46 messages in com.xensource.lists.xen-develRe: [Xen-devel] /proc/xen/xenbus supp...
FromSent OnAttachments
NAHieu08 Sep 2005 01:02 
Christian Limpach08 Sep 2005 03:37 
Rusty Russell08 Sep 2005 17:43 
Christian Limpach13 Sep 2005 02:42 
Rusty Russell13 Sep 2005 17:20 
Christian Limpach14 Sep 2005 01:23 
Rusty Russell14 Sep 2005 02:18 
Christian Limpach14 Sep 2005 05:54 
Rusty Russell14 Sep 2005 18:39 
Keir Fraser15 Sep 2005 03:52 
Christian Limpach15 Sep 2005 04:02 
Rusty Russell17 Sep 2005 01:25 
Keir Fraser17 Sep 2005 01:33 
Christian Limpach17 Sep 2005 10:39 
Rusty Russell18 Sep 2005 17:10 
Rusty Russell18 Sep 2005 17:19 
Keir Fraser19 Sep 2005 01:54 
Rusty Russell20 Sep 2005 04:00 
Keir Fraser21 Sep 2005 02:35 
Keir Fraser21 Sep 2005 02:39 
harry21 Sep 2005 04:41 
Rusty Russell21 Sep 2005 19:07 
Rusty Russell21 Sep 2005 19:22 
Keir Fraser22 Sep 2005 02:34 
Keir Fraser22 Sep 2005 02:35 
Rusty Russell22 Sep 2005 15:54 
Rusty Russell22 Sep 2005 16:50 
Andrew Warfield22 Sep 2005 18:01 
Keir Fraser23 Sep 2005 02:16 
Keir Fraser23 Sep 2005 02:24 
Rusty Russell24 Sep 2005 17:56 
Rusty Russell24 Sep 2005 18:09 
Rusty Russell24 Sep 2005 20:29 
Keir Fraser25 Sep 2005 04:02 
Keir Fraser25 Sep 2005 04:08 
Keir Fraser25 Sep 2005 04:33 
Christian Limpach25 Sep 2005 11:55 
Rusty Russell25 Sep 2005 15:51 
Rusty Russell25 Sep 2005 16:05 
Rusty Russell25 Sep 2005 23:35 
Keir Fraser26 Sep 2005 00:33 
Christian Limpach26 Sep 2005 11:51 
Keir Fraser26 Sep 2005 12:29 
Rusty Russell26 Sep 2005 23:47 
Rusty Russell27 Sep 2005 00:15 
David Hopwood27 Sep 2005 16:31 
Subject:Re: [Xen-devel] /proc/xen/xenbus supports watch?
From:Keir Fraser (Keir@cl.cam.ac.uk)
Date:09/25/2005 04:08:53 AM
List:com.xensource.lists.xen-devel

On 25 Sep 2005, at 01:57, Rusty Russell wrote:

Now, we already have this "domain won't save until transactions are done" simply because we use a single big lock, but this discussion started because we want to get rid of that lock for /proc/xen/xenbus (it's fine for drivers). I think we should do so, but keep this wont-save-during-transactions semantic; it means a waitqueue etc, but I don't think it's too bad. As you say, our transactions are pretty small.

Do people like this more?

Blocking system progress on arbitrary user apps doesn't sound particularly attractive to me, but I guess it is at least simple. I'm more inclined to EAGAIN on read/write, maybe sugared by exceptions in languages that support them, or setjmp/longjmp to get you back to the outermost scope of the transaction. I agree that's not pretty either, though.

How will you handle 'xenstored restart'? You can't really guarantee that to always happen at opportune moments with no transactions in flight.

-- Keir

_______________________________________________ Xen-devel mailing list Xen-@lists.xensource.com http://lists.xensource.com/xen-devel