On 11.05.2017 12:33, Carsten Haitzler (The Rasterman) wrote:
> On Wed, 10 May 2017 09:48:19 +0200 PaulTT <[email protected]> said:
> 
>> i just posted a message about this... (sorry, i've seen now this thread)
>>
>> as i said there, there's also a problem with unlocking (so, pam related, i
>> assume ?)
>> via console su and sudo worked like a charm (i've got error messages about
>> cpufreq and backlight too)
> 
> pam would be executing a setuid root binary to do the password check... so 
> it's
> the same issue. something has decided that e and app processes below it in the
> process tree "cant run setuid (root) binaries" and has disabled that feature.
> that feature seems to only kick in with 4.11 kernel. it certainly is not e
> doing this. it has relied on this working for many years. it's something new
> security-wise that is being enabled by a new kernel.
> 
> maybe some parent process is using setpriv? CAP_SETUID  disabled? man
> capabilities ... for info ... maybe run captest ?
>e
> 12:20PM ~ > captest
> User  credentials uid:1000 euid:1000 suid:1000
> Group credentials gid:1000 egid:1000 sgid:1000
> Current capabilities: none
> securebits flags: none
> Attempting direct access to shadow...FAILED (Permission denied)
> Attempting to access shadow by child process...FAILED
> Child User  credentials uid:1000 euid:1000 suid:1000
> Child Group credentials gid:1000 egid:1000 sgid:1000
> Child capabilities: none
> Child securebits flags: none
> 
> is what i get. which is normal.

I get the same as you on my system here:

florian@washu:~ # uname -a
Linux washu 4.11.0 #2 SMP PREEMPT Tue May 2 12:12:51 JST 2017 i686 GNU/Linux
florian@washu:~ # captest
User  credentials uid:500 euid:500 suid:500
Group credentials gid:100 egid:100 sgid:100
Current capabilities: none
securebits flags: none
Attempting direct access to shadow...FAILED (Permission denied)
Attempting to access shadow by child process...FAILED
Child User  credentials uid:500 euid:500 suid:500
Child Group credentials gid:100 egid:100 sgid:100
Child capabilities: none
Child securebits flags: none

Cheers,
Florian

>> could the problem be related to some new sh**y systemd operation????
>> i saw that also using wayland, i coulnd't access halt/reboot/suspend menu
>> items too (this happens to me also with previous kernels)
> 
> works for me with enlightenment + wl + arch (+systemd)... i can do all the
> power off etc. stuff...
> 
>> On Thu, May 4, 2017 at 6:19 AM, Carsten Haitzler <[email protected]>
>> wrote:
>>
>>> On Thu, 04 May 2017 11:09:13 +0900 <[email protected]> said:
>>>
>>>> Hi,
>>>>
>>>> Carsten Haitzler (The Rasterman) <[email protected]> wrote:
>>>>
>>>>> On Wed, 3 May 2017 12:09:21 +0900 Florian Schaefer <[email protected]>
>>> said:
>>>>
>>>>>> Hi!
>>>>>>
>>>>>> On 03.05.2017 10:04, Carsten Haitzler (The Rasterman) wrote:
>>>>>>> On Tue, 02 May 2017 21:16:40 +0900 [email protected] said:
>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> I got the source of kernel 4.11, then comipled, and check the
>>>>>>>> behaviour of enlightenment (efl 1.18.1, enlightment 0.21.7).
>>>>>>>>
>>>>>>>> At the start up of enlightenment, I got an error message;
>>>>>>>>
>>>>>>>>  There was an error trying to set the cpu power state setting via
>>> the
>>>>>>>>  module's setfreq utility.
>>>>>>>>
>>>>>>>> On the kernel 4.10.x, I never see such a message.
>>>>>>>>
>>>>>>>> And, I tried to use su and sudo command in the terminal, I got a
>>>>>>>> strange message;
>>>>>>>>
>>>>>>>> fulwood@linux-uw5l:~> sudo
>>>>>>>> sudo: effective uid is not 0, is sudo installed setuid root
>>>>>>>>
>>>>>>>> fulwood@linux-uw5l:~> su
>>>>>>>> Password:
>>>>>>>> su: incorrect password
>>>>>>>>
>>>>>>>> This means, there is a problem concerning uid treatment in the
>>>>>>>> enlightenment, doesn't it.
>>>>>>>>
>>>>>>>> Moreover, VirtualBox indicate the problem of enlightenment
>>> directly;
>>>>>>>>
>>>>>>>> fulwood@linux-uw51:~> VirtualBox
>>>>>>>>
>>>>>>>> VirtualBox: Error -10 in SUPRHardenedMain!
>>>>>>>> VirtualBox: Effective UID is not root (euid=1000, egid=100,
>>> uid=1000,
>>>>>>>>             gid=100)
>>>>>>>> VirtualBox: Tip! It may help to reintall VirtualBox.
>>>>>>>>
>>>>>>>> Why does uid=1000?
>>>>>>>
>>>>>>> that's a common uid to start with for users added to a system -
>>> first
>>>>>>> user added commonly is uid 1000... that's probably ... you.
>>>>>>>
>>>>>>>> So, we can't use enlightenment on the kernel 4.11.
>>>>>>>
>>>>>>> from the above it seems like since you compiled your own kernel it
>>>>>>> seemingly has disabled setuid root binaries. i assume this is some
>>> new
>>>>>>> feature of kernels since 4.11 that has been turned on. i suggest
>>> you
>>>>>>> turn it off to allow them again. your kernel broke far more than
>>>>>>> enlightenment. it broke sudo. probably even broke su. it broke
>>>>>>> virtualbox... it broke stuff. what that option is - i don't know.
>>> this
>>>>>>> is news to me.
>>>>>>
>>>>>> Just for the record I'd like to add that I observe the same behavior.
>>>>>>
>>>>>> Since switching from 4.9 to 4.11 yesterday I cannot do suid requiring
>>>>>> operations (like su or mount.cifs) from within E (using terminology
>>> or
>>>>>> xterm) any more. Interestingly, if I am right at the console (so no
>>> Xorg
>>>>>> and e in-between) all those commands work like a charm.
>>>>>>
>>>>>> I could not find any setuid related option in the kernel
>>> configuration
>>>>>> so I cannot really imagine where it is misconfigured.
>>>>
>>>>> it'll likely be some security option that ends up doing this for child
>>>>> processes ... whatever/however it is... but its certainly a change in
>>> the
>>>>> kernel and "security options" of some sort.
>>>>
>>>> But, why the kernel's change has an impact on enlightenment only?
>>>> On e16 and kde-plasma, no impact.
>>>
>>> neither controls cpu frequency/governor or don't use setuid root binaries
>>> or
>>> they come from packages with specific selinux rules to allow setuid root
>>> binaries... or something. but it's a kernel change that creates the issue.
>>> what
>>> - i don't know. ask your friendly neighbourhood kernel developer. the
>>> setuid
>>> root binaries are specifically erroring out unable to assume root privs
>>> where
>>> they could before.
>>>
>>>
>>> --
>>> ------------- Codito, ergo sum - "I code, therefore I am" --------------
>>> The Rasterman (Carsten Haitzler)    [email protected]
>>>
>>>
>>> ------------------------------------------------------------
>>> ------------------
>>> Check out the vibrant tech community on one of the world's most
>>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>>> _______________________________________________
>>> enlightenment-devel mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
>>>
>> ------------------------------------------------------------------------------
>> Check out the vibrant tech community on one of the world's most
>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>> _______________________________________________
>> enlightenment-devel mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
>>
> 
> 

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
enlightenment-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to