Re: [opensuse-factory] Next snapshot with updated systemd and plymouth
Ohh, there are more than that one. I remember, that I also commented on one of these: no graphical input, instead textmode input is shown. The question is, whether these are indeed related... This time, no imput is shown at all (here in approx. 20% of the reboots). Cheers, Robby. -- On Dienstag, 31. Januar 2017 15:55:56 CET Jeroen Mathon wrote:
Hey Robby,
It seems that Chan Ju Ping also made a bug report in 2016.
https://bugzilla.opensuse.org/show_bug.cgi?id=997200
Perhaps also worth looking at or marking as duplicate to avoid people recreating the wheel.
On 31-01-17 15:48, Robby Engelmann wrote:
You may want to have a look here: https://bugzilla.novell.com/show_bug.cgi?id=1020283
Am Dienstag, den 31.01.2017, 16:06 +0100 schrieb Robby Engelmann:
Ohh, there are more than that one. I remember, that I also commented on one of these: no graphical input, instead textmode input is shown. The question is, whether these are indeed related... This time, no imput is shown at all (here in approx. 20% of the reboots).
Hi, given the endless number of bugs we see with plymouth I think we should just drop it. Regards Oliver -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, 2017-02-01 at 09:51 +0100, Oliver Neukum wrote:
Hi,
given the endless number of bugs we see with plymouth I think we should just drop it.
What alternative are you proposing? Just reverting to 'old school' text mode during boot up? Cheers, Dominique
Dominique Leuenberger / DimStar composed on 2017-02-01 09:57 (UTC+0100):
On Wed, 2017-02-01 at 09:51 +0100, Oliver Neukum wrote:
given the endless number of bugs we see with plymouth I think we should just drop it.
What alternative are you proposing? Just reverting to 'old school' text mode during boot up?
It's one of the nice distinguishing features of Linux over alternative operating systems, watching things happen and passing the time between POST and login, being aware of the many things comprising startup, and seeing quickly if something has or is going wrong. GUI coverups of Windows, Mac and OS/2 startup processes are annoying, like Plymouth. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hello, SCNR: On Feb 1 09:57 Dominique Leuenberger / DimStar wrote (excerpt):
On Wed, 2017-02-01 at 09:51 +0100, Oliver Neukum wrote:
given the endless number of bugs we see with plymouth I think we should just drop it.
What alternative are you proposing? Just reverting to 'old school' text mode during boot up?
For me: Yes please! Probably a bit off-topic but seriously: Personally I prefer to see "what a thingy is doing" when something is working in the background for a noticeable amount of time. I.e. I like continuously reported info messages while something is working in the background for some time regardless whether or not I understand the messages. As long as such messages are continuously reported it is obvious "that thingy is still doing something". When it fails one can directly see its last messages regardless whether or not one understands the meaning. But with the last messages one can at least try to find someone who understands what goes on. I do not necessarily mean a full screen of such messages. Only one or a few status message lines (at the bottom) are perfectly sufficient as user information. Unfortunately it seems usually it is considered to be better to not tell the user anything about what goes on behind when something is working in the background for some time cf. "no news is good news". Accordingly SUSE and Fedora and Ubuntu show basically empty screens - sometimes even with some meaningless animated graphical stuff that tells exactly nothing where the user is forced to do dull waiting until that nonsense (I mean it literally: "without sense") is gone. Kind Regards Johannes Meixner -- SUSE LINUX GmbH - GF: Felix Imendoerffer, Jane Smithard, Graham Norton - HRB 21284 (AG Nuernberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wednesday 2017-02-01 10:36, Johannes Meixner wrote:
I like continuously reported info messages[...] As long as such messages are continuously reported it is obvious "that thingy is still doing something". When it fails one can directly see its last messages regardless whether or not one understands the meaning. [...] I do not necessarily mean a full screen of such messages. Only one or a few status message lines (at the bottom) are perfectly sufficient as user information.
That sounds just exactly like what plymouth is doing. The progress bar is the message. It meets the criteria of "not occupying the entire screen". It meets the criteria that while "that thingy is still doing something", it will update. And when it fails or does nothing (hello overly long systemd timeouts), one can see its last (textual) message with ESC, whether one understands the meaning of the text or not. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hello, On Feb 1 13:02 Jan Engelhardt wrote:
On Wednesday 2017-02-01 10:36, Johannes Meixner wrote:
I like continuously reported info messages[...] As long as such messages are continuously reported it is obvious "that thingy is still doing something". When it fails one can directly see its last messages regardless whether or not one understands the meaning. [...] I do not necessarily mean a full screen of such messages. Only one or a few status message lines (at the bottom) are perfectly sufficient as user information.
That sounds just exactly like what plymouth is doing. The progress bar is the message. It meets the criteria of "not occupying the entire screen". It meets the criteria that while "that thingy is still doing something", it will update. And when it fails or does nothing (hello overly long systemd timeouts), one can see its last (textual) message with ESC, whether one understands the meaning of the text or not.
You left out an important part from what I had written which makes it look different compared to what I meant and a progress bar is no info messages. I had also written: ----------------------------------------------------------- ... SUSE and Fedora and Ubuntu show basically empty screens - sometimes even with some meaningless animated graphical stuff that tells exactly nothing ... ----------------------------------------------------------- Additionally: Let the user do this or that (e.g. pressing whatever key or so) to get some meaningful output on a basically empty screen is not what I meant. FYI: I cannot make statements what is good or bad by default for openSUSE users because I think I am not a "normal openSUSE user". (I know how to change it to make it work o.k. for my own needs.) Kind Regards Johannes Meixner -- SUSE LINUX GmbH - GF: Felix Imendoerffer, Jane Smithard, Graham Norton - HRB 21284 (AG Nuernberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Jan Engelhardt wrote:
On Wednesday 2017-02-01 10:36, Johannes Meixner wrote:
I like continuously reported info messages[...] As long as such messages are continuously reported it is obvious "that thingy is still doing something". When it fails one can directly see its last messages regardless whether or not one understands the meaning. [...] I do not necessarily mean a full screen of such messages. Only one or a few status message lines (at the bottom) are perfectly sufficient as user information.
That sounds just exactly like what plymouth is doing. The progress bar is the message. It meets the criteria of "not occupying the entire screen". It meets the criteria that while "that thingy is still doing something", it will update. And when it fails or does nothing (hello overly long systemd timeouts), one can see its last (textual) message with ESC, whether one understands the meaning of the text or not.
If the system hangs during boot and does not react to ESC key anymore the last message on a text mode screen could be useful. Not always but sometimes. Ciao, Michael.
Dominique Leuenberger / DimStar wrote:
On Wed, 2017-02-01 at 09:51 +0100, Oliver Neukum wrote:
Hi,
given the endless number of bugs we see with plymouth I think we should just drop it.
What alternative are you proposing? Just reverting to 'old school' text mode during boot up?
Yes. Why not just text mode? None of my friends (non-experienced Tumbleweed notebook users) ever complained about text mode during boot. Ciao, Michael.
On Wednesday, February 1, 2017 8:00:15 PM CET Michael Ströder wrote:
Dominique Leuenberger / DimStar wrote:
On Wed, 2017-02-01 at 09:51 +0100, Oliver Neukum wrote:
Hi,
given the endless number of bugs we see with plymouth I think we should just drop it.
What alternative are you proposing? Just reverting to 'old school' text mode during boot up?
Yes. Why not just text mode?
None of my friends (non-experienced Tumbleweed notebook users) ever complained about text mode during boot.
Ciao, Michael. Some users(Non Technical) even liked it becaouse they saw a little of what goes on inside an OS(Services whise)
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Le mercredi 01 février 2017 à 09:51 +0100, Oliver Neukum a écrit :
Am Dienstag, den 31.01.2017, 16:06 +0100 schrieb Robby Engelmann:
Ohh, there are more than that one. I remember, that I also commented on one of these: no graphical input, instead textmode input is shown. The question is, whether these are indeed related... This time, no imput is shown at all (here in approx. 20% of the reboots).
Hi,
given the endless number of bugs we see with plymouth I think we should just drop it.
How about trying to work with upstream on fixing the issues ? -- Frederic Crozat Enterprise Desktop Release Manager SUSE -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, 2017-02-01 at 10:18 +0100, Frederic Crozat wrote:
Le mercredi 01 février 2017 à 09:51 +0100, Oliver Neukum a écrit :
Am Dienstag, den 31.01.2017, 16:06 +0100 schrieb Robby Engelmann:
Ohh, there are more than that one. I remember, that I also commented on one of these: no graphical input, instead textmode input is shown. The question is, whether these are indeed related... This time, no imput is shown at all (here in approx. 20% of the reboots).
Hi,
given the endless number of bugs we see with plymouth I think we should just drop it.
How about trying to work with upstream on fixing the issues ?
Fabian has been very active in the last couple days to sort out some things in our plymouth package - to the biggest surprise, almost all hist fixes were 'drop an openSUSE patch here and there'. Seems quite some patches are just not up-do the task there. From hearing around, it seems Fedora is not struggling that much with plymouth - so either their usecases are just less critical than ours or some of our 'bugs' are not considered there . In any case, I'd say it shows once again: downstream patches are evil! Work with upstream to get our ideas merged there. Let's not develop the same applications in silos but SHARE the work. So, from my side: a total +1 to work with upstream on fixing the issues. and as for text mode (Sorry Frederic, it's out of context in this reply...): the ones that want it CAN have it - you are free to disable plymouth on your machine! But I honestly do not think that a screen full of text is very appealing to Joe Average. openSUSE has the 'name' of being a polished distro - from boot to desktop: let's keep it that way and IMPROVE on that fact, not revert to the stone ages. Cheers, Dominique
Le mercredi 01 février 2017 à 10:54 +0100, Dominique Leuenberger / DimStar a écrit :
On Wed, 2017-02-01 at 10:18 +0100, Frederic Crozat wrote:
Le mercredi 01 février 2017 à 09:51 +0100, Oliver Neukum a écrit :
Am Dienstag, den 31.01.2017, 16:06 +0100 schrieb Robby Engelmann:
Ohh, there are more than that one. I remember, that I also commented on one of these: no graphical input, instead textmode input is shown. The question is, whether these are indeed related... This time, no imput is shown at all (here in approx. 20% of the reboots).
Hi,
given the endless number of bugs we see with plymouth I think we should just drop it.
How about trying to work with upstream on fixing the issues ?
Fabian has been very active in the last couple days to sort out some things in our plymouth package - to the biggest surprise, almost all hist fixes were 'drop an openSUSE patch here and there'. Seems quite some patches are just not up-do the task there.
Yes, this is extremely unfortunate :(
From hearing around, it seems Fedora is not struggling that much with plymouth - so either their usecases are just less critical than ours or some of our 'bugs' are not considered there . In any case, I'd say it shows once again: downstream patches are evil! Work with upstream to get our ideas merged there. Let's not develop the same applications in silos but SHARE the work.
I also think part of the issues we are seeing are caused by a different in the defaults between Fedora and openSUSE Leap (and SLE ): - for quite some time, Fedora has been using VT1 for boot splash AND for graphical session. - openSUSE Leap (and SLE) stayed with the old default where VT1 is still a text console and VT7 is graphical. To ensure there is no flickering between boot splash and X starting, this required boot to start on VT7. And unfortunately, this VT switch can cause deadlock (which is not visible on Fedora, of course). I don't remember if TW is affected. But, with my SLE Release Manager hat, changing this default is "difficult" because it breaks what users (or customers) are expecting.. Even if it means less bugs in the end. Maybe this is something we need to revisit for next major Leap (43.x) and SLE13..
-- Frederic Crozat Enterprise Desktop Release Manager SUSE
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 01.02.2017 10:54, Dominique Leuenberger / DimStar wrote:
and as for text mode (Sorry Frederic, it's out of context in this reply...): the ones that want it CAN have it - you are free to disable plymouth on your machine! But I honestly do not think that a screen full of text is very appealing to Joe Average.
try "zypper rm libply4; reboot" and if you really see a screen full of text after GRUB is done with the default "quiet" boot option, something is probably very wrong on your machine. 5 Seconds of mostly black screen (the grub output is in the left upper corner, then the desktop fires up -- that's on a 7 year old Thinkpad X200s with SSD, so a modern machine will probably boot even faster.
openSUSE has the 'name' of being a polished distro - from boot to desktop: let's keep it that way and IMPROVE on that fact, not revert to the stone ages.
Well, but I'd rather have a working, fast booting distro than a "polished", slow bootin one :-) -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Have you guys tried blog-plymouth? Works fine for me. But I won't mind using plymouth. As soon as it will reliably work with mounting encrypted home. On 1.2.2017 17:32, Stefan Seyfried wrote:
On 01.02.2017 10:54, Dominique Leuenberger / DimStar wrote:
and as for text mode (Sorry Frederic, it's out of context in this reply...): the ones that want it CAN have it - you are free to disable plymouth on your machine! But I honestly do not think that a screen full of text is very appealing to Joe Average.
try "zypper rm libply4; reboot" and if you really see a screen full of text after GRUB is done with the default "quiet" boot option, something is probably very wrong on your machine.
5 Seconds of mostly black screen (the grub output is in the left upper corner, then the desktop fires up -- that's on a 7 year old Thinkpad X200s with SSD, so a modern machine will probably boot even faster.
openSUSE has the 'name' of being a polished distro - from boot to desktop: let's keep it that way and IMPROVE on that fact, not revert to the stone ages.
Well, but I'd rather have a working, fast booting distro than a "polished", slow bootin one :-)
-- Vit Pelcak vpelcak@suse.cz Team Lead in QA/Maintenance SUSE LINUX, s.r.o. CORSO IIa Krizikova 148/34 186 00 Prague 8 Czech Republic
On Wed, Feb 01, 2017 at 05:34:52PM +0100, Vit Pelcak wrote:
Have you guys tried blog-plymouth?
Welcome :) More users and testers out there? Any reports about blog? Werner -- https://github.com/bitstreamout/showconsole --------------------------------------------------------------------------- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr
*From:* Dr. Werner Fink *Sent:* 01/02/2017 15:21:53 (-0200 UTC) *To:* Opensuse-factory *Subject:* Re: [opensuse-factory] Next snapshot with updated systemd and plymouth
On Wed, Feb 01, 2017 at 05:34:52PM +0100, Vit Pelcak wrote:
Have you guys tried blog-plymouth? Welcome :)
More users and testers out there? Any reports about blog?
Werner
Sorry to add a dumb request but it would be really welcomed to add Plymouth Splash image with OPENSUSE logo instead of BIZ.COM for example. I customized my splash-screen modifying the default solar flares picture but I think that the distribution would looks more professional and polished by using brand logos! Cheers, Marco N�����r��y隊Z)z{.���r�+�맲��r��z�^�ˬz��N�(�֜��^� ޭ隊Z)z{.���r�+��0�����Ǩ�
On 01/02/17 12:45 PM, Marco Calistri wrote:
*From:* Dr. Werner Fink *Sent:* 01/02/2017 15:21:53 (-0200 UTC) *To:* Opensuse-factory *Subject:* Re: [opensuse-factory] Next snapshot with updated systemd and plymouth
On Wed, Feb 01, 2017 at 05:34:52PM +0100, Vit Pelcak wrote:
Have you guys tried blog-plymouth? Welcome :)
More users and testers out there? Any reports about blog?
Werner
Sorry to add a dumb request but it would be really welcomed to add Plymouth Splash image with OPENSUSE logo instead of BIZ.COM for example.
I customized my splash-screen modifying the default solar flares picture but I think that the distribution would looks more professional and polished by using brand logos!
Cheers,
Marco N�����r��y隊Z)z{.���r�+�맲��r��z�^�ˬz��N�(�֜��^� ޭ隊Z)z{.���r�+��0�����Ǩrg==
Don't forget to suppress the "login prompt" that's been haunting openSUSE for a number of distro releases. ;-) I don't see this "login prompt" appearing with Fedora during the boot process or shutdown. Roman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am Mittwoch, den 01.02.2017, 10:18 +0100 schrieb Frederic Crozat:
Le mercredi 01 février 2017 à 09:51 +0100, Oliver Neukum a écrit :
Am Dienstag, den 31.01.2017, 16:06 +0100 schrieb Robby Engelmann:
Ohh, there are more than that one. I remember, that I also commented on one of these: no graphical input, instead textmode input is shown. The question is, whether these are indeed related... This time, no imput is shown at all (here in approx. 20% of the reboots).
Hi,
given the endless number of bugs we see with plymouth I think we should just drop it.
How about trying to work with upstream on fixing the issues ?
That would be inefficient. The basic interfaces for the console were not designed for more than one owner of the console actually doing graphics. Thus we'd work hard on making more beautiful a gap we are working to eliminate. And booting just has to work. The amount of unreliability we can accept there is much lower than in other areas. Regards Oliver -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Le mercredi 01 février 2017 à 11:19 +0100, Oliver Neukum a écrit :
Am Mittwoch, den 01.02.2017, 10:18 +0100 schrieb Frederic Crozat:
Le mercredi 01 février 2017 à 09:51 +0100, Oliver Neukum a écrit :
Am Dienstag, den 31.01.2017, 16:06 +0100 schrieb Robby Engelmann:
Ohh, there are more than that one. I remember, that I also commented on one of these: no graphical input, instead textmode input is shown. The question is, whether these are indeed related... This time, no imput is shown at all (here in approx. 20% of the reboots).
Hi,
given the endless number of bugs we see with plymouth I think we should just drop it.
How about trying to work with upstream on fixing the issues ?
That would be inefficient. The basic interfaces for the console were not designed for more than one owner of the console actually doing graphics. Thus we'd work hard on making more beautiful a gap we are working to eliminate.
Unfortunately, even with recent hardware and SSD, this gap is still there and visible. For a desktop system, reverting to text mode is definitively a regression. -- Frederic Crozat Enterprise Desktop Release Manager SUSE -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, Feb 01, 2017 at 02:03:38PM +0100, Frederic Crozat wrote:
Le mercredi 01 février 2017 à 11:19 +0100, Oliver Neukum a écrit :
Am Mittwoch, den 01.02.2017, 10:18 +0100 schrieb Frederic Crozat:
Le mercredi 01 février 2017 à 09:51 +0100, Oliver Neukum a écrit :
Am Dienstag, den 31.01.2017, 16:06 +0100 schrieb Robby Engelmann:
Ohh, there are more than that one. I remember, that I also commented on one of these: no graphical input, instead textmode input is shown. The question is, whether these are indeed related... This time, no imput is shown at all (here in approx. 20% of the reboots).
Hi,
given the endless number of bugs we see with plymouth I think we should just drop it.
How about trying to work with upstream on fixing the issues ?
That would be inefficient. The basic interfaces for the console were not designed for more than one owner of the console actually doing graphics. Thus we'd work hard on making more beautiful a gap we are working to eliminate.
Unfortunately, even with recent hardware and SSD, this gap is still there and visible. For a desktop system, reverting to text mode is definitively a regression.
The main problem is that plymouthd starts on vt1 and switch over to vt7 and at stop of plymouthd the vt1 is reached again. Then with start of X then we will switch back to vt7. Hence this large gap IMHO plymouthd should start on vt7 ... that is a small line ExecStartPre=/usr/bin/chvt 7 together with the option --tty=/dev/tty7 on the comamnd line of /usr/sbin/plymouthd in /usr/lib/systemd/system/plymouth-start.service might help here. Maybe the line TTYPath=/dev/tty7 could be an further option in /usr/lib/systemd/system/plymouth-start.service At least plymouthd may not access vt7 at the same time a X does. Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Dr. Werner Fink writes:
The main problem is that plymouthd starts on vt1 and switch over to vt7 and at stop of plymouthd the vt1 is reached again. Then with start of X then we will switch back to vt7. Hence this large gap
I actually think that all these messages should be on their own VT so one can read at least the last screenful of them just by switching away from one of the console VT or the X session. Bonus points if that "messages VT" would have a scrollback buffer. Would that be possible? Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ Samples for the Waldorf Blofeld: http://Synth.Stromeko.net/Downloads.html#BlofeldSamplesExtra -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wednesday 2017-02-01 20:33, Achim Gratz wrote:
Dr. Werner Fink writes:
The main problem is that plymouthd starts on vt1 and switch over to vt7 and at stop of plymouthd the vt1 is reached again. Then with start of X then we will switch back to vt7. Hence this large gap
I actually think that all these messages should be on their own VT so one can read at least the last screenful of them just by switching away from one of the console VT or the X session. Bonus points if that "messages VT" would have a scrollback buffer. Would that be possible?
Already implemented. suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK=y suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK_SIZE=64 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
02.02.2017 02:56, Jan Engelhardt пишет:
On Wednesday 2017-02-01 20:33, Achim Gratz wrote:
Dr. Werner Fink writes:
The main problem is that plymouthd starts on vt1 and switch over to vt7 and at stop of plymouthd the vt1 is reached again. Then with start of X then we will switch back to vt7. Hence this large gap
I actually think that all these messages should be on their own VT so one can read at least the last screenful of them just by switching away from one of the console VT or the X session. Bonus points if that "messages VT" would have a scrollback buffer. Would that be possible?
Already implemented.
suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK=y suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK_SIZE=64
My experience is that scroll buffer is available only as long as you stay on vt. When you switch from it and back again you cannot scroll past screen content anymore. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Andrei Borzenkov composed on 2017-02-02 06:55 (UTC+0300):
Jan Engelhardt composed:
Achim Gratz wrote:
"messages VT" would have a scrollback buffer. Would that be possible?
Already implemented.
suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK=y suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK_SIZE=64
My experience is that scroll buffer is available only as long as you stay on vt. When you switch from it and back again you cannot scroll past screen content anymore.
Agree. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Thu, 02 Feb 2017 05:03:00 +0100, Felix Miata wrote:
Andrei Borzenkov composed on 2017-02-02 06:55 (UTC+0300):
Jan Engelhardt composed:
Achim Gratz wrote:
"messages VT" would have a scrollback buffer. Would that be possible?
Already implemented.
suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK=y suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK_SIZE=64
My experience is that scroll buffer is available only as long as you stay on vt. When you switch from it and back again you cannot scroll past screen content anymore.
Agree.
FWIW, recent a patchset was merged to have a persistent scroll buffer, so it'll be fixed in 4.11. Takashi -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Jan Engelhardt writes:
I actually think that all these messages should be on their own VT so one can read at least the last screenful of them just by switching away from one of the console VT or the X session. Bonus points if that "messages VT" would have a scrollback buffer. Would that be possible?
Already implemented.
I was about to say it doesn't work for me, but I see later in the thread why that is so.
suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK=y suse/master:config/x86_64/default:CONFIG_VGACON_SOFT_SCROLLBACK_SIZE=64
Hmm. That's less than a half a screen on my monitor. Better than nothing, but somwhow I was hoping that you'd be able to scroll through the full bootup sequence. Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ Samples for the Waldorf Blofeld: http://Synth.Stromeko.net/Downloads.html#BlofeldSamplesExtra -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am Mittwoch, den 01.02.2017, 14:03 +0100 schrieb Frederic Crozat:
Le mercredi 01 février 2017 à 11:19 +0100, Oliver Neukum a écrit :
Am Mittwoch, den 01.02.2017, 10:18 +0100 schrieb Frederic Crozat:
Le mercredi 01 février 2017 à 09:51 +0100, Oliver Neukum a écrit :
given the endless number of bugs we see with plymouth I think we should just drop it.
How about trying to work with upstream on fixing the issues ?
That would be inefficient. The basic interfaces for the console were not designed for more than one owner of the console actually doing graphics. Thus we'd work hard on making more beautiful a gap we are working to eliminate.
Unfortunately, even with recent hardware and SSD, this gap is still there and visible. For a desktop system, reverting to text mode is definitively a regression.
Absolutely. But so is not booting. And we have done so with unnerving regularity. We might reintroduce the kernel splash. Regards Oliver -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 01.02.2017 14:03, Frederic Crozat wrote:
Unfortunately, even with recent hardware and SSD, this gap is still there and visible.
I have alternative facts here. Booting without plymouth is faster (initrd, which is loaded by slow legacy BIOS code is much smaller). Once kernel has booted a small amount of time is saved additionally by not needing to display fancy graphics. » For a desktop system, reverting to text mode is
definitively a regression.
susi:~ # systemd-analyze Startup finished in 2.614s (kernel) + 858ms (initrd) + 13.088s (userspace) = 16.562s This is with cryptohome, entering the passphrase take some time: susi:~ # systemd-analyze blame|head -1 5.400s systemd-cryptsetup@cr_sda7.service So it's really less than 10 seconds from Kernel startup (which does not show splash anyway) to Desktop. On ancient, slow (for nowadays standards) hardware. Oh, and I run tor, autofs, postfix and ntpd, without those I could shave another 5 seconds off the boot time. -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, 01 Feb 2017 17:43:25 +0100, Stefan Seyfried wrote:
On 01.02.2017 14:03, Frederic Crozat wrote:
Unfortunately, even with recent hardware and SSD, this gap is still there and visible.
I have alternative facts here.
As the meaning Kellyanne Conway used? :) But I agree with your point. And on a modern hardware, the gap is much shorter. Currently we can't achieve the complete seamless boot splash from GRUB to X, and the flickering shall happen a few times. In such a situation, I'd rather like to have a blank screen for one or two seconds. And on my server, I'd rather like to see the actual boot messages. Takashi -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, 2017-02-01 at 17:43 +0100, Stefan Seyfried wrote:
On 01.02.2017 14:03, Frederic Crozat wrote:
Unfortunately, even with recent hardware and SSD, this gap is still there and visible.
I have alternative facts here.
Booting without plymouth is faster (initrd, which is loaded by slow legacy BIOS code is much smaller). Once kernel has booted a small amount of time is saved additionally by not needing to display fancy graphics.
» For a desktop system, reverting to text mode is
definitively a regression.
susi:~ # systemd-analyze Startup finished in 2.614s (kernel) + 858ms (initrd) + 13.088s (userspace) = 16.562s
homer:/root # systemd-analyze Startup finished in 9.208s (firmware) + 24.633s (loader) + 1.819s (kernel) + 4.976s (initrd) + 32.996s (userspace) = 1min 13.633s homer:/root # systemd-analyze blame|head -1 17.343s wicked.service Heh, wicked alone (bridge) chews up more time here than your everything combined. -Mike -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 1. Februar 2017 18:26:15 MEZ schrieb Mike Galbraith <mgalbraith@suse.de>:
On Wed, 2017-02-01 at 17:43 +0100, Stefan Seyfried wrote:
On 01.02.2017 14:03, Frederic Crozat wrote:
Unfortunately, even with recent hardware and SSD, this gap is still there and visible.
I have alternative facts here.
Booting without plymouth is faster (initrd, which is loaded by slow legacy BIOS code is much smaller). Once kernel has booted a small amount of time is saved additionally by not needing to display fancy graphics.
» For a desktop system, reverting to text mode is
definitively a regression.
susi:~ # systemd-analyze Startup finished in 2.614s (kernel) + 858ms (initrd) + 13.088s (userspace) = 16.562s
homer:/root # systemd-analyze Startup finished in 9.208s (firmware) + 24.633s (loader) + 1.819s (kernel) + 4.976s (initrd) + 32.996s (userspace) = 1min 13.633s homer:/root # systemd-analyze blame|head -1 17.343s wicked.service
Heh, wicked alone (bridge) chews up more time here than your everything combined.
Yes, and there is a bug reported against that for Leap as well ..... -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 01.02.2017 um 18:26 schrieb Mike Galbraith:
Heh, wicked alone (bridge) chews up more time here than your everything combined.
Well yeah, I of course do not use wicked but NetworkManager. But even wicked can be tamed (on Leap, but still) with a static ip setup: 6.234s wicked.service server:~ # systemctl status -n 0 wickedd-auto4.service wickedd-dhcp4.service wickedd-dhcp6.service ● wickedd-auto4.service Loaded: masked (/dev/null; bad) Active: inactive (dead) ● wickedd-dhcp4.service Loaded: masked (/dev/null; bad) Active: inactive (dead) ● wickedd-dhcp6.service Loaded: masked (/dev/null; bad) Active: inactive (dead) server:~ # grep CHECK_DUP /etc/sysconfig/network/config CHECK_DUPLICATE_IP="no" server:~ # ip a 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet6 ::1/128 scope host valid_lft forever preferred_lft forever 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:19:99:65:b7:9e brd ff:ff:ff:ff:ff:ff inet 192.168.200.1/24 brd 192.168.200.255 scope global eth0 valid_lft forever preferred_lft forever inet6 fe80::219:99ff:fe65:b79e/64 scope link valid_lft forever preferred_lft forever 3: br0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN group default qlen 1000 link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff inet 192.168.202.1/24 brd 192.168.202.255 scope global br0 valid_lft forever preferred_lft forever inet6 fe80::bc0b:1ff:fe80:bda4/64 scope link valid_lft forever preferred_lft forever But yeah, I was almost ready to just ditch it and replace it with a dumb 20 line shell script :-) -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wednesday 2017-02-01 17:43, Stefan Seyfried wrote:
On 01.02.2017 14:03, Frederic Crozat wrote:
Booting without plymouth is faster. Once kernel has booted a small amount of time is saved additionally by not needing to display fancy graphics.
Ironically, booting with splash on can be faster for some hw/drivers. On vesafb where scrolling init messages usually mean a full repaint, they can be much more of a penalty than updating a measly progress bar bbox. :) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Jan Engelhardt composed on 2017-02-02 00:54 (UTC+0100):
Ironically, booting with splash on can be faster for some hw/drivers. On vesafb where scrolling init messages usually mean a full repaint, they can be much more of a penalty than updating a measly progress bar bbox. :)
"Measly" is its own problem. Progress bars are notorious for taking up a dearth of display space and providing crude accuracy. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hi Jan, On 02.02.2017 00:54, Jan Engelhardt wrote:
On Wednesday 2017-02-01 17:43, Stefan Seyfried wrote:
Booting without plymouth is faster. Once kernel has booted a small amount of time is saved additionally by not needing to display fancy graphics.
Ironically, booting with splash on can be faster for some hw/drivers. On vesafb where scrolling init messages usually mean a full repaint, they can be much more of a penalty than updating a measly progress bar bbox. :)
If you would have tried it, you would have noticed, that with the current default, but plymouth uninstalled, there is no scrolling init messages. There is not any output at all apart from kernel error messages. Default setting is "quiet" on the kernel command line. Even without "quiet", systemd is not putting out lots of information, only one line per service, so you will hardly notice lots of scrolling with vesafb (well, maybe if you have 640x400 vesafb mode, but there scrolling should be reasonably fast again). Try it. Present numbers, not alternative facts. -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
participants (19)
-
Achim Gratz -
Andrei Borzenkov -
Axel Braun -
Dominique Leuenberger / DimStar -
Dr. Werner Fink -
Felix Miata -
Frederic Crozat -
Jan Engelhardt -
jeroenmathonmac＠gmail.com -
Johannes Meixner -
Marco Calistri -
Michael Ströder -
Mike Galbraith -
Oliver Neukum -
Robby Engelmann -
Roman Bysh -
Stefan Seyfried -
Takashi Iwai -
Vit Pelcak