From mboxrd@z Thu Jan 1 00:00:00 1970 From: paul.winalski@gmail.com (Paul Winalski) Date: Wed, 25 Apr 2018 17:17:49 -0400 Subject: [TUHS] Disk data layout (was: /dev/drum) In-Reply-To: <20180425013134.GM31055@eureka.lemis.com> References: <7wfu3nuqeb.fsf@junk.nocrew.org> <3A18DFEC-42B7-4234-9DD1-367733270D50@tfeb.org> <0abe01d3db28$b6573660$2305a320$@ronnatalie.com> <866bbea1-3a26-20a4-e233-1b8dc0ea2683@spamtrap.tnetconsulting.net> <20180425013134.GM31055@eureka.lemis.com> Message-ID: On 4/24/18, Greg 'groggy' Lehey wrote: > > "LBA" is newer than the time we're talking of. In those days, disk > data was addressed physically, by cylinder, head and sector, terms > that only died out round the turn of the century. IBM DASD--Direct Access Storage Device, a term that encompassed drums, disks and the data cell drive--addressed data on the media physically by bin, cylinder, head, and record, as a hexadecimal number BBCCHHR. Bin number was zero except on the IBM 2321 data cell drive. CKD drives supported a variable number of records on each track, hence the term "record" rather than "sector". Logical block addressing (LBA) for sector-oriented disks allowed the OS (or later, the disk controller) to hide bad block replacement from programs. VAX/VMS used a protected system file ([sysexe]badblock.sys) to keep track of the bad blocks on each disk volume. DEC once got a customer bug report complaining that if a privileged user gave the command "TYPE SYS$SYSTEM:BADBLOCK.SYS", the console logged bad block errors. How does/did Unix handle bad block replacement? -Paul W.