Re: New ARM elfloader patch

[Michel Dänzer]

On Fri, 2001-10-12 at 00:48, Adam C Powell IV wrote:

> Okay, out of six attempts, it has signal 4'd twice, one of those with
> gpm off.

We used to get seemingly non-deterministic signal 4s on PPC when the
cache flushing code in the loader was broken.


-- 
Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
XFree86 and DRI project member   /  CS student, Free Software enthusiast

Re: New ARM elfloader patch

[Adam C Powell IV]

[EMAIL PROTECTED]  wrote:


In message <[EMAIL PROTECTED]>, Adam C Powell IV writes:

Ah, yes.  You're right.  If I tell it 8-bit, then it gets past that, and 
gets signal 4 later, after setting up the mouse.  Oh well.  Config and 
log at the same place as last time.  But yes, the ELF loader might just 
be working... :-)



Signal 4 is "illegal instruction", which unfortunately points the
finger of blame back at the loader.  The log on its own isn't
especially illuminating; can you run XFree86 under the debugger and
find out what instruction it's tripping up on?

Hmm.  Must be an intermittent thing.  I ran it in the debugger, and it 
came up just fine!  Well, almost, the mouse was not happy, but that's 
because I was running gpm too.  And the ctrl-alt commands didn't seem to 
work.  With gpm turned off, startx ran perfectly!  Well, slow as 
anything because of all the debugging info it's generating (though 
according to top, it uses 98% of CPU, so it's not a disk issue), it 
takes 3-4 minutes to start!


Okay, out of six attempts, it has signal 4'd twice, one of those with 
gpm off.  I'm afraid I've run out of time to work on this, I guess a few 
more runs under the debugger would do it.


But in the meantime...  I can log in via gdm, and run GNOME and 
Enlightenment, with ripples, and transparent gnome-terminal, and 
alpha-blended icons, and, and, it's so beautiful!  Eight-bit color isn't 
the greatest, but, well, I'm pretty happy.  Once the server is up, it 
seems pretty stable (okay, so I've just tested it for 30 minutes :-).


I've moved the packages to http://lyre.mit.edu/~powell/debs/ for others 
to help test, and just share and enjoy!  I'm short of time for the next 
couple of weeks, so it will be a while before I can work on this again.


Zeen,
--

-Adam P.

GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Welcome to the best software in the world today cafe! 

SID X problem

[Derek J Witt]

Just edit /etc/X11/Xsession.d/99xsession-options.sh (not sure on the exact
name). Remove the quote marks. Voila! Problem fixed.

**  Derek J Witt  **
*   Email: mailto:[EMAIL PROTECTED]   *
*   Home Page: http://www.flinthills.com/~djw/ *
*** "...and on the eighth day, God met Bill Gates." - Unknown **

Re: New ARM elfloader patch

[Michel Dänzer]

On Fri, 2001-10-12 at 00:48, Adam C Powell IV wrote:

> Okay, out of six attempts, it has signal 4'd twice, one of those with
> gpm off.

We used to get seemingly non-deterministic signal 4s on PPC when the
cache flushing code in the loader was broken.


-- 
Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
XFree86 and DRI project member   /  CS student, Free Software enthusiast


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: XFree-4.1 SID

[Marc Wilson]

Please read:

http://lists.debian.org/debian-x/2001/debian-x-200110/msg4.html

This will tell you what's wrong with the Xsession.d file in question.

On Thu, Oct 11, 2001 at 12:18:15PM +0200, Clemens Wohld wrote:
> 
> Hallo,
> 
> I have big problems to get my xserver back after the last 
> "apt-get dist-upgrade".
> 
> Somebody told me that the /etc/X11/Xsession have any bugs,
> any escape is not theredont know.
> 
> Can find nothing.
> Also google tells me nothing about this error.
> 
> My Debian "Sid-update" I made on 10.10.2001, there are KDE include.
> 
> Someone can gimmi any hints/help to run xinit normal?
> 
> 
> 
> Thx and Greetings 
> 
> Clemens Wohld
> 
> PS: I speak better german, so when your can write the mail in
> german will be happy ;)
> 
> 
> -- 
>  _  _ _ http://urlz.de/xpage
>   __| | ___| |__ (_) __ _ _ __ _ __   __ __  __ ___  ___
>  / _` |/ _ \ '_ \| |/ _` | '_ \   | '_ \ /  \\ \ /\ / // _ \|  _|
> | (_| |  __/ |_) | | (_| | | | |  |  _)-| () |\ V  V /|  __/| |
>  \__,_|\___|_.__/|_|\__,_|_| |_|  |_|\__/  \_/\_/  \___||_|
>   Debian GNU/Linux: The OS So Big Its Scary-*-senc-*-
> --
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

-- 
Marc Wilson
[EMAIL PROTECTED]
[EMAIL PROTECTED]

Re: New ARM elfloader patch

[Adam C Powell IV]

[EMAIL PROTECTED]  wrote:

>In message <[EMAIL PROTECTED]>, Adam C Powell IV writes:
>
>>Ah, yes.  You're right.  If I tell it 8-bit, then it gets past that, and 
>>gets signal 4 later, after setting up the mouse.  Oh well.  Config and 
>>log at the same place as last time.  But yes, the ELF loader might just 
>>be working... :-)
>>
>Signal 4 is "illegal instruction", which unfortunately points the
>finger of blame back at the loader.  The log on its own isn't
>especially illuminating; can you run XFree86 under the debugger and
>find out what instruction it's tripping up on?
>
Hmm.  Must be an intermittent thing.  I ran it in the debugger, and it 
came up just fine!  Well, almost, the mouse was not happy, but that's 
because I was running gpm too.  And the ctrl-alt commands didn't seem to 
work.  With gpm turned off, startx ran perfectly!  Well, slow as 
anything because of all the debugging info it's generating (though 
according to top, it uses 98% of CPU, so it's not a disk issue), it 
takes 3-4 minutes to start!

Okay, out of six attempts, it has signal 4'd twice, one of those with 
gpm off.  I'm afraid I've run out of time to work on this, I guess a few 
more runs under the debugger would do it.

But in the meantime...  I can log in via gdm, and run GNOME and 
Enlightenment, with ripples, and transparent gnome-terminal, and 
alpha-blended icons, and, and, it's so beautiful!  Eight-bit color isn't 
the greatest, but, well, I'm pretty happy.  Once the server is up, it 
seems pretty stable (okay, so I've just tested it for 30 minutes :-).

I've moved the packages to http://lyre.mit.edu/~powell/debs/ for others 
to help test, and just share and enjoy!  I'm short of time for the next 
couple of weeks, so it will be a while before I can work on this again.

Zeen,
-- 

-Adam P.

GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Welcome to the best software in the world today cafe! 






-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New ARM elfloader patch

[pb]

In message <[EMAIL PROTECTED]>, Adam C Powell IV writes:
>Ah, yes.  You're right.  If I tell it 8-bit, then it gets past that, and 
>gets signal 4 later, after setting up the mouse.  Oh well.  Config and 
>log at the same place as last time.  But yes, the ELF loader might just 
>be working... :-)

Signal 4 is "illegal instruction", which unfortunately points the
finger of blame back at the loader.  The log on its own isn't
especially illuminating; can you run XFree86 under the debugger and
find out what instruction it's tripping up on?

p.

SID X problem

[Derek J Witt]

Just edit /etc/X11/Xsession.d/99xsession-options.sh (not sure on the exact
name). Remove the quote marks. Voila! Problem fixed.

**  Derek J Witt  **
*   Email: mailto:[EMAIL PROTECTED]   *
*   Home Page: http://www.flinthills.com/~djw/ *
*** "...and on the eighth day, God met Bill Gates." - Unknown **


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New ARM elfloader patch

[pb]

In message <[EMAIL PROTECTED]>, Adam C Powell IV writes:
>Next question: why can't I have 16-bit?  fbset -depth 16 gives me the 
>same error.

I don't know.  The kernel experts hang out on
[EMAIL PROTECTED]; try asking them there.

p.

Re: New ARM elfloader patch

[Adam C Powell IV]

Philip Blundell wrote:


FBIOPUT_VSCREENINFO: Invalid argument

That looks like a "good" error, in the sense that I don't think it's at all 
related to the ELF loader.  Instead you seem to have some kind of fbcon 
problem.  I think this can happen if you are trying to start up X in a 
different colour depth to the one you currently have set, or that kind of 
thing.  I think "fbset" will tell you the current situation.


Ah, yes.  You're right.  If I tell it 8-bit, then it gets past that, and 
gets signal 4 later, after setting up the mouse.  Oh well.  Config and 
log at the same place as last time.  But yes, the ELF loader might just 
be working... :-)


Next question: why can't I have 16-bit?  fbset -depth 16 gives me the 
same error.  But I have 2 MB VRAM (standard Netwinder), and at 1024x768, 
16-bit should use just 1.5 MB, right?  I've had 16-bit success with 
previous 2.4.x kernels, but can't get 2.4.5 to do it.


Here's what fbset -i has to say:

mode "1024x768-60"
   # D: 64.998 MHz, H: 48.362 kHz, V: 60.002 Hz
   geometry 1024 768 1024 768 8
   timings 15385 160 24 29 3 136 6
   accel true
   rgba 8/0,8/0,8/0,0/0
endmode

Frame buffer device information:
   Name: CyberPro2000
   Address : 0xc000
   Size: 2097152
   Type: PACKED PIXELS
   Visual  : PSEUDOCOLOR
   XPanStep: 0
   YPanStep: 1
   YWrapStep   : 0
   LineLength  : 1024
   MMIO Address: 0xc080
   MMIO Size   : 786432
   Accelerator : Unknown (33)

Thanks,
--

-Adam P.

GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Welcome to the best software in the world today cafe! 

Re: XFree-4.1 SID

[Marc Wilson]

Please read:

http://lists.debian.org/debian-x/2001/debian-x-200110/msg4.html

This will tell you what's wrong with the Xsession.d file in question.

On Thu, Oct 11, 2001 at 12:18:15PM +0200, Clemens Wohld wrote:
> 
> Hallo,
> 
> I have big problems to get my xserver back after the last 
> "apt-get dist-upgrade".
> 
> Somebody told me that the /etc/X11/Xsession have any bugs,
> any escape is not theredont know.
> 
> Can find nothing.
> Also google tells me nothing about this error.
> 
> My Debian "Sid-update" I made on 10.10.2001, there are KDE include.
> 
> Someone can gimmi any hints/help to run xinit normal?
> 
> 
> 
> Thx and Greetings 
> 
> Clemens Wohld
> 
> PS: I speak better german, so when your can write the mail in
> german will be happy ;)
> 
> 
> -- 
>  _  _ _ http://urlz.de/xpage
>   __| | ___| |__ (_) __ _ _ __ _ __   __ __  __ ___  ___
>  / _` |/ _ \ '_ \| |/ _` | '_ \   | '_ \ /  \\ \ /\ / // _ \|  _|
> | (_| |  __/ |_) | | (_| | | | |  |  _)-| () |\ V  V /|  __/| |
>  \__,_|\___|_.__/|_|\__,_|_| |_|  |_|\__/  \_/\_/  \___||_|
>   Debian GNU/Linux: The OS So Big Its Scary-*-senc-*-
> --
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

-- 
Marc Wilson
[EMAIL PROTECTED]
[EMAIL PROTECTED]


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New ARM elfloader patch

[Philip Blundell]

>FBIOPUT_VSCREENINFO: Invalid argument

That looks like a "good" error, in the sense that I don't think it's at all 
related to the ELF loader.  Instead you seem to have some kind of fbcon 
problem.  I think this can happen if you are trying to start up X in a 
different colour depth to the one you currently have set, or that kind of 
thing.  I think "fbset" will tell you the current situation.

p.

Re: New ARM elfloader patch

[pb]

In message <[EMAIL PROTECTED]>, Adam C Powell IV writes:
>Ah, yes.  You're right.  If I tell it 8-bit, then it gets past that, and 
>gets signal 4 later, after setting up the mouse.  Oh well.  Config and 
>log at the same place as last time.  But yes, the ELF loader might just 
>be working... :-)

Signal 4 is "illegal instruction", which unfortunately points the
finger of blame back at the loader.  The log on its own isn't
especially illuminating; can you run XFree86 under the debugger and
find out what instruction it's tripping up on?

p.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New ARM elfloader patch

[Adam C Powell IV]

Guess what...  it built!  (Had to split it among the two local 
partitions and put relatively inactive stuff on NFS.)  So the patches 
were successful that far; they're at:


Adam C Powell IV wrote:


Christian T. Steigies wrote:


On Tue, Oct 09, 2001 at 04:04:51PM -0400, Adam C Powell IV wrote:


http://lyre.mit.edu/~powell/debs/311_arm_compiler_h.diff
http://lyre.mit.edu/~powell/debs/312_arm_elfloader.diff
http://lyre.mit.edu/~powell/debs/350_mips_compiler_h.diff
http://lyre.mit.edu/~powell/debs/400_hppa_support.diff


Unpack the X source, then put all of these in debian/patches, and 
remove 310_arm_compiler_h.diff from that directory.



Build logs are at:
http://lyre.mit.edu/~powell/debs/xfree86_4.1.0-7.8.log
http://lyre.mit.edu/~powell/debs/xfree86_4.1.0-7.8b.log
(the first completed build and ran out of space in install; the second 
has install and binary-arch.)


There was one warning worth noting, during the "dh_movefiles" substitute:

cp: cannot stat 
`/redhat/packages/xfree86-4.1.0/debian/tmp/usr/X11R6/lib/modules/drivers/tdfx.o': 
No such file or directory


Now the bad news: it doesn't work. :-(  The debugging loader churns out 
over 30 MB of stuff, then dies with:


ElfGetSymbolNameIndex(6b,a) afbCreateDefColormap 1 0 0
FBIOPUT_VSCREENINFO: Invalid argument

Fatal server error:
AddScreen/ScreenInit failed for driver 0

The (huge!) log and XF86Config (generated by debconf) are at:
http://lyre.mit.edu/~powell/debs/XFree86.0.log
http://lyre.mit.edu/~powell/debs/XF86Config-4

Does this much (this little) progress justify including the above 
patches in the X package?


Zeen,
--

-Adam P.

GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Welcome to the best software in the world today cafe! 

Re: New ARM elfloader patch

[pb]

In message <[EMAIL PROTECTED]>, Adam C Powell IV writes:
>Next question: why can't I have 16-bit?  fbset -depth 16 gives me the 
>same error.

I don't know.  The kernel experts hang out on
[EMAIL PROTECTED]; try asking them there.

p.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New ARM elfloader patch

[Adam C Powell IV]

Philip Blundell wrote:

>>FBIOPUT_VSCREENINFO: Invalid argument
>>
>That looks like a "good" error, in the sense that I don't think it's at all 
>related to the ELF loader.  Instead you seem to have some kind of fbcon 
>problem.  I think this can happen if you are trying to start up X in a 
>different colour depth to the one you currently have set, or that kind of 
>thing.  I think "fbset" will tell you the current situation.
>
Ah, yes.  You're right.  If I tell it 8-bit, then it gets past that, and 
gets signal 4 later, after setting up the mouse.  Oh well.  Config and 
log at the same place as last time.  But yes, the ELF loader might just 
be working... :-)

Next question: why can't I have 16-bit?  fbset -depth 16 gives me the 
same error.  But I have 2 MB VRAM (standard Netwinder), and at 1024x768, 
16-bit should use just 1.5 MB, right?  I've had 16-bit success with 
previous 2.4.x kernels, but can't get 2.4.5 to do it.

Here's what fbset -i has to say:

mode "1024x768-60"
# D: 64.998 MHz, H: 48.362 kHz, V: 60.002 Hz
geometry 1024 768 1024 768 8
timings 15385 160 24 29 3 136 6
accel true
rgba 8/0,8/0,8/0,0/0
endmode

Frame buffer device information:
Name: CyberPro2000
Address : 0xc000
Size: 2097152
Type: PACKED PIXELS
Visual  : PSEUDOCOLOR
XPanStep: 0
YPanStep: 1
YWrapStep   : 0
LineLength  : 1024
MMIO Address: 0xc080
MMIO Size   : 786432
Accelerator : Unknown (33)

Thanks,
-- 

-Adam P.

GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Welcome to the best software in the world today cafe! 





-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New ARM elfloader patch

[Philip Blundell]

>FBIOPUT_VSCREENINFO: Invalid argument

That looks like a "good" error, in the sense that I don't think it's at all 
related to the ELF loader.  Instead you seem to have some kind of fbcon 
problem.  I think this can happen if you are trying to start up X in a 
different colour depth to the one you currently have set, or that kind of 
thing.  I think "fbset" will tell you the current situation.

p.



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New ARM elfloader patch

[Adam C Powell IV]

Guess what...  it built!  (Had to split it among the two local 
partitions and put relatively inactive stuff on NFS.)  So the patches 
were successful that far; they're at:

Adam C Powell IV wrote:

> Christian T. Steigies wrote:
>
>> On Tue, Oct 09, 2001 at 04:04:51PM -0400, Adam C Powell IV wrote:
>>
 http://lyre.mit.edu/~powell/debs/311_arm_compiler_h.diff
 http://lyre.mit.edu/~powell/debs/312_arm_elfloader.diff
 http://lyre.mit.edu/~powell/debs/350_mips_compiler_h.diff
 http://lyre.mit.edu/~powell/debs/400_hppa_support.diff
>>>
>>> Unpack the X source, then put all of these in debian/patches, and 
>>> remove 310_arm_compiler_h.diff from that directory.
>>
Build logs are at:
http://lyre.mit.edu/~powell/debs/xfree86_4.1.0-7.8.log
http://lyre.mit.edu/~powell/debs/xfree86_4.1.0-7.8b.log
(the first completed build and ran out of space in install; the second 
has install and binary-arch.)

There was one warning worth noting, during the "dh_movefiles" substitute:

cp: cannot stat 
`/redhat/packages/xfree86-4.1.0/debian/tmp/usr/X11R6/lib/modules/drivers/tdfx.o': 
No such file or directory

Now the bad news: it doesn't work. :-(  The debugging loader churns out 
over 30 MB of stuff, then dies with:

ElfGetSymbolNameIndex(6b,a) afbCreateDefColormap 1 0 0
FBIOPUT_VSCREENINFO: Invalid argument

Fatal server error:
AddScreen/ScreenInit failed for driver 0

The (huge!) log and XF86Config (generated by debconf) are at:
http://lyre.mit.edu/~powell/debs/XFree86.0.log
http://lyre.mit.edu/~powell/debs/XF86Config-4

Does this much (this little) progress justify including the above 
patches in the X package?

Zeen,
-- 

-Adam P.

GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Welcome to the best software in the world today cafe! 





-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

XFree-4.1 SID

[Clemens Wohld]

Hallo,

I have big problems to get my xserver back after the last 
"apt-get dist-upgrade".

Somebody told me that the /etc/X11/Xsession have any bugs,
any escape is not theredont know.

Can find nothing.
Also google tells me nothing about this error.

My Debian "Sid-update" I made on 10.10.2001, there are KDE include.

Someone can gimmi any hints/help to run xinit normal?



Thx and Greetings 

Clemens Wohld

PS: I speak better german, so when your can write the mail in
german will be happy ;)


-- 
 _  _ _ http://urlz.de/xpage
  __| | ___| |__ (_) __ _ _ __ _ __   __ __  __ ___  ___
 / _` |/ _ \ '_ \| |/ _` | '_ \   | '_ \ /  \\ \ /\ / // _ \|  _|
| (_| |  __/ |_) | | (_| | | | |  |  _)-| () |\ V  V /|  __/| |
 \__,_|\___|_.__/|_|\__,_|_| |_|  |_|\__/  \_/\_/  \___||_|
  Debian GNU/Linux: The OS So Big Its Scary-*-senc-*-
--

XFree-4.1 SID

[Clemens Wohld]

Hallo,

I have big problems to get my xserver back after the last 
"apt-get dist-upgrade".

Somebody told me that the /etc/X11/Xsession have any bugs,
any escape is not theredont know.

Can find nothing.
Also google tells me nothing about this error.

My Debian "Sid-update" I made on 10.10.2001, there are KDE include.

Someone can gimmi any hints/help to run xinit normal?



Thx and Greetings 

Clemens Wohld

PS: I speak better german, so when your can write the mail in
german will be happy ;)


-- 
 _  _ _ http://urlz.de/xpage
  __| | ___| |__ (_) __ _ _ __ _ __   __ __  __ ___  ___
 / _` |/ _ \ '_ \| |/ _` | '_ \   | '_ \ /  \\ \ /\ / // _ \|  _|
| (_| |  __/ |_) | | (_| | | | |  |  _)-| () |\ V  V /|  __/| |
 \__,_|\___|_.__/|_|\__,_|_| |_|  |_|\__/  \_/\_/  \___||_|
  Debian GNU/Linux: The OS So Big Its Scary-*-senc-*-
--


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]