Messages in this thread | | | Date | Thu, 7 Apr 2005 16:51:50 +0200 | From | Ingo Molnar <> | Subject | Re: crash in entry.S restore_all, 2.6.12-rc2, x86, PAGEALLOC |
| |
* Linus Torvalds <torvalds@osdl.org> wrote:
> > > pushfl
> After all, I very strongly suspect that we don't actually really > _care_ if eflags stays the same over a system call, and I could see > that some dynamic CPU's might prefer not having to push an eflags > value that just got changed by the "sti", so it _might_ save several > cycles to avoid that dependency, and also obviously avoid a subtle > dependency on a sw level that the previous patch clearly introduced. > > Anybody willing to time it? ;)
i can tell you without any measurement that pushfl is slower by a couple of cycles than a simple pushl $0x00010046, on basically all x86 CPUs. And since we only support SYENTER from 32-bit mode and we dont guarantee flags to be saved, it isnt all that incorrect to do?
Ingo - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |