Commit Graph

16 Commits

Author SHA1 Message Date
Jim Warner
86b155b83f library: make the pids_sort_order enums more intuitive
The values of PROCPS_SORT_ASCEND & PROCPS_SORT_DESCEND
were a tad unintuitive. This patch will just make them
a more natural +1 for ascending and -1 for descending.

[ plus it still allows that fast path multiplication ]
[ instead of a comparison for signed numbers/strings ]

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-09-21 22:39:40 +10:00
Jim Warner
007e032805 library: introduce a fatal 'proc not mounted' function
Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-09-21 22:38:38 +10:00
Jim Warner
42af8ffbc1 library: enable an 'invariant' stack for that pids API
After peeking at possible conversion of the ps program
it appeared that we may ultimately need the concept of
a 'static' pids_stack in suport of look_up_our_self().
In other words, one not impacted by procps_pids_reset.

This patch simply sets the stage for that possibility.

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-09-21 22:37:40 +10:00
Jim Warner
f0e447cc66 library: exchange <pids> 'addr' for that 'ul_int' type
Work on converting ps has revealed the desirability of
trading a void pointer for that ul_int type. There was
much arithmetic employed against such values and casts
would otherwise have been required. Even pmap needed a
cast on occasions when comparing an internal variable.

Besides, there is much to be said for reducing demands
on (and the complexity of) the result structure union.

[ we choose ul_int over ull_int since that former is ]
[ the exact same size and capacity as a void pointer ]
[ regardless of whether compiled as 32-bit or 64-bit ]

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-09-21 22:36:05 +10:00
Jim Warner
17d77f8016 library: add PROCPS_PIDS_TTY_NAME to compliment number
OK, ok, this was kind of a huge omission. So please do
not select the TTY field for display in top quite yet,
at least until a next patch has been pushed to GitLab.

And to produce a correct sort order for this new field
the GNU 'strverscmp' routine was a necessary addition.

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-09-09 21:22:43 +10:00
Jim Warner
5e46a145ba library: ensure any 'flags' is consistently 'unsigned'
This commit is an outgrowth of the research into a bug
that recently surfaced with the 'w' program. And while
that program was just a victim several inconsistencies
were found in the handling of library flags during the
research. This patch just address such irregularities.

Reference(s):
http://www.freelists.org/post/procps/newlib-at-the-precipice,4

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-09-07 18:11:48 +10:00
Jim Warner
c39f4735c0 library: swat (now) obvious bug from PROCPS_PIDS_extra
I can't remember during what sleepy hour that enum was
added to pids.h, but it damn sure fell short of a full
implementation by approximately 1,000,000 miles or so.

Strangely it surfaced on Craig's system seemingly ggdb
related whereas on mine, a segmentation fault was only
seen when that -h (no header) argument was being used.

Oh well, the road to its resolution also offered us an
opportunity to cleanup some other items along the way.

Reference(s):
http://www.freelists.org/post/procps/newlib-at-the-precipice,4

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-09-07 18:11:48 +10:00
Jim Warner
1b38b751ff library: more miscellaneous pids implementation tweaks
A patch containing the following miscellaneous tweaks:

. make a supposedly robust parameter test truly robust
[ ensure the largest enum value used with validation ]

. remove duplicate item test in cleanup_stack function
[ is already subordinate to test of PROCPS_PIDS_noop ]

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-09-03 22:37:50 +10:00
Craig Small
a61f78d6e0 library: rework namespace calls
Functions related to namespaces were half-in half-out of the
procps library and didn't fit the standard naming scheme.

While struct { long ns[x]} is a bit clunky, its the only way
to "lock in" x. The alternative is to use ns_* variables.

This work was needed before pgrep could be converted.
2015-09-03 22:32:19 +10:00
Jim Warner
eba58ec17a library: beef up 'enum pids_item' parameter validation
I was surprised to find that ol' gcc silently converts
a single (different) enum into an address where one or
more enums were expected to be dereferenced. Of course
this was just yet another way to generate an old SEGV.

So this commit will strengthen those parameter checks.

[ we will *not* blame Craig for a failure to consult ]
[ the documentation, since it doesn't even exist yet ]

Reference(s):
http://www.freelists.org/post/procps/newlib-ps-fix,8

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-08-31 17:49:19 +10:00
Jim Warner
e169b41a85 library: accumulated changes to <pids> code & comments
A patch containing the following miscellaneous tweaks:

. avoided distortions unique to PROCPS_PIDS_TICS_DELTA
. addressed several smatch warnings and/or suggestions
. ensured oldproc_close invoked should tally_proc fail
. keeping that namespace clean, added a missing #undef
. added 2 comments acknowledging pids_item as unsigned
. added/clarified comments regarding proc flags & strv

From smatch analysis:
. these were indeed boo-boos
pids.c:580 make_hist() warn: the 'Hr' macro might need parens
pids.c:1058 procps_pids_reap() warn: variable dereferenced before check 'info' (see line 1056)
. these were not errors (and we did double check)
pids.c:1067 procps_pids_reap() warn: double check that we're allocating correct size: 8 vs 128
pids.c:1068 procps_pids_reap() warn: double check that we're allocating correct size: 8 vs 128

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-08-31 17:46:53 +10:00
Jim Warner
f69c421295 library: add real string vectors to the pids interface
After experimenting with an adaptation of pidof to the
new pids interface, it became apparent that vectorized
versions of those command lines would be necessary. So
this commit adds that option and the strv result type.

And since the stage had been set, a vectorized version
of PROCPS_PIDS_ENVIRON & PROCPS_PIDS_CGROUP was added.

Lastly, any use of 'const' in the result structure was
removed so callers need not be bothered with casts and
compiler warnings. Hopefully, they'll respect a stack.

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-08-27 22:21:30 +10:00
Jim Warner
9ecc0c8473 library: add sequential access to that pids collection
To ease the transition to the new interface, for other
than that top program, individual read provisions have
been added to the <proc/pids.h> API. This represents a
refinement of a position stated in a post noted below.

Reference(s):
http://www.freelists.org/post/procps/newlib-ps-fix

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-08-27 22:20:51 +10:00
Jim Warner
def5858810 library: fix major oops in procps_pids_reap() function
In my zeal to finalize the initial pids implementation
I omitted some quite important parameter checking from
the above function. Thank goodness top was kind to us.

Also, in anticipation of the additions of single stack
read and supporting functions some items were renamed.

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-08-27 22:19:54 +10:00
Jim Warner
387d4030f4 library: miscellaneous tweaks to pid code and comments
. traded a complex misaligned memory allocation scheme
in the make_hist function for a simple aligned scheme.
plus memory allocation increases are globally defined.

. changed 1 parameter for procps_pids_stacks_sort() to
better reflect the 'array of pointers', not an address
of a pointer as is used with guys such as 'new/unref'.

. the pids_reap struct was changed slightly to make it
more reflective of it's actual implementation details.

. the Item_table member .mustfree is now .needfree and
that .makehist was now made .needhist for consistency.

. reduced the number of separate 'return NULL;' source
statements in that primary procps_pids_reap() routine.

. ensured consistent reference to sizeof(void *) & not
occasional reference to sizeof(void*) without a space.

. rather than enable/disable validate_stacks via a #if
in the function body, it is now handled via a #define.

. some comments in the procps_pids_reset function were
adjusted to reflect this current implementation. shown
originally, they reflected an aborted attempt to avoid
a testing aberration not fully understood at the time.

. added a summary of the memory overhead cost of HST_t
processing to that UNREF_RPTHASH output at unref time.

. a 'PIDs at max depth:' portion of that UNREF_RPTHASH
enabled #define is now published only when the maximum
depth of hash table entry chains exceed depths of one.

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-08-23 21:08:18 +10:00
Jim Warner
7e6a371d8a library: implement task/thread support via the new api
This commit is the culmination of efforts to modernize
the library api. It should be treated as a first blush
attempt, especially since I have absolutely no library
design experience. But I did have a very strong desire
to lessen the new library's impact on the top program.

Under this new api, a 'stack' is the equivalent of the
old proc_t. It can be seen as a variable length record
whose contents & order is under complete user control.

That initial stack/record configuration is established
at procps_pids_new() time and will probably serve most
program needs. But, a dynamic & demanding program like
top will later change a stack via procps_pids_reset().

For programs like top & ps, procps_pids_reap() will be
the function that will retrieve all tasks and threads.

Any program that needs to filter / select only certain
processes or users have available other functions that
can be used: procps_pids_stacks_alloc, fill & dealloc.

This implementation attempts to maximize that existing
proven libprocps code base. As we gain more experience
such actual code can be migrated into the pids.c file.

Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-08-23 21:05:06 +10:00