system revision 1.10
1 $NetBSD: system,v 1.10 2012/01/15 03:11:13 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 will appear in NetBSD 6.0: 10 1. 64-bit time values supported 11 2. Better Kernel Modules support 12 5. namei() tactical changes 13 9. namei() strategic changes 14 10. emap (ephemeral mapping) support for i386 and amd64 15 11. support for mDNSResponder 16 17 The following projects may make it into future releases: 18 3. Full kernel preemption for real-time threads 19 4. POSIX shared memory 20 6. Better resource controls 21 7. Improved observability: online crashdumps, remote debugging 22 8. Processor and cache topology aware scheduler 23 24 We'll continue to update this roadmap as features and dates get firmed up. 25 26 27 Some explanations 28 ================= 29 30 1. 64-bit time_t support 31 ------------------------- 32 33 The Unix 32-bit time_t value will overflow in 2037 - any mortgage calculations 34 which use a time_t value are in danger of overflowing at the present time - 35 and to address this, 64-bit time_t values will be used to contain the number 36 of seconds since 1970. This was completed in 5.99.7 by Christos (with no libc 37 major bump - kudos and respect), and will be in 6.0. 38 39 Responsible: christos 40 41 42 2. Better Kernel Module Support 43 ------------------------------- 44 45 Starting with 5.99.2, the kernel support for modules was enhanced by 46 ad, and GENERIC was switched over to be a MODULAR kernel. Support 47 from booting from modules, like ffs, was introduced at the same time. 48 Some work has been done by Luke Mewburn in this area to define module 49 locations and paths so that effective kernel development can be done 50 using modules. Kernel modules have been moved out of base.tgz and a new 51 modules.tgz has been created. 52 53 Responsible: ad, lukem 54 55 56 3. Full kernel preemption for real-time threads on non-x86 57 ---------------------------------------------------------- 58 59 With the revamp of the kernel concurrency model, much of the kernel is 60 fully multi-threaded and can therefore be preempted at any time. In 61 support of lower context switch and dispatch times for real-time 62 threads, full kernel preemption is being implemented. This has been 63 implemented already for i386 and x86_64 (and is in 5.0), but needs to 64 be extended to support ARM and other ports. MIPS has this, as does 65 PowerPC, but it is not yet enabled. 66 67 Responsible: rmind 68 69 70 4. POSIX shared memory 71 ---------------------- 72 73 Implement POSIX shared memory facilities, which can be used to create 74 the shared memory objects and add the memory locations to the address 75 space of a process. Changes were proposed on tech-kern, although 76 there were some concerns with the kernel implementation, and so a 77 different approach using wrapper functions on tmpfs is being aimed at 78 for 6.0. 79 80 Responsible: rmind 81 82 83 5. Incremental namei improvements, Phase 1 84 ------------------------------------------ 85 86 In NetBSD 5.99.15, some changes were made to split the namei() routine 87 up into logical parts, so that changes can be made to the constituent 88 parts in a less intrusive way. This is in the repository now, and will 89 be in 6.0. 90 91 Responsible: dholland 92 93 94 6. Better resource controls 95 --------------------------- 96 97 A resource provisioning and control framework that extends beyond the 98 traditional Unix process limits. 99 100 Responsible: TBD 101 102 103 7. Improved observability: online crashdumps, remote debugging 104 -------------------------------------------------------------- 105 106 XXX crashdumps while the system is running 107 XXX firewire support in libkvm 108 109 Responsible: TBD 110 111 112 8. Processor and cache topology aware scheduler 113 ----------------------------------------------- 114 115 Implement the detection of the topology of the processors and caches. 116 Improve the scheduler to make decisions about thread migration 117 according to the topology, to get better thread affinity and less 118 cache thrashing, and thus improve overall performance in modern SMP 119 systems. Code has been written, but did not show any performance 120 impreovment. We will continue to monitor this area. 121 122 Responsible: rmind 123 124 125 9. Incremental namei improvements, Phase 2 126 ------------------------------------------ 127 128 Building on the namei() split which was introduced in 5.99.15 (see (5) 129 above), further changes will be introduced: see the changes to namei 130 outlined in Message-ID: <20080319053709.GB3951 (a] netbsd.org> for more 131 information. This will simplify the locking and behavior of namei() 132 calls within the kernel to resolve path names within file systems. 133 134 Responsible: dholland 135 136 137 10. Ephemeral Mapping 138 --------------------- 139 140 Responsible: rmind 141 142 143 11. Multicast DNS and DNS Service Discovery 144 ------------------------------------------- 145 146 mDNSResponder (also known as mdnsd on some systems) is a daemon invoked 147 at boot time to implement Multicast DNS and DNS Service Discovery. On Mac 148 OS X 10.6 (Snow Leopard), mDNSResponder is also the system-wide Unicast 149 DNS Resolver. Ty Sarna added support for mdnsd to NetBSD-current, and it 150 will appear in NetBSD 6.0. 151 152 153 Alistair Crooks 154 Sat Jan 14 11:40:49 PST 2012 155
Indexes created Fri Sep 26 20:09:58 GMT 2025