This message is from the T13 list server.
> -----Original Message----- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On > Behalf Of Curtis Stevens > > Where in the universe has Spocks brain gone?? > > I am not certain about SATA behavior, but in PATA, when the > DF bit is set, > the BSY but needs to be cleared in order to let the host read > the condition. This is the point I was trying to make, that a command cannot be aborted without clearing busy to post that status. > I also believe that if somebody is advocating setting BSY to > one and going > off-line as a solution, they will find that the host may totally > misunderstand the problem. Going off-line in that manner could mean > anything. When the device returns a 20h or 21h status, the device has > indicated that the registers are valid to read, it is not > ready, and it has > a fault. The device prevents new commands from being issued > by the host > when it leaves RDY cleared. That being said, a status other than 0x50 to indicate this condition is probably the best answer, but whether it can be recovered by a reset (hard or soft) is far from a sure thing. Requirements of maintaining cache states etc mean that these events cannot just branch your CPU through address zero, and some internal faults may not recover without a power cycle. A RDY=0 BSY=0 task file is a possibly good state, as well as the RDY=1 WF=1 to indicate that it cannot write for some reason. --eric
