Jim Warner [Wed, 13 May 2020 05:00:00 +0000 (00:00 -0500)]
top: add '4' toggle for 2 abreast cpu display, program
In the back of my mind, I've always wanted to enable a
two abreast cpu display. Folks with massively parallel
machines must surely have been frustrated with the '1'
toggle when Off (individual cpus in the Summary Area).
So, I'll use that recently raised issue shown below as
a justification for finally implementing this feature.
pgrep.c: change to allow 'older than' style selection
pgrep.1: man page addition for above selection
Makefile.am: addition of global.c to include sections since boot in pgrep.
Dylan Swiggett [Tue, 29 Nov 2016 22:34:37 +0000 (22:34 +0000)]
Fixes small bug in struct proc_t documentation.
From http://man7.org/linux/man-pages/man5/proc.5.html:
(22) starttime %llu
The time the process started after system boot. In
kernels before Linux 2.6, this value was expressed
in jiffies. Since Linux 2.6, the value is expressed
in clock ticks (divide by sysconf(_SC_CLK_TCK)).
Craig Small [Tue, 28 Apr 2020 09:40:07 +0000 (19:40 +1000)]
build-sys: Enable testing of sigqueue
The referenced commits enavled both pkill and kill to send an integer to
the killed or signalled process. The test_process now will report on the
integer if sent and the testsuite changes take advantage of this
new feature.
Another process make/destroy set had to be made as using spawn
instead of exec changes both the SID and TTY for the underlying
process, making other tests fail.
Craig Small [Fri, 24 Apr 2020 07:42:58 +0000 (17:42 +1000)]
free: Adjust space to really use 9 chars
@steffhip found that while the translation hint said use 9 characters in
the free headers, it really was only 7.
Currently each line is constructed with the following (in non wide format):
Header + 6 Columns. The header takes 7 characters and each column is 11
characters wide and prefixed with one space. Thus we have
7 + (1 + 11) * 6 = 79 characters for each line
By dropping the leading space for the first column after the header -the
header is already terminated by a colon- one could indeed provide the needed
9 letters for the header and thus have 9 + 11 * 1 + (1 + 11) * 5 = 80 Chars
per line which would fit into one line.
Jim Warner [Tue, 14 Apr 2020 05:00:00 +0000 (00:00 -0500)]
library: adapted to the latest lxc conventions (again)
Well, shit! With release 4.0 on March 25th the lxc/lxd
folks have stuck it to us once again. They changed the
cgroup lxc prefix used to identify the container name.
Signed-off-by: Jim Warner <james.warner@comcast.net>
Jan Rybar [Tue, 7 Apr 2020 11:41:48 +0000 (13:41 +0200)]
pidof: show worker threads
Reimplementation of pidof for procps toolset contains sort of deactivated code and does not return results for processes without task.cmdline entry (usually kernel worker threads). Old pidof and pgrep do that in comparison. Despite all perks provided by using pgrep instead, pidof should show those workers again.
Jim Warner [Sun, 29 Mar 2020 05:00:00 +0000 (00:00 -0500)]
top: for symmetry with 'E' add 'e' command line switch
Several years after the 'e' & 'E' interactive commands
were introduce to affect memory scaling, an 'E' switch
was added. This was after discovering a dropped Redhat
patch which provided a unique 'M' command line switch.
If only for symmetry it makes sense to offer a similar
command switch ('e') for the Task Area memory scaling.
As was true with 'E', top's help text will show 'e' as
if it were a switch without arguments in order to keep
help text displayed without wrap in an 80x24 terminal.
The man page, however, will show all of the arguments.
The Debian bug referenced below has nothing to do with
locales. In fact, top was made locale independent back
in release 3.3.13 (April, 2018). However, that bug did
reveal some misplaced logic which this patch corrects.
Prompted by the Qualys audit, all rcfile field strings
were checked for potential duplicates which could only
have resulted from some user's manual/malicious edits.
Unfortunately, that code was executed before top had a
chance to enforce the proper/maximum string length (in
the event an extremely old rcfile had just been read).
This created some potential string overrun references.
In top's original 3.3.15 implementation, the potential
overrun extended for 15 characters. That is the number
of field characters added with 3.3.9 (December, 2013).
But, since strchr() was used, no error exit was taken.
In the revised 3.3.16 implementation, the strchr() was
replaced with '&w->rc.fieldscur[n]'. This held overrun
to a single position while producing an error message.
So, this commit just moves that logic to a point where
fieldscur is guaranteed to be longer than EU_MAXPFLGS.
Jim Warner [Wed, 22 Jan 2020 06:00:00 +0000 (00:00 -0600)]
top: restore one line of code to sig_endpgm() function
When that potential abend at program end was addressed
in the patch shown below, one line of code was removed
in error. That line served to suppress some end-of-job
reports should ATEOJ_RPTSTD or ATEOJ_RPTHSH be active.
So, this patch restores that previously deleted logic.
Jim Warner [Sat, 4 Jan 2020 06:00:00 +0000 (00:00 -0600)]
ps: for abnormal end allow core dumps (fix qualys bug)
A Qualys audit patch, represented in the commit below,
added the _exit() call to our abnormal signal handler.
Unfortunately, that disabled the associated core dump.
This patch restores expected behavior of those signals
whose default produces a core dump file + termination.
Jim Warner [Fri, 3 Jan 2020 06:00:00 +0000 (00:00 -0600)]
top: at abnormal end allow core dumps (fix qualys bug)
A Qualys audit patch, represented in the commit below,
added the _exit() call to our abnormal signal handler.
Unfortunately, that disabled the associated core dump.
This patch restores expected behavior of those signals
whose default produces a core dump file + termination.
Jim Warner [Thu, 2 Jan 2020 05:00:00 +0000 (00:00 -0500)]
top: reposition some data due to a translation comment
In a translator hint, under a caution, a reference was
made to the "next three items". Unfortunately however,
there was one intervening 'item' to which that caution
did not apply. This commit just relocates that 'item'.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit duplicates a change to that newlib branch.
However, it should be noted that such a change was not
really necessary under this master branch since proc_t
data remains valid much longer. It is being duplicated
here as documentation only. Below is the original msg.
------------------------------------------------------
This patch will eliminate a bug which is unique to our
newlib branch. It's extremely rare and only happens if
a search ('L'/'&') is initiated during the period when
fields are currently being auto-sized (AUTOX_MODE on).
This bug surfaces as either all zero results for tasks
displayed or a segmentation fault, depending upon what
fields were activated. It is caused by the timing of a
call to the <pids> 'reset' function. When called after
a task refresh, but before do_key(), this bug appears.
So this patch just ensures that 'reset' will be called
after do_key() & before the tasks have been refreshed.
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
Craig Small [Sun, 5 Jan 2020 04:05:55 +0000 (15:05 +1100)]
pgrep: check sanity of SC_ARG_MAX
A kernel change means we cannot trust what sysconf(SC_ARG_MAX)
returns. We clamp it so its more than 4096 and less than 128*1024
which is what findutils does.
Craig Small [Sun, 8 Dec 2019 03:56:53 +0000 (14:56 +1100)]
library: Adjust version for release
Previous release incremented the age instead of incrementing the
revision. The age can only increment if revision (internal changes
only) or current (API changes) also increments.
A C:R:A of 8.0.2 means its the latest revision of 8.0 library and
any binary linked against 8.0 will work. This is our third revision
of the 8.0 library.
Craig Small [Tue, 12 Nov 2019 10:15:33 +0000 (21:15 +1100)]
docs: more Makefile updates
The install target was using the wrong directory and would throw
some errors where a language couldn't translate man pages. This
one should be hopefully more robust.
Craig Small [Mon, 11 Nov 2019 10:52:21 +0000 (21:52 +1100)]
docs: Update translations
Downloaded the updated translations of the binaries and the manpages
from translation project. Then fixed the typos that caused po4a to
fail, mainly errant spaces.
Craig Small [Mon, 11 Nov 2019 10:41:00 +0000 (21:41 +1100)]
build-sys: Use po4a for manpage translations
Using the newer po4a tool for manpage translations. Also removing
the manpage po file update from dist target because it should be
something the is explicitly done.
The git repository will hold the original man pages and the
po translation files. The distribution tarball will hold those
and the translated manpages. This means most people won't need po4a
as the distribution fill will have these translated manpages.
Craig Small [Mon, 28 Oct 2019 20:48:50 +0000 (07:48 +1100)]
nls: Add all man pages to pot file
The previous Makefile rule would only put the first required file
into the pot file because it used $< Unfortunately, due to how
po4a tools work, its not just a simple matter of changing it to $^
and you're done, but needs a foreach loop to add -m to each manpage
file.
This is a temporary fix, after some more work looking into po4a the
unified tool (called po4a) will be used.
Jim Warner [Sun, 13 Oct 2019 05:00:00 +0000 (00:00 -0500)]
top: update copyright notations & fix a man page error
Beyond the copyrights, the single oops in the man page
was introduced in the commit which is referenced below
dealing with some cleanup following that Qualys audit.
Craig Small [Mon, 21 Oct 2019 21:14:35 +0000 (08:14 +1100)]
snice: Fix matching on PID
@MarsChan correctly pointed out that the read() always returns 128
bytes, so skipping on >= 128 will always mean we skip. Their suggestion
was to remove the equality, but read will never go past 128 bytes so
I just removed that part of the check.
Jim Warner [Tue, 6 Aug 2019 05:00:00 +0000 (00:00 -0500)]
top: avoid a potential SEGV during program termination
The backtrace shown in the bug report referenced below
illustrates a 'normal' program termination interrupted
with some signal, ultimately then causing a top crash.
So this commit just rearranges a little code such that
all signals will be blocked during that rather lengthy
end of program processing regardless of how initiated.
[ in that report, ignore the assertion regarding the ]
[ '-n' option. it obviously was not '1' since do_key ]
[ had been called, which otherwise wouldn't be true. ]
[ and when it is '1' the -d option would be ignored. ]
Jim Warner [Sat, 3 Aug 2019 05:00:00 +0000 (00:00 -0500)]
top: tweak logic associated with keyboard translations
This commit most significant change is the elimination
of the kbd_ENTER entry from that tinfo_tab in iokey().
That entry was a useless artifact left from the commit
which is shown below. It makes no sense to 'translate'
a keystroke into something it already was (i.e. '\n').
The remaining changes just reorder those table entries
for a progression consistent with vim keys: h,j,k & l.
Jim Warner [Sat, 29 Jun 2019 05:00:00 +0000 (00:00 -0500)]
top: standardize PgUp/PgDn management within task area
This commit standardizes the behavior of the PgUp/PgDn
keys when on the main top display. With PgDn, the last
process will become the first process. With PgUp, that
first task will now appear as the last task displayed.
[ this also eliminates some quirks that were evident ]
[ when paging at or near the end of the process list ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
Jim Warner [Mon, 24 Jun 2019 05:00:00 +0000 (00:00 -0500)]
top: attempt to provide missing xterm vim keys support
A recent issue (and merge request) reminded me of gaps
in top's alternate 'vim' navigation keys support. Some
xterm emulators do not pass the customary strings when
keys were used with the <Ctrl> and/or <Alt> modifiers.
While it was a known problem, this issue/merge request
prompted research into the root cause. As it turns out
the problem is traceable to an X resource known by the
name 'eightBitInput'. When 'true' (the default), a key
pressed in combination with <Alt> will not be preceded
by the <Esc> character. Rather, a single character was
presented (modified via an 'eightBitMeta' X resource).
The following approaches would eliminate this problem:
. use: ~/.Xresources with 'Xterm*eightBitInput: false'
. build xterm with 'configure --enable-meta-sends-esc'
( apparently used for CentOS, Fedora, openSUSE, etc. )
. enable xterm's menu via 'configure --enable-toolbar'
( so the user can set the 'Meta Sends Escape' option )
Of course, none of the above steps is desirable from a
user's perspective. So, this patch will add additional
entries to the iokey function's tinfo_tab to represent
strings passed when the <Alt> key does not send <Esc>.
[ hopefully they'll be the same across all platforms ]
Lastly, this patch will also eliminate those redundant
<Atl> + '\', '/', '<' & '>' provisions, which now seem
like overkill and suffer from that same 'eightBitMeta'
xterm problem. And we might as well say goodbye to the
4 '<Alt> + arrow key' table entries (which do not seem
to currently work with any emulator which I can find).
[ what in the world was I thinking way back in 2011? ]
Jim Warner [Wed, 13 Feb 2019 06:00:00 +0000 (00:00 -0600)]
library: tweak that lxc container support a final time
Since the patch referenced below traded a compile-time
'sizeof' directive for a run-time 'strlen' call, there
is no need to declare lxc patterns as explicit arrays.
We'll also use the actual lxc patterns by omitting the
beginning slashes ('/') for both of those definitions.
And, looking to the future when most/all lxc users are
using the most recent lxc release, we will make things
slightly more efficient by reversing those two pattern
literals so the most recent pattern was checked first.
Of course, such a change only benefits tasks which are
running in a container. For the majority of processes,
both literals will be compared in that 'if' statement,
assuming the 'LXC' field is currently being displayed.
[ plus, a leftover parenthesis pair has been removed ]
sysctl: do not report set key in case `close_stream` fails
As we're using buffered I/O when writing kernel parameters, write errors
may get delayed until we close the `FILE` stream. As we are currently
outputting the key that is to be set disregarding the return value of
`close_stream`, we may end up in a situation where we report error and
success:
procio: fix potential out-of-bounds access when write fails
When writing to procfs via `proc_write` fails, we try to chunk the
buffer into smaller pieces to work around that issue. When searching for
the next location to split the buffer, though, we can underflow the
buffer in case the current offset is smaller than `LINELEN`. Fix the
issue by passing `cookie->offset` instead of `LINELEN` into `memrchr` in
case `cookie->offset` is smaller than `LINELEN`.
This bug can be triggered on musl-based systems, e.g. by executing
As the value is out-of-range, `write` will return an error and set
`errno` to `EINVAL`. As we're only trying to write a smallish buffer
with a length smaller than `LINELEN` and as the buffer does not contain
any newlines, the call
procio: use the user-supplied delimiter to split large input
The `fprocopen` function allows users to specify a delimiter chacter
that is used to split very large input lines into smaller chunks. While
the code checks that the caller did actually supply the delimiter, it is
in fact never used to split the string. Instead, the hardcoded default
character ',' is always used to split the string.
I thank Guido Jäkel for raising the issue cited in the
merge request referenced below. While restoring 1 line
of code would produce the desired results, it does not
address the root cause of that problem he experienced.
The variable 'smp_num_cpus' was set by libprocps via a
sysconf(_SC_NPROCESSORS_ONLN) call. It was supposed to
represent total number of processors currently online.
It also served as the position in the Cpu_tics[] array
where the /proc/stat line #1 (cpu summary) was stored.
The variable 'Cpu_faux_tot' was valued by top based on
total individual cpus parsed from the /proc/stat file.
It serves as a fence post for Cpu_tics[] array access.
The problem Guido experienced results from a disparity
between those 2 variables, plus one instance where the
wrong variable was used in the summary_show() routine.
. Here is the real culprit, the actual incorrect code:
. summary_hlp(&Cpu_tics[Cpu_faux_tot], N_txt(WORD_a...
Which always should have been represented in this way:
. summary_hlp(&Cpu_tics[smp_num_cpus], N_txt(WORD_a...
------------------------------------------------------
The above 'disparity' might arise in any system when a
cpu is taken offline since there's a 3 second delay in
cpu and memory refreshes in an effort to reduce costs.
Usually this particular condition will be short lived.
However, there is a more persistent problem under lxc.
If a host cpu is taken offline and then brought online
again, within the container sysconf returns the proper
number of online processors. But, /proc/stat does not!
Sadly, I've yet to find a way to coax a container into
refreshing its /proc/stat, short of reboting the host.
[ might that represent a potential bug in lxc logic? ]
Jim Warner [Thu, 10 Jan 2019 06:00:00 +0000 (00:00 -0600)]
library: adapt readproc for the latest lxc conventions
The merge request shown below prompted (thankfully) an
examination of our lxc containers logic in readproc.c.
As it turns out, the lxc folks changed that eyecatcher
used to identify containers within a task cgroup file.
So this patch, with little extra cost, will enable the
libprocps lxc_containers() guy to handle both strings.
[ additionally, I was shocked to find lxc allows the ]
[ eyecatcher to be changed at ./configure time. such ]
[ a provision has always existed. unfortunately, the ]
[ changed value was only available to root, assuming ]
[ one wished to tackle that undocumented liblxc api. ]
Reference(s):
. what prompted lxc support reevaluation
https://gitlab.com/procps-ng/procps/merge_requests/82
. original lxc support introduced
commit 0557504f9cb84987f9d9038755404be017bdb7d1
Signed-off-by: Jim Warner <james.warner@comcast.net>
Jim Warner [Mon, 8 Oct 2018 05:00:00 +0000 (00:00 -0500)]
top: harden management of 'Hide_pid' array allocations
While setting the size of that Hide_pid array to equal
total pids high water mark was probably safe, in truth
there is no real relationship. At some point one could
exceed that HWM if the 'v' toggle was used extensively
and at least 1 of those entries remained non-negative.
This commit simply divorces Hide_tot from the pids HWM
and bases Hide_pid array size on actual run-time need.
Signed-off-by: Jim Warner <james.warner@comcast.net>
Jim Warner [Mon, 1 Oct 2018 05:00:00 +0000 (00:00 -0500)]
top: enable alternate '+' placement with collapsed pid
Currently, except for tasks that have no parents, when
a process' children are collapsed the '+' indicator is
shown in the first position within that COMMAND field.
This commit simply provides for indenting the '+' char
so it displays next to that program name/command line.
Signed-off-by: Jim Warner <james.warner@comcast.net>
Jim Warner [Tue, 18 Sep 2018 05:00:00 +0000 (00:00 -0500)]
top: plug a minor hole in the vertical scrolling logic
In that commit referenced below, a few edge cases were
addressed regarding vertical positioning involving any
'hidden' tasks. But, 2 additional edge cases remained.
In a running top, if the user employed 'other filters'
(o/O) or 'user filters' (u/U) proper vertical position
was not ensured. And, while this could be easily fixed
by striking the home/end or up/down arrow keys, it was
very poor etiquette to shift this burden to the users.
So, this patch plugs that gap, automating the process.
Jim Warner [Tue, 21 Aug 2018 05:00:00 +0000 (00:00 -0500)]
top: a tweak to the forest view collapsed code (again)
From the outset, top has tried to provide some minimal
garbage collection in support of forest view collapse.
For example, with every 'v' keystroke, a check is made
of the currently targeted pids. If all were negative,
which means expanded, that Hide_pid array was emptied.
Recently, yet another efficiency was added wherein the
continuing scan for a targeted pid was terminated when
a match was found. But, one more inefficiency existed.
When a task which was subject to collapse under forest
view mode has disappeared (ended), repeatedly scanning
for such a pid with each iteration makes little sense.
So this commit will negate such targeted pids and thus
avoid scanning every current task looking for a match.
Then, if 'v' is ever stuck at some point in the future
there will be a chance to empty that Hide_pid[] array.
[ hopefully this will be a final tweak of the forest ]
[ view collapse stuff, but cross your fingers anyway ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
Jim Warner [Tue, 14 Aug 2018 05:00:00 +0000 (00:00 -0500)]
top: the '#define PRETEND2_5_X' was found to be broken
Our newlib branch has already dropped support for such
old kernels. However, the master branch still supports
them. So this patch will correct a broken #define that
is used to influence the top Summary Area information.
Signed-off-by: Jim Warner <james.warner@comcast.net>
Jim Warner [Thu, 9 Aug 2018 05:00:00 +0000 (00:00 -0500)]
top: speed up the collapsed children forest view logic
In forest view mode, once a collapsible parent process
and all of its children (if any) have been identified,
there is no longer a need to scan the remaining tasks.
So this patch will just force a new scan for any other
'Hide_pid' entries which might remain to be identified
after a targeted parent has been completely processed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
Craig Small [Wed, 8 Aug 2018 10:13:58 +0000 (20:13 +1000)]
docs: Update ps.1 to warn about command name length
Previous versions of ps used to only match on the first 15 characters
because that's what the kernel used to provide. Newer kernels have a
longer length for this field so procps has been updated to suit.
Jim Warner [Wed, 25 Jul 2018 05:00:00 +0000 (00:00 -0500)]
top: existing 'Inspect' pipe feature now more flexible
Currently, it isn't possible to establish an 'Inspect'
pipe that relies on SIGINT to end. That's because this
signal will also end the parent process (top) as well.
So this patch will temporarily ignore that signal when
processing any 'Inspect' pipe, allowing one like this:
Upon startup there exists the potential for some minor
memory leakage should some rcfile 'Inspect' entries be
invalid. By delaying any malloc/strdup until after the
entries are completely validated we will prevent that.
Signed-off-by: Jim Warner <james.warner@comcast.net>
Jim Warner [Wed, 18 Jul 2018 05:00:00 +0000 (00:00 -0500)]
top: ensure collapsed children cpu reported accurately
Parent tasks with collapsed children should have their
cpu reflect any unseen tasks only under the following:
1) When built without TREE_VCPUOFF having been defined
2) Exclusively when 'Show_FOREST' display mode was set
3) And only under the current window when in alternate
display mode (except if TREE_VWINALL has been defined)
So, this commit just ensures these objectives are met.
Reference(s):
. issue that began odyssey
https://gitlab.com/procps-ng/procps/issues/99
. original cpu implementation
commit 3da7318683d2fea10526384e0a4368a378b486a5
Signed-off-by: Jim Warner <james.warner@comcast.net>
Jim Warner [Tue, 17 Jul 2018 05:00:00 +0000 (00:00 -0500)]
top: fix the collapsed children cpu segmentation fault
While that 'Hide_cpu' value will always be zero unless
there are collapsed children, the damn array will only
be present when a window's in 'Show_FOREST' view mode.
projects / procps-ng / log