Re: Thanks Thomas!

2017-04-05 Thread songbird
Thomas Schmitt wrote:
>songbird wrote:
...
>>   ok, looks like the two versions are the same in
>> the first sector (netinst for i386 and amd64) so
>> the fix should work...
>
> The fix should apply to all Debian i386 and amd64 ISOs which were made
> with isohybrid functionality. The bug was introduced in may 2009. Steve
> McIntyre announced the new capability of Debian testing ISOs in january
> 2011. Debian 6 came out in february.
> I have a debian-6.0.5-amd64-businesscard.iso which already is isohybrid.
>
> One possible drawback is that it does not preserve the Apple Partition
> Map of the EFI-capable Debian ISOs.
> I am not aware of any machine which would boot Debian ISOs with APM
> and would not if APM is missing. To my understanding it is necessary
> to anounce HFS+ boot images to certain old Macs. But Debian ISOs do
> not provide HFS+ images.

...

  i just realised this morning that for the new
machine i should not need to apply this fix to
get it to boot from a USB stick.  heh...


  songbird



Re: Thanks Thomas!

2017-04-05 Thread Thomas Schmitt
Hi,

i wrote:
> > (It would be embarrassing if a different Thomas was meant.)

songbird wrote:
>   context is good...  i dislike posting last names to usenet/mailing lists.

Well, germany is full of baby-boomer Thomases. :))


> >   http://www.ludd.ltu.se/~ams/tmp/isohdpfx.bin.170324

> i hope the author released it as open source/bytes!  :)

The bug fix is not worth an own copyright.
So it is provided under the BSD-ish license of isohdpfx.S by H. Peter
Anvin and Intel Corporation:
  http://git.zytor.com/syslinux/syslinux.git/tree/mbr/isohdpfx.S


> i'm not sure what else i can do.

Nothing more for now. We have to wait for a decision of debian-cd
how to handle this not-so-exotic-any-more bug.


>   ok, looks like the two versions are the same in
> the first sector (netinst for i386 and amd64) so
> the fix should work...

The fix should apply to all Debian i386 and amd64 ISOs which were made
with isohybrid functionality. The bug was introduced in may 2009. Steve
McIntyre announced the new capability of Debian testing ISOs in january
2011. Debian 6 came out in february.
I have a debian-6.0.5-amd64-businesscard.iso which already is isohybrid.

One possible drawback is that it does not preserve the Apple Partition
Map of the EFI-capable Debian ISOs.
I am not aware of any machine which would boot Debian ISOs with APM
and would not if APM is missing. To my understanding it is necessary
to anounce HFS+ boot images to certain old Macs. But Debian ISOs do
not provide HFS+ images.


> diff new amd
> < 000   3 355 220 220 220 220 220 220 220 220 220 220 220 220 220 220
> < 020 220 220 220 220 220 220 220 220 220 220 220 220 220 220 220 220
> ---
> > 000   E   R  \b  \0  \0  \0 220 220  \0  \0  \0  \0  \0  \0  \0  \0
> > 020  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0

Upper without APM signature, lower with APM signature.
Both groups of 32 bytes are supposed to do nothing harmful when executed
as x86 machine code (which PC-BIOS does).

The other differences are due to the new instructions of the fixed version.
They change positions of older instructions and cause changes in relative
memory addresses.


Have a nice day :)

Thomas



Re: Thanks Thomas!

2017-04-04 Thread songbird
Thomas Schmitt wrote:

> (It would be embarrassing if a different Thomas was meant.)

  context is good...  i dislike posting
last names to usenet/mailing lists.


> songbird wrote:
>> your recent efforts helped me get an
>> install going via USB stick on this ancient 
>> machine (the bug with the cd image not being
>> able to find isolinux.bin also was affecting
>> the netinst i386 cd-image i had downloaded).
>
> I assume you mean
>   https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=857597
>
> In this case i have to share the thanks with David Christensen, who
> reported the bug and bravely tested, and Martin Str|mberg, who produced
> the fixed code. 

  ack'd in moreinfo followup.  :)


>>  the dd to copy the bytes as posted to the
>> bug (see cd-image bug list for those who
>> care) took care of it and i was off and 
>> running...
>
> The vulnerable BIOSes seem to be more widespread than thought.

  old machines are still being used and people 
who might want to upgrade may try to use them
to write a new USB stick.

  i wasn't even sure at first that any new USB
stick would work (it did, i picked up a sandisk
cruzer the other day for a few $) on such an
ancient machine.

uses the Intel D865GVHZ chipset.


> The fixed MBR code piece
>   http://www.ludd.ltu.se/~ams/tmp/isohdpfx.bin.170324
> is said to be available only for a limited time.

  well now there are other people who have it.  i
hope the author released it as open source/bytes!  :)


> Be so kind and write a mail to  857...@bugs.debian.org  which tells
> that David Christensen is not the only one who has a needy computer.
> Consider to support my proposal in
>   https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=857597#55
>   "I propose that Debian gets it too and offers it for download together
>with a description what problem it might fix and how to apply it.
>If it gets a Debian URL and if i get pointed to an empty wiki page,
>i would volunteer to write the description."

  i added a moreinfo to it and a thank you.  i'm
not sure what else i can do.

  i have been searching for a local computer store
to visit because i dislike doing online hardware
purchases where i'm not sure if the hardware is
even found by the debian installer.  having a bootable
USB stick i can take with me is the goal.

  now i have to get the 64 bit version sorted out.

  ok, looks like the two versions are the same in
the first sector (netinst for i386 and amd64) so
the fix should work...

  here is the diff between fixed and unfixed



diff new amd
1,2c1,2
< 000   3 355 220 220 220 220 220 220 220 220 220 220 220 220 220 220
< 020 220 220 220 220 220 220 220 220 220 220 220 220 220 220 220 220
---
> 000   E   R  \b  \0  \0  \0 220 220  \0  \0  \0  \0  \0  \0  \0  \0
> 020  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0
7,21c7,21
< 140 203 341 001   t  \v   f 307 006 363 006 264   B 353 025 353 002
< 160   1 311   Z   Q 264  \b 315 023   [ 017 266 306   @   P 203 341
< 200   ?   Q 367 341   S   R   P 273  \0   | 271 004  \0   f 241 260
< 220  \a 350   D  \0 017 202 200  \0   f   @ 200 307 002 342 362   f
< 240 201   >   @   | 373 300   x   p   u  \t 372 274 354   { 352   D
< 260   |  \0  \0 350 203  \0   i   s   o   l   i   n   u   x   .   b
< 300   i   n   m   i   s   s   i   n   g   o   r   c   o
< 320   r   r   u   p   t   .  \r  \n   f   `   f   1 322   f 003 006
< 340 370   {   f 023 026 374   {   f   R   f   P 006   S   j 001   j
< 360 020 211 346   f 367   6 350   { 300 344 006 210 341 210 305 222
< 400 366   6 356   { 210 306  \b 341   A 270 001 002 212 026 362   {
< 420 315 023 215   d 020   f   a 303 350 036  \0   O   p   e   r   a
< 440   t   i   n   g   s   y   s   t   e   m   l   o   a   d
< 460   e   r   r   o   r   .  \r  \n   ^ 254 264 016 212   >   b
< 500 004 263  \a 315 020   <  \n   u 361 315 030 364 353 375  \0  \0
---
> 140 203 341 001   t  \v   f 307 006 361 006 264   B 353 025 353  \0
> 160   Z   Q 264  \b 315 023 203 341   ?   [   Q 017 266 306   @   P
> 200 367 341   S   R   P 273  \0   | 271 004  \0   f 241 260  \a 350
> 220   D  \0 017 202 200  \0   f   @ 200 307 002 342 362   f 201   >
> 240   @   | 373 300   x   p   u  \t 372 274 354   { 352   D   |  \0
> 260  \0 350 203  \0   i   s   o   l   i   n   u   x   .   b   i   n
> 300   m   i   s   s   i   n   g   o   r   c   o   r   r
> 320   u   p   t   .  \r  \n   f   `   f   1 322   f 003 006 370   {
> 340   f 023 026 374   {   f   R   f   P 006   S   j 001   j 020 211
> 360 346   f 367   6 350   { 300 344 006 210 341 210 305 222 366   6
> 400 356   { 210 306  \b 341   A 270 001 002 212 026 362   { 315 023
> 420 215   d 020   f   a 303 350 036  \0   O   p   e   r   a   t   i
> 440   n   g   s   y   s   t   e   m   l   o   a   d   e
> 460   r   r   o   r   .  \r  \n   ^ 254 264 016 

Re: Thanks Thomas!

2017-04-04 Thread Thomas Schmitt
Hi,

(It would be embarrassing if a different Thomas was meant.)

songbird wrote:
> your recent efforts helped me get an
> install going via USB stick on this ancient 
> machine (the bug with the cd image not being
> able to find isolinux.bin also was affecting
> the netinst i386 cd-image i had downloaded).

I assume you mean
  https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=857597

In this case i have to share the thanks with David Christensen, who
reported the bug and bravely tested, and Martin Str|mberg, who produced
the fixed code. 


>  the dd to copy the bytes as posted to the
> bug (see cd-image bug list for those who
> care) took care of it and i was off and 
> running...

The vulnerable BIOSes seem to be more widespread than thought.

The fixed MBR code piece
  http://www.ludd.ltu.se/~ams/tmp/isohdpfx.bin.170324
is said to be available only for a limited time.

Be so kind and write a mail to  857...@bugs.debian.org  which tells
that David Christensen is not the only one who has a needy computer.
Consider to support my proposal in
  https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=857597#55
  "I propose that Debian gets it too and offers it for download together
   with a description what problem it might fix and how to apply it.
   If it gets a Debian URL and if i get pointed to an empty wiki page,
   i would volunteer to write the description."


Have a nice day :)

Thomas



Thanks Thomas!

2017-04-04 Thread songbird
  your recent efforts helped me get an
install going via USB stick on this ancient 
machine (the bug with the cd image not being
able to find isolinux.bin also was affecting
the netinst i386 cd-image i had downloaded).

  the dd to copy the bytes as posted to the
bug (see cd-image bug list for those who
care) took care of it and i was off and 
running...

  it actually works.  :)


  songbird