MICRO-KERNELS

1
MICRO-KERNELS

• New Generation Innovation

2
The contents

• Traditional OS view its problems.
• Micro-kernel introduction to concept.
• First generation micro-kernels.
• Obvious advantages.
• Some Implementation problems.
• Second generation micro-kernels.
• L4 Kernel.
• Exo-Kernel.
• SPIN OS.

3
The contents (contd)
4
Traditional OS view

• Prerequisite component TCB
• Kernel is OS part of TCB.
• Kernel is monolithic.(includes sched, fs, n/w,
  mm, dd etc..)
• Kernel abstracts protects system
  resources.(PM?VM, DISK?FILES, CPU?processes)

5
Advantages of this kernel

• Portable interface to underlying machine.
• Large, generic, default functionality
  base.(Applications need not worry about device
  drivers or memory management policies )
• Provides protection.(kernel controls use of all
  resources.)

6
Serious disadvantages

• Huge kernel size.
• ONLY privileged ones can access and manage system
  resources. (kernel itself, or servers.)
• Applications are forced to use implementation
  of these privileged ones.
• This generic implementation is bound to be
  imperfect for specific application needs.
• Why ?
• Because of typical performance/cost tradeoffs.

7
What to do ? 1st generation micro-kernels.

• Remove un-wanted components.
• Implement all services as external servers.(fs,
  mm, dd, n/w etc )
• All servers run in user mode.

8
What are the benefits ?

• Smaller kernel. (hand-held devices ?)
• OS is more modular, flexible, extensible and
  customizable.
• More than one implementation possible of various
  services. (can co-exist).
• May even run concurrently, if needed.
• Easy crash recovery (only servers crash)

9
Minimal implementation primitives required.

• Assumptions
• Support for un-trustworthy applications
• Page based virtual memory implementation by h/w.
• Issues for such a system
• Protection scheme.
• Guarantee of Independence.
• Interrupt handling.
• IPC

10
Implementation

• Two concepts solves all above problems the
  concept of Address spaces threads.
• Traditional address space mapping associating
  virtual page ? physical page.
• New Initially, only 1 global address space,
  representing the physical memory.
• New address spaces can be recursively constructed
  outside the kernel.

11
Address space construction

• For this support, 3 operations provided in the
  micro-kernel.
• GRANT, MAP, FLUSH.
• Using these kernel provided primitives, memory
  management and paging schemes can be implemented
  outside kernel.

12
Definitions of Primitives

• GRANT
• Owner can grant its pages to anyone else,
  provided receiver agrees to take it.
• Granted page removed from granters addr space.
• MAP
• Owner can map its pages to anyone elses
  address space, provided receiver agrees for it.
• Page can now be accessed in both addr spaces.

13
Definitions contd

• FLUSH
• Owner can flush any of its pages.
• Flushed pages remain accessible to owner.
• But once flushed, the page is removed from any
  other address space (non-owner), where it may
  possibly have been mapped.
• Note that owner do NOT need explicit permission
  from other mappers before flushing.
• Idea is that anyway page belongs to self only,
  and users of this page had already accepted
  this .

14
Where are these reqd.?

• MAP FLUSH used for implementing memory managers
  and pagers on top of micro-kernel.
• GRANT only used in special circumstances
• When page mappings should be passed through a
  controlling sub-system without burdening the
  controllers address space.
• (See figure below for explanations.)

15
Explaining primitives
User A
User N
grant
Pager - F
map
Pager f2 (fs)
Pager f1 (fs)
map
disk
Standard pager
16
Threads concept

• Micro-kernel support the notion of threads at the
  lowest level of abstraction, within the kernel.
• A thread t is an activity executing inside an
  address-space as.
• t is characterized by registers, stack-ptr,
  state information (which as it belongs to)

17
How this satisfies everything ?

• Protection Independence are obvious due to very
  structure of separate AS concept.
• IPC can be realized by transferring messages
  between kernel threads. Note that this forms
  the basis of communication and is used for
  implementing operations like map and grant.
• Interrupts are treated as IPC messages. Hardware
  is regarded as set of threads.

18
Questions / thoughts ?

• How can this model help provide controlled
  access? think of ACLs. !!
• Provide access rights concept on pages.
• Map/Grant can copy/move sources rights.
• Flushing can revoke permissions.

19
What can be built on top ?

• Memory manager (stacked ?)
• Pager can implement virtual mem., resident
  memory for device drivers etc.
• Device drivers.
• They will directly access hardware I/O ports
  which are mapped into its address space.
  Messages are recvd from h/w (interrupts) through
  IPC.
• Secondary cache / TLB handlers.
• R-IPC (RPC)
• UNIX server ( system calls.)

20
1st generation problems

• Performance was terrible.
• IPC costs (200 sec vs. 40 sec)(All user level
  servers would be accessed by RPC mechanism.)
• High MCPI costs (memory cycle per inst) Even
  upto .25 CPI (additional). !!
• Non-Portability.
• WHY ?

21
MCPI problems due to

• Increased cache misses due to
• Worse locality properties of the combined
  micro-kernel OS code.
• System self-interference, incorrectly
  invalidating cache lines, due to more modularity
  of the OS.
• More inter-module copying due to the higher
  modularity of the OS.

22
Why Non-Portability ?

• Problems with Portable kernels -
• Can not take advantage of specific hardware
• Can not take precautions to circumvent or avoid
  perfromance problems of specific hardware
• Additional layer ( abstraction of hardware)
  costs performance.
• WE SHOULD ACCEPT THAT micro-kernel is going to be
  h/w dependent.

23
Example 486 Vs Pentium
486 Pentium
TLB entries, ways 32 32i 64d
Cache size Line, thru 8K 16B, through 8Ki 8Kd 32B, back
Fast instr. Segmt reg. 1 cycle 9 cycles 0.5-1 cycle 3 cycles.
24
Differences

• Pentium micro-kernel will use segment registers
  for implementing user address space
• 486 will have to use conventional
  hardware-address-space switch because of
  expensive segment register operations.
• This is Due to points 2 3 in table above.

25
Conclusions

• Micro-kernel design offers great flexibility and
  innovation in way world percieves OS.
• Micro-kernels must be developed per-processor,
  but performance achieved is probably worth the
  effort.
• More work needed in this area.
• STEP TOWARDS 2nd generation micro-kernels.