[systemd-devel] nspawn: No Return key in machinectl login?

[Tobias Hunger]

Hi Lennart,

sorry, this took a bit longer than expected, but I took the time to
upgrade to systemd 221, so the results should be a bit closer to the
current state than before (which was still using systemd 219).

Inside container (broken shell/fish):

speed 38400 baud; rows 60; columns 184; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = <undef>;
eol2 = <undef>; swtch = <undef>; start = ^Q; stop = ^S; susp = ^Z;
rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O;
min = 1; time = 0;
-parenb -parodd -cmspar cs8 hupcl -cstopb cread -clocal -crtscts
-ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl
-ixon -ixoff -iuclc -ixany -imaxbel iutf8
opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0
isig icanon -iexten echo echoe echok -echonl -noflsh -xcase -tostop
-echoprt -echoctl echoke

Inside container (working shell/bash):

speed 38400 baud; rows 60; columns 184; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = <undef>;
eol2 = <undef>; swtch = <undef>; start = ^Q; stop = ^S; susp = ^Z;
rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O;
min = 1; time = 0;
-parenb -parodd -cmspar cs8 hupcl -cstopb cread -clocal -crtscts
-ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl
-ixon -ixoff -iuclc -ixany -imaxbel iutf8
opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0
isig icanon -iexten echo echoe echok -echonl -noflsh -xcase -tostop
-echoprt -echoctl echoke

Outside container:

speed 38400 baud; rows 60; columns 184; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = <undef>;
eol2 = <undef>; swtch = <undef>; start = ^Q; stop = ^S; susp = ^Z;
rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O;
min = 1; time = 0;
-parenb -parodd -cmspar cs8 -hupcl -cstopb cread -clocal -crtscts
-ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl
-ixon -ixoff -iuclc -ixany -imaxbel iutf8
-opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0
-isig -icanon -iexten -echo echoe echok -echonl -noflsh -xcase -tostop
-echoprt echoctl echoke


Diff inside/outside:
opost/-opost isig/-isig icanon/-icanon echo/-echo -echoctl/echoctl


I am not 100% sure I did gather this data correctly though: Systemd
had already brought up the machine during boot, so I just ran
machinectl login and then did the stty commands there.

I also did not log out to switch between fish and bash, I just started
bash from fish, so I am not surprised that the results are -- to my
untrained eye -- identical for these two cases.

Running ssty -a -F /dev/pts/2 in another konsole window did produce
some differences.

It does not seem to effect the result in any way whether or not I run
this in konsole or xterm.

I got this on a laptop, so if you want to play with a machine that
shows this behavior I can demonstrate it if that help:-)

Best Regards,
Tobias


On Fri, Jun 19, 2015 at 10:15 AM, Tobias Hunger <tobias.hunger at gmail.com> wrote:
> Thanks for the reply!
>
> I'll try to collect all requested info tonight or over the weekend.
>
> Best Regards,
> Tobias
>
> On Thu, Jun 18, 2015 at 9:24 PM, Lennart Poettering
> <lennart at poettering.net> wrote:
>> On Tue, 26.05.15 21:40, Tobias Hunger (tobias.hunger at gmail.com) wrote:
>>
>>> This is stty -a from outside the container:
>>>
>>> speed 38400 baud; rows 46; columns 114; line = 0;
>>> intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = M-^?;
>>> eol2 = M-^?; swtch = <undef>; start = ^Q;
>>> stop = ^S; susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O;
>>> min = 1; time = 0;
>>> -parenb -parodd -cmspar cs8 -hupcl -cstopb cread -clocal -crtscts
>>> -ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr icrnl
>>> -ixon -ixoff -iuclc ixany imaxbel -iutf8
>>> opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0
>>> isig icanon iexten echo echoe echok -echonl -noflsh -xcase -tostop
>>> -echoprt echoctl echoke
>>>
>>> This is stty -a inside the nspawn-container:
>>>
>>> speed 38400 baud; rows 46; columns 114; line = 0;
>>> intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = <undef>;
>>> eol2 = <undef>; swtch = <undef>; start = ^Q;
>>> stop = ^S; susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O;
>>> min = 1; time = 0;
>>> -parenb -parodd -cmspar cs8 hupcl -cstopb cread -clocal -crtscts
>>> -ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl
>>> -ixon -ixoff -iuclc -ixany -imaxbel iutf8
>>> opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0
>>> isig icanon -iexten echo echoe echok -echonl -noflsh -xcase -tostop
>>> -echoprt -echoctl echoke
>>>
>>> The difference is:
>>> eol, eol2, -icrnl -ixany -imaxbel iutf8 -iexten -echoctl
>>
>> Sorry for the late reply...
>>
>> Hmm, I think the most interesting info would actually be to see stty
>> -a from a working instance, and from a non-working
>> instance. I.e. start the container, log into it, type "stty -a" command when
>> everything works, and when it doesn't, and let me know the diff of it.
>>
>> Also, right after doing the "stty -a" in the container, please run the
>> same commands on the host, in a seperate xterm, but connect to the
>> host side container tty using "stty -a -F /dev/pts/xyz", where
>> /dev/pts/xyz is the pts that nspawn itself is running on.
>>
>> Or to explain it in more steps:
>>
>> a) open an xterm of some form
>>
>> b) type "tty" into it, and remember the pty name it responds. It
>>    should be something like "/dev/pts/xyz".
>>
>> c) now run systemd-nspawn inside the xterm, and login there, then type
>>    "stty -a" in it, and save the output that command generated
>>    somewhere.
>>
>> d) now, leave everything as it is now, open a second xterm. In it run
>>    "stty -a -F /dev/pts/xyz", replacing "/dev/pts/xyz" with the pty
>>    name from step b) and save the output somwhere.
>>
>> Then, close both xterms. Do these steps once for a container where
>> things work, and once for a container where things are borked. Then
>> let me know the diffs between the working and non-working outputs from
>> both runs of c), as we as the diffs between the working and
>> non-working outputs from both runs of d).
>>
>> Make sure you take the stty snapshots at the exact same states each
>> time, because shells and so on tend to toggle some bits of it
>> depending on whether they are in the fg or not...
>>
>> Also, it would be good, to check if different xterm implementations
>> (gnome, kde, original xterm) behave differently.
>>
>> Lennart
>>
>> --
>> Lennart Poettering, Red Hat