Messages in this thread | | | Date | Fri, 7 Nov 1997 22:50:45 -0500 (EST) | From | "Richard B. Johnson" <> | Subject | Re: Pentium DEATH in user-mode |
| |
On Fri, 7 Nov 1997, Robert R. Collins wrote:
> Richard B. Johnson, you wrote: > > > >The following user-mode program will stop the internal clock > >of a Pentium. > > Actually, it doesn't stop the internal clock. It sends out two > (back-to-back) locked memory read cycles, without any intervening locked > memory write cycles. This (most likely) causes the state machines of the > chipsets to go bonkers. The net effect is a crashed system. > > > > >char x[5]={0xf0,0x0f,0xc7,0xc8,0x00};
Well, according to my bus analyzer and an educated finger, the chip stops __cold__ literly. The external chip temperature goes to ambient from a nominal 22-28 degC above ambient and the power supply load current goes way down.
All of these code-sequences cause the same effect.
0xf0,0x0f,0xc7,0xc8 0xf0,0x0f,0xc7,0xc9 0xf0,0x0f,0xc7,0xca 0xf0,0x0f,0xc7,0xcb 0xf0,0x0f,0xc7,0xcc 0xf0,0x0f,0xc7,0xcd 0xf0,0x0f,0xc7,0xce 0xf0,0x0f,0xc7,0xcf
I made a simple program to perform an exhaustive search of bad opcodes that can crash a protected-mode system from user-mode.
Cheers, Dick Johnson
Richard B. Johnson Project Engineer Analogic Corporation Penguin : Linux version 2.1.60 on an i586 machine (66.15 BogoMips). Warning : It's hard to remain at the trailing edge of technology.
| |