cagney@chook.ua.oz (Andrew Cagney - aka Noid) (03/19/90)
Subject: Re: ibm-1.5.5:build.c: what happened to the 0xAA55 problem? In article <6007@star.cs.vu.nl> Andy Tanenbaum writes: > In article <759@augean.OZ> Andrew Cagney writes: > >A while ago there was some discussion about the minix boot block and how > >some BIOS's require the word `0xAA55' at the end of the boot block. > > 2. I think I missed this stuff. What has to be done to 1.5.5 to keep these > BIOSes happy? Last year there were two discussions to do with 0xAA55 at the end of the boot block. The first was to do with dosread and the second was to do with problems booting minix on (I think,) a Sun386. MESSYDOG, when it formats a disk, puts the value 0xAA55 in the last word of the boot block. Some BIOS's (eg Sun386) verify a boot disk by checking for the presence of 0xAA55 at the end of a disk before jumping to the start of the boot code. [can someone confirm it was the Sun386?]. I've tested a version of build/bootblok, written by Bruce Evans, that contained the required changes. It should be noted that adding 0xAA55 to the end of bootblock.s makes it incompatible with the existing versions of build as the build table at the end of the boot block must move :-(. Have I got this all wrong? :-) Andrew Cagney cagney@cs.ua.oz.au BTW: dosread, the other program changed at the same time now checks for 0xAA55 when looking for a dos disk.