| Message ID | 20200413190843.044112674@gmail.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Preparing to phase out uswsusp | expand |
On Monday, April 13, 2020 9:08:43 PM CEST Domenico Andreoli wrote: > Dear maintainers, > > it is my understanding that the usermode software suspend is not any > more the first choice for suspend/hibernate needs (uswsusp tool did > not receive any update in years). While the last part is true, the user space utilities are still usable. Moreover, the kernel-based hibernation is still missing one important piece with respect to the user space variant, which is image encryption. That gap needs to be removed before the user space variant can be regarded as deprecated. > This small series aims to better separate the kernel side in view of > a future phase out. > > Marking the new option as deprecated and disabled by default is quite > extreme but I've not any specific agenda for pushing it that far, > it's just my interpretation and could be totally wrong. It is not totally wrong, but a bit premature at this point. Thanks!
On Sun, Apr 26, 2020 at 05:46:37PM +0200, Rafael J. Wysocki wrote: > On Monday, April 13, 2020 9:08:43 PM CEST Domenico Andreoli wrote: > > Dear maintainers, > > > > it is my understanding that the usermode software suspend is not any > > more the first choice for suspend/hibernate needs (uswsusp tool did > > not receive any update in years). > > While the last part is true, the user space utilities are still usable. > Indeed I've been using them until now and work perfectly. > Moreover, the kernel-based hibernation is still missing one important > piece with respect to the user space variant, which is image encryption. > Interesting, is it advertised anywhere? > That gap needs to be removed before the user space variant can be > regarded as deprecated. > > > This small series aims to better separate the kernel side in view of > > a future phase out. > > > > Marking the new option as deprecated and disabled by default is quite > > extreme but I've not any specific agenda for pushing it that far, > > it's just my interpretation and could be totally wrong. > > It is not totally wrong, but a bit premature at this point. Thanks for clarifying. I CCed the Debian bug which requested the removal of uswsusp from the archive. > > Thanks!
On Mon, Apr 27, 2020 at 11:53 AM Domenico Andreoli <domenico.andreoli@linux.com> wrote: > > On Sun, Apr 26, 2020 at 05:46:37PM +0200, Rafael J. Wysocki wrote: > > On Monday, April 13, 2020 9:08:43 PM CEST Domenico Andreoli wrote: > > > Dear maintainers, > > > > > > it is my understanding that the usermode software suspend is not any > > > more the first choice for suspend/hibernate needs (uswsusp tool did > > > not receive any update in years). > > > > While the last part is true, the user space utilities are still usable. > > > > Indeed I've been using them until now and work perfectly. > > > Moreover, the kernel-based hibernation is still missing one important > > piece with respect to the user space variant, which is image encryption. > > > > Interesting, is it advertised anywhere? I'm not sure what you mean by "advertised". There is a way to combine hibernation with dm-crypt which is kind of described by Documentation/power/swsusp-dmcrypt.rst, but the lack of encryption support in the built-in hibernation code is not documented otherwise AFAICS.
On Wed, Apr 29, 2020 at 01:06:01PM +0200, Rafael J. Wysocki wrote: > On Mon, Apr 27, 2020 at 11:53 AM Domenico Andreoli > <domenico.andreoli@linux.com> wrote: > > > > On Sun, Apr 26, 2020 at 05:46:37PM +0200, Rafael J. Wysocki wrote: > > > On Monday, April 13, 2020 9:08:43 PM CEST Domenico Andreoli wrote: > > > > Dear maintainers, > > > > > > > > it is my understanding that the usermode software suspend is not any > > > > more the first choice for suspend/hibernate needs (uswsusp tool did > > > > not receive any update in years). > > > > > > While the last part is true, the user space utilities are still usable. > > > > > > > Indeed I've been using them until now and work perfectly. > > > > > Moreover, the kernel-based hibernation is still missing one important > > > piece with respect to the user space variant, which is image encryption. > > > > > > > Interesting, is it advertised anywhere? > > I'm not sure what you mean by "advertised". Wrong and misinformed wording, encryption is also mentioned on uswsusp web page http://suspend.sourceforge.net/. > > There is a way to combine hibernation with dm-crypt which is kind of > described by Documentation/power/swsusp-dmcrypt.rst, but the lack of > encryption support in the built-in hibernation code is not documented > otherwise AFAICS. I use a full disk encryption on Debian, swap included. I've to say that everything works flawlessly, after the pass phrase is entered it resumes from hibernation whatever system I used to hibernate: it works either with uswsusp or in-kernel mechanism. Is there any advantage in the approach of swsusp-dmcrypt.rst? Thanks, Domenico