[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#632627: Re[2]: resume file



On Thu, 2013-07-25 at 21:31 +0200, Rodolfo García Peñas wrote:
> On Thu, 25 Jul 2013, Ben Hutchings escribió:
> 
> > On Thu, 2013-07-25 at 09:54 +0000, Rodolfo García Peñas (kix) wrote:
> > > Ben Hutchings <ben@decadent.org.uk> escribió:
> > > 
> > > > On Wed, 2013-07-24 at 15:57 +0400, Askar Safin wrote:
> > > >> > I think this file should be updated every time the package  
> > > >> "initramfs-tools" is reconfigured.
> > > >>
> > > >> Also, ideally, initramfs should be rebuilded every time user changes
> > > >> fstab. But, I think this is very hard to reach.
> > > >> Also, I think, every tool which updates fstab, should call initramfs
> > > >> rebuilding (if swap is changed and hibernation is enabled).
> > > >> In particular, every package configuration script which updates fstab
> > > >> should trigger initramfs-tools reconfiguring.
> > > >
> > > > I think there is only one such script, which is the postinst script for
> > > > linux-base which made a one-time update of configuration files to use
> > > > UUIDs.  It updated the initramfs-tools resume file too.
> > > >
> > > > But I wonder whether this configuration file is really needed for most
> > > > system.  Couldn't we use blkid to find the swap partition automagically?
> > > > (The configuration file would still be necessary if there are multiple
> > > > swap partitions.)
> > > >
> > > > Ben.
> > > >
> > > > --
> > > > Ben Hutchings
> > > > Once a job is fouled up, anything done to improve it makes it worse.
> > > 
> > > Hi Ben, Askar,
> > > 
> > > thanks a lot for your replies. I will update the uswsusp.config and  
> > > uswsusp.postinst scripts to write that file and then call  
> > > update-initramfs script. I will use UUIDs in that file (when available).
> > 
> > uswsusp is not allowed to change an initramfs-tools config file
> > directly:
> > <file:///usr/share/doc/debian-policy/policy.html/ch-files.html#s10.7.4>
> 
> Yes, sorry about that. Then initramfs-tools should update that file,
> checking if the /etc/uswsusp.conf file exists and reading the value of
> "resume device". Should I forward the bug to initramfs-tools package?

Sorry, I think there's some confusion here.

initramfs-tools has its own resume configuration file, but so far as I
can see the uswsusp package does not use it.  So there would be no
reason for you to modify it - would there?

The uswsusp package can edit its own configuration file though I do not
recommend it.  I think it would be better to leave the resume device
unspecified in /etc/uswsusp.conf by default, and then to select it
automatically (both at suspend and resume time) if possible.

> > And why should it make a difference whether uswsusp or kernel-only
> > suspend is used?
> 
> uswsusp has three tools:
[...]
> s2disk and s2both tools use the resume file, so your question is about
> these tools. These tools can write the image to the swap file selected
> by the user (the resume file), they can compress the image (using less
> swap and doing the suspend faster (less disk I/O)), they can encrypt
> the image file and other less important things.

Sure, uswsusp has some advantages.  But if we agree it is a good idea to
select the resume device automatically, isn't that true for both
suspend-to-disk methods?

Ben.

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: