Rev 155 | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed
Rev 155 | Rev 156 | ||
---|---|---|---|
Line 1... | Line 1... | ||
1 | \chapter{Specification} |
1 | \chapter{Goals and Achievements} |
2 | 2 | ||
3 | \section{Overall Conception} |
3 | \section{Overall Conception} |
4 | 4 | ||
5 | General-purpose and portable operating system with elements of |
5 | General-purpose and portable operating system with elements of |
6 | microkernel design and fully preemptive kernel. |
6 | microkernel design and fully preemptive kernel. |
Line 32... | Line 32... | ||
32 | management and protection, hardware resource management, IPC). |
32 | management and protection, hardware resource management, IPC). |
33 | Device drivers, filesystems, network stacks, etc. will be |
33 | Device drivers, filesystems, network stacks, etc. will be |
34 | implemented in user-space. |
34 | implemented in user-space. |
35 | \end{itemize} |
35 | \end{itemize} |
36 | 36 | ||
- | 37 | {\em The overall conecption of the kernel design was completely met. The kernel |
|
- | 38 | is fully preemptible, SMP ready with fine-grained locking. If possible, |
|
- | 39 | device drivers are implemented as standalone userspace tasks. HelenOS |
|
- | 40 | fully supports statically linked tasks. Both userspace tasks and kernel |
|
- | 41 | tasks are supported (N:M multithreading model). |
|
- | 42 | ||
- | 43 | The kernel was successfully ported to 5 architectures with one other |
|
- | 44 | architecture to come. The interfaces in the kernel are designed in such |
|
- | 45 | a way to fully utilize specifics of every platform, e.g. ASID and RID |
|
- | 46 | allocation in MIPS and IA64, two stacks for IA64 and SMP routines. |
|
- | 47 | } |
|
- | 48 | ||
37 | \section[1a]{Research Domains} |
49 | \section[1a]{Research Domains} |
38 | 50 | ||
39 | Following features can be eventually implemented as research subjects, |
51 | Following features can be eventually implemented as research subjects, |
40 | but are optional to the overall design of the system: |
52 | but are optional to the overall design of the system: |
41 | |
53 | |
42 | \begin{itemize} |
54 | \begin{itemize} |
43 | \item Kernel-level virtualization: Apart from some standard security |
55 | \item Kernel-level virtualization: Apart from some standard security |
44 | model (i.e. unix-like or any other) the OS might support |
56 | model (i.e. unix-like or any other) the OS might support |
45 | kernel-level context separation allowing to run more virtual |
57 | kernel-level context separation allowing to run more virtual |
46 | operating environments on a single physical machine. |
58 | operating environments on a single physical machine. |
- | 59 | ||
- | 60 | {\em Kernel-level virtualization was not attempted, although the microkernel |
|
- | 61 | design by itself allows completely different namespace simply by connecting |
|
- | 62 | the task to different name service daemon. Because new IPC connections can |
|
- | 63 | be created only through existing paths in the graph of the connections, |
|
- | 64 | messages can never flow between unconnected components of the graph. |
|
- | 65 | } |
|
47 | \item Framework for running GNU/Linux applications: There should be no |
66 | \item Framework for running GNU/Linux applications: There should be no |
48 | syscall or native API compatibility, but rather some kind of |
67 | syscall or native API compatibility, but rather some kind of |
49 | compile-time layer (libc and other shared libraries) allowing to |
68 | compile-time layer (libc and other shared libraries) allowing to |
50 | compile common GNU/Linux applications from sources. |
69 | compile common GNU/Linux applications from sources. |
- | 70 | ||
- | 71 | {\em Two applications were ported with little effort - libpci and tetris. The |
|
- | 72 | porting of the tetris consisted mainly in rewriting termios dependent |
|
- | 73 | code. The libc library contains emulation layer for the most common functions.} |
|
- | 74 | ||
51 | \item Object/message paradigm: In the contrary to Unix file paradigm |
75 | \item Object/message paradigm: In the contrary to Unix file paradigm |
52 | (where every object in the system is represented by a file - even |
76 | (where every object in the system is represented by a file - even |
53 | if there is no consistent mapping from the given object's methods |
77 | if there is no consistent mapping from the given object's methods |
54 | to generic file methods), HelenOS might have a tree of objects |
78 | to generic file methods), HelenOS might have a tree of objects |
55 | instead of a tree of files. Each object in the tree can support an |
79 | instead of a tree of files. Each object in the tree can support an |
56 | arbitrary set of messages and files are those objects which |
80 | arbitrary set of messages and files are those objects which |
57 | support the set of messages representing file methods (i.e. open, |
81 | support the set of messages representing file methods (i.e. open, |
58 | close, read, write, seek, etc.). All objects might support several |
82 | close, read, write, seek, etc.). All objects might support several |
59 | compulsory messages (GetName, GetSupportedMessages, etc.). The |
83 | compulsory messages (GetName, GetSupportedMessages, etc.). The |
60 | message passing mechanism will be synchronous. |
84 | message passing mechanism will be synchronous. |
- | 85 | ||
- | 86 | {\em Every IPC message contains a field that specifies method number. However, |
|
- | 87 | tree of objects or any more complex functionality were not implemented. } |
|
61 | \end{itemize} |
88 | \end{itemize} |
- | 89 | {\em However, because we have decided to use asynchronous message passing, |
|
- | 90 | a framework was needed to facilitate reasonably synchronous application view. |
|
- | 91 | This framework, heavily using userspace thread switching, allows writing |
|
- | 92 | transparent applications without the hassle usually connected with |
|
- | 93 | asynchronous applications, at the same time being easily portable to |
|
- | 94 | kernel-threaded environment. } |
|
62 | 95 | ||
63 | \section{Particular features} |
96 | \section{Particular features} |
64 | 97 | ||
65 | \begin{itemize} |
98 | \begin{itemize} |
66 | \item Kernel features |
99 | \item Kernel features |
67 | \begin{itemize} |
100 | \begin{itemize} |
68 | \item Preemptive multiprocessing, SMP support, threads (tasks) |
101 | \item Preemptive multiprocessing, SMP support, threads (tasks) |
69 | \begin{itemize} |
102 | \begin{itemize} |
70 | \item Simple scheduler (but more complex than round-robin), |
103 | \item Simple scheduler (but more complex than round-robin), |
71 | with threads as basic scheduling element |
104 | with threads as basic scheduling element. {\em Achieved.} |
72 | \item Support for thread priorities (possibly classes of |
105 | \item Support for thread priorities (possibly classes of |
73 | priorities for user-space tasks) |
106 | priorities for user-space tasks). {\em Achieved} |
- | 107 | ||
74 | \item Support for SMP CPU bounding |
108 | \item Support for SMP CPU bounding. {\em Achieved.} |
75 | \item Utilization of non-boot CPU(s) |
109 | \item Utilization of non-boot CPU(s). {\em Achieved.} |
76 | \item Support for user-space threads (tasks as sets of |
110 | \item Support for user-space threads (tasks as sets of |
77 | threads) |
111 | threads). {\em Achieved.} |
- | 112 | ||
78 | \item Support for kernel threads (independent code executed |
113 | \item Support for kernel threads (independent code executed |
79 | within the kernel) |
114 | within the kernel) {\em Achieved.} |
80 | \end{itemize} |
115 | \end{itemize} |
81 | \item Kernel synchronization primitives, small granularity |
116 | \item Kernel synchronization primitives, small granularity |
82 | synchronization (preemptive kernel) |
117 | synchronization (preemptive kernel) |
83 | \begin{itemize} |
118 | \begin{itemize} |
84 | \item Semaphores, mutexes, condition variables, RW-locks, |
119 | \item Semaphores, mutexes, condition variables, RW-locks, |
85 | spin-locks, etc. |
120 | spin-locks, etc. {\em Achieved.} |
86 | \item No "big kernel lock" |
121 | \item No "big kernel lock". {\em Achieved.} |
87 | \end{itemize} |
122 | \end{itemize} |
- | 123 | ||
88 | \item Physical and virtual memory management |
124 | \item Physical and virtual memory management |
89 | \begin{itemize} |
125 | \begin{itemize} |
90 | \item Proper handling of physical memory regions |
126 | \item Proper handling of physical memory regions. {\em Achieved.} |
- | 127 | |
|
91 | \item Physical memory heap (allocating of continuous blocks of |
128 | \item Physical memory heap (allocating of continuous blocks of |
92 | physical memory) |
129 | physical memory). {\em Achieved.} |
93 | \item Arbitrary number of independent virtual memory mappings |
130 | \item Arbitrary number of independent virtual memory mappings |
94 | (both for threads and internal kernel usage) |
131 | (both for threads and internal kernel usage). {\em Achieved.} |
95 | \item Kernel allocator in virtual memory (buddy/slab) |
132 | \item Kernel allocator in virtual memory (buddy/slab). {\em Achieved.} |
96 | \item Named (text, stack, heap) and unnamed virtual memory |
133 | \item Named (text, stack, heap) and unnamed virtual memory |
97 | areas |
134 | areas. {\em Achieved.} |
98 | \item Copying and sharing pages between different memory |
135 | \item Copying and sharing pages between different memory |
99 | mappings |
136 | mappings. {\em Achieved.} |
100 | \end{itemize} |
137 | \end{itemize} |
101 | \item Basic hardware handling |
138 | \item Basic hardware handling |
102 | \begin{itemize} |
139 | \begin{itemize} |
103 | \item Handling of basic boot-time hardware (CPU, PCI buses, |
140 | \item Handling of basic boot-time hardware (CPU, PCI buses, |
104 | memory, display, keyboard, RTC, etc.) in kernel |
141 | memory, display, keyboard, RTC, etc.) in kernel. {\em Achieved.} |
105 | \item Handling of specific hardware resources which are |
142 | \item Handling of specific hardware resources which are |
106 | fundamentaly unreachable from user-space on given |
143 | fundamentaly unreachable from user-space on given |
107 | platform |
144 | platform. {\em Achieved.} |
108 | \end{itemize} |
145 | \end{itemize} |
109 | \item IPC, user-space hardware access framework |
146 | \item IPC, user-space hardware access framework |
110 | \begin{itemize} |
147 | \begin{itemize} |
111 | \item Abstraction for implementing inter-process communication |
148 | \item Abstraction for implementing inter-process communication |
112 | (message passing, etc.) |
149 | (message passing, etc.). {\em Achieved.} |
113 | \item Interface for enabling the user-space threads to gain |
150 | \item Interface for enabling the user-space threads to gain |
114 | access and manage hardware resources (with kernel |
151 | access and manage hardware resources (with kernel |
115 | modules where needed) |
152 | modules where needed). {\em Achieved.} |
116 | \end{itemize} |
153 | \end{itemize} |
117 | \item User-space features |
154 | \item User-space features |
118 | \begin{itemize} |
155 | \begin{itemize} |
119 | \item Basic API |
156 | \item Basic API |
120 | \begin{itemize} |
157 | \begin{itemize} |
121 | \item Memory management API (memory regions creation, |
158 | \item Memory management API (memory regions creation, |
122 | descruction, resizing) |
159 | descruction, resizing). {\em Achieved.} |
123 | \item Task/thread management API |
160 | \item Task/thread management API. {\em Achieved.} |
124 | \item Synchronization API |
161 | \item Synchronization API. {\em Futexes implemented.} |
125 | \end{itemize} |
162 | \end{itemize} |
126 | \end{itemize} |
163 | \end{itemize} |
127 | \end{itemize} |
164 | \end{itemize} |
128 | \end{itemize} |
165 | \end{itemize} |
129 | 166 | ||
Line 133... | Line 170... | ||
133 | \item Supported platforms |
170 | \item Supported platforms |
134 | \begin{itemize} |
171 | \begin{itemize} |
135 | \item Real hardware support |
172 | \item Real hardware support |
136 | \begin{itemize} |
173 | \begin{itemize} |
137 | \item IA-32 (will be tested on multiple consumer Intel Pentium~4, |
174 | \item IA-32 (will be tested on multiple consumer Intel Pentium~4, |
138 | Intel Pentium~M, AMD Athlon~XP and AMD Athlon~MP machines) |
175 | Intel Pentium~M, AMD Athlon~XP and AMD Athlon~MP machines) |
- | 176 | ||
- | 177 | {\em Runs on comodity hardware. Tested on several multiprocessor computers.} |
|
139 | \item PowerPC (will be tested on a consumer IBM PowerPC G5 machine) |
178 | \item PowerPC (will be tested on a consumer IBM PowerPC G5 machine) |
- | 179 | {\em To some extent runs on the G4 machine. G5 machine is a 64-bit architecture completely different from 32-bit port that was attampted. } |
|
140 | \end{itemize} |
180 | \end{itemize} |
141 | \item Emulated support |
181 | \item Emulated support |
142 | \begin{itemize} |
182 | \begin{itemize} |
143 | \item MIPS (will be tested in MSIM R4000 simulator) |
183 | \item MIPS (will be tested in MSIM R4000 simulator) |
- | 184 | {\em Tested in msim, gxemul and partially in simics simulators. Booted kernel on SGI Indy, however no real hardware input/output support was attempted.} |
|
144 | \item IA-64 (will be tested in Ski simulator) |
185 | \item IA-64 (will be tested in Ski simulator) |
- | 186 | {\em Tested in Ski simulator.} |
|
145 | \item AMD64 (will be tested in Simics simulator) |
187 | \item AMD64 (will be tested in Simics simulator) |
- | 188 | {\em Tested on single-processor computer. Runs in simics, bochs and qemu simulators.} |
|
146 | \end{itemize} |
189 | \end{itemize} |
147 | \end{itemize} |
190 | \end{itemize} |
148 | \end{itemize} |
191 | \end{itemize} |