[rfc] use ext3 instead of ext2
spyro at f2s.com
Fri Jul 16 10:36:13 PDT 2004
On Fri, 16 Jul 2004 12:55:15 -0400
Bryan Kadzban <bryan at kadzban.is-a-geek.net> wrote:
> > yes. what preempt alleviates is the problem of one task making
> > others waiting to enter the kernel.
> I thought preempt allowed other tasks to execute (in any mode) while
> one was in kernel mode?
Correct. my comment wasnt too well written there.
> I didn't think it had anything to do with *entering*
> the kernel (executing an INT 80h (on x86 anyway) to run a syscall), I
> thought it allowed syscalls (or whatever) to be interrupted by higher-
> priority processes, and then resumed later.
well, in the past once a process entered the kernel nothing else could
run, so nothing else could enter the kernel. Now that can happen.
> Also, let me play dumb here -- what does low-latency alleviate? All
> that the low-latency patches do is make the kernel functions give up
> their locks at certain predefined "OK" points, right?
Basically yeah - they were just 'cleanups' to get drivers to 'do the
right thing' and give up their locks when they werent needed.
More information about the lfs-dev