[script splash] missed passphrase keypress for encrypted HD

Could you try plymouth 0.8.0~-14 and see whether that has the same issue

Which output plugin are you using?

Problem is still here, and I've seen it with both the framebuffer renderer and the drm renderer. Again, this is only in the MSI Wind U100, not in my desktop machine.

After trying everything as installed by default, I then extended the passphrase box so I could verify every character. This works, but everyone doing this needs to be sure to extend the box well beyond the length needed, so as to avoid giving any adversary a clue as to the length of the passphrase.

> Date: Sat, 13 Mar 2010 10:40:06 +0000
> From: <email address hidden>
> To: <email address hidden>
> Subject: [Bug 537793] Re: missed passphrase keypress bug is back on MSI Wind U100
>
> Could you try plymouth 0.8.0~-14 and see whether that has the same issue
>
> Which output plugin are you using?
>
> ** Changed in: plymouth (Ubuntu)
> Status: New => Incomplete
>
> ** Changed in: plymouth (Ubuntu)
> Importance: Undecided => Medium
>
> --
> missed passphrase keypress bug is back on MSI Wind U100
> https://bugs.launchpad.net/bugs/537793
> You received this bug notification because you are a direct subscriber
> of the bug.

_________________________________________________________________
Hotmail has tools for the New Busy. Search, chat and e-mail from your inbox.
http://www.windowslive.com/campaign/thenewbusy?ocid=PID27925::T:WLMTAGL:ON:WL:en-US:WM_HMP:032010_1

On Sun, 2010-03-14 at 01:41 +0000, Luke wrote:

> After trying everything as installed by default, I then extended the
> passphrase box so I could verify every character. This works, but
> everyone doing this needs to be sure to extend the box well beyond the
> length needed, so as to avoid giving any adversary a clue as to the
> length of the passphrase.
>
I'm confused.

You're saying that simply increasing the size of the box in the theme
means that your passphrase *works*, but if you type the exact same
passphrase when the box is smaller, your passphrase isn't recognised?

Scott
--
Scott James Remnant
<email address hidden>

No-increasing the length of the passphrase box is just so I get a dot for every character, so I can see if any of the last ones were missed. Has NO effect on the missed keypresses themselves, just allows me to detect them. If I can see a missed keypress, I can strike that character again-and I have come to expect to do this on the netbook.

> Date: Mon, 15 Mar 2010 12:03:27 +0000
> From: <email address hidden>
> To: <email address hidden>
> Subject: RE: [Bug 537793] Re: missed passphrase keypress bug is back on MSI Wind U100
>
> On Sun, 2010-03-14 at 01:41 +0000, Luke wrote:
>
> > After trying everything as installed by default, I then extended the
> > passphrase box so I could verify every character. This works, but
> > everyone doing this needs to be sure to extend the box well beyond the
> > length needed, so as to avoid giving any adversary a clue as to the
> > length of the passphrase.
> >
> I'm confused.
>
> You're saying that simply increasing the size of the box in the theme
> means that your passphrase *works*, but if you type the exact same
> passphrase when the box is smaller, your passphrase isn't recognised?
>
> Scott
> --
> Scott James Remnant
> <email address hidden>
>
> --
> missed passphrase keypress bug is back on MSI Wind U100
> https://bugs.launchpad.net/bugs/537793
> You received this bug notification because you are a direct subscriber
> of the bug.

_________________________________________________________________
The New Busy is not the old busy. Search, chat and e-mail from your inbox.
http://www.windowslive.com/campaign/thenewbusy?ocid=PID27925::T:WLMTAGL:ON:WL:en-US:WM_HMP:032010_3

On Mon, 2010-03-15 at 18:10 +0000, Luke wrote:

> No-increasing the length of the passphrase box is just so I get a dot
> for every character, so I can see if any of the last ones were missed.
> Has NO effect on the missed keypresses themselves, just allows me to
> detect them. If I can see a missed keypress, I can strike that
> character again-and I have come to expect to do this on the netbook.
>
Got ya.

What is the output of "cat /proc/fb" ?

Scott
--
Scott James Remnant
<email address hidden>

Here's the output of cat /proc/fb run as sudo:

0 inteldrmfb
1 VGA16 VGA

> What is the output of "cat /proc/fb" ?
>
> Scott
> --

_________________________________________________________________
Hotmail has tools for the New Busy. Search, chat and e-mail from your inbox.
http://www.windowslive.com/campaign/thenewbusy?ocid=PID27925::T:WLMTAGL:ON:WL:en-US:WM_HMP:032010_1

This may be related to the TTY raw/cooked issues - could you try the latest updates to plymouth and upstart, and see whether this fixes the problem

  A quick test boot (I'm in a park on batteries) showed no sign of the keypress bug with the new plymouth, smoothly took every character. I will let you know if the problem resurfaces, but the odds that this boot would have been this clean with the last version would have been less than 1 in 10.

> Date: Thu, 18 Mar 2010 17:02:51 +0000
> From: <email address hidden>
> To: <email address hidden>
> Subject: [Bug 537793] Re: missed passphrase keypress bug is back on MSI Wind U100
>
> This may be related to the TTY raw/cooked issues - could you try the
> latest updates to plymouth and upstart, and see whether this fixes the
> problem
>
> --
> missed passphrase keypress bug is back on MSI Wind U100
> https://bugs.launchpad.net/bugs/537793
> You received this bug notification because you are a direct subscriber
> of the bug.

_________________________________________________________________
The New Busy is not the old busy. Search, chat and e-mail from your inbox.
http://www.windowslive.com/campaign/thenewbusy?ocid=PID27925::T:WLMTAGL:ON:WL:en-US:WM_HMP:032010_3

On Fri, 2010-03-19 at 00:50 +0000, Luke wrote:

> A quick test boot (I'm in a park on batteries) showed no sign of the
> keypress bug with the new plymouth, smoothly took every character. I
> will let you know if the problem resurfaces, but the odds that this
> boot would have been this clean with the last version would have been
> less than 1 in 10.
>
This would make sense to me.

Since you're using encrypted filesystems, you must have cryptsetup
installed, so Plymouth is starting from the initramfs. Plymouth will
set its TTY up to be in "raw mode", and make it the active VT on the
console.

Upstart starts later, to bring up the full system; and some code existed
in there (the last thing done only because sysvinit did) to reset the
console with "sane" settings. This put the console back into "cooked
mode", ie. Plymouth's active VT.

Since Plymouth is now receiving cooked key codes not raw key codes, it's
not what it expects; and thus we have the bug where keypresses are
missed or mis-interpreted.

Later on, this same problem would cause X to crash on the Enter key. So
it should be fixed as a result of that fix, we not only fixed bits in
Plymouth but dropped that code from Upstart too. I'll mark it fixed
now, but please do reopen if you have the issue again, and we'll walk
through the code and see what else is going screwy.

 status fixreleased

Scott
--
Scott James Remnant
<email address hidden>

  In the past few days of use, I've still had an occasional missed keypress, down to maybe one keypress every two or three entries of the (long) passphrase. Plymouth and cryptsetup are being run from the initramfs, since this is full disk encryption.

  I have tested the keyboard itself after booting by firing up Leafpad and looking for missed presses, finding none. Still, I cannot rule out a keyboard driver bug that early in the boot process, so plymouth may or many not even be at fault for those.

Still, it's been a hell of a lot easier to boot the MSI since the last round of plymouth fixes-thanks!

> Date: Fri, 19 Mar 2010 02:49:50 +0000
> From: <email address hidden>
> To: <email address hidden>
> Subject: RE: [Bug 537793] Re: missed passphrase keypress bug is back on MSI Wind U100
>
> On Fri, 2010-03-19 at 00:50 +0000, Luke wrote:
>
> > A quick test boot (I'm in a park on batteries) showed no sign of the
> > keypress bug with the new plymouth, smoothly took every character. I
> > will let you know if the problem resurfaces, but the odds that this
> > boot would have been this clean with the last version would have been
> > less than 1 in 10.
> >
> This would make sense to me.
>
> Since you're using encrypted filesystems, you must have cryptsetup
> installed, so Plymouth is starting from the initramfs. Plymouth will
> set its TTY up to be in "raw mode", and make it the active VT on the
> console.
>
> Upstart starts later, to bring up the full system; and some code existed
> in there (the last thing done only because sysvinit did) to reset the
> console with "sane" settings. This put the console back into "cooked
> mode", ie. Plymouth's active VT.
>
> Since Plymouth is now receiving cooked key codes not raw key codes, it's
> not what it expects; and thus we have the bug where keypresses are
> missed or mis-interpreted.
>
>
> Later on, this same problem would cause X to crash on the Enter key. So
> it should be fixed as a result of that fix, we not only fixed bits in
> Plymouth but dropped that code from Upstart too. I'll mark it fixed
> now, but please do reopen if you have the issue again, and we'll walk
> through the code and see what else is going screwy.
>
> status fixreleased
>
> Scott
> --
> Scott James Remnant
> <email address hidden>
>
>
> ** Changed in: plymouth (Ubuntu)
> Status: Incomplete => Fix Released
>
> --
> missed passphrase keypress bug is back on MSI Wind U100
> https://bugs.launchpad.net/bugs/537793
> You received this bug notification because you are a direct subscriber
> of the bug.

_________________________________________________________________
Hotmail is redefining busy with tools for the New Busy. Get more from your inbox.
http://www.windowslive.com/campaign/thenewbusy?ocid=PID27925::T:WLMTAGL:ON:WL:en-US:WM_HMP:032010_2

Bad news-this is now back, worse than ever, with Plymouth 0.8.1-1ubuntu3, and kernel 2.8.32-18-generic. Came back with the last couple day's updates. Some boot attempts OK, others miss so many keypresses it's better to power-cycle and start over.

> Date: Wed, 24 Mar 2010 19:45:00 +0000
> From: <email address hidden>
> To: <email address hidden>
> Subject: RE: [Bug 537793] Re: missed passphrase keypress bug is back on MSI Wind U100
>
>
> In the past few days of use, I've still had an occasional missed keypress, down to maybe one keypress every two or three entries of the (long) passphrase. Plymouth and cryptsetup are being run from the initramfs, since this is full disk encryption.
>
> I have tested the keyboard itself after booting by firing up Leafpad
> and looking for missed presses, finding none. Still, I cannot rule out a
> keyboard driver bug that early in the boot process, so plymouth may or
> many not even be at fault for those.
>
> Still, it's been a hell of a lot easier to boot the MSI since the last
> round of plymouth fixes-thanks!
>
>
>
> > Date: Fri, 19 Mar 2010 02:49:50 +0000
> > From: <email address hidden>
> > To: <email address hidden>
> > Subject: RE: [Bug 537793] Re: missed passphrase keypress bug is back on MSI Wind U100
> >
> > On Fri, 2010-03-19 at 00:50 +0000, Luke wrote:
> >
> > > A quick test boot (I'm in a park on batteries) showed no sign of the
> > > keypress bug with the new plymouth, smoothly took every character. I
> > > will let you know if the problem resurfaces, but the odds that this
> > > boot would have been this clean with the last version would have been
> > > less than 1 in 10.
> > >
> > This would make sense to me.
> >
> > Since you're using encrypted filesystems, you must have cryptsetup
> > installed, so Plymouth is starting from the initramfs. Plymouth will
> > set its TTY up to be in "raw mode", and make it the active VT on the
> > console.
> >
> > Upstart starts later, to bring up the full system; and some code existed
> > in there (the last thing done only because sysvinit did) to reset the
> > console with "sane" settings. This put the console back into "cooked
> > mode", ie. Plymouth's active VT.
> >
> > Since Plymouth is now receiving cooked key codes not raw key codes, it's
> > not what it expects; and thus we have the bug where keypresses are
> > missed or mis-interpreted.
> >
> >
> > Later on, this same problem would cause X to crash on the Enter key. So
> > it should be fixed as a result of that fix, we not only fixed bits in
> > Plymouth but dropped that code from Upstart too. I'll mark it fixed
> > now, but please do reopen if you have the issue again, and we'll walk
> > through the code and see what else is going screwy.
> >
> > status fixreleased
> >
> > Scott
> > --
> > Scott James Remnant
> > <email address hidden>
> >
> >
> > ** Changed in: plymouth (Ubuntu)
> > Status: Incomplete => Fix Released
> >
> > --
> > missed passphrase keypress bug is back on MSI Wind U100
> > https://bugs.launchpad.net/bugs/537793
> > You received this bug notification because you are a direct subscriber
> > of the bug.
>
> _________________________________________...

Confirming bug is here on Plymouth 0.8.1-1ubuntu3 and EeePC 1005PE (Intel Atom as well). I've seen the bug with all Plymouth versions since Lucid Beta so far.

I would describe it this way : On Plymouth passphrase dialog, I have to type SLOWLY and keep keys pressed firmly and long enough and check that a "bullet" is displayed for every key I press. If I type "normally fast", Plymouth misses 10-20% of the keys pressed.

Of course my machine's keyboard works perfectly once machine is booted...

Shouldn't this bug be considered a duplicate of #497115 ? They really look the same.

No, because bug #497115 is filed by a user who no longer experiences the problem. Same symptom, different cause.

I can confirm this bug, too.
Same behaviour as described by Swâmi Petaramesh.
lucid beta1
Plymouth 0.8.1-1ubuntu3
Acer Aspire 5740G (core i5)
If I type very slow I have a good chance, that every keystroke is recognized.
If I type faster, keystrokes are randomly missing.

Maybe the title of the bug should be changed, because it seems that it is not affecting only the MSI Wind.

I believe I am seeing this on an ASUS 1001P pinetrail netbook.

I'm on a Core 2 Quad desktop machine (so no Atom), using EFI VGA framebuffer and experiencing the same thing. However,
as a workaround, I have removed the "quiet splash" options from the kernel command line, so now I use text mode
to enter my LUKS password -- miss-free.

Using linux 2.6.32-20-generic, plymouth 0.8.1-4ubuntu1 and upstart 0.6.5-6.

HTH, Jan

Should the importance of this bug be bumped to High? My rationale for doing so would be the following two points from the High category on http://wiki.kubuntu.org/Bugs/Importance :
    * Has a severe impact on a small portion of Ubuntu users (estimated)
    * Makes a default Ubuntu installation generally unusable for some users
          o For example, if the system fails to boot, or X fails to start, on a certain make and model of computer

This bug has a severe impact on a small to moderate portion of Ubuntu users (everybody with an encrypted hard disk, so what I'm guessing is a large portion of people with laptops or netbooks). It makes the system nearly unbootable after a fresh install. The user needs to know that (s)he'll need to boot with recovery mode or manually edit the grub line the first time around, then edit /etc/default/grub and set the GRUB_CMDLINE_LINUX_DEFAULT option to "" to work around the issue.

IMHO bumping it to High is adequate as the "This bug affects n people" counter keeps rising and soon it could even meet "Has a severe impact on a moderate portion of Ubuntu users". And they say "High: A bug which fulfills _one_ of the following criteria".

   Since kernel 2.6.32-21-generic came out, I have seen exactly one missed keypress in days of use on the MSI Wind U100 with Intel Atom. Possibly a more efficient keyboard driver helped that machine.

      Does the newest kernel update help on any of the other machines with this problem? If thgis isn't fixed, a LOT of people are going to want to disable Plymouth in the initramfs and run Cryptsetup from console. I use a custom initramfs configured that way in single-user mode and it works fine. You can put splash back on the grub command line and Plymouth will launch on the root fs init.

This fix is ugly, so I only use it in single-user mode. Since a lot of machines (like my AMD) never get the missed keypress bug, some easy way that non-hackers could choose to use or not use Plymouth in the initramfs with Cryptsetup would be nice. A whiptail dialog box running during the initial configuration of either cryptsetup or plymouth would work fine for this-and it would need to be easy to reconfigure. I do this stuff by hand, manually disabling execution on Plymouth and making an initramfs labelled "single" for that purpose, but how many thousands of laptop users need encryption and don't know how to manually work in the initramfs?

> Date: Thu, 15 Apr 2010 18:44:36 +0000
> From: <email address hidden>
> To: <email address hidden>
> Subject: [Bug 537793] Re: missed passphrase keypress for encrypted HD
>
> IMHO bumping it to High is adequate as the "This bug affects n people"
> counter keeps rising and soon it could even meet "Has a severe impact on
> a moderate portion of Ubuntu users". And they say "High: A bug which
> fulfills _one_ of the following criteria".
>
> --
> missed passphrase keypress for encrypted HD
> https://bugs.launchpad.net/bugs/537793
> You received this bug notification because you are a direct subscriber
> of the bug.

_________________________________________________________________
The New Busy is not the old busy. Search, chat and e-mail from your inbox.
http://www.windowslive.com/campaign/thenewbusy?ocid=PID28326::T:WLMTAGL:ON:WL:en-US:WM_HMP:042010_3

  This is the first I heard about the grub command line workaround. I have simply checked each dot to verify that the characters were accepted. If I were unaware that the keypresses were being ignored, it would have been unbootable. Some boot attempts missed enough attempts that if I got two failures to accept the finished passphrase or too many missed dots I would power cycle the machine and it would then usually work better, as the severity varied from boot to boot.

> Date: Thu, 15 Apr 2010 10:21:40 +0000
> From: <email address hidden>
> To: <email address hidden>
> Subject: [Bug 537793] Re: missed passphrase keypress for encrypted HD
>
> Should the importance of this bug be bumped to High? My rationale for doing so would be the following two points from the High category on http://wiki.kubuntu.org/Bugs/Importance :
> * Has a severe impact on a small portion of Ubuntu users (estimated)
> * Makes a default Ubuntu installation generally unusable for some users
> o For example, if the system fails to boot, or X fails to start, on a certain make and model of computer
>
> This bug has a severe impact on a small to moderate portion of Ubuntu
> users (everybody with an encrypted hard disk, so what I'm guessing is a
> large portion of people with laptops or netbooks). It makes the system
> nearly unbootable after a fresh install. The user needs to know that
> (s)he'll need to boot with recovery mode or manually edit the grub line
> the first time around, then edit /etc/default/grub and set the
> GRUB_CMDLINE_LINUX_DEFAULT option to "" to work around the issue.
>
> --
> missed passphrase keypress for encrypted HD
> https://bugs.launchpad.net/bugs/537793
> You received this bug notification because you are a direct subscriber
> of the bug.

_________________________________________________________________
Hotmail has tools for the New Busy. Search, chat and e-mail from your inbox.
http://www.windowslive.com/campaign/thenewbusy?ocid=PID28326::T:WLMTAGL:ON:WL:en-US:WM_HMP:042010_1

I too can confirm that this issue has disappeared with the latest upgrade to
Linux qoppa 2.6.32-21-generic #32-Ubuntu SMP Fri Apr 16 08:10:02 UTC 2010 i686
GNU/Linux

This is on an MSI Wind U123

Cheers,
Ryan

Looks like it's fixed here on EeePC 1005PE with latest updates (Plymouth Version: 0.8.2-2 and kernel 2.6.32-21-generic i686), confirming Ryan note above, so we *might* consider this issue possibly fixed ?

0.8.2-2 doesn't have any changes that would have fixed this, and it was reproducible for me here with 0.8.2-1 (finally have a reproducible test case) - so I don't believe this is fxed. Will test.

Steve, have you checked if this happens still for you (in your test case).

From the other comments it looks resolved.

Yes, this was still present in lucid final.

post-lucid I have not encountered this issue, so I think it's probably resolved now.