ac@gpu.utcs.toronto.edu (04/03/87)
Pete Lyall posted a "reprint" of a bug report by Kenneth Scales. This bug was essentially identical to the CHD bug I have. Two points - Firstly, CHD only fails for me when I slow the processor to 1MHz to accommodate my slow controller. If I switch to 2 MHz the problem doesn't occur although my system will eventually crash at that speed if I do alot of disk accessing. When I borrowed my friends controller my system works fine at 2MHz but still exibits the CHD bug at 1MHz. Was Mr Scales using a 'stock' and unpatched level II system. I know some people have talked about patches to run the disk I/O at 1MHz. Could he have been running one of these? Secondly, the patch listed causes CHD to use read instead of update mode from the shell. Trouble is that the shell is by no means the only user of the CHD service. Basic09 exibits a similar failure with its CHD command. In fact the problem is even more fundamental than that. Any open for output or update to a write protected disk causes the problem. CHD is indeed the most commom case doing such an open and is the only case that is expected to occur but not be regarded as an error. To go even further, it is not really open at fault. It appears that any time a write command is sent to the controller and that command fails on a write protect error, the system may crash. I'm not sure if other types of controller detected errors cause similar problems. -- Name: Mark Acfield (University of Toronto Computing Services) Path: ihnp4!utgpu!ac Alias: ac@utoronto.bitnet