Re: xlock permission problem

[Vadim Zhukov]

2016-07-17 22:20 GMT+03:00 Jyri Hovila [iki.fi] :
> Hi, and thanks for your time!
>
> On Sun, 17 Jul 2016 21:52:53 +0300
> Vadim Zhukov  wrote:
>
>> Check the following:
>>
>> - you have root:auth mode rwxr-sr-x set for /usr/X11R6/bin/xlock;
>
> # ls -la /usr/X11R6/bin/xlock
> -rwxr-sr-x  1 root  auth  2970216 Jun  2 15:00 /usr/X11R6/bin/xlock

Jun, 2... is your base in sync? I mean, userland corresponds to kernel
and to xenocara?

>> - you do NOT have nosuid option set for /usr/X11R6.
>
> # mount
> /dev/sd1g on /usr/X11R6 type ffs (local, nodev)
>
> Still a mystery. =o

Then check what owner/group/rights do /etc/*passwd and /etc/login.conf
files, and /usr/libexec/auth directory have. They basically should be
(according to /etc/mtree/special):

-rw-r--r--  1 root  wheel   2770 Jan 14  2016 /etc/login.conf
-rw---  1 root  wheel  10350 Jul 12 17:25 /etc/master.passwd
-rw-r--r--  1 root  wheel   8522 Jul 12 16:42 /etc/passwd
drwxr-x---  2 root  auth 512 Jul  3 17:49 /usr/libexec/auth/

Most likely you have /etc/login.conf non-world-readable... If this
won't help, you'll probably need to use ktrace to find out the
problematic syscall.

--
  WBR,
  Vadim Zhukov

Re: xlock permission problem

[Ali Farzanrad]

Jyri Hovila [iki.fi] wrotes:
>
>Hello, world!
>
>I'm having an issue with xlock being unable to unlock a locked session.
>
>I'm running a CURRENT version of OpenBSD on amd64 architecture.
>
>I can lock my X session with xlock just fine, but when I enter my
>password the unlock, xlock says the password is invalid. However, the
>password I've entered (several times) is 100% correct.
>
It might be ugly, but are you sure that you are in rigth keyboard layout?

I have similar issue when using another layout (in my case ir) and lock
my system without changing back layout.

Re: xlock permission problem

[Jyri Hovila [iki.fi]]

Hi, and thanks for your time!

On Sun, 17 Jul 2016 21:52:53 +0300
Vadim Zhukov  wrote:

> Check the following:
> 
> - you have root:auth mode rwxr-sr-x set for /usr/X11R6/bin/xlock;

# ls -la /usr/X11R6/bin/xlock
-rwxr-sr-x  1 root  auth  2970216 Jun  2 15:00 /usr/X11R6/bin/xlock

> - you do NOT have nosuid option set for /usr/X11R6.

# mount
/dev/sd1g on /usr/X11R6 type ffs (local, nodev)

Still a mystery. =o

- Jyri

Trying to get Wacom CTL-490 to work

[Frank Groeneveld]

I'm trying to get my new Wacom tablet working with X11 in -current.
It's a Wacom Intuos Pen (CTL-490) and it shows up in dmesg like this:

uhidev0 at uhub3 port 1 configuration 1 interface 0 "Wacom Co.,Ltd.
Intuos PS" rev 2.00/1.00 addr 5
uhidev0: iclass 3/0, 192 report ids
uhid0 at uhidev0 reportid 2: input=0, output=0, feature=1
uhid1 at uhidev0 reportid 3: input=0, output=0, feature=1
uhid2 at uhidev0 reportid 4: input=0, output=0, feature=1
uhid3 at uhidev0 reportid 5: input=0, output=0, feature=1
uhid4 at uhidev0 reportid 7: input=0, output=0, feature=9
uhid5 at uhidev0 reportid 8: input=0, output=0, feature=9
uhid6 at uhidev0 reportid 16: input=9, output=0, feature=2
uhid7 at uhidev0 reportid 17: input=0, output=0, feature=16
uhid8 at uhidev0 reportid 19: input=0, output=0, feature=1
uhid9 at uhidev0 reportid 20: input=0, output=0, feature=31
uhid10 at uhidev0 reportid 32: input=0, output=0, feature=5
uhid11 at uhidev0 reportid 33: input=0, output=0, feature=1
uhid12 at uhidev0 reportid 34: input=0, output=0, feature=1
uhid13 at uhidev0 reportid 35: input=0, output=0, feature=14
uhid14 at uhidev0 reportid 36: input=0, output=0, feature=31
uhid15 at uhidev0 reportid 37: input=0, output=0, feature=4
uhid16 at uhidev0 reportid 48: input=0, output=0, feature=2
uhid17 at uhidev0 reportid 49: input=0, output=0, feature=33
uhid18 at uhidev0 reportid 50: input=0, output=0, feature=33
uhid19 at uhidev0 reportid 51: input=0, output=0, feature=1
uhid20 at uhidev0 reportid 64: input=0, output=0, feature=10
uhid21 at uhidev0 reportid 192: input=9, output=0, feature=0
uhidev1 at uhub3 port 1 configuration 1 interface 1 "Wacom Co.,Ltd.
Intuos PS" rev 2.00/1.00 addr 5
uhidev1: iclass 3/0, 3 report ids
uhid22 at uhidev1 reportid 2: input=63, output=0, feature=0
uhid23 at uhidev1 reportid 3: input=63, output=0, feature=0
uhidev2 at uhub3 port 1 configuration 1 interface 2 "Wacom Co.,Ltd.
Intuos PS" rev 2.00/1.00 addr 5
uhidev2: iclass 3/1, 1 report id
ums0 at uhidev2 reportid 1: 5 buttons
wsmouse1 at ums0 mux 0

When I run `cat /dev/uhid6` (after making that device) I get output that
seems to suggest it is related to the pen movement. When not moving the
pen there is less output etc.
However, `cat /dev/wsmouse1` doesn't give any output.

Searching the archives I found this suggested xorg.conf InputDevice
section, which I tried:

Section "InputDevice"
Identifier "pen"
Driver "usbtablet"
Option "Type" "stylus"
Option "Device" "/dev/uhid6"
Option "Mode" "Absolute"
EndSection

Running this gives me the following error in my Xorg log file though.
Maybe because it was old and only meant for serial-over-usb tablets?

[46.602] (II) Using input driver 'usbtablet' for 'pen'
[46.602] (**) Option "CorePointer"
[46.602] (**) pen: always reports core events
[46.602] (**) Stylus: always reports core events
[46.602] (**) Stylus device is /dev/uhid6
[46.602] (**) Stylus is in absolute mode
[46.602] (==) Stylus: threshold = 5
[46.602] (==) Stylus: suppress = 2
[46.602] (II) XINPUT: Adding extended input device "Stylus" (type:
TABLET, id 6)
[46.602] (**) Option "Device" "/dev/uhid6"
[46.602] (EE) xf86OpenSerial: Cannot open device /dev/uhid6
Operation not permitted.
[46.602] (EE) Error opening /dev/uhid6: Operation not permitted
[46.602] (EE) Couldn't init device "Stylus"
[46.602] (II) UnloadModule: "usbtablet"

Finally I've modified the kernel and called uhidev_set_report with the
same parameters as for the Graphire 3 & 4, because the Linux kernel
driver seems to suggest that without that call the device will run in a
more generic mouse mode. However, it did not change anything.

Does anybody know how to get X working with the uhid device? Or is that
not the suggested method? Based on my findings that uhid6 is for pen
movement, might ums attach to the wrong device?

Thanks for any help.

Frank

Re: xlock permission problem

[Vadim Zhukov]

2016-07-17 21:31 GMT+03:00 Jyri Hovila [iki.fi] :
> Hello, world!
>
> I'm having an issue with xlock being unable to unlock a locked session.
>
> I'm running a CURRENT version of OpenBSD on amd64 architecture.
>
> I can lock my X session with xlock just fine, but when I enter my
> password the unlock, xlock says the password is invalid. However, the
> password I've entered (several times) is 100% correct.
>
> Authlog provides the following hints:
>
> ==
> xlock[82700]: Start: myusername, myusername, :0
> xlock[82700]: staff: getting class information: Permission denied
> xlock[82700]: daemon: getting classinformation: Permission denied
> xlock[82700]: xlock: failed unlock attempt on user myusername
> xlock[82700]: xlock: xio_error
> xlock[82700]: Stop: myusername, myusername, :0, 6m 5s
> ==
>
> I've googled thoroughly, checked the permissions of /etc/group etc, but
> have not found either a solution nor any hints on how to further
> diagnose the problem.
>
> Any ideas, anyone?

Check the following:

- you have root:auth mode rwxr-sr-x set for /usr/X11R6/bin/xlock;
- you do NOT have nosuid option set for /usr/X11R6.

--
  WBR,
  Vadim Zhukov

xlock permission problem

[Jyri Hovila [iki.fi]]

Hello, world!

I'm having an issue with xlock being unable to unlock a locked session.

I'm running a CURRENT version of OpenBSD on amd64 architecture.

I can lock my X session with xlock just fine, but when I enter my
password the unlock, xlock says the password is invalid. However, the
password I've entered (several times) is 100% correct.

Authlog provides the following hints:

==
xlock[82700]: Start: myusername, myusername, :0
xlock[82700]: staff: getting class information: Permission denied
xlock[82700]: daemon: getting classinformation: Permission denied
xlock[82700]: xlock: failed unlock attempt on user myusername
xlock[82700]: xlock: xio_error
xlock[82700]: Stop: myusername, myusername, :0, 6m 5s
==

I've googled thoroughly, checked the permissions of /etc/group etc, but
have not found either a solution nor any hints on how to further
diagnose the problem.

Any ideas, anyone?

- Jyri

Re: long name functions in bc

[Otto Moerbeek]

On Sun, Jul 17, 2016 at 03:23:14PM -0300, Henrique N. Lengler wrote:

> On Sun, Jul 17, 2016 at 07:20:48PM +0200, Otto Moerbeek wrote:
> > On Sun, Jul 17, 2016 at 12:51:03PM -0300, Henrique N. Lengler wrote:
> > 
> > > Hi,
> > > 
> > > I'm new on bc, I'm writing a library with some math functions I intend to 
> > > use
> > > on calculations.
> > > I would like to use long names for variables and functions on my library
> > > but also use functions from the main math library, which I do by calling 
> > > bc
> > > with '-l' flag. But once I start bc with '-l' I can't use any long name, 
> > > why?
> > > 
> > > Is there a way to change this behaviour?
> > > Why it works like this?
> > > 
> > > Regards,
> > > 
> > > Henrique N. Lengler
> > 
> > This is a bug introduced by not execing dc anymore, but calling the
> > dc code as a function.
> > 
> > In that case getopt is called a second time for a new set of
> > parameters and it has to be reinitialized,
> > 
> > The code below should fix it. Recompile both dc and bc after patching.
> > 
> > -Otto
> > 
> > Index: dc.c
> > ===
> > RCS file: /cvs/src/usr.bin/dc/dc.c,v
> > retrieving revision 1.17
> > diff -u -p -r1.17 dc.c
> > --- dc.c3 Nov 2015 04:58:58 -   1.17
> > +++ dc.c17 Jul 2016 17:20:19 -
> > @@ -50,6 +50,8 @@ dc_main(int argc, char *argv[])
> > if ((buf = strdup("")) == NULL)
> > err(1, NULL);
> > /* accept and ignore a single dash to be 4.4BSD dc(1) compatible */
> > +   optind = 1;
> > +   optreset = 1;
> > while ((ch = getopt(argc, argv, "e:x-")) != -1) {
> > switch (ch) {
> > case 'e':
> 
> Will try it, shouldn't this patch go to the repository?
> 
> Regards,
> 
> Henrique N. Lengler

Yes, in the meantime I committed it in the -current tree.

-Otto

Re: long name functions in bc

[Henrique N. Lengler]

On Sun, Jul 17, 2016 at 07:20:48PM +0200, Otto Moerbeek wrote:
> On Sun, Jul 17, 2016 at 12:51:03PM -0300, Henrique N. Lengler wrote:
> 
> > Hi,
> > 
> > I'm new on bc, I'm writing a library with some math functions I intend to 
> > use
> > on calculations.
> > I would like to use long names for variables and functions on my library
> > but also use functions from the main math library, which I do by calling bc
> > with '-l' flag. But once I start bc with '-l' I can't use any long name, 
> > why?
> > 
> > Is there a way to change this behaviour?
> > Why it works like this?
> > 
> > Regards,
> > 
> > Henrique N. Lengler
> 
> This is a bug introduced by not execing dc anymore, but calling the
> dc code as a function.
> 
> In that case getopt is called a second time for a new set of
> parameters and it has to be reinitialized,
> 
> The code below should fix it. Recompile both dc and bc after patching.
> 
>   -Otto
> 
> Index: dc.c
> ===
> RCS file: /cvs/src/usr.bin/dc/dc.c,v
> retrieving revision 1.17
> diff -u -p -r1.17 dc.c
> --- dc.c  3 Nov 2015 04:58:58 -   1.17
> +++ dc.c  17 Jul 2016 17:20:19 -
> @@ -50,6 +50,8 @@ dc_main(int argc, char *argv[])
>   if ((buf = strdup("")) == NULL)
>   err(1, NULL);
>   /* accept and ignore a single dash to be 4.4BSD dc(1) compatible */
> + optind = 1;
> + optreset = 1;
>   while ((ch = getopt(argc, argv, "e:x-")) != -1) {
>   switch (ch) {
>   case 'e':

Will try it, shouldn't this patch go to the repository?

Regards,

Henrique N. Lengler

Re: long name functions in bc

[Otto Moerbeek]

On Sun, Jul 17, 2016 at 12:51:03PM -0300, Henrique N. Lengler wrote:

> Hi,
> 
> I'm new on bc, I'm writing a library with some math functions I intend to use
> on calculations.
> I would like to use long names for variables and functions on my library
> but also use functions from the main math library, which I do by calling bc
> with '-l' flag. But once I start bc with '-l' I can't use any long name, why?
> 
> Is there a way to change this behaviour?
> Why it works like this?
> 
> Regards,
> 
> Henrique N. Lengler

This is a bug introduced by not execing dc anymore, but calling the
dc code as a function.

In that case getopt is called a second time for a new set of
parameters and it has to be reinitialized,

The code below should fix it. Recompile both dc and bc after patching.

-Otto

Index: dc.c
===
RCS file: /cvs/src/usr.bin/dc/dc.c,v
retrieving revision 1.17
diff -u -p -r1.17 dc.c
--- dc.c3 Nov 2015 04:58:58 -   1.17
+++ dc.c17 Jul 2016 17:20:19 -
@@ -50,6 +50,8 @@ dc_main(int argc, char *argv[])
if ((buf = strdup("")) == NULL)
err(1, NULL);
/* accept and ignore a single dash to be 4.4BSD dc(1) compatible */
+   optind = 1;
+   optreset = 1;
while ((ch = getopt(argc, argv, "e:x-")) != -1) {
switch (ch) {
case 'e':

long name functions in bc

[Henrique N. Lengler]

Hi,

I'm new on bc, I'm writing a library with some math functions I intend to use
on calculations.
I would like to use long names for variables and functions on my library
but also use functions from the main math library, which I do by calling bc
with '-l' flag. But once I start bc with '-l' I can't use any long name, why?

Is there a way to change this behaviour?
Why it works like this?

Regards,

Henrique N. Lengler

Re: USB and Intel Bay Trail

[Martin Pieuchot]

On 16/07/16(Sat) 13:08, Tony Sarendal wrote:
> Hola,
> 
> I got a pair of mini-pc's to play with for the summer vacation, small
> fanless
> thingies with 4xGE and wifi.
> 
> http://www.qotom.net/goods-129-QOTOM-Q190G4+4+LAN+Mini+PC.html
> 
> When testing with the latest snapshot USB wont play.
> Any ideas ?

This is certainly and interrupt problem, please send a full report to
bugs@ including an ACPI dump.