Download Process Management and FreeBSD Process Structure - Prof. Stephen Chapin and more Study notes Operating Systems in PDF only on Docsity!
Process Management Process Management
CIS 657CIS 657
Definitions
Definitions
Process: program under executionProcess: program under execution
User mode:User mode: ““normalnormal”” execution modeexecution mode
Kernel mode: privileged execution mode; required for,Kernel mode: privileged execution mode; required for,
e.g., I/O operationse.g., I/O operations
User state: registers, PC, PSW, stack pointers, plusUser state: registers, PC, PSW, stack pointers, plus
memory contentsmemory contents
Kernel state:Kernel state:
Process structure (always resident)Process structure (always resident)
User structure (only resident when process is; very little left inUser structure (only resident when process is; very little left in
FreeBSD 5.2.1) FreeBSD 5.2.1)
Multiprogramming and Multiprogramming and
Multitasking Support
Multitasking Support
Context switching: changing the currentlyContext switching: changing the currently
running process (changing execution context)running process (changing execution context)
Hardware dependent; some ISA define instructions
Hardware dependent; some ISA define instructions
to save processor context; others must explicitly to save processor context; others must explicitly
save save
Kernel software state saved/loaded explicitlyKernel software state saved/loaded explicitly
Optimization is a very good idea
Optimization is a very good idea …
…
Scheduling: deciding which process to run nextScheduling: deciding which process to run next
Related to CS, in that the CS just picks the process
Related to CS, in that the CS just picks the process
at the head of the ready queue; scheduler maintains at the head of the ready queue; scheduler maintains
ready queue ready queue
Historical Comment Historical Comment
Early UEarly UNIXNIX didndidn’’t have multiprogrammingt have multiprogramming
support: only one process in memory at a timesupport: only one process in memory at a time
MMU support allowed multiple processes, butMMU support allowed multiple processes, but
synchronous disk I/O meant no multitasking
synchronous disk I/O meant no multitasking
(no automatically switching between(no automatically switching between
processes)processes)
In 1972 asynchronous disk I/O made itIn 1972 asynchronous disk I/O made it
possible to switch between processes, sopossible to switch between processes, so
UUNIXNIX got real multitasking.got real multitasking.
Scheduling
Scheduling
Processes haveProcesses have ““burstsbursts”” of activity (both CPUof activity (both CPU
bursts and I/O bursts)bursts and I/O bursts)
We characterize processes by the ratio of
We characterize processes by the ratio of
these bursts:
these bursts:
Lots of I/O, little CPU: I/O bound (interactive)Lots of I/O, little CPU: I/O bound (interactive)
Lots of CPU, little I/O: CPU bound (
Lots of CPU, little I/O: CPU bound ( noninteractive
noninteractive )
)
Real-time processes must haveReal-time processes must have subportionssubportions
completed by particular deadlines; think ofcompleted by particular deadlines; think of
factory-floor automation.factory-floor automation.
BSD Scheduling
BSD Scheduling
BSD uses a priority-based round robinBSD uses a priority-based round robin
scheme favoring interactive processesscheme favoring interactive processes
Processes have dynamic prioritiesProcesses have dynamic priorities
Jobs get a time slice
Jobs get a time slice
Complete your time slice: priority dropsComplete your time slice: priority drops
Give up the CPU before the end: priority
Give up the CPU before the end: priority
stays the same
stays the same
Inactive processes have priority increasedInactive processes have priority increased
5.2.1 FreeBSD Process Structure 5.2.1 FreeBSD Process Structure
Process
Structure
Thread list
process group
credential
VM space
file descriptors
resource limits
statistics
signal actions
process control block
process kernel stack
session
region list
file entries
syscall vector user structure
Thread List Entries
Thread List Entries
Pointer from
Process Struct
Scheduling info
Thread control block
Thread kernel stack
Machine-dependent
Thread information
Scheduling info
Thread control block
Thread kernel stack
Machine-dependent
Thread information
Process Structure Main Entries
Process Structure Main Entries
Process ID: process id and parentProcess ID: process id and parent
PIDPID
Signal state: signals pendingSignal state: signals pending
delivery, current mask, and actions
delivery, current mask, and actions
Tracing information
Tracing information
Real-time timer and CPU utilization
Real-time timer and CPU utilization
counterscounters
Process Structure SubstructuresProcess Structure Substructures
Process group ID: process group andProcess group ID: process group and
session for this processsession for this process
Credentials: real, effective, and savedCredentials: real, effective, and saved
UID and GIDUID and GID
Memory Management: virtual addressMemory Management: virtual address
space management (page tables)space management (page tables)
File descriptors: pointers to open file
File descriptors: pointers to open file
entries and
entries and
cwd
cwd
Process Structure Substructures II
Process Structure Substructures II
System call vector: FreeBSD supports systemSystem call vector: FreeBSD supports system
call numbering for Linux, OSF/1, SVR4.call numbering for Linux, OSF/1, SVR4.
Resource Accounting: system/user time used,
Resource Accounting: system/user time used,
memory allocation, paging and disk I/O, etc.memory allocation, paging and disk I/O, etc.
Statistics (kept in user structure): accountingStatistics (kept in user structure): accounting
informationinformation
Signal actions (kept in user structure): pointersSignal actions (kept in user structure): pointers
to signal handlers, list of blocked signals, etc.to signal handlers, list of blocked signals, etc.
Process States
Process States
NEW: process is partially createdNEW: process is partially created
NORMAL: process has resourcesNORMAL: process has resources
allocated for executionallocated for execution
RUNNABLE: threads ready to runRUNNABLE: threads ready to run
SLEEPING: threads awaiting eventSLEEPING: threads awaiting event
STOPPED: threads stopped by signal orSTOPPED: threads stopped by signal or
parent process parent process
ZOMBIE: process is partially terminatedZOMBIE: process is partially terminated
256 Priorities and 5 Classes256 Priorities and 5 Classes
224-255224-255IDLEIDLE Idle userIdle user
160-223160-223TIMESHARETIMESHARE Time-sharing userTime-sharing user
128-159128-159REALTIMEREALTIME Real-time userReal-time user
64-12764-127 KERNKERN Top half of kernelTop half of kernel
Bottom half of kernel
Bottom half of kernel
(interrupt)(interrupt)
ITHD
ITHD
RangeRange ClassClass TypeType
Meaning of
Meaning of
Priorities
Priorities
First, classes arenFirst, classes aren’’t really meaningfult really meaningful
Highest priority (lowest number) threadHighest priority (lowest number) thread
always* runs
always* runs
Bottom half > Top half >Bottom half > Top half > RealtimeRealtime >>
Timeshare > IdleTimeshare > Idle
rtpriortprio() system call sets priority for() system call sets priority for
RealtimeRealtime and Idle threads--not adjustedand Idle threads--not adjusted
by kernel
by kernel
User
User
vs
vs
. Kernel priority
. Kernel priority
kg_user_prikg_user_pri field holds current userfield holds current user
priority (that in the TIMESHARE class)priority (that in the TIMESHARE class)
td_prioritytd_priority may be differentmay be different
AA thread running in the kernel (e.g., afterthread running in the kernel (e.g., after
system call) has its priority temporarily system call) has its priority temporarily
changed to be in KERN class changed to be in KERN class
When asleep, and immediately afterWhen asleep, and immediately after
awakened awakened
Priority for service over user-priority threadsPriority for service over user-priority threads
Queues Queues
Run queues: hold threads ready to executeRun queues: hold threads ready to execute
Not a single ready queue;
Not a single ready queue; 64 queues
64 queues
All
All threads in same queue
threads in same queue are treated as same
are treated as same
priority priority
Sleep queues: hold threads waiting on events
Sleep queues: hold threads waiting on events
Hashed data structure, optimized for searching onHashed data structure, optimized for searching on
wait channel wait channel
In older versions of the OS, these heldIn older versions of the OS, these held
processes, notprocesses, not threadsthreads
Sleep() and Sleep Queues
Sleep() and Sleep Queues
LOOKUP(x) hashes an identifier to a sleep
LOOKUP(x) hashes an identifier to a sleep
queue (
queue ( kern_synch
kern_synch .c)
.c)
IdentIdent can be a void *,can be a void *, intint, or other (generally a, or other (generally a
pointer).pointer).
The sleep routines do not determine howThe sleep routines do not determine how
many identifiers there are in the systemmany identifiers there are in the system
Threads can give up the CPU by
Threads can give up the CPU by making a
making a
sleep() call (passing wait channel ID and
sleep() call (passing wait channel ID and
priority at which to be awakened)priority at which to be awakened)
Algorithm for Sleep
Algorithm for Sleep
1.1. AcquireAcquire sched_mutexsched_mutex locklock
Record the wait channel in the thread
Record the wait channel in the thread
structure, and hash the channel to find a
structure, and hash the channel to find a
sleep queuesleep queue
3.3. Set the threadSet the thread’’s prioritys priority for when it isfor when it is
awakened, set SLEEPING flagawakened, set SLEEPING flag
4.4. Place thread at end of sleep queuePlace thread at end of sleep queue
5.5. CallCall mi_switchmi_switch() to context switch() to context switch
((sched_locksched_lock automatically released)automatically released)
Process Priority Calculation Process Priority Calculation
From proc.h:From proc.h:
kg_estcpukg_estcpu provides estimate of recent CPU utilizationprovides estimate of recent CPU utilization
kg_nicekg_nice user-settable weighting factor betweenuser-settable weighting factor between – – 2020
and 20 (default 0)
and 20 (default 0)
Recalculate every 4 ticks:Recalculate every 4 ticks:
kg_user_prikg_user_pri = PRI_MIN_TIMESHARE += PRI_MIN_TIMESHARE +
((kg_estcpukg_estcpu / 4) + 2 x/ 4) + 2 x kg_nicekg_nice;;
kg_user_pri
kg_user_pri = max(
= max( kg_user_pri
kg_user_pri , PRI_MIN_TIMESHARE)
, PRI_MIN_TIMESHARE)
kg_user_prikg_user_pri = min(= min(kg_user_prikg_user_pri, PRI_MAX_TIMESHARE), PRI_MAX_TIMESHARE)
PRI_MIN_TIMESHAREPRI_MIN_TIMESHARE = 160,= 160, PRI_MAX_TIMESHAREPRI_MAX_TIMESHARE = 223= 223
Scheduling Calculations II
Scheduling Calculations II
kg_estcpukg_estcpu incremented every clock tick forincremented every clock tick for
the current processthe current process
kg_estcpu
kg_estcpu
aged (decayed) every second;
aged (decayed) every second;
forgets about 90% of the history in aforgets about 90% of the history in a
variable time periodvariable time period
Load_weight = (2 * load) / (2 * load + 1);Load_weight = (2 * load) / (2 * load + 1);
kg_estcpukg_estcpu = Load_weight *= Load_weight * kg_estcpukg_estcpu ++ kg_nicekg_nice;;
