Mailinglist Archive: opensuse-mobile (36 mails)
| < Previous | Next > |
Re: [opensuse-mobile] two issues with suspend to disk in Suse 10.2
- From: Wilfried Maschtera <wilfried.maschtera@xxxxxxxxxxxx>
- Date: Wed, 2 May 2007 11:40:50 +0200
- Message-id: <20070502094050.GA16555@xxxxxxxxxxxx>
On Tue, May 01, 2007 at 11:43:38PM +0200, Martin Hofius wrote:
> Hi Wilfried,
>
> Am Dienstag, 1. Mai 2007 19:03 schrieb Wilfried Maschtera:
> > On Tue, May 01, 2007 at 02:04:30PM +0200, Martin Hofius wrote:
> > > Am Dienstag, 1. Mai 2007 12:03 schrieb Pramod:
> ...
> > > > 1) I would like to have the option of suspending Linux and resuming
> > > > Windows and vice versa
> > > > without booting either OS (as it used to be with Suse 10.0). At the
> ...
> > > Hi Pramod,
> > > sorry I have no other useful hints, but: are Your shure to know what You
> > > are doing??? I remember the choice after suspend was disabled by SuSE to
> > > prevent the destroying of Your data...
> > > When one of the both systems is suspended, also the state information of
> > > the file systems are saved. If You would start an other system and change
> > > files in that filesystem, You may image what happens?
> > > If it was possible in SuSE 10.0 I gues it was a bug!?!
> >
> > Why should it be a bug? In my opinion if the state information of the
> > systems are stored in different locations suspending one system and
> > resuming another one is not risky but of course you should mount files
I should have reread before sending it, the text should be:
you should not mount
Nevertheless some comments
> > from a suspend system read/writeable. I think, windows will not use
> > the linux swap.
> Do You really mount your partitions from suspended system read/writeable???
> One example:
> I assume to have windows with a fat32 partition (to exchange data with linux)
> and run linux with the linux partitions (of course...) and the fat32
> partition mounted. Then I suspend linux and use windows to write some files
> to the fat32 partition . I'm not shure that all changes made by linux are
> completly written to the disk... ok, linux kernel sources are available,
>From the point of (end) user I would say:
I expect that linux will sync and write all buffers etc. before suspending.
The information in the swap is used to speed the boot process up and - mainly -
to restore the status with respect to running applications, open data sets
(of the end user) and so on during resumption. It can not be that I will have
a system with uncertain status or iprobably lost data if the information on the
swap is lost for some reason. If this should be the case I have to think which
unix I will use in the future.
> perhaps can one of the kernel developers answer this...
>
> Then I suspend windows and start linux. Every operating system stores
> information about nodes, trees or fat sectors in RAM. So I think this
> information goes to the swap partition on suspend. And now - the changes made
> by windows are different to the state stored in the linux swap.
>
> Maybe the linux kernel is very good and compares this stored information with
> the actual information on the windows partition - and cleans the cache?
>
> What's about the next step? Linux writes somes files to the fat32 partition
> and windows resumes? Windows does'nt know anything about other operating
> systems ("I am windows, your operation system. You may not use other systems
> than windows ;-))
see above: a small but important error - I did not type the word "not".
But principially it should be the same. The storage of additional data has the
purpose to put the system into the status the system had when the user "stopped
working". It is a service for the user and not for the boot process or the OS.
Therefore the suspension must be clean and safe. This is important especially
for suspension to RAM. The additional infos are stored in RAM only and no
syncing took place and power will fail ?
>
> I cannot exclude that it can work without problems for a log time - but
> nevertheless I'm NOT shure that it is safe to do it except all comon data
> partitions are mounted readonly (but than I can't exchange data between the
> systems).
>
> I for myself solved this problem by using two computers one for windows (video
> processing ...) and one for linux (for all other work). And since a few month
> I use vmware and can have to systems on one computer - at the very same time,
> not for video processing, but to proof web site layouts with ie or programm
> with visual studio (grrrrrr, but some people seem to need really).
Yes of course the situation has changed. I have a similar situation. I have never
used windows for my "normal" work.
Kind regards, Wilfried
>
> Greetings
> Martin
> >
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: opensuse-mobile+unsubscribe@xxxxxxxxxxxx
> For additional commands, e-mail: opensuse-mobile+help@xxxxxxxxxxxx
>
---------------------------------------------------------------------
To unsubscribe, e-mail: opensuse-mobile+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-mobile+help@xxxxxxxxxxxx
> Hi Wilfried,
>
> Am Dienstag, 1. Mai 2007 19:03 schrieb Wilfried Maschtera:
> > On Tue, May 01, 2007 at 02:04:30PM +0200, Martin Hofius wrote:
> > > Am Dienstag, 1. Mai 2007 12:03 schrieb Pramod:
> ...
> > > > 1) I would like to have the option of suspending Linux and resuming
> > > > Windows and vice versa
> > > > without booting either OS (as it used to be with Suse 10.0). At the
> ...
> > > Hi Pramod,
> > > sorry I have no other useful hints, but: are Your shure to know what You
> > > are doing??? I remember the choice after suspend was disabled by SuSE to
> > > prevent the destroying of Your data...
> > > When one of the both systems is suspended, also the state information of
> > > the file systems are saved. If You would start an other system and change
> > > files in that filesystem, You may image what happens?
> > > If it was possible in SuSE 10.0 I gues it was a bug!?!
> >
> > Why should it be a bug? In my opinion if the state information of the
> > systems are stored in different locations suspending one system and
> > resuming another one is not risky but of course you should mount files
I should have reread before sending it, the text should be:
you should not mount
Nevertheless some comments
> > from a suspend system read/writeable. I think, windows will not use
> > the linux swap.
> Do You really mount your partitions from suspended system read/writeable???
> One example:
> I assume to have windows with a fat32 partition (to exchange data with linux)
> and run linux with the linux partitions (of course...) and the fat32
> partition mounted. Then I suspend linux and use windows to write some files
> to the fat32 partition . I'm not shure that all changes made by linux are
> completly written to the disk... ok, linux kernel sources are available,
>From the point of (end) user I would say:
I expect that linux will sync and write all buffers etc. before suspending.
The information in the swap is used to speed the boot process up and - mainly -
to restore the status with respect to running applications, open data sets
(of the end user) and so on during resumption. It can not be that I will have
a system with uncertain status or iprobably lost data if the information on the
swap is lost for some reason. If this should be the case I have to think which
unix I will use in the future.
> perhaps can one of the kernel developers answer this...
>
> Then I suspend windows and start linux. Every operating system stores
> information about nodes, trees or fat sectors in RAM. So I think this
> information goes to the swap partition on suspend. And now - the changes made
> by windows are different to the state stored in the linux swap.
>
> Maybe the linux kernel is very good and compares this stored information with
> the actual information on the windows partition - and cleans the cache?
>
> What's about the next step? Linux writes somes files to the fat32 partition
> and windows resumes? Windows does'nt know anything about other operating
> systems ("I am windows, your operation system. You may not use other systems
> than windows ;-))
see above: a small but important error - I did not type the word "not".
But principially it should be the same. The storage of additional data has the
purpose to put the system into the status the system had when the user "stopped
working". It is a service for the user and not for the boot process or the OS.
Therefore the suspension must be clean and safe. This is important especially
for suspension to RAM. The additional infos are stored in RAM only and no
syncing took place and power will fail ?
>
> I cannot exclude that it can work without problems for a log time - but
> nevertheless I'm NOT shure that it is safe to do it except all comon data
> partitions are mounted readonly (but than I can't exchange data between the
> systems).
>
> I for myself solved this problem by using two computers one for windows (video
> processing ...) and one for linux (for all other work). And since a few month
> I use vmware and can have to systems on one computer - at the very same time,
> not for video processing, but to proof web site layouts with ie or programm
> with visual studio (grrrrrr, but some people seem to need really).
Yes of course the situation has changed. I have a similar situation. I have never
used windows for my "normal" work.
Kind regards, Wilfried
>
> Greetings
> Martin
> >
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: opensuse-mobile+unsubscribe@xxxxxxxxxxxx
> For additional commands, e-mail: opensuse-mobile+help@xxxxxxxxxxxx
>
---------------------------------------------------------------------
To unsubscribe, e-mail: opensuse-mobile+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-mobile+help@xxxxxxxxxxxx
| < Previous | Next > |