Package: linux-2.6
Version: consider enabling CONFIG_RAMZSWAP in 2.6.33+
Severity: wishlist
please consider setting CONFIG_RAMZSWAP (a.k.a. compcache), which enables a ram
based compressed swap device. i think it can be enabled as a module, but i'm
not sure.
this feature is very useful to thin clients, embedded devices, or other
environments with limited ram and no conventional swap device available.
live well,
vagrant
--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20100314222759.GA2083@claws.fglan">http://lists.debian.org/20100314222759.GA2083@claws.fglan
Vagrant Cascadian wrote:
> Please consider setting CONFIG_RAMZSWAP (a.k.a. compcache), which
> enables a ram based compressed swap device. i think it can be
> enabled as a module, but i'm not sure.
>
> This feature is very useful to thin clients, embedded devices, or
> other environments with limited ram and no conventional swap device
> available.
This is also useful for Debian Live images generated using the
live-helper and live-initramfs packages. Using Sid's live-helper to
build an Ubuntu 10.04 live image already results in a 25% ramzswap, so
I guess Ubuntu have patched compcache into their default .32 kernel.
--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20100318033616.GA4430@Clio">http://lists.debian.org/20100318033616.GA4430@Clio
On Thu, 2010-03-18 at 14:36 +1100, Trent W. Buck wrote:
> Vagrant Cascadian wrote:
> > Please consider setting CONFIG_RAMZSWAP (a.k.a. compcache), which
> > enables a ram based compressed swap device. i think it can be
> > enabled as a module, but i'm not sure.
> >
> > This feature is very useful to thin clients, embedded devices, or
> > other environments with limited ram and no conventional swap device
> > available.
>
> This is also useful for Debian Live images generated using the
> live-helper and live-initramfs packages. Using Sid's live-helper to
> build an Ubuntu 10.04 live image already results in a 25% ramzswap, so
> I guess Ubuntu have patched compcache into their default .32 kernel.
Would you like me to add this to 2.6.32 for squeeze, then?
Ben.
--
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.
On Sun, Apr 04, 2010 at 07:16:36AM +0100, Ben Hutchings wrote:
> On Thu, 2010-03-18 at 14:36 +1100, Trent W. Buck wrote:
> > Vagrant Cascadian wrote:
> > > Please consider setting CONFIG_RAMZSWAP (a.k.a. compcache), which
> > > enables a ram based compressed swap device. i think it can be
> > > enabled as a module, but i'm not sure.
> > >
> > > This feature is very useful to thin clients, embedded devices, or
> > > other environments with limited ram and no conventional swap device
> > > available.
> >
> > This is also useful for Debian Live images generated using the
> > live-helper and live-initramfs packages. Using Sid's live-helper to
> > build an Ubuntu 10.04 live image already results in a 25% ramzswap, so
> > I guess Ubuntu have patched compcache into their default .32 kernel.
>
> Would you like me to add this to 2.6.32 for squeeze, then?
if it's not too difficult, sure! i had thought it only made it upstream as
recently as 2.6.33.
live well,
vagrant
--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20100404182112.GV27790@claws.fglan">http://lists.debian.org/20100404182112.GV27790@claws.fglan
On Sun, 2010-04-04 at 11:21 -0700, Vagrant Cascadian wrote:
> On Sun, Apr 04, 2010 at 07:16:36AM +0100, Ben Hutchings wrote:
> > On Thu, 2010-03-18 at 14:36 +1100, Trent W. Buck wrote:
> > > Vagrant Cascadian wrote:
> > > > Please consider setting CONFIG_RAMZSWAP (a.k.a. compcache), which
> > > > enables a ram based compressed swap device. i think it can be
> > > > enabled as a module, but i'm not sure.
> > > >
> > > > This feature is very useful to thin clients, embedded devices, or
> > > > other environments with limited ram and no conventional swap device
> > > > available.
> > >
> > > This is also useful for Debian Live images generated using the
> > > live-helper and live-initramfs packages. Using Sid's live-helper to
> > > build an Ubuntu 10.04 live image already results in a 25% ramzswap, so
> > > I guess Ubuntu have patched compcache into their default .32 kernel.
> >
> > Would you like me to add this to 2.6.32 for squeeze, then?
>
> if it's not too difficult, sure! i had thought it only made it upstream as
> recently as 2.6.33.
Indeed, but it seems trivial to backport. Please test it when 2.6.32-11
is released.
Ben.
--
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.
On Sun, Apr 04, 2010 at 08:06:23PM +0100, Ben Hutchings wrote:
> On Sun, 2010-04-04 at 11:21 -0700, Vagrant Cascadian wrote:
> > On Sun, Apr 04, 2010 at 07:16:36AM +0100, Ben Hutchings wrote:
> > > On Thu, 2010-03-18 at 14:36 +1100, Trent W. Buck wrote:
> > > > Vagrant Cascadian wrote:
> > > > > Please consider setting CONFIG_RAMZSWAP (a.k.a. compcache), which
> > > > > enables a ram based compressed swap device. i think it can be
> > > > > enabled as a module, but i'm not sure.
...snip...
> > > Would you like me to add this to 2.6.32 for squeeze, then?
> >
> > if it's not too difficult, sure! i had thought it only made it upstream as
> > recently as 2.6.33.
>
> Indeed, but it seems trivial to backport. Please test it when 2.6.32-11
> is released.
thanks!
it works, although needs rzscontrol from compcache-tools (in the NEW queue
since december) in order to initialize the ramzswap device.
i had the impression from the documentation that the module could initialize
the device itself with the defaults, but this apparently isn't the case, at
least not with my initial testing, or the version that was backported.
live well,
vagrant
--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20100407041811.GD27790@claws.fglan">http://lists.debian.org/20100407041811.GD27790@claws.fglan
Is the notify patch included on this module backport? It helps a lot
to free compressed RAM not used anymore. Acording to the description,
This will enable 'swap free notify' feature which allows kernel to send
callback to
ramzswap as soon as a swap slot becomes free. So, we can immediately
free memory allocated for this page, eliminating any stale data in
(compressed) memory.
--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 4BC37F6B.5060902@terra.es">http://lists.debian.org/4BC37F6B.5060902@terra.es
On Mon, 2010-04-12 at 22:15 +0200, Alberto wrote:
> Is the notify patch included on this module backport? It helps a lot
> to free compressed RAM not used anymore. Acording to the description,
This does not appear to have been accepted upstream yet. If it has,
please provide a reference (commit hash and repository URL).
Ben.
--
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.
I don't have the skills to evaluate if backporting this feature
would be feassible, or too intrusive.
--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 4C03DDC5.6050106@terra.es">http://lists.debian.org/4C03DDC5.6050106@terra.es
On Mon, 2010-05-31 at 18:03 +0200, Alberto wrote:
> On Sun, 18 Apr 2010 at 18:13:55 +0100, Ben Hutchings wrote:
> >> Is the notify patch included on this module backport? It helps a lot
> >> to free compressed RAM not used anymore. Acording to the description,
>
> >This does not appear to have been accepted upstream yet. If it has,
> >please provide a reference (commit hash and repository URL).
>
>
> It seems that some patches have been accepted for 2.6.35:
>
> http://www.mail-archive.com/linux-mm-cc@lists.laptop.org/msg00182.html
>
> I don't have the skills to evaluate if backporting this feature
> would be feassible, or too intrusive.
This looks possible to backport, but it will involve an ABI bump (hence
a change of binary package names) so I won't apply the changes straight
away.
Bug#573912: linux-2.6: please enable ramzswap/compcache
