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-17 08:52:50 +10:00
Craig Small
bf7e728098 Updated Chinese po file 2016-07-17 08:45:48 +10:00
Dr. Werner Fink
4ed44ab58e misc: fix strtod_nol_err tests
A better way of implementing the string to double
conversion and a better way of testing it.

Signed-off-by: Craig Small <csmall@enc.com.au>
2016-07-13 20:08:51 +10:00
Craig Small
100afbc149 misc: Remove strtod tests
Due to the interesting properties of floating points, the strtod
tests have been removed. One day I may bother enabling them again.
2016-07-11 09:25:43 +10:00
Craig Small
e564ddcb01 watch: define HOST_NAME_MAX
Those infernal arches of kfreebsd-i386 (not -amd64) don't define
HOST_NAME_MAX. This patch is a work-around for those systems with
lacking include files.
2016-07-11 08:41:21 +10:00
Craig Small
e0784ddaed misc: Update translations
Update the translations to the latest versions.
We also now have simplified Chinese language too!
2016-07-10 09:09:32 +10:00
Craig Small
6118be808c misc: testsuite to source aux scripts correctly
Because I don't know how to redirect properly in tcl, there are
some small scripts that do this for me. With your standard
make check, all is good because the scripts and the binary are
in the usual spots.

make distcheck however puts them all over the place. The binary is
in a different tree to the test and aux scripts. The change now
tells where aux script where its binary is.
2016-07-10 09:04:44 +10:00
Craig Small
96f151a4f8 6:0:0 lbirary API bump 2016-07-10 08:55:45 +10:00
Craig Small
9069db6c5d misc: Add note for Debian bug about ANSI 2016-07-10 07:59:20 +10:00
Craig Small
9115ad7655 watch: fix process_ansi typo
When I had to apply Josh's ansi fix a few commits below I put the
return before the setattr

References:
 commit 261c571aca
2016-07-09 15:57:11 +10:00
Craig Small
2f5e0da5bb watch: Remove second [m check
With the previous watch fixes, there is no need for the second
[m check as these are covered in previous checks.
2016-07-09 15:16:57 +10:00
Josh Triplett
0236d278f2 watch: Don't attempt to ungetc parts of unknown ANSI escape sequences
If process_ansi encountered an unknown character when processing an ANSI
escape sequence, it would ungetc all the characters read so far, except
for the character just read, and the opening '\033['.  ungetting the
middle of the escape sequence does not produce useful results, and also
relies on the unportable assumption that ungetc works on multiple
characters (which glibc does not support).  Discard the characters
instead.

Signed-off-by: Josh Triplett <josh@joshtriplett.org>
2016-07-09 14:56:29 +10:00
Josh Triplett
f9a4fef6a9 watch: Fix ANSI escape sequence termination
process_ansi stopped processing an ANSI escape sequence if
(c < '0' && c > '9' && c != ';'), which will never happen.  Fix the
range check to use || instead.

Signed-off-by: Josh Triplett <josh@joshtriplett.org>
2016-07-09 14:55:11 +10:00
Josh Triplett
261c571aca watch: Don't process additional numbers in unknown ANSI color escapes
process_ansi assumed all numbers in a color control sequence correspond
to colors or attributes, which breaks badly if it encounters a
ISO-8613-3 escape sequence (such as for truecolor RGB).  For instance,
the sequence "\x1b[38;2;10;20;30m" sets the foreground color to
rgb(10,20,30), but watch will interpret all five numbers in the sequence
as colors or attributes themselves.

Stop processing the entire escape sequence if watch encounters any
number it doesn't understand, as that number may change the meaning of
the rest of the sequence.
2016-07-09 14:52:54 +10:00
Craig Small
ab0b00a239 build-sys: Re vert noinst and check programs
Previously there was a commit to change all noinst_PROGRAMS into
check_PROGRAMS. This was not a good idea.

check_PROGRAMS are built before TESTS are run. However they are
NOT build before the dejagnu tests are run, causing those tests
to fail.

So:
If the program is required for dejagnu, it needs to go into
noinst_PROGRAMS
If the program is required for TESTS or is one of those TESTS,
it needs to go into check_PROGRAMS
2016-07-09 14:35:06 +10:00
Craig Small
cf1326052b build-sys: Make check programs before check
For some unknown reason, check_PROGRAMS are not built before check.
They are built before recheck and after check, which isn't very
useful.

This means any tests by dejagnu that need those programs will fail.
On my console I get a build error, the CI merrily reports the error
but considers the build OK; go figure.

The kludge adds check_PROGRAMS to be a dependency to check.
Note, TESTS don't need to be included in this, because they are
properly compliled after the dejagnu tests but before they are
run.
2016-07-09 14:11:06 +10:00
Craig Small
3ead0207ae watch: fix 8bit regression
As part of the fix to truncate the command in non-8bit, watch had
the function for output_header changed (much for scope cleanliness
and cohesiveness than anything; so I'm going to blame Meyer)...

Anyhow the 8bit enabled version did not have that update which
meant watch failed to compile. Thanks to @asavah for issue #37
and the patch.

References:
 commit 5a40c7970d
2016-07-09 13:27:23 +10:00
Jim Warner
c69b02438f NEWS: update (and reorganize) with accumulated changes
For what could be our last oldlib release, why not try
to make the NEWS a little more readable. So, it's been
reorganized by category and bug fixes were out-dented.

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-09 13:23:27 +10:00
Jim Warner
400bbcb059 build sys: update configure.ac for the latest autoscan
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-09 13:23:27 +10:00
Jim Warner
ecff2213bd misc: repair a single instance of 'whitespace' boo-boo
[ this is just the case of the final newline missing ]

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-09 13:23:27 +10:00
Jim Warner
25a6ecdbfb library: avoid QUICK_THREADS, swat Redhat bug #1284091
For this release, what's likely to be our last oldlib,
we really no longer care when top or ps threads access
is made a little less efficient. This is especially so
when efficiency was gained at the expense of accuracy.

The newlib branch already has turned off QUICK_THREADS
so this patch just brings the master branch into line.

Reference(s):
https://bugzilla.redhat.com/show_bug.cgi?id=1284091

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-09 13:23:27 +10:00
Craig Small
639c58312b library: rename clashing pwcache functions
On MacOS the system already has user_from_uid and group_from_gid.
These are renamed pwcache_get_user and pwcache_get_group.

For the old library, pwcache_get_user needs to be exported
for skill.

References:
 https://gitlab.com/procps-ng/procps/issues/34

Signed-off-by: Craig Small <csmall@enc.com.au>
2016-07-06 22:26:02 +10:00
Jim Warner
eca9168617 misc: adapted others to that new proc_t 'cgname' field
This patch adapts the ps program to a newly add proc_t
field and provides for new support in that top program
along with his man document (ps was already ok there).

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-06 21:33:17 +10:00
Jim Warner
5d54dc0d2e library: play catchup with a new proc_t 'cgname' field
Since we're about to break the ABI/API anyway, why not
go ahead and add yet another field to our proc_t which
the newlib branch has had for awhile. This then allows
the top program to offer 'control group name' and will
also permit a few reductions in that ps program logic.

And let's also clean up some unrelated warnings below.

Clang warnings:
proc/readproc.c:1178:50: warning: address of array 'ent->d_name' will always evaluate to 'true' [-Wpointer-bool-conversion]
    if(unlikely(unlikely(!ent) || unlikely(!ent->d_name))) return 0;
                                           ~~~~~~^~~~~~
proc/readproc.c:1205:50: warning: address of array 'ent->d_name' will always evaluate to 'true' [-Wpointer-bool-conversion]
    if(unlikely(unlikely(!ent) || unlikely(!ent->d_name))) return 0;
                                           ~~~~~~^~~~~~

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-06 21:33:17 +10:00
Jim Warner
c15aea9d08 top: eliminate 2 author sections from the man document
We'll follow Craig's lead and whack some author stuff.

[ and we'll honor the SEE ALSO guideline for periods ]
[ but essentially ignore all the other busybody crap ]
[ which, to be honest, we pretty much follow already ]

[ actually, if you're told to follow a certain style ]
[ in program examples, you've gone way past busybody ]
[ crap and have entered the realm of anal retentive! ]

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-06 21:33:17 +10:00
Jim Warner
5e33325f9a top: improved that #define PRETENDNOCAP implementation
This development (only) define can be used to turn top
into a simple text program, disabling termcap effects.
But input (at screen bottom) suffers from a line wrap.

So, this commit just makes the input prompt processing
a little more effective by adding one leading newline.

Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-06 21:33:17 +10:00
Craig Small
19f0ae79bc documentation: Update watch man page
Added locale details and fixed this manual page to follow standards
including ordering it the right way, keeping the names of things
consistent and removing authors section.

Signed-off-by: Craig Small <csmall@enc.com.au>
2016-07-03 16:16:28 +10:00
Craig Small
1310a76848 watch,free: interpet intervals in non-locale way
Both watch and free used the locale to determine the required delay
interval for subsequent updates. It's preferable to not care about
locale and accept both 12.34 and 12,34 as meaning 12 seconds and
340 microseconds.

References:
 https://bugs.debian.org/692113

Signed-off-by: Craig Small <csmall@enc.com.au>
2016-07-03 16:14:36 +10:00
Craig Small
d936f62eec documentation: rewrite for shared in free
Minor tweaks in the free manual page
* Removed author section. It's wrong and strongly discouraged
  by the man page standards.
* Moved note about shared not in old kernels into bugs sections
as this is not relevant for most people and declutters the top
sections.

References:
 https://bugs.debian.org/755233

Signed-off-by: Craig Small <csmall@enc.com.au>
2016-07-03 10:36:58 +10:00
Simon Tatham
365cb37029 library: find tty device name of process quicker
The procps library attempts to work out the tty of a process
through several methods. For things like /dev/tty123 or
/dev/foo it works fine.

For tty devices that put the minor number in a directory
of the major name this fails. So then we have to fallback
to stating things like the processes STDERR and try again.

Considering a lot of processes sit on ttys such as
/dev/pts/3 this is a lot of wasted time. At the point of
entering driver_name we know "/dev/pts" and we know "3"
we just didn't join them up the right way as this is old
code.

This change now looks for /dev/pts/3 as well. It does it
after looking for /dev/pts3 so the behaviour is the same.

References:
 https://bugs.debian.org/770215

Signed-off-by: Craig Small <csmall@enc.com.au>
2016-07-03 10:08:18 +10:00
Craig Small
5a40c7970d watch: truncate command in non-8 bit mode
When the screen width is not long enough to display the entire
command, watch puts three dots ... like elipses at the end of
the truncated line.

It's been like that for years, perhaps noone uses non 8-bit
watch?

References:
 commit 367b8bb616

Signed-off-by: Craig Small <csmall@enc.com.au>
2016-07-02 16:07:56 +10:00