Jump to content

Talk:Double fault

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Going to drop the merge, if anything the triple fault should be merged into the double fault article. MadnessASAP (talk) 00:12, 26 November 2007 (UTC)[reply]

TSS links to the wrong thing! Argh, I hate article links by acronyms.

Can someone tell me what TSS really stands for in this context and link to it here? Graue 05:51, 22 May 2005 (UTC)[reply]

TSS = Task state segment. — Bcat (talkemail) 19:07, 14 August 2005 (UTC)[reply]

"As double faults can only happen due to kernel bugs, they are rarely[clarify] caused by userland programs."

A double fault is imho _always_ a kernel bug. If a userland program can trigger a double fault, than this is nevertheless a kernel bug, because the kernel shouldn't allow the userland program to do so. --Solaristhesun 06:13, 18 October 2007 (UTC)[reply]
Not entirely. I am, and should be, able to cause a double fault by trashing /proc/kcore. In general, it can happen whenever the kernel places significant trust in a userland program, which is not necessarily a bug (though one may consider it unwise under all circumstances). 85.146.241.224 00:18, 10 November 2007 (UTC)[reply]
If the processor doesn't have a complete memory management unit (MMU), but can still have execution protection, then you could theoretically cause a double fault with a user program. An example is the TI89 graphing calculator; it doesn't have a MMU, but it does have execution and write protections. If a program doesn't restore an interrupt vector, you could cause the system to crash, but I don't know if that would even be considered a double fault. Joeyadams (talk) 03:04, 28 November 2007 (UTC)[reply]