Home | History | Annotate | Line # | Download | only in roadmaps
system revision 1.8
      1 $NetBSD: system,v 1.8 2009/11/17 21:09:54 agc Exp $
      2 
      3 NetBSD System Roadmap
      4 =====================
      5 
      6 This is a small roadmap document, and deals with the main system
      7 aspects of the operating system.
      8 
      9 The following projects are already in NetBSD-current and will appear
     10 in NetBSD 6.0:
     11 
     12 1. 64-bit time values supported
     13 2. Better Kernel Modules support
     14 5. namei() tactical changes
     15 10. emap (ephemeral mapping) support for i386 and amd64
     16 11. support for mDNSResponder
     17 
     18 The following projects are expected to be included in NetBSD 6.0
     19 
     20 3. Full kernel preemption for real-time threads
     21 4. POSIX shared memory
     22 6. Better resource controls
     23 7. Improved observability: online crashdumps, remote debugging
     24 8. Processor and cache topology aware scheduler
     25 9. namei() strategic changes
     26 
     27 We currently expect to branch 6.0 in the March 2010 timeframe, with a view
     28 to a 6.0 release later in 2010.
     29 
     30 We'll continue to update this roadmap as features and dates get firmed up.
     31 
     32 
     33 Some explanations
     34 =================
     35 
     36 1. 64-bit time_t support
     37 -------------------------
     38 
     39 The Unix 32-bit time_t value will overflow in 2037 - any mortgage calculations
     40 which use a time_t value are in danger of overflowing at the present time -
     41 and to address this, 64-bit time_t values will be used to contain the number
     42 of seconds since 1970. This was completed in 5.99.7 by Christos (with no libc
     43 major bump - kudos and respect), and will be in 6.0.
     44 
     45 Responsible: christos
     46 
     47 2. Better Kernel Module Support
     48 -------------------------------
     49 
     50 Starting with 5.99.2, the kernel support for modules was enhanced by
     51 ad, and GENERIC was switched over to be a MODULAR kernel.  Support
     52 from booting from modules, like ffs, was introduced at the same time. 
     53 Some work has been done by Luke Mewburn in this area to define module
     54 locations and paths so that effective kernel development can be done
     55 using modules. Kernel modules have been moved out of base.tgz and a new
     56 modules.tgz has been created.
     57 
     58 Responsible: ad, lukem
     59 
     60 3. Full kernel preemption for real-time threads on non-x86
     61 ----------------------------------------------------------
     62 
     63 With the revamp of the kernel concurrency model, much of the kernel is
     64 fully multi-threaded and can therefore be preempted at any time.  In
     65 support of lower context switch and dispatch times for real-time
     66 threads, full kernel preemption is being implemented.  This has been
     67 implemented already for i386 and x86_64 (and is in 5.0), but needs to
     68 be extended to support the ARM, MIPS and SuperH ports.
     69 
     70 Responsible: rmind
     71 
     72 4. POSIX shared memory
     73 ----------------------
     74 
     75 Implement POSIX shared memory facilities, which can be used to create the
     76 shared memory objects and add the memory locations to the address space of
     77 a process.
     78 
     79 Responsible: rmind
     80 
     81 5. Incremental namei improvements, Phase 1
     82 ------------------------------------------
     83 
     84 In NetBSD 5.99.15, some changes were made to split the namei() routine
     85 up into logical parts, so that changes can be made to the constituent
     86 parts in a less intrusive way. This is in the repository now, and will
     87 be in 6.0.
     88 
     89 Responsible: dholland
     90 
     91 6. Better resource controls
     92 ---------------------------
     93 
     94 A resource provisioning and control framework that extends beyond the
     95 traditional Unix process limits.
     96 
     97 7. Improved observability: online crashdumps, remote debugging
     98 --------------------------------------------------------------
     99 
    100 XXX crashdumps while the system is running
    101 XXX firewire support in libkvm
    102 
    103 8. Processor and cache topology aware scheduler
    104 -----------------------------------------------
    105 
    106 Implement the detection of the topology of the processors and caches. 
    107 Improve the scheduler to make decisions about thread migration
    108 according to the topology, to get better thread affinity and less
    109 cache thrashing, and thus improve overall performance in modern SMP
    110 systems.
    111 
    112 Responsible: rmind
    113 
    114 9. Incremental namei improvements, Phase 2
    115 ------------------------------------------
    116 
    117 Building on the namei() split which was introduced in 5.99.15 (see (5)
    118 above), further changes will be introduced:  see the changes to namei
    119 outlined in Message-ID:  <20080319053709.GB3951 (a] netbsd.org> for more
    120 information.  This will simplify the locking and behavior of namei()
    121 calls within the kernel to resolve path names within file systems.
    122 
    123 Responsible: dholland
    124 
    125 
    126 10. Ephemeral Mapping
    127 ---------------------
    128 
    129 Responsible: rmind
    130 
    131 
    132 11. Multicast DNS and DNS Service Discovery
    133 -------------------------------------------
    134 
    135 mDNSResponder (also known as mdnsd on some systems) is a daemon invoked
    136 at boot time to implement Multicast DNS and DNS Service Discovery. On Mac
    137 OS X 10.6 (Snow Leopard), mDNSResponder is also the system-wide Unicast
    138 DNS Resolver. Ty Sarna added support for mdnsd to NetBSD-current, and it
    139 will appear in NetBSD 6.0.
    140 
    141 
    142 Alistair Crooks
    143 Thu Nov  5 07:47:34 PST 2009
    144