dougp@voodoo.ucsb.edu (05/06/90)
-Message-Text-Follows- In article <1410046@hpcvca.CV.HP.COM>, charles@hpcvca.CV.HP.COM (Charles Brown) writes... >> Handle floppies the way it is done now, untill you get an error. If >> you find an error on reading (the only time you can detect an error) >> go back and retry, but do the retry syncronous to index. > >But by then you have already lost data. My intention is that you >should never loose data. But a read is done immediatly after a write (otherwise you could never detect an error while writing), this is commonly known as verifing. Come to think of it, I don't know if the Amiga verifies the write or not. It should at liest be an option for those willing to take a performance hit. If an error was detected on a verify read, the track could be rewritten in index syncronous mode, if it failed again, the block could be mapped out. >> Douglas Peale > Charles Brown charles@cv.hp.com or charles%hpcvca@hplabs.hp.com > or hplabs!hpcvca!charles or "Hey you!" Douglas Peale
gmreise@cs.vu.nl (Reisenstadt G M) (05/08/90)
dougp@voodoo.ucsb.edu writes: >In article <1410046@hpcvca.CV.HP.COM>, charles@hpcvca.CV.HP.COM (Charles Brown) writes... >>> Handle floppies the way it is done now, untill you get an error. If >>> you find an error on reading (the only time you can detect an error) >>> go back and retry, but do the retry syncronous to index. >> >>But by then you have already lost data. My intention is that you >>should never loose data. >But a read is done immediatly after a write (otherwise you could never >detect an error while writing), this is commonly known as verifing. Come >to think of it, I don't know if the Amiga verifies the write or not. The Amiga does not verify writes. >It should at liest be an option for those willing to take a performance >hit. You need a program called 'TrackSalve' by Dirk Reisig (of PCopy fame). I beleive he send an update (v1.3) to Fred Fish for inclusion in his library. TrackSalve gives you the option to verify your writes on a per-drive basis, so you could decide to have a 'save' and a 'fast' disk-drive if you have more than one drive. The new version even gives you something called 'MFM Update'. You'll have to check the docs for that, but from memory I can tell that it has some- thing to do with unused sectors in tracks becoming trashed after a while. >If an error was detected on a verify read, the track could be rewritten >in index syncronous mode, if it failed again, the block could be mapped >out. It does give you a requester with 'rewrite' or 'ignore'. If rewrite does not work for some reason, you can save that file to another disk. I don't think it does map blocks out (discussed a while back as to why not). >>> Douglas Peale >> Charles Brown charles@cv.hp.com or charles%hpcvca@hplabs.hp.com >> or hplabs!hpcvca!charles or "Hey you!" >Douglas Peale Marco Niese. E-Mail: gmreise@cs.vu.nl (Put: TO BMR in subject line.)