zeldon
/
htop
mirror of https://github.com/xzeldon/htop.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
htop/linux/LinuxProcessList.h

140 lines
2.7 KiB
C
Raw Normal View History

Move platform-dependent parts of Linux battery meter.
2014-11-27 23:02:52 +00:00
#ifndef HEADER_LinuxProcessList
#define HEADER_LinuxProcessList
/*
htop - LinuxProcessList.h
(C) 2014 Hisham H. Muhammad
Released under the GNU GPL, see the COPYING file
in the source distribution for its full text.
*/
Updates to generated header files
2018-02-26 14:44:00 +00:00
#ifdef MAJOR_IN_MKDEV
Remove unnecessary HAVE_SYS_SYSMACROS_H check HAVE_SYS_SYSMACROS_H is always true if MAJOR_IN_SYSMACROS. This way of checking is recommended in autoconf 2.70 documentation: https://git.savannah.gnu.org/gitweb/?p=autoconf.git;a=blobdiff;f=doc/autoconf.texi;h=4f041bd4e;hp=9ad7dc1c5f02c8ba25b2fe1218bf931c7113a5d5;hb=e17a30e987d7ee695fb4294a82d987ec3dc9b974;hpb=565a6dc50cfa01cec2fb4db894026689cdf4970c NOTE: currently https://www.gnu.org/software/autoconf/manual/autoconf.html is the doc for autoconf 2.69.
2018-12-15 13:06:00 +00:00
#elif defined(MAJOR_IN_SYSMACROS)
Updates to generated header files
2018-02-26 14:44:00 +00:00
#endif
Adds support for linux delay accounting (#667) Adds support for showing columns with linux delay accounting. This information can be read from the netlink interface, and thus we set up a socket to read from that when initializing the LinuxProcessList (LinuxProcessList_initNetlinkSocket). After that, for each process we call LinuxProcessList_readDelayAcctData, which sends a message thru the socket after setting up a callback to get the answer from the Kernel. That callback sets the process total delay time attribute. We then set the delay percent as the percentage of time process cpu time since last scan.
2017-12-04 02:15:29 +00:00
#ifdef HAVE_DELAYACCT
#endif
Move platform-dependent parts of Linux battery meter.
2014-11-27 23:02:52 +00:00
#include "ProcessList.h"
Refactor common OpenZFS sysctl access Darwin and FreeBSD export zfs kstats through the same APIs, so moving functions into a common file.
2019-07-07 23:27:00 +00:00
#include "zfs/ZfsArcStats.h"
Move platform-dependent parts of Linux battery meter.
2014-11-27 23:02:52 +00:00
Linux: add process->starttime and use it for STARTTIME column (#700) this way a remount of /proc will not reset starttimes and we can also see startup times for processes started before the mount of /proc also record btime (boot time in seconds since epoch) as Linux semi-global
2018-08-19 04:29:03 +00:00
extern long long btime;
Sorry about the mega-patch. This is a work-in-progress, code is currently broken. (Some actions, and notably, the header, are missing.)
2015-01-22 01:27:31 +00:00
typedef struct CPUData_ {
unsigned long long int totalTime;
unsigned long long int userTime;
unsigned long long int systemTime;
unsigned long long int systemAllTime;
unsigned long long int idleAllTime;
unsigned long long int idleTime;
unsigned long long int niceTime;
unsigned long long int ioWaitTime;
unsigned long long int irqTime;
unsigned long long int softIrqTime;
unsigned long long int stealTime;
unsigned long long int guestTime;
Clean up existing whitespace
2019-10-31 16:39:12 +00:00
Sorry about the mega-patch. This is a work-in-progress, code is currently broken. (Some actions, and notably, the header, are missing.)
2015-01-22 01:27:31 +00:00
unsigned long long int totalPeriod;
unsigned long long int userPeriod;
unsigned long long int systemPeriod;
unsigned long long int systemAllPeriod;
unsigned long long int idleAllPeriod;
unsigned long long int idlePeriod;
unsigned long long int nicePeriod;
unsigned long long int ioWaitPeriod;
unsigned long long int irqPeriod;
unsigned long long int softIrqPeriod;
unsigned long long int stealPeriod;
unsigned long long int guestPeriod;
Move sysfs-reading code to LinuxProcessList.c and add average frequency. This way the frequency is read from sysfs only once per update cycle instead of every time the UI is redrawn. This also changes the code to read from /proc/cpuinfo instead. This is because reading from scaling_cur_freq stalls for 10ms if the previous read for the file was more than one second ago. [1] Since htop's update cycle is longer than that, it would cause the read of each CPU's scaling_cur_freq file to block the UI for 20ms. This easily led to a noticeable half-second lag on a 20+ CPU machine. /proc/cpuinfo also has a 10ms delay, but this applies for the whole file so the delay does not scale with the number of CPUs. [2] [1]: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=4815d3c56d1e10449a44089a47544d9ba84fad0d [2]: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=7d5905dc14a87805a59f3c5bf70173aac2bb18f8
2019-08-11 05:19:32 +00:00
double frequency;
Sorry about the mega-patch. This is a work-in-progress, code is currently broken. (Some actions, and notably, the header, are missing.)
2015-01-22 01:27:31 +00:00
} CPUData;
Interpret TTY_NR column on Linux, translate dev_t to major:minor on other platforms. Closes #316.
2016-10-01 06:09:04 +00:00
typedef struct TtyDriver_ {
char* path;
unsigned int major;
unsigned int minorFrom;
unsigned int minorTo;
} TtyDriver;
Sorry about the mega-patch. This is a work-in-progress, code is currently broken. (Some actions, and notably, the header, are missing.)
2015-01-22 01:27:31 +00:00
typedef struct LinuxProcessList_ {
ProcessList super;
Clean up existing whitespace
2019-10-31 16:39:12 +00:00
Sorry about the mega-patch. This is a work-in-progress, code is currently broken. (Some actions, and notably, the header, are missing.)
2015-01-22 01:27:31 +00:00
CPUData* cpus;
Interpret TTY_NR column on Linux, translate dev_t to major:minor on other platforms. Closes #316.
2016-10-01 06:09:04 +00:00
TtyDriver* ttyDrivers;
Linux: Use /proc/*/smaps_rollup for improved PSS parsing speed
2018-10-16 18:08:23 +00:00
bool haveSmapsRollup;
Clean up existing whitespace
2019-10-31 16:39:12 +00:00
Adds support for linux delay accounting (#667) Adds support for showing columns with linux delay accounting. This information can be read from the netlink interface, and thus we set up a socket to read from that when initializing the LinuxProcessList (LinuxProcessList_initNetlinkSocket). After that, for each process we call LinuxProcessList_readDelayAcctData, which sends a message thru the socket after setting up a callback to get the answer from the Kernel. That callback sets the process total delay time attribute. We then set the delay percent as the percentage of time process cpu time since last scan.
2017-12-04 02:15:29 +00:00
#ifdef HAVE_DELAYACCT
struct nl_sock *netlink_socket;
int netlink_family;
#endif
ZFS arcstats for Linux If no pools are imported (ARC size == 0) or the ZFS module is not in the kernel (/proc/spl/kstat/zfs/arcstats does not exist), then the Meter reports "Unavailable".
2019-07-07 02:37:02 +00:00
Refactor common OpenZFS sysctl access Darwin and FreeBSD export zfs kstats through the same APIs, so moving functions into a common file.
2019-07-07 23:27:00 +00:00
ZfsArcStats zfs;
Sorry about the mega-patch. This is a work-in-progress, code is currently broken. (Some actions, and notably, the header, are missing.)
2015-01-22 01:27:31 +00:00
} LinuxProcessList;
Move platform-dependent parts of Linux battery meter.
2014-11-27 23:02:52 +00:00
#ifndef PROCDIR
#define PROCDIR "/proc"
#endif
Move sysfs-reading code to LinuxProcessList.c and add average frequency. This way the frequency is read from sysfs only once per update cycle instead of every time the UI is redrawn. This also changes the code to read from /proc/cpuinfo instead. This is because reading from scaling_cur_freq stalls for 10ms if the previous read for the file was more than one second ago. [1] Since htop's update cycle is longer than that, it would cause the read of each CPU's scaling_cur_freq file to block the UI for 20ms. This easily led to a noticeable half-second lag on a 20+ CPU machine. /proc/cpuinfo also has a 10ms delay, but this applies for the whole file so the delay does not scale with the number of CPUs. [2] [1]: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=4815d3c56d1e10449a44089a47544d9ba84fad0d [2]: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=7d5905dc14a87805a59f3c5bf70173aac2bb18f8
2019-08-11 05:19:32 +00:00
#ifndef PROCCPUINFOFILE
#define PROCCPUINFOFILE PROCDIR "/cpuinfo"
#endif
Move platform-dependent parts of Linux battery meter.
2014-11-27 23:02:52 +00:00
#ifndef PROCSTATFILE
#define PROCSTATFILE PROCDIR "/stat"
#endif
#ifndef PROCMEMINFOFILE
#define PROCMEMINFOFILE PROCDIR "/meminfo"
#endif
ZFS arcstats for Linux If no pools are imported (ARC size == 0) or the ZFS module is not in the kernel (/proc/spl/kstat/zfs/arcstats does not exist), then the Meter reports "Unavailable".
2019-07-07 02:37:02 +00:00
#ifndef PROCARCSTATSFILE
#define PROCARCSTATSFILE PROCDIR "/spl/kstat/zfs/arcstats"
#endif
Interpret TTY_NR column on Linux, translate dev_t to major:minor on other platforms. Closes #316.
2016-10-01 06:09:04 +00:00
#ifndef PROCTTYDRIVERSFILE
#define PROCTTYDRIVERSFILE PROCDIR "/tty/drivers"
#endif
Extend buffer for reading lines from /proc. Apparently a line longer than 255 chars was spotted in the wild: http://serverfault.com/questions/577939/linux-ps-htop-show-processes-running-for-hundreds-or-thousands-of-days-though-h#comment676098_577939
2015-12-14 15:27:11 +00:00
#ifndef PROC_LINE_LENGTH
Linux: add process->starttime and use it for STARTTIME column (#700) this way a remount of /proc will not reset starttimes and we can also see startup times for processes started before the mount of /proc also record btime (boot time in seconds since epoch) as Linux semi-global
2018-08-19 04:29:03 +00:00
#define PROC_LINE_LENGTH 4096
Extend buffer for reading lines from /proc. Apparently a line longer than 255 chars was spotted in the wild: http://serverfault.com/questions/577939/linux-ps-htop-show-processes-running-for-hundreds-or-thousands-of-days-though-h#comment676098_577939
2015-12-14 15:27:11 +00:00
#endif
Introduce CLAMP macro. Unify all MIN(MAX(a,b),c) uses. With the CLAMP macro replacing the combination of MIN and MAX, we will have at least two advantages: 1. It's more obvious semantically. 2. There are no more mixes of confusing uses like MIN(MAX(a,b),c) and MAX(MIN(a,b),c) and MIN(a,MAX(b,c)) appearing everywhere. We unify the 'clamping' with a single macro. Note that the behavior of this CLAMP macro is different from the combination `MAX(low,MIN(x,high))`. * This CLAMP macro expands to two comparisons instead of three from MAX and MIN combination. In theory, this makes the code slightly smaller, in case that (low) or (high) or both are computed at runtime, so that compilers cannot optimize them. (The third comparison will matter if (low)>(high); see below.) * CLAMP has a side effect, that if (low)>(high) it will produce weird results. Unlike MIN & MAX which will force either (low) or (high) to win. No assertion of ((low)<=(high)) is done in this macro, for now. This CLAMP macro is implemented like described in glib <http://developer.gnome.org/glib/stable/glib-Standard-Macros.html> and does not handle weird uses like CLAMP(a++, low++, high--) .
2016-01-15 12:26:01 +00:00
#ifndef CLAMP
#define CLAMP(x,low,high) (((x)>(high))?(high):(((x)<(low))?(low):(x)))
#endif
Interpret TTY_NR column on Linux, translate dev_t to major:minor on other platforms. Closes #316.
2016-10-01 06:09:04 +00:00
Adds support for linux delay accounting (#667) Adds support for showing columns with linux delay accounting. This information can be read from the netlink interface, and thus we set up a socket to read from that when initializing the LinuxProcessList (LinuxProcessList_initNetlinkSocket). After that, for each process we call LinuxProcessList_readDelayAcctData, which sends a message thru the socket after setting up a callback to get the answer from the Kernel. That callback sets the process total delay time attribute. We then set the delay percent as the percentage of time process cpu time since last scan.
2017-12-04 02:15:29 +00:00
#ifdef HAVE_DELAYACCT
#endif
Re-generate all headers with latest scripts/MakeHeader.py Sync-up missing extern declarations for many functions.
2020-08-18 07:41:49 +00:00
extern ProcessList* ProcessList_new(UsersTable* usersTable, Hashtable* pidWhiteList, uid_t userId);
Move platform-dependent parts of Linux battery meter.
2014-11-27 23:02:52 +00:00
Re-generate all headers with latest scripts/MakeHeader.py Sync-up missing extern declarations for many functions.
2020-08-18 07:41:49 +00:00
extern void ProcessList_delete(ProcessList* pl);
Move platform-dependent parts of Linux battery meter.
2014-11-27 23:02:52 +00:00
#ifdef HAVE_TASKSTATS
#endif
#ifdef HAVE_OPENVZ
#endif
#ifdef HAVE_CGROUP
#endif
#ifdef HAVE_VSERVER
#endif
Adds support for linux delay accounting (#667) Adds support for showing columns with linux delay accounting. This information can be read from the netlink interface, and thus we set up a socket to read from that when initializing the LinuxProcessList (LinuxProcessList_initNetlinkSocket). After that, for each process we call LinuxProcessList_readDelayAcctData, which sends a message thru the socket after setting up a callback to get the answer from the Kernel. That callback sets the process total delay time attribute. We then set the delay percent as the percentage of time process cpu time since last scan.
2017-12-04 02:15:29 +00:00
#ifdef HAVE_DELAYACCT
#endif
Re-generate all headers with latest scripts/MakeHeader.py Sync-up missing extern declarations for many functions.
2020-08-18 07:41:49 +00:00
extern void ProcessList_goThroughEntries(ProcessList* super);
Move platform-dependent parts of Linux battery meter.
2014-11-27 23:02:52 +00:00
#endif