April 2010 - Oberon - mailman3.inf.ethz.ch

persistence-across-reboot
by Paul Reed 31 May '10

31 May '10

Hi folks, Another reason to use a filesystem driver with flash memory is to take advantage of the implementation of wear-levelling (assuming it isn't handled by the hardware), although this would of course be less important if the memory was solely used for persistence across reboots. Checkpointing the system state every couple of minutes would be a useful defence against power failure, too - but then you would definitely want to look at how many write cycles your device claims to last for :) Cheers, Paul > Message: 1 > Date: Wed, 14 Apr 2010 18:29:55 +0200 > From: Peter Matthias <PeterMatthias(a)web.de> > Subject: Re: [Oberon] Re (3): persistence-across-reboot > To: ETH Oberon and related systems <oberon(a)lists.inf.ethz.ch> > Message-ID: <201004141829.55153.PeterMatthias(a)web.de> > Content-Type: Text/Plain; charset="iso-8859-15" > > Am Mittwoch 14 April 2010 schrieb peasthope(a)shaw.ca: >> All the non-volatile storage I've heard of organizes >> data as files in a filesystem. Does anyone sell >> flash memory which is accessed by an address just >> as RAM is? > > You can use a block device driver and mmap it to the address space, as it > it > done in Linux. > > Peter > > > ------------------------------ > > -- > Oberon(a)lists.inf.ethz.ch mailing list for ETH Oberon and related systems > https://lists.inf.ethz.ch/mailman/listinfo/oberon > > > End of Oberon Digest, Vol 76, Issue 2 > ************************************* >

3 2

Re (3): persistence-across-reboot
by peasthope＠shaw.ca 14 Apr '10

14 Apr '10

From: Jack Johnson <knap...(a)gmail.com> Date: Tue, 13 Apr 2010 16:23:12 -0800 > I've often thought that as traditional storage approaches access times > of RAM from a decade or two ago, why wouldn't we start designing more > operating systems to assume 200 GB or so of persistent state? All the non-volatile storage I've heard of organizes data as files in a filesystem. Does anyone sell flash memory which is accessed by an address just as RAM is? Really the traditional filesystem should be obsolete and used only for legacy data. Regards, ... Peter E. -- Google "pathology workshop". In ETHNO click here -> Desktops.OpenDoc "http://carnot.yi.org/".

2 1

Re (2): persistence-across-reboot; was Re: [Oberon] Re. HOW2 'magnify view' for BB?Aos?A2?
by peasthope＠shaw.ca 14 Apr '10

14 Apr '10

From: Jack Johnson <knap....(a)gmail.com> Date: Tue, 13 Apr 2010 12:24:39 -0800 > ... Smalltalk VM ... Virtual Machine? Anyhow, I guess Chris' wish list contribution would be hibernation for A2. Thanks, ... Peter E. -- Google "pathology workshop". In ETHNO click here -> Desktops.OpenDoc "http://carnot.yi.org/".

2 1

persistence-across-reboot; was Re: [Oberon] Re. HOW2 'magnify view' for BB?Aos?A2?
by peasthope＠shaw.ca 13 Apr '10

13 Apr '10

Chris & all, From: Chris Glur <easlab(a)absamail.co.za> Date: Wed, 17 Feb 2010 23:21:32 -0700 (MST) > It's amazing that no one has yet made a utility to restore all the > settings to allow a PC to be switched off. > That's a perfect example of Alvin Toffler's 70's "replacing material > by intelligence". Eg. so that we don't need the PCs to run 24/7. FWIIW, some of the handheld computers do have persistence-across-reboot functionality. Two which I own are the Sony PSP and the Sony Mylo COM-2. The Mylo is mentioned here. http://carnot.yi.org/ExampleSystems.html The iPod Touch and iPhone and the Nokia N810 and N900 and many other handhelds might also have this capability. I haven't checked. Why is such persistence absent from desktop and laptop systems? I can only speculate that it involves behaviour of third party software which Apple and MS prefer not to involve. Yet there is no such concern with A2 and other systems where all relevant software is released through one entity. I might be completely wrong. If someone with more understanding can explain, please do. Regards, ... Peter E. -- Google "pathology workshop". In ETHNO click here -> Desktops.OpenDoc "http://carnot.yi.org/".

2 1