Re: [Suspend2][ 2/2] [Suspend2] Freezer upgrade.

[Nigel Cunningham <nigel@suspend2.net>]

[-- Attachment #1: Type: text/plain, Size: 1476 bytes --]

Hi.

On Tuesday 27 June 2006 21:09, Pavel Machek wrote:
> Hi!
>
> > > > This patch represents the Suspend2 upgrades to the freezer
> > > > implementation. Due to recent changes in the vanilla version, I
> > > > should be able to greatly reduce the size of this patch. TODO.
> > >
> > > So I assume the patch will change in the future.
> >
> > This is after the changes. Sorry - forgot to update the comment.
>
> Also please explain why we want those patches. "upgrades the freezer"
> is not good enough reason to apply a patch.

I guess you missed the reply to Rafael. In it, I wrote:

"XFS. Did you see Nathan's reply not long ago, confirming that it doesn't stop 
all activity if you don't freeze bdevs? That isn't critical for swsusp 
(although I guess you could end up with some filesystem inconsistency if XFS 
writes something after the atomic copy), but keeping the LRU static is 
important for suspend2."

In another email, I mentioned that trying to free memory with swap on a 
journalled filesystem is a potential deadlock situation without the 
capability of thawing kernel threads alone. You could potentially thaw all 
threads while eating memory, but then there's a greater chance of racing 
against another program that's trying to allocate memory (if you're in this 
situation, you're low on memory to start with).

Regards,

Nigel

-- 
See http://www.suspend2.net for Howtos, FAQs, mailing
lists, wiki and bugzilla info.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]