17 messages in org.freebsd.freebsd-arch64bit ticks, was Re: Changing p_swtim...| From | Sent On | Attachments |
|---|---|---|
| Jeff Roberson | Sep 17, 2007 5:03 pm | |
| Kostik Belousov | Sep 17, 2007 9:33 pm | |
| Julian Elischer | Sep 17, 2007 10:48 pm | |
| Jeff Roberson | Sep 18, 2007 1:23 am | |
| Jeff Roberson | Sep 18, 2007 1:25 am | |
| Andre Oppermann | Sep 18, 2007 8:12 am | |
| Jeff Roberson | Sep 18, 2007 2:19 pm | |
| Jeff Roberson | Sep 18, 2007 3:34 pm | |
| Julian Elischer | Sep 18, 2007 4:55 pm | |
| Jeff Roberson | Sep 18, 2007 5:08 pm | |
| Sam Leffler | Sep 18, 2007 5:33 pm | |
| Constantine A. Murenin | Sep 18, 2007 7:03 pm | |
| Garrett Wollman | Sep 18, 2007 7:13 pm | |
| Jeff Roberson | Sep 18, 2007 7:14 pm | |
| Doug Ambrisko | Sep 18, 2007 8:08 pm | |
| Bruce Evans | Sep 18, 2007 9:57 pm | |
| Peter Jeremy | Sep 19, 2007 1:58 am |
| Subject: | 64bit ticks, was Re: Changing p_swtime and td_slptime to ticks | |
|---|---|---|
| From: | Julian Elischer (jul...@elischer.org) | |
| Date: | Sep 18, 2007 4:55:07 pm | |
| List: | org.freebsd.freebsd-arch | |
Sam Leffler wrote:
Jeff Roberson wrote:
On Tue, 18 Sep 2007, Jeff Roberson wrote:
On Tue, 18 Sep 2007, Andre Oppermann wrote:
Jeff Roberson wrote:
On Mon, 17 Sep 2007, Julian Elischer wrote:
Jeff Roberson wrote:
Enclosed is a patch that fixes swapping with ULE. ULE has never properly set p_swtime and td_slptime which are used by the swapout/swapin code to select the appropriate thread to swap.
I have not looked at in the depth required, but 2 points that I was unable to check to my satisfaction before I got called away for work....
1/ the source of the ticks is a monotonically increasing count that never goes backwards or changes?
ticks is incremented each time hardclock() is called. That's it.
2/ nothing that used to be accounted in seconds becomes accounted for in ticks?
I scale back to seconds where it is required. Really I think ticks would be the better metric in vm_glue.c but didn't want to make any drastic changes.
ticks is 2^31 on x86 and at HZ=1000 is wraps within a reasonable short uptime. You have to make sure that your code handles that correctly or you run into lots of strange effects which are almost impossible to reproduce. In TCP we've got bitten by that.
Thanks Andre, this is a good point. For the td_slptime I don't think it's of practical concern. However, for swtime I think I will convert it then to seconds from boot.
Is there a good reason for not making ticks 64bit? math involving this value is relatively infrequent. Bruce? Any comments? It'd sure let us forget all of these counter wrapping problems.
ticks is used a lot. I'd rather set hz back to 100 by default. This approach is a perfect example of ignoring low-end platforms.
but it still needs to work on systems with high hz values.
Sam