darkrealm Wiki - Benutzerbeiträge [de]

Dreamcast

2026-05-05T21:55:50Z

Chris: /* Draft */

== CDR Sizes ==

80min/700MB: 1st Session 674MB Maximum

90min/800MB: max 748MB when 45000 Sectors in 1st Session with dummy burn

90min/800MB: max ~735MB when 45000 Sectors in 1st Session

== Access Files on CDI Images ==

# Extract the seperate Track first

<pre>
cdirip <CDIFILE>.cdi
</pre>

# In case of 2nd Data Track, fix the LBA first

<pre>
isofix <TRACK02>.iso <LBA Value reported by cdirip>
</pre>

== What binhack does ==

Changes LBA Value at offset 1461904 ( 8 Bytes before CD001 string) from 6EB0 (45166) to LBA Value+166 specified during binhack.

Adds Region Codes for JAP,EUR and USA and strings for that.
Adds some bootcode, probably SH4 ASM to make the GDROM accept CDRs.

== What cdda does ==

Changes some bytes, probably SH4 ASM to enable playback of CDDA tracks.

== What dahack does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to 9600 (150). Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== What hack3 does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to LBA Value+150 specified during dahack. Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== Another Copy Protection ==

Searching for CDE4 036A (or CDE4 436A) and replacing with 0900 0900 is another Copy Protection defeat.

CDE4 436A -> 0900 0900
1032 0D8B -> 0800 0D8B

== LBA 45000 Trick ==

69120000 bytes in first session will result in LBA 45000.

== Fill 1st Session ISO File to 69120000 bytes with zeros ==

6912000 bytes = 33750 2048-byte-sectors

CORRECTION : Must be 33598 Sectors !!! = 68808703 bytes

1st Session Leadout = 11250 Sectors = 23040000 Bytes

2nd Session will start at Sector 45000

When using extract.exe, add 150 Sectors to LBA reported in gdinfo_FG.exe track05

<pre>
dd if=/dev/zero of=<ISOFILE> bs=1 count=1 seek=69119999
</pre>

== Creating & Burning DATA/DATA ISOs ==

== Merging track03 & track05 into one file ==

Create dummy file first
Calculate 1185760800 - <SIZE OF track03.bin + SIZE OF track05.bin>

<pre>
dd if=/dev/zero bs=1 count=0 seek=<RESULT> of=dummy.bin
cat track03.bin dummy.bin track05.bin > combined.bin
</pre>

This also helps when extract.exe only outputs garbage out of track03.iso along with track05.iso. This way only the combined.iso needs to be extracted.

== Burning DATA/DATA CDIs ==

<pre>
cdirip <CDIFile> <Destination Directory>
cdrecord dev=/dev/sr0 -v -multi -tao -xa tdata01.iso
cdrecord dev=/dev/sr0 -v -tao -xa tdata02.iso
</pre>

== Full rip process from GD-ROM to CD-ROM ==

=== Get GDROM Data from Dreamcast with http-ack ===

https://www.gotwalls.com

https://github.com/sega-dreamcast/httpd-ack

# Boot HTTP-ack from CD
# Swap HTTP-ack CD with Original GD-ROm
# Connect via Browser to http://192.168.1.30
# Download all files starting with track03.bin
# Download disc.gdi

=== Extract Data from Tracks ===

==== Use gditools ====

https://sourceforge.net/projects/dcisotools/

Either use GUI or gditools.py

<pre>
python2.7 ./gditools.py
Usage: ./gditools.py -i input_gdi [options]

-h, --help Display this help
-l, --list List all files in the filesystem and exit
-o [outdir] Output directory. Default: gdi folder
-s [filename] Create a sorttxt file with custom name
(It uses *data-folder* as prefix)
-b [ipname] Dump the ip.bin with custom name
-e [filename] Dump a single file from the filesystem
--extract-all Dump all the files in the *data-folder*
--data-folder [name] *data-folder* subfolder. Default: data
(__volume_label__ --> Use ISO9660 volume label)
--sort-spacer [num] Sorttxt entries are sperated by num
--silent Minimal verbosity mode
[no option] Display gdi infos if not silent

gditools.py by FamilyGuy, http://sourceforge.net/p/dcisotools/
Licensed under GPLv3, see licences folder.

iso9660.py by Barney Gale, http://github.com/barneygale
Licensed under a BSD-based license, see licences folder.
</pre>

<pre>
python2.7 ./gditools.py -i <disc.gdi> --extract-all -o /mnt/development/dreamcast/rips/orgs/<GAME_NAME>/<CD Number> --data-folder __volume_label__ -s sort.txt -b ip.bin
</pre>

==== Old way ====
# If the game contains more than track03.bin, combine the 2 binfiles into 1:
<pre>
cat track03.bin dummy.bin track05.bin > <BINFILE>
</pre>
# Convert .BIN Files to .ISO Files first with bin2iso.exe
<pre>
bin2iso.exe <BINFILE> <ISOFILE>
</pre>
# Extract Data from .ISO files
<pre>
extract.exe <ISOFILE>
</pre>
If extracted 1ST_READ.BIN has filesize of 0, LBA of track05 is wrong.

=== Downsampling ===

==== AFS Files ====

https://sizious.com/download/dreamcast/

AFS files contain multiple ADX files combined into one AFS file.

compact AFS with:

<pre>
AFSCompactor.exe
</pre>

for Alone in the Dark - The New Nightmare .XA files are AFS files, rename first with:

<pre>
rename .XA .afs *.XA
</pre>

then compact and rename back to .XA:

<pre>
rename .afs .XA *.afs
</pre>

AFSCompactor.exe will rename files inside .afs to file0001 etc... AITD4 doesnt seem to care.

extract AFS with:

<pre>
afs_extract.exe *.afs|*.XA
</pre>
or
<pre>
afsextract.exe -sc afsfile|XAfile
</pre>
or
<pre>
afsutils_v2.exe | rename .XA files to .afs beforehand.
</pre>

==== SFD Movies ====

http://www.dc-swat.ru/page/download/

Demux video to .m1v and audio to .wav first

Example (play around with bitrates):

<pre>
for f in <SFD Files>; do ffmpeg -i $f -vcodec mpeg1video -b:v 2600k -minrate 2600k -maxrate 3600k -y $f.m1v -acodec adpcm_ima_wav -y $f.wav ; done
</pre>

Open all .wavs in single Audacity project and mass export again, 16bit and 44100 Hz.

Start Dreamcast Movie Converter and remux/reencode to .SFD. Dont use batch processing, videos get truncated!

==== 4XM Movies ====

http://samples.mplayerhq.hu/game-formats/4xm/

Convert to AVI first with ffmpeg

<pre>
for i in ../rips/unmodified/aitd4/cd2/AITD4_GD_2/MOVIES/*.4XM; do ffmpeg -i "$i" -vf scale=320x160 -y "${i%.*}.avi"; done
</pre>

Downsample with 4xm_comp.exe

Speed : 30 fps
Data Rate : 110000
Resize Width 320 Height 160

=== CDDA Audio ===

Generate file list starting in GD-ROM HD Area after first data Track

<pre>
for f in ../backups/GDI/Mortal\ Kombat\ Gold/track*.raw; do f="${f}"; [[ "$f" == *track04.raw ]] && start=1; [[ $start -eq 1 ]] && files+=("$f"); done
</pre>

Burn CDDA Tracks first
Note: -swab is only neccessary when CDDA is static, depends on the burner.

<pre>
cdrecord -v -multi -tao -audio -swab "${files[@]}"
</pre>

Then get MSINFO Value for 2nd Session.

=== Create ISO Images ===

For the 1st Session, chosen files must not exceed total of 69120000 bytes!

==== 1st Session ====

===== Create first ISOFile =====

<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <CHOSEN FILE> <CHOSEN FILE> <CHOSEN FILE>...
</pre>

Fill first ISOFile with zeros so it takes up exactly 68808703 bytes = LBA 45000

<pre>
dd if=/dev/zero of=<OUTPUT1.ISO> bs=1 count=1 seek=68808703
</pre>

===== Burn 1st session =====

<pre>
cdrecord -v -multi -tao -xa <OUTPUT1.ISO>
</pre>

Check MSINFO value & TOC

<pre>
cdrecord -msinfo

0,45000
</pre>

<pre>
cdrecord -toc

first: 1 last 1
track: 1 lba: 0 ( 0) 00:02:00 adr: 1 control: 4 mode: -1
track:lout lba: 33600 ( 134400) 07:30:00 adr: 1 control: 4 mode: -1
</pre>

To include subdirectories into the root directory of the CD, use
<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <SUBDIRECTORY>/=<PATH_OF_SUBDIRECTORY>
</pre>
Example:
<pre>
mkisofs -no-allow-lowercase -graft-points -l -duplicates-once -V AITD4_GD_2 -o aitd4_d2_1.iso MOVIES/=../rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES/
</pre>

==== 2nd Session ====

===== Hack 1ST_READ.BIN & IP.BIN =====

Get MSINFO Value

<pre>
cdrecord -msinfo
</pre>

<pre>
binhack

Enter name of binary: 1ST_READ.BIN
Enter name of bootsector: ip.bin
Enter msinfo value: <MSINFO Value>
File 1ST_READ.BIN successfully hacked.
Bincon detected, OS flag set to 0.
File ip.bin successfully created.
</pre>

===== Insert custom Logo =====

<pre>
logoinsert darkrealm.mr IP.BIN
</pre>

===== Burn 2nd Session =====

<pre>
mkisofs -sort sort.txt -no-allow-lowercase -l -duplicates-once -V <DISCNAME> -G <IP.BIN> -M <OUTPUT1.ISO> -C <MSINFO VALUE> -o <OUTPUT2.ISO> <GAME_DATA_DIRECTORY>
cdrecord -v -tao -xa <OUTPUT2.ISO>
</pre>

== Testing ==

* Burn whole stuff to a CD-RW
* Generate CDI from CD-RW
** Launch DiscJuggler from emulated Windows Machine and connect via iSCSI
* Test CDI with ReiCast oder ReDream

== Create CDI Images from ISO ==

* Create ISO first

<pre>
mkisofs -l -duplicates-once -graft-points -V <DISCNAME> -G <IP.BIN> -o <GAME.ISO> <GAME DIRECTORY>
</pre>

* Create DATA/DATA CDI

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

* or create AUDIO/DATA CDI, with 2nd Session starting at LBA 11702
** Note: 1ST_READ.BIN must eventually be binhacked to LBA 11702

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

== 60Hz <-> 50Hz Binary Comparisations ==

=== Dino Crisis GER ===

1ST_READ.BIN:

<pre>
Offset : 0x001545C0
50Hz : 01 E4 03 A0 02 E4 01 A0 03
60Hz : 00 E4 03 A0 00 E4 01 A0 00
</pre>

== Tools ==

=== Efficient 1st Session Filler ===

<pre>
#!/bin/bash

# Prüfen, ob ein Verzeichnis übergeben wurde
if [ -z "$1" ]; then
echo "Usage: $0 <Verzeichnis>"
exit 1
fi

DIR="$1"

# Prüfen, ob es ein gültiges Verzeichnis ist
if [ ! -d "$DIR" ]; then
echo "Fehler: '$DIR' ist kein gültiges Verzeichnis."
exit 1
fi

# Maximal erlaubte Größe in Bytes
MAX_SIZE=68808703

# Ausgabedatei für die Liste
OUTPUT="dateiliste.txt"

# Leere die Ausgabedatei
> "$OUTPUT"

# Variable für die aktuelle Gesamtsumme
total_size=0

# Durchlaufe die Dateien im Verzeichnis, sortiert nach Größe absteigend
while IFS= read -r file; do
filesize=$(stat -c%s "$file")

if (( total_size + filesize <= MAX_SIZE )); then
# echo "$filesize $file" >> "$OUTPUT"
echo "$file" >> "$OUTPUT"
total_size=$((total_size + filesize))
fi
done < <(find "$DIR" -maxdepth 1 -type f -printf "%s %p\n" | sort -nr | awk '{print $2}')

echo "Liste erstellt: $OUTPUT"
echo "Gesamtgröße: $total_size Bytes"
</pre>

== Draft ==

* ./find1stsessionfiles.sh ~/media/development/dreamcast/Rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES
* mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V AITD4_GD2 -o aitd4_cd2_s1.iso -path-list ~/media/chris/development/dreamcast/tools/dateiliste.txt
* dd if=/dev/zero of=aitd4_d2_s1.iso bs=1 count=1 seek=68808703
* cdrecord -v -tao -multi -xa aitd4_d2_s1.iso
* binhack64 1ST_READ.BIN & IP.BIN to LBA 45000
* mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V AITD4_GD2 -G ip.bin -sort ../Rips/unmodified/aitd4/cd2/sort.txt -M aitd4_d2_s1.iso -C 0,45000 -o aitd4_d2_s2.iso ../Rips/modified/aitd4/cd2/AITD4_GD_2/ 1ST_READ.BIN
* cdrecord -v -tao -xa aitd4_d2_s2.iso

Dreamcast

2026-05-05T21:34:31Z

Chris: /* Draft */

Dreamcast

2026-05-05T21:22:59Z

Chris: /* Draft */

== CDR Sizes ==

80min/700MB: 1st Session 674MB Maximum

90min/800MB: max 748MB when 45000 Sectors in 1st Session with dummy burn

90min/800MB: max ~735MB when 45000 Sectors in 1st Session

== Access Files on CDI Images ==

# Extract the seperate Track first

<pre>
cdirip <CDIFILE>.cdi
</pre>

# In case of 2nd Data Track, fix the LBA first

<pre>
isofix <TRACK02>.iso <LBA Value reported by cdirip>
</pre>

== What binhack does ==

Changes LBA Value at offset 1461904 ( 8 Bytes before CD001 string) from 6EB0 (45166) to LBA Value+166 specified during binhack.

Adds Region Codes for JAP,EUR and USA and strings for that.
Adds some bootcode, probably SH4 ASM to make the GDROM accept CDRs.

== What cdda does ==

Changes some bytes, probably SH4 ASM to enable playback of CDDA tracks.

== What dahack does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to 9600 (150). Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== What hack3 does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to LBA Value+150 specified during dahack. Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== Another Copy Protection ==

Searching for CDE4 036A (or CDE4 436A) and replacing with 0900 0900 is another Copy Protection defeat.

CDE4 436A -> 0900 0900
1032 0D8B -> 0800 0D8B

== LBA 45000 Trick ==

69120000 bytes in first session will result in LBA 45000.

== Fill 1st Session ISO File to 69120000 bytes with zeros ==

6912000 bytes = 33750 2048-byte-sectors

CORRECTION : Must be 33598 Sectors !!! = 68808703 bytes

1st Session Leadout = 11250 Sectors = 23040000 Bytes

2nd Session will start at Sector 45000

When using extract.exe, add 150 Sectors to LBA reported in gdinfo_FG.exe track05

<pre>
dd if=/dev/zero of=<ISOFILE> bs=1 count=1 seek=69119999
</pre>

== Creating & Burning DATA/DATA ISOs ==

== Merging track03 & track05 into one file ==

Create dummy file first
Calculate 1185760800 - <SIZE OF track03.bin + SIZE OF track05.bin>

<pre>
dd if=/dev/zero bs=1 count=0 seek=<RESULT> of=dummy.bin
cat track03.bin dummy.bin track05.bin > combined.bin
</pre>

This also helps when extract.exe only outputs garbage out of track03.iso along with track05.iso. This way only the combined.iso needs to be extracted.

== Burning DATA/DATA CDIs ==

<pre>
cdirip <CDIFile> <Destination Directory>
cdrecord dev=/dev/sr0 -v -multi -tao -xa tdata01.iso
cdrecord dev=/dev/sr0 -v -tao -xa tdata02.iso
</pre>

== Full rip process from GD-ROM to CD-ROM ==

=== Get GDROM Data from Dreamcast with http-ack ===

https://www.gotwalls.com

https://github.com/sega-dreamcast/httpd-ack

# Boot HTTP-ack from CD
# Swap HTTP-ack CD with Original GD-ROm
# Connect via Browser to http://192.168.1.30
# Download all files starting with track03.bin
# Download disc.gdi

=== Extract Data from Tracks ===

==== Use gditools ====

https://sourceforge.net/projects/dcisotools/

Either use GUI or gditools.py

<pre>
python2.7 ./gditools.py
Usage: ./gditools.py -i input_gdi [options]

-h, --help Display this help
-l, --list List all files in the filesystem and exit
-o [outdir] Output directory. Default: gdi folder
-s [filename] Create a sorttxt file with custom name
(It uses *data-folder* as prefix)
-b [ipname] Dump the ip.bin with custom name
-e [filename] Dump a single file from the filesystem
--extract-all Dump all the files in the *data-folder*
--data-folder [name] *data-folder* subfolder. Default: data
(__volume_label__ --> Use ISO9660 volume label)
--sort-spacer [num] Sorttxt entries are sperated by num
--silent Minimal verbosity mode
[no option] Display gdi infos if not silent

gditools.py by FamilyGuy, http://sourceforge.net/p/dcisotools/
Licensed under GPLv3, see licences folder.

iso9660.py by Barney Gale, http://github.com/barneygale
Licensed under a BSD-based license, see licences folder.
</pre>

<pre>
python2.7 ./gditools.py -i <disc.gdi> --extract-all -o /mnt/development/dreamcast/rips/orgs/<GAME_NAME>/<CD Number> --data-folder __volume_label__ -s sort.txt -b ip.bin
</pre>

==== Old way ====
# If the game contains more than track03.bin, combine the 2 binfiles into 1:
<pre>
cat track03.bin dummy.bin track05.bin > <BINFILE>
</pre>
# Convert .BIN Files to .ISO Files first with bin2iso.exe
<pre>
bin2iso.exe <BINFILE> <ISOFILE>
</pre>
# Extract Data from .ISO files
<pre>
extract.exe <ISOFILE>
</pre>
If extracted 1ST_READ.BIN has filesize of 0, LBA of track05 is wrong.

=== Downsampling ===

==== AFS Files ====

https://sizious.com/download/dreamcast/

AFS files contain multiple ADX files combined into one AFS file.

compact AFS with:

<pre>
AFSCompactor.exe
</pre>

for Alone in the Dark - The New Nightmare .XA files are AFS files, rename first with:

<pre>
rename .XA .afs *.XA
</pre>

then compact and rename back to .XA:

<pre>
rename .afs .XA *.afs
</pre>

AFSCompactor.exe will rename files inside .afs to file0001 etc... AITD4 doesnt seem to care.

extract AFS with:

<pre>
afs_extract.exe *.afs|*.XA
</pre>
or
<pre>
afsextract.exe -sc afsfile|XAfile
</pre>
or
<pre>
afsutils_v2.exe | rename .XA files to .afs beforehand.
</pre>

==== SFD Movies ====

http://www.dc-swat.ru/page/download/

Demux video to .m1v and audio to .wav first

Example (play around with bitrates):

<pre>
for f in <SFD Files>; do ffmpeg -i $f -vcodec mpeg1video -b:v 2600k -minrate 2600k -maxrate 3600k -y $f.m1v -acodec adpcm_ima_wav -y $f.wav ; done
</pre>

Open all .wavs in single Audacity project and mass export again, 16bit and 44100 Hz.

Start Dreamcast Movie Converter and remux/reencode to .SFD. Dont use batch processing, videos get truncated!

==== 4XM Movies ====

http://samples.mplayerhq.hu/game-formats/4xm/

Convert to AVI first with ffmpeg

<pre>
for i in ../rips/unmodified/aitd4/cd2/AITD4_GD_2/MOVIES/*.4XM; do ffmpeg -i "$i" -vf scale=320x160 -y "${i%.*}.avi"; done
</pre>

Downsample with 4xm_comp.exe

Speed : 30 fps
Data Rate : 110000
Resize Width 320 Height 160

=== CDDA Audio ===

Generate file list starting in GD-ROM HD Area after first data Track

<pre>
for f in ../backups/GDI/Mortal\ Kombat\ Gold/track*.raw; do f="${f}"; [[ "$f" == *track04.raw ]] && start=1; [[ $start -eq 1 ]] && files+=("$f"); done
</pre>

Burn CDDA Tracks first
Note: -swab is only neccessary when CDDA is static, depends on the burner.

<pre>
cdrecord -v -multi -tao -audio -swab "${files[@]}"
</pre>

Then get MSINFO Value for 2nd Session.

=== Create ISO Images ===

For the 1st Session, chosen files must not exceed total of 69120000 bytes!

==== 1st Session ====

===== Create first ISOFile =====

<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <CHOSEN FILE> <CHOSEN FILE> <CHOSEN FILE>...
</pre>

Fill first ISOFile with zeros so it takes up exactly 68808703 bytes = LBA 45000

<pre>
dd if=/dev/zero of=<OUTPUT1.ISO> bs=1 count=1 seek=68808703
</pre>

===== Burn 1st session =====

<pre>
cdrecord -v -multi -tao -xa <OUTPUT1.ISO>
</pre>

Check MSINFO value & TOC

<pre>
cdrecord -msinfo

0,45000
</pre>

<pre>
cdrecord -toc

first: 1 last 1
track: 1 lba: 0 ( 0) 00:02:00 adr: 1 control: 4 mode: -1
track:lout lba: 33600 ( 134400) 07:30:00 adr: 1 control: 4 mode: -1
</pre>

To include subdirectories into the root directory of the CD, use
<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <SUBDIRECTORY>/=<PATH_OF_SUBDIRECTORY>
</pre>
Example:
<pre>
mkisofs -no-allow-lowercase -graft-points -l -duplicates-once -V AITD4_GD_2 -o aitd4_d2_1.iso MOVIES/=../rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES/
</pre>

==== 2nd Session ====

===== Hack 1ST_READ.BIN & IP.BIN =====

Get MSINFO Value

<pre>
cdrecord -msinfo
</pre>

<pre>
binhack

Enter name of binary: 1ST_READ.BIN
Enter name of bootsector: ip.bin
Enter msinfo value: <MSINFO Value>
File 1ST_READ.BIN successfully hacked.
Bincon detected, OS flag set to 0.
File ip.bin successfully created.
</pre>

===== Insert custom Logo =====

<pre>
logoinsert darkrealm.mr IP.BIN
</pre>

===== Burn 2nd Session =====

<pre>
mkisofs -sort sort.txt -no-allow-lowercase -l -duplicates-once -V <DISCNAME> -G <IP.BIN> -M <OUTPUT1.ISO> -C <MSINFO VALUE> -o <OUTPUT2.ISO> <GAME_DATA_DIRECTORY>
cdrecord -v -tao -xa <OUTPUT2.ISO>
</pre>

== Testing ==

* Burn whole stuff to a CD-RW
* Generate CDI from CD-RW
** Launch DiscJuggler from emulated Windows Machine and connect via iSCSI
* Test CDI with ReiCast oder ReDream

== Create CDI Images from ISO ==

* Create ISO first

<pre>
mkisofs -l -duplicates-once -graft-points -V <DISCNAME> -G <IP.BIN> -o <GAME.ISO> <GAME DIRECTORY>
</pre>

* Create DATA/DATA CDI

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

* or create AUDIO/DATA CDI, with 2nd Session starting at LBA 11702
** Note: 1ST_READ.BIN must eventually be binhacked to LBA 11702

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

== 60Hz <-> 50Hz Binary Comparisations ==

=== Dino Crisis GER ===

1ST_READ.BIN:

<pre>
Offset : 0x001545C0
50Hz : 01 E4 03 A0 02 E4 01 A0 03
60Hz : 00 E4 03 A0 00 E4 01 A0 00
</pre>

== Tools ==

=== Efficient 1st Session Filler ===

<pre>
#!/bin/bash

# Prüfen, ob ein Verzeichnis übergeben wurde
if [ -z "$1" ]; then
echo "Usage: $0 <Verzeichnis>"
exit 1
fi

DIR="$1"

# Prüfen, ob es ein gültiges Verzeichnis ist
if [ ! -d "$DIR" ]; then
echo "Fehler: '$DIR' ist kein gültiges Verzeichnis."
exit 1
fi

# Maximal erlaubte Größe in Bytes
MAX_SIZE=68808703

# Ausgabedatei für die Liste
OUTPUT="dateiliste.txt"

# Leere die Ausgabedatei
> "$OUTPUT"

# Variable für die aktuelle Gesamtsumme
total_size=0

# Durchlaufe die Dateien im Verzeichnis, sortiert nach Größe absteigend
while IFS= read -r file; do
filesize=$(stat -c%s "$file")

if (( total_size + filesize <= MAX_SIZE )); then
# echo "$filesize $file" >> "$OUTPUT"
echo "$file" >> "$OUTPUT"
total_size=$((total_size + filesize))
fi
done < <(find "$DIR" -maxdepth 1 -type f -printf "%s %p\n" | sort -nr | awk '{print $2}')

echo "Liste erstellt: $OUTPUT"
echo "Gesamtgröße: $total_size Bytes"
</pre>

== Draft ==

* ./find1stsessionfiles.sh ~/media/development/dreamcast/Rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES
* mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V AITD4_GD2 -o aitd4_cd2_s1.iso -path-list ~/media/chris/development/dreamcast/tools/dateiliste.txt
* cdrecord -v -tao -multi -xa aitd4_d2_s1.iso
* dd if=/dev/zero of=aitd4_d2_s1.iso bs=1 count=1 seek=68808703

Dreamcast

2026-05-05T21:21:23Z

Chris: /* Draft */

== CDR Sizes ==

80min/700MB: 1st Session 674MB Maximum

90min/800MB: max 748MB when 45000 Sectors in 1st Session with dummy burn

90min/800MB: max ~735MB when 45000 Sectors in 1st Session

== Access Files on CDI Images ==

# Extract the seperate Track first

<pre>
cdirip <CDIFILE>.cdi
</pre>

# In case of 2nd Data Track, fix the LBA first

<pre>
isofix <TRACK02>.iso <LBA Value reported by cdirip>
</pre>

== What binhack does ==

Changes LBA Value at offset 1461904 ( 8 Bytes before CD001 string) from 6EB0 (45166) to LBA Value+166 specified during binhack.

Adds Region Codes for JAP,EUR and USA and strings for that.
Adds some bootcode, probably SH4 ASM to make the GDROM accept CDRs.

== What cdda does ==

Changes some bytes, probably SH4 ASM to enable playback of CDDA tracks.

== What dahack does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to 9600 (150). Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== What hack3 does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to LBA Value+150 specified during dahack. Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== Another Copy Protection ==

Searching for CDE4 036A (or CDE4 436A) and replacing with 0900 0900 is another Copy Protection defeat.

CDE4 436A -> 0900 0900
1032 0D8B -> 0800 0D8B

== LBA 45000 Trick ==

69120000 bytes in first session will result in LBA 45000.

== Fill 1st Session ISO File to 69120000 bytes with zeros ==

6912000 bytes = 33750 2048-byte-sectors

CORRECTION : Must be 33598 Sectors !!! = 68808703 bytes

1st Session Leadout = 11250 Sectors = 23040000 Bytes

2nd Session will start at Sector 45000

When using extract.exe, add 150 Sectors to LBA reported in gdinfo_FG.exe track05

<pre>
dd if=/dev/zero of=<ISOFILE> bs=1 count=1 seek=69119999
</pre>

== Creating & Burning DATA/DATA ISOs ==

== Merging track03 & track05 into one file ==

Create dummy file first
Calculate 1185760800 - <SIZE OF track03.bin + SIZE OF track05.bin>

<pre>
dd if=/dev/zero bs=1 count=0 seek=<RESULT> of=dummy.bin
cat track03.bin dummy.bin track05.bin > combined.bin
</pre>

This also helps when extract.exe only outputs garbage out of track03.iso along with track05.iso. This way only the combined.iso needs to be extracted.

== Burning DATA/DATA CDIs ==

<pre>
cdirip <CDIFile> <Destination Directory>
cdrecord dev=/dev/sr0 -v -multi -tao -xa tdata01.iso
cdrecord dev=/dev/sr0 -v -tao -xa tdata02.iso
</pre>

== Full rip process from GD-ROM to CD-ROM ==

=== Get GDROM Data from Dreamcast with http-ack ===

https://www.gotwalls.com

https://github.com/sega-dreamcast/httpd-ack

# Boot HTTP-ack from CD
# Swap HTTP-ack CD with Original GD-ROm
# Connect via Browser to http://192.168.1.30
# Download all files starting with track03.bin
# Download disc.gdi

=== Extract Data from Tracks ===

==== Use gditools ====

https://sourceforge.net/projects/dcisotools/

Either use GUI or gditools.py

<pre>
python2.7 ./gditools.py
Usage: ./gditools.py -i input_gdi [options]

-h, --help Display this help
-l, --list List all files in the filesystem and exit
-o [outdir] Output directory. Default: gdi folder
-s [filename] Create a sorttxt file with custom name
(It uses *data-folder* as prefix)
-b [ipname] Dump the ip.bin with custom name
-e [filename] Dump a single file from the filesystem
--extract-all Dump all the files in the *data-folder*
--data-folder [name] *data-folder* subfolder. Default: data
(__volume_label__ --> Use ISO9660 volume label)
--sort-spacer [num] Sorttxt entries are sperated by num
--silent Minimal verbosity mode
[no option] Display gdi infos if not silent

gditools.py by FamilyGuy, http://sourceforge.net/p/dcisotools/
Licensed under GPLv3, see licences folder.

iso9660.py by Barney Gale, http://github.com/barneygale
Licensed under a BSD-based license, see licences folder.
</pre>

<pre>
python2.7 ./gditools.py -i <disc.gdi> --extract-all -o /mnt/development/dreamcast/rips/orgs/<GAME_NAME>/<CD Number> --data-folder __volume_label__ -s sort.txt -b ip.bin
</pre>

==== Old way ====
# If the game contains more than track03.bin, combine the 2 binfiles into 1:
<pre>
cat track03.bin dummy.bin track05.bin > <BINFILE>
</pre>
# Convert .BIN Files to .ISO Files first with bin2iso.exe
<pre>
bin2iso.exe <BINFILE> <ISOFILE>
</pre>
# Extract Data from .ISO files
<pre>
extract.exe <ISOFILE>
</pre>
If extracted 1ST_READ.BIN has filesize of 0, LBA of track05 is wrong.

=== Downsampling ===

==== AFS Files ====

https://sizious.com/download/dreamcast/

AFS files contain multiple ADX files combined into one AFS file.

compact AFS with:

<pre>
AFSCompactor.exe
</pre>

for Alone in the Dark - The New Nightmare .XA files are AFS files, rename first with:

<pre>
rename .XA .afs *.XA
</pre>

then compact and rename back to .XA:

<pre>
rename .afs .XA *.afs
</pre>

AFSCompactor.exe will rename files inside .afs to file0001 etc... AITD4 doesnt seem to care.

extract AFS with:

<pre>
afs_extract.exe *.afs|*.XA
</pre>
or
<pre>
afsextract.exe -sc afsfile|XAfile
</pre>
or
<pre>
afsutils_v2.exe | rename .XA files to .afs beforehand.
</pre>

==== SFD Movies ====

http://www.dc-swat.ru/page/download/

Demux video to .m1v and audio to .wav first

Example (play around with bitrates):

<pre>
for f in <SFD Files>; do ffmpeg -i $f -vcodec mpeg1video -b:v 2600k -minrate 2600k -maxrate 3600k -y $f.m1v -acodec adpcm_ima_wav -y $f.wav ; done
</pre>

Open all .wavs in single Audacity project and mass export again, 16bit and 44100 Hz.

Start Dreamcast Movie Converter and remux/reencode to .SFD. Dont use batch processing, videos get truncated!

==== 4XM Movies ====

http://samples.mplayerhq.hu/game-formats/4xm/

Convert to AVI first with ffmpeg

<pre>
for i in ../rips/unmodified/aitd4/cd2/AITD4_GD_2/MOVIES/*.4XM; do ffmpeg -i "$i" -vf scale=320x160 -y "${i%.*}.avi"; done
</pre>

Downsample with 4xm_comp.exe

Speed : 30 fps
Data Rate : 110000
Resize Width 320 Height 160

=== CDDA Audio ===

Generate file list starting in GD-ROM HD Area after first data Track

<pre>
for f in ../backups/GDI/Mortal\ Kombat\ Gold/track*.raw; do f="${f}"; [[ "$f" == *track04.raw ]] && start=1; [[ $start -eq 1 ]] && files+=("$f"); done
</pre>

Burn CDDA Tracks first
Note: -swab is only neccessary when CDDA is static, depends on the burner.

<pre>
cdrecord -v -multi -tao -audio -swab "${files[@]}"
</pre>

Then get MSINFO Value for 2nd Session.

=== Create ISO Images ===

For the 1st Session, chosen files must not exceed total of 69120000 bytes!

==== 1st Session ====

===== Create first ISOFile =====

<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <CHOSEN FILE> <CHOSEN FILE> <CHOSEN FILE>...
</pre>

Fill first ISOFile with zeros so it takes up exactly 68808703 bytes = LBA 45000

<pre>
dd if=/dev/zero of=<OUTPUT1.ISO> bs=1 count=1 seek=68808703
</pre>

===== Burn 1st session =====

<pre>
cdrecord -v -multi -tao -xa <OUTPUT1.ISO>
</pre>

Check MSINFO value & TOC

<pre>
cdrecord -msinfo

0,45000
</pre>

<pre>
cdrecord -toc

first: 1 last 1
track: 1 lba: 0 ( 0) 00:02:00 adr: 1 control: 4 mode: -1
track:lout lba: 33600 ( 134400) 07:30:00 adr: 1 control: 4 mode: -1
</pre>

To include subdirectories into the root directory of the CD, use
<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <SUBDIRECTORY>/=<PATH_OF_SUBDIRECTORY>
</pre>
Example:
<pre>
mkisofs -no-allow-lowercase -graft-points -l -duplicates-once -V AITD4_GD_2 -o aitd4_d2_1.iso MOVIES/=../rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES/
</pre>

==== 2nd Session ====

===== Hack 1ST_READ.BIN & IP.BIN =====

Get MSINFO Value

<pre>
cdrecord -msinfo
</pre>

<pre>
binhack

Enter name of binary: 1ST_READ.BIN
Enter name of bootsector: ip.bin
Enter msinfo value: <MSINFO Value>
File 1ST_READ.BIN successfully hacked.
Bincon detected, OS flag set to 0.
File ip.bin successfully created.
</pre>

===== Insert custom Logo =====

<pre>
logoinsert darkrealm.mr IP.BIN
</pre>

===== Burn 2nd Session =====

<pre>
mkisofs -sort sort.txt -no-allow-lowercase -l -duplicates-once -V <DISCNAME> -G <IP.BIN> -M <OUTPUT1.ISO> -C <MSINFO VALUE> -o <OUTPUT2.ISO> <GAME_DATA_DIRECTORY>
cdrecord -v -tao -xa <OUTPUT2.ISO>
</pre>

== Testing ==

* Burn whole stuff to a CD-RW
* Generate CDI from CD-RW
** Launch DiscJuggler from emulated Windows Machine and connect via iSCSI
* Test CDI with ReiCast oder ReDream

== Create CDI Images from ISO ==

* Create ISO first

<pre>
mkisofs -l -duplicates-once -graft-points -V <DISCNAME> -G <IP.BIN> -o <GAME.ISO> <GAME DIRECTORY>
</pre>

* Create DATA/DATA CDI

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

* or create AUDIO/DATA CDI, with 2nd Session starting at LBA 11702
** Note: 1ST_READ.BIN must eventually be binhacked to LBA 11702

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

== 60Hz <-> 50Hz Binary Comparisations ==

=== Dino Crisis GER ===

1ST_READ.BIN:

<pre>
Offset : 0x001545C0
50Hz : 01 E4 03 A0 02 E4 01 A0 03
60Hz : 00 E4 03 A0 00 E4 01 A0 00
</pre>

== Tools ==

=== Efficient 1st Session Filler ===

<pre>
#!/bin/bash

# Prüfen, ob ein Verzeichnis übergeben wurde
if [ -z "$1" ]; then
echo "Usage: $0 <Verzeichnis>"
exit 1
fi

DIR="$1"

# Prüfen, ob es ein gültiges Verzeichnis ist
if [ ! -d "$DIR" ]; then
echo "Fehler: '$DIR' ist kein gültiges Verzeichnis."
exit 1
fi

# Maximal erlaubte Größe in Bytes
MAX_SIZE=68808703

# Ausgabedatei für die Liste
OUTPUT="dateiliste.txt"

# Leere die Ausgabedatei
> "$OUTPUT"

# Variable für die aktuelle Gesamtsumme
total_size=0

# Durchlaufe die Dateien im Verzeichnis, sortiert nach Größe absteigend
while IFS= read -r file; do
filesize=$(stat -c%s "$file")

if (( total_size + filesize <= MAX_SIZE )); then
# echo "$filesize $file" >> "$OUTPUT"
echo "$file" >> "$OUTPUT"
total_size=$((total_size + filesize))
fi
done < <(find "$DIR" -maxdepth 1 -type f -printf "%s %p\n" | sort -nr | awk '{print $2}')

echo "Liste erstellt: $OUTPUT"
echo "Gesamtgröße: $total_size Bytes"
</pre>

== Draft ==

* ./find1stsessionfiles.sh ~/media/development/dreamcast/Rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES
* mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V AITD4_GD2 -o aitd4_cd2_s1.iso -path-list ~/media/chris/development/dreamcast/tools/dateiliste.txt
* cdrecord -v -tao -multi -xa aitd4_d2_s1.iso
* dd if=/dev/zero of=aitd4_d2_s1.iso bs=1 count=1 seek=69119999

Dreamcast

2026-05-05T21:15:11Z

Chris: /* Draft */

Dreamcast

2026-05-05T21:08:27Z

Chris: /* Tools */

Dreamcast

2026-05-05T21:03:36Z

Chris:

Dreamcast

2026-04-25T19:56:17Z

Chris: /* Full rip process from GD-ROM to CD-ROM */

Dreamcast

2026-04-25T19:54:20Z

Chris: /* 4XM Movies */

== CDR Sizes ==

80min/700MB: 1st Session 674MB Maximum

90min/800MB: max 748MB when 45000 Sectors in 1st Session with dummy burn

90min/800MB: max ~735MB when 45000 Sectors in 1st Session

== Access Files on CDI Images ==

# Extract the seperate Track first

<pre>
cdirip <CDIFILE>.cdi
</pre>

# In case of 2nd Data Track, fix the LBA first

<pre>
isofix <TRACK02>.iso <LBA Value reported by cdirip>
</pre>

== What binhack does ==

Changes LBA Value at offset 1461904 ( 8 Bytes before CD001 string) from 6EB0 (45166) to LBA Value+166 specified during binhack.

Adds Region Codes for JAP,EUR and USA and strings for that.
Adds some bootcode, probably SH4 ASM to make the GDROM accept CDRs.

== What cdda does ==

Changes some bytes, probably SH4 ASM to enable playback of CDDA tracks.

== What dahack does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to 9600 (150). Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== What hack3 does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to LBA Value+150 specified during dahack. Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== Another Copy Protection ==

Searching for CDE4 036A (or CDE4 436A) and replacing with 0900 0900 is another Copy Protection defeat.

CDE4 436A -> 0900 0900
1032 0D8B -> 0800 0D8B

== LBA 45000 Trick ==

69120000 bytes in first session will result in LBA 45000.

== Fill 1st Session ISO File to 69120000 bytes with zeros ==

6912000 bytes = 33750 2048-byte-sectors

CORRECTION : Must be 33598 Sectors !!! = 68808703 bytes

1st Session Leadout = 11250 Sectors = 23040000 Bytes

2nd Session will start at Sector 45000

When using extract.exe, add 150 Sectors to LBA reported in gdinfo_FG.exe track05

<pre>
dd if=/dev/zero of=<ISOFILE> bs=1 count=1 seek=69119999
</pre>

== Creating & Burning DATA/DATA ISOs ==

== Merging track03 & track05 into one file ==

Create dummy file first
Calculate 1185760800 - <SIZE OF track03.bin + SIZE OF track05.bin>

<pre>
dd if=/dev/zero bs=1 count=0 seek=<RESULT> of=dummy.bin
cat track03.bin dummy.bin track05.bin > combined.bin
</pre>

This also helps when extract.exe only outputs garbage out of track03.iso along with track05.iso. This way only the combined.iso needs to be extracted.

== Burning DATA/DATA CDIs ==

<pre>
cdirip <CDIFile> <Destination Directory>
cdrecord dev=/dev/sr0 -v -multi -tao -xa tdata01.iso
cdrecord dev=/dev/sr0 -v -tao -xa tdata02.iso
</pre>

== Full rip process from GD-ROM to CD-ROM ==

=== Get GDROM Data from Dreamcast with http-ack ===

https://www.gotwalls.com

https://github.com/sega-dreamcast/httpd-ack

# Boot HTTP-ack from CD
# Swap HTTP-ack CD with Original GD-ROm
# Connect via Browser to http://192.168.1.30
# Download all files starting with track03.bin
# Download disc.gdi

=== Extract Data from Tracks ===

==== Use gditools ====

https://sourceforge.net/projects/dcisotools/

Either use GUI or gditools.py

<pre>
python2.7 ./gditools.py
Usage: ./gditools.py -i input_gdi [options]

-h, --help Display this help
-l, --list List all files in the filesystem and exit
-o [outdir] Output directory. Default: gdi folder
-s [filename] Create a sorttxt file with custom name
(It uses *data-folder* as prefix)
-b [ipname] Dump the ip.bin with custom name
-e [filename] Dump a single file from the filesystem
--extract-all Dump all the files in the *data-folder*
--data-folder [name] *data-folder* subfolder. Default: data
(__volume_label__ --> Use ISO9660 volume label)
--sort-spacer [num] Sorttxt entries are sperated by num
--silent Minimal verbosity mode
[no option] Display gdi infos if not silent

gditools.py by FamilyGuy, http://sourceforge.net/p/dcisotools/
Licensed under GPLv3, see licences folder.

iso9660.py by Barney Gale, http://github.com/barneygale
Licensed under a BSD-based license, see licences folder.
</pre>

<pre>
python2.7 ./gditools.py -i <disc.gdi> --extract-all -o /mnt/development/dreamcast/rips/orgs/<GAME_NAME>/<CD Number> --data-folder __volume_label__ -s sort.txt -b ip.bin
</pre>

==== Old way ====
# If the game contains more than track03.bin, combine the 2 binfiles into 1:
<pre>
cat track03.bin dummy.bin track05.bin > <BINFILE>
</pre>
# Convert .BIN Files to .ISO Files first with bin2iso.exe
<pre>
bin2iso.exe <BINFILE> <ISOFILE>
</pre>
# Extract Data from .ISO files
<pre>
extract.exe <ISOFILE>
</pre>
If extracted 1ST_READ.BIN has filesize of 0, LBA of track05 is wrong.

=== Downsampling ===

==== AFS Files ====

https://sizious.com/download/dreamcast/

AFS files contain multiple ADX files combined into one AFS file.

compact AFS with:

<pre>
AFSCompactor.exe
</pre>

for Alone in the Dark - The New Nightmare .XA files are AFS files, rename first with:

<pre>
rename .XA .afs *.XA
</pre>

then compact and rename back to .XA:

<pre>
rename .afs .XA *.afs
</pre>

AFSCompactor.exe will rename files inside .afs to file0001 etc... AITD4 doesnt seem to care.

extract AFS with:

<pre>
afs_extract.exe *.afs|*.XA
</pre>
or
<pre>
afsextract.exe -sc afsfile|XAfile
</pre>
or
<pre>
afsutils_v2.exe | rename .XA files to .afs beforehand.
</pre>

==== SFD Movies ====

http://www.dc-swat.ru/page/download/

Demux video to .m1v and audio to .wav first

Example (play around with bitrates):

<pre>
for f in <SFD Files>; do ffmpeg -i $f -vcodec mpeg1video -b:v 2600k -minrate 2600k -maxrate 3600k -y $f.m1v -acodec adpcm_ima_wav -y $f.wav ; done
</pre>

Open all .wavs in single Audacity project and mass export again, 16bit and 44100 Hz.

Start Dreamcast Movie Converter and remux/reencode to .SFD. Dont use batch processing, videos get truncated!

==== 4XM Movies ====

http://samples.mplayerhq.hu/game-formats/4xm/

Convert to AVI first with ffmpeg

<pre>
for i in ../rips/unmodified/aitd4/cd2/AITD4_GD_2/MOVIES/*.4XM; do ffmpeg -i "$i" -vf scale=320x160 -y "${i%.*}.avi"; done
</pre>

Downsample with 4xm_comp.exe

Speed : 30 fps
Data Rate : 110000
Resize Width 320 Height 160

==== CDDA Audio ====

Generate file list starting in GD-ROM HD Area after first data Track

<pre>
for f in ../backups/GDI/Mortal\ Kombat\ Gold/track*.raw; do f="${f}"; [[ "$f" == *track04.raw ]] && start=1; [[ $start -eq 1 ]] && files+=("$f"); done
</pre>

Burn CDDA Tracks first
Note: -swab is only neccessary when CDDA is static, depends on the burner.

<pre>
cdrecord -v -multi -tao -audio -swab "${files[@]}"
</pre>

Then get MSINFO Value for 2nd Session.

=== Create ISO Images ===

For the 1st Session, chosen files must not exceed total of 69120000 bytes!

==== 1st Session ====

===== Create first ISOFile =====

<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <CHOSEN FILE> <CHOSEN FILE> <CHOSEN FILE>...
</pre>

Fill first ISOFile with zeros so it takes up exactly 68808703 bytes = LBA 45000

<pre>
dd if=/dev/zero of=<OUTPUT1.ISO> bs=1 count=1 seek=68808703
</pre>

===== Burn 1st session =====

<pre>
cdrecord -v -multi -tao -xa <OUTPUT1.ISO>
</pre>

Check MSINFO value & TOC

<pre>
cdrecord -msinfo

0,45000
</pre>

<pre>
cdrecord -toc

first: 1 last 1
track: 1 lba: 0 ( 0) 00:02:00 adr: 1 control: 4 mode: -1
track:lout lba: 33600 ( 134400) 07:30:00 adr: 1 control: 4 mode: -1
</pre>

To include subdirectories into the root directory of the CD, use
<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <SUBDIRECTORY>/=<PATH_OF_SUBDIRECTORY>
</pre>
Example:
<pre>
mkisofs -no-allow-lowercase -graft-points -l -duplicates-once -V AITD4_GD_2 -o aitd4_d2_1.iso MOVIES/=../rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES/
</pre>

==== 2nd Session ====

===== Hack 1ST_READ.BIN & IP.BIN =====

Get MSINFO Value

<pre>
cdrecord -msinfo
</pre>

<pre>
binhack

Enter name of binary: 1ST_READ.BIN
Enter name of bootsector: ip.bin
Enter msinfo value: <MSINFO Value>
File 1ST_READ.BIN successfully hacked.
Bincon detected, OS flag set to 0.
File ip.bin successfully created.
</pre>

===== Insert custom Logo =====

<pre>
logoinsert darkrealm.mr IP.BIN
</pre>

===== Burn 2nd Session =====

<pre>
mkisofs -sort sort.txt -no-allow-lowercase -l -duplicates-once -V <DISCNAME> -G <IP.BIN> -M <OUTPUT1.ISO> -C <MSINFO VALUE> -o <OUTPUT2.ISO> <GAME_DATA_DIRECTORY>
cdrecord -v -tao -xa <OUTPUT2.ISO>
</pre>

== Testing ==

* Burn whole stuff to a CD-RW
* Generate CDI from CD-RW
** Launch DiscJuggler from emulated Windows Machine and connect via iSCSI
* Test CDI with ReiCast oder ReDream

== Create CDI Images from ISO ==

* Create ISO first

<pre>
mkisofs -l -duplicates-once -graft-points -V <DISCNAME> -G <IP.BIN> -o <GAME.ISO> <GAME DIRECTORY>
</pre>

* Create DATA/DATA CDI

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

* or create AUDIO/DATA CDI, with 2nd Session starting at LBA 11702
** Note: 1ST_READ.BIN must eventually be binhacked to LBA 11702

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

== 60Hz <-> 50Hz Binary Comparisations ==

=== Dino Crisis GER ===

1ST_READ.BIN:

<pre>
Offset : 0x001545C0
50Hz : 01 E4 03 A0 02 E4 01 A0 03
60Hz : 00 E4 03 A0 00 E4 01 A0 00
</pre>

Dreamcast

2026-04-25T13:44:56Z

Chris:

== CDR Sizes ==

80min/700MB: 1st Session 674MB Maximum

90min/800MB: max 748MB when 45000 Sectors in 1st Session with dummy burn

90min/800MB: max ~735MB when 45000 Sectors in 1st Session

== Access Files on CDI Images ==

# Extract the seperate Track first

<pre>
cdirip <CDIFILE>.cdi
</pre>

# In case of 2nd Data Track, fix the LBA first

<pre>
isofix <TRACK02>.iso <LBA Value reported by cdirip>
</pre>

== What binhack does ==

Changes LBA Value at offset 1461904 ( 8 Bytes before CD001 string) from 6EB0 (45166) to LBA Value+166 specified during binhack.

Adds Region Codes for JAP,EUR and USA and strings for that.
Adds some bootcode, probably SH4 ASM to make the GDROM accept CDRs.

== What cdda does ==

Changes some bytes, probably SH4 ASM to enable playback of CDDA tracks.

== What dahack does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to 9600 (150). Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== What hack3 does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to LBA Value+150 specified during dahack. Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== Another Copy Protection ==

Searching for CDE4 036A (or CDE4 436A) and replacing with 0900 0900 is another Copy Protection defeat.

CDE4 436A -> 0900 0900
1032 0D8B -> 0800 0D8B

== LBA 45000 Trick ==

69120000 bytes in first session will result in LBA 45000.

== Fill 1st Session ISO File to 69120000 bytes with zeros ==

6912000 bytes = 33750 2048-byte-sectors

CORRECTION : Must be 33598 Sectors !!! = 68808703 bytes

1st Session Leadout = 11250 Sectors = 23040000 Bytes

2nd Session will start at Sector 45000

When using extract.exe, add 150 Sectors to LBA reported in gdinfo_FG.exe track05

<pre>
dd if=/dev/zero of=<ISOFILE> bs=1 count=1 seek=69119999
</pre>

== Creating & Burning DATA/DATA ISOs ==

== Merging track03 & track05 into one file ==

Create dummy file first
Calculate 1185760800 - <SIZE OF track03.bin + SIZE OF track05.bin>

<pre>
dd if=/dev/zero bs=1 count=0 seek=<RESULT> of=dummy.bin
cat track03.bin dummy.bin track05.bin > combined.bin
</pre>

This also helps when extract.exe only outputs garbage out of track03.iso along with track05.iso. This way only the combined.iso needs to be extracted.

== Burning DATA/DATA CDIs ==

<pre>
cdirip <CDIFile> <Destination Directory>
cdrecord dev=/dev/sr0 -v -multi -tao -xa tdata01.iso
cdrecord dev=/dev/sr0 -v -tao -xa tdata02.iso
</pre>

== Full rip process from GD-ROM to CD-ROM ==

=== Get GDROM Data from Dreamcast with http-ack ===

https://www.gotwalls.com

https://github.com/sega-dreamcast/httpd-ack

# Boot HTTP-ack from CD
# Swap HTTP-ack CD with Original GD-ROm
# Connect via Browser to http://192.168.1.30
# Download all files starting with track03.bin
# Download disc.gdi

=== Extract Data from Tracks ===

==== Use gditools ====

https://sourceforge.net/projects/dcisotools/

Either use GUI or gditools.py

<pre>
python2.7 ./gditools.py
Usage: ./gditools.py -i input_gdi [options]

-h, --help Display this help
-l, --list List all files in the filesystem and exit
-o [outdir] Output directory. Default: gdi folder
-s [filename] Create a sorttxt file with custom name
(It uses *data-folder* as prefix)
-b [ipname] Dump the ip.bin with custom name
-e [filename] Dump a single file from the filesystem
--extract-all Dump all the files in the *data-folder*
--data-folder [name] *data-folder* subfolder. Default: data
(__volume_label__ --> Use ISO9660 volume label)
--sort-spacer [num] Sorttxt entries are sperated by num
--silent Minimal verbosity mode
[no option] Display gdi infos if not silent

gditools.py by FamilyGuy, http://sourceforge.net/p/dcisotools/
Licensed under GPLv3, see licences folder.

iso9660.py by Barney Gale, http://github.com/barneygale
Licensed under a BSD-based license, see licences folder.
</pre>

<pre>
python2.7 ./gditools.py -i <disc.gdi> --extract-all -o /mnt/development/dreamcast/rips/orgs/<GAME_NAME>/<CD Number> --data-folder __volume_label__ -s sort.txt -b ip.bin
</pre>

==== Old way ====
# If the game contains more than track03.bin, combine the 2 binfiles into 1:
<pre>
cat track03.bin dummy.bin track05.bin > <BINFILE>
</pre>
# Convert .BIN Files to .ISO Files first with bin2iso.exe
<pre>
bin2iso.exe <BINFILE> <ISOFILE>
</pre>
# Extract Data from .ISO files
<pre>
extract.exe <ISOFILE>
</pre>
If extracted 1ST_READ.BIN has filesize of 0, LBA of track05 is wrong.

=== Downsampling ===

==== AFS Files ====

https://sizious.com/download/dreamcast/

AFS files contain multiple ADX files combined into one AFS file.

compact AFS with:

<pre>
AFSCompactor.exe
</pre>

for Alone in the Dark - The New Nightmare .XA files are AFS files, rename first with:

<pre>
rename .XA .afs *.XA
</pre>

then compact and rename back to .XA:

<pre>
rename .afs .XA *.afs
</pre>

AFSCompactor.exe will rename files inside .afs to file0001 etc... AITD4 doesnt seem to care.

extract AFS with:

<pre>
afs_extract.exe *.afs|*.XA
</pre>
or
<pre>
afsextract.exe -sc afsfile|XAfile
</pre>
or
<pre>
afsutils_v2.exe | rename .XA files to .afs beforehand.
</pre>

==== SFD Movies ====

http://www.dc-swat.ru/page/download/

Demux video to .m1v and audio to .wav first

Example (play around with bitrates):

<pre>
for f in <SFD Files>; do ffmpeg -i $f -vcodec mpeg1video -b:v 2600k -minrate 2600k -maxrate 3600k -y $f.m1v -acodec adpcm_ima_wav -y $f.wav ; done
</pre>

Open all .wavs in single Audacity project and mass export again, 16bit and 44100 Hz.

Start Dreamcast Movie Converter and remux/reencode to .SFD. Dont use batch processing, videos get truncated!

==== 4XM Movies ====

http://samples.mplayerhq.hu/game-formats/4xm/

Convert to AVI first with ffmpeg

<pre>
for i in ../rips/unmodified/aitd4/cd2/AITD4_GD_2/MOVIES/*.4XM; do ffmpeg -i "$i" -vf scale=320x160 -y "${i%.*}.avi"; done
</pre>

Downsample with 4xm_comp.exe

Speed : 30 fps
Data Rate : 110000
Resize Width 320 Height 160

=== Create ISO Images ===

For the 1st Session, chosen files must not exceed total of 69120000 bytes!

==== 1st Session ====

===== Create first ISOFile =====

<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <CHOSEN FILE> <CHOSEN FILE> <CHOSEN FILE>...
</pre>

Fill first ISOFile with zeros so it takes up exactly 68808703 bytes = LBA 45000

<pre>
dd if=/dev/zero of=<OUTPUT1.ISO> bs=1 count=1 seek=68808703
</pre>

===== Burn 1st session =====

<pre>
cdrecord -v -multi -tao -xa <OUTPUT1.ISO>
</pre>

Check MSINFO value & TOC

<pre>
cdrecord -msinfo

0,45000
</pre>

<pre>
cdrecord -toc

first: 1 last 1
track: 1 lba: 0 ( 0) 00:02:00 adr: 1 control: 4 mode: -1
track:lout lba: 33600 ( 134400) 07:30:00 adr: 1 control: 4 mode: -1
</pre>

To include subdirectories into the root directory of the CD, use
<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <SUBDIRECTORY>/=<PATH_OF_SUBDIRECTORY>
</pre>
Example:
<pre>
mkisofs -no-allow-lowercase -graft-points -l -duplicates-once -V AITD4_GD_2 -o aitd4_d2_1.iso MOVIES/=../rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES/
</pre>

==== 2nd Session ====

===== Hack 1ST_READ.BIN & IP.BIN =====

Get MSINFO Value

<pre>
cdrecord -msinfo
</pre>

<pre>
binhack

Enter name of binary: 1ST_READ.BIN
Enter name of bootsector: ip.bin
Enter msinfo value: <MSINFO Value>
File 1ST_READ.BIN successfully hacked.
Bincon detected, OS flag set to 0.
File ip.bin successfully created.
</pre>

===== Insert custom Logo =====

<pre>
logoinsert darkrealm.mr IP.BIN
</pre>

===== Burn 2nd Session =====

<pre>
mkisofs -sort sort.txt -no-allow-lowercase -l -duplicates-once -V <DISCNAME> -G <IP.BIN> -M <OUTPUT1.ISO> -C <MSINFO VALUE> -o <OUTPUT2.ISO> <GAME_DATA_DIRECTORY>
cdrecord -v -tao -xa <OUTPUT2.ISO>
</pre>

== Testing ==

* Burn whole stuff to a CD-RW
* Generate CDI from CD-RW
** Launch DiscJuggler from emulated Windows Machine and connect via iSCSI
* Test CDI with ReiCast oder ReDream

== Create CDI Images from ISO ==

* Create ISO first

<pre>
mkisofs -l -duplicates-once -graft-points -V <DISCNAME> -G <IP.BIN> -o <GAME.ISO> <GAME DIRECTORY>
</pre>

* Create DATA/DATA CDI

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

* or create AUDIO/DATA CDI, with 2nd Session starting at LBA 11702
** Note: 1ST_READ.BIN must eventually be binhacked to LBA 11702

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

== 60Hz <-> 50Hz Binary Comparisations ==

=== Dino Crisis GER ===

1ST_READ.BIN:

<pre>
Offset : 0x001545C0
50Hz : 01 E4 03 A0 02 E4 01 A0 03
60Hz : 00 E4 03 A0 00 E4 01 A0 00
</pre>

Dreamcast

2026-04-25T10:52:31Z

Chris: /* Burn 2nd Session */

== CDR Sizes ==

80min/700MB: 1st Session 674MB Maximum

90min/800MB: max 748MB when 45000 Sectors in 1st Session with dummy burn

90min/800MB: max ~735MB when 45000 Sectors in 1st Session

== Access Files on CDI Images ==

# Extract the seperate Track first

<pre>
cdirip <CDIFILE>.cdi
</pre>

# In case of 2nd Data Track, fix the LBA first

<pre>
isofix <TRACK02>.iso <LBA Value reported by cdirip>
</pre>

== What binhack does ==

Changes LBA Value at offset 1461904 ( 8 Bytes before CD001 string) from 6EB0 (45166) to LBA Value+166 specified during binhack.

Adds Region Codes for JAP,EUR and USA and strings for that.
Adds some bootcode, probably SH4 ASM to make the GDROM accept CDRs.

== What cdda does ==

Changes some bytes, probably SH4 ASM to enable playback of CDDA tracks.

== What dahack does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to 9600 (150). Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== What hack3 does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to LBA Value+150 specified during dahack. Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== Another Copy Protection ==

Searching for CDE4 036A (or CDE4 436A) and replacing with 0900 0900 is another Copy Protection defeat.

CDE4 436A -> 0900 0900
1032 0D8B -> 0800 0D8B

== LBA 45000 Trick ==

69120000 bytes in first session will result in LBA 45000.

== Fill 1st Session ISO File to 69120000 bytes with zeros ==

6912000 bytes = 33750 2048-byte-sectors

CORRECTION : Must be 33598 Sectors !!! = 68808703 bytes

1st Session Leadout = 11250 Sectors = 23040000 Bytes

2nd Session will start at Sector 45000

When using extract.exe, add 150 Sectors to LBA reported in gdinfo_FG.exe track05

<pre>
dd if=/dev/zero of=<ISOFILE> bs=1 count=1 seek=69119999
</pre>

== Creating & Burning DATA/DATA ISOs ==

== Merging track03 & track05 into one file ==

Create dummy file first
Calculate 1185760800 - <SIZE OF track03.bin + SIZE OF track05.bin>

<pre>
dd if=/dev/zero bs=1 count=0 seek=<RESULT> of=dummy.bin
cat track03.bin dummy.bin track05.bin > combined.bin
</pre>

This also helps when extract.exe only outputs garbage out of track03.iso along with track05.iso. This way only the combined.iso needs to be extracted.

== Burning DATA/DATA CDIs ==

<pre>
cdirip <CDIFile> <Destination Directory>
cdrecord dev=/dev/sr0 -v -multi -tao -xa tdata01.iso
cdrecord dev=/dev/sr0 -v -tao -xa tdata02.iso
</pre>

== Full rip process from GD-ROM to CD-ROM ==

=== Get GDROM Data from Dreamcast with http-ack ===

https://www.gotwalls.com

https://github.com/sega-dreamcast/httpd-ack

# Boot HTTP-ack from CD
# Swap HTTP-ack CD with Original GD-ROm
# Connect via Browser to http://192.168.1.30
# Download all files starting with track03.bin
# Download disc.gdi

=== Extract Data from Tracks ===

==== Use gditools ====

https://sourceforge.net/projects/dcisotools/

Either use GUI or gditools.py

<pre>
python2.7 ./gditools.py
Usage: ./gditools.py -i input_gdi [options]

-h, --help Display this help
-l, --list List all files in the filesystem and exit
-o [outdir] Output directory. Default: gdi folder
-s [filename] Create a sorttxt file with custom name
(It uses *data-folder* as prefix)
-b [ipname] Dump the ip.bin with custom name
-e [filename] Dump a single file from the filesystem
--extract-all Dump all the files in the *data-folder*
--data-folder [name] *data-folder* subfolder. Default: data
(__volume_label__ --> Use ISO9660 volume label)
--sort-spacer [num] Sorttxt entries are sperated by num
--silent Minimal verbosity mode
[no option] Display gdi infos if not silent

gditools.py by FamilyGuy, http://sourceforge.net/p/dcisotools/
Licensed under GPLv3, see licences folder.

iso9660.py by Barney Gale, http://github.com/barneygale
Licensed under a BSD-based license, see licences folder.
</pre>

<pre>
python2.7 ./gditools.py -i <disc.gdi> --extract-all -o /mnt/development/dreamcast/rips/orgs/<GAME_NAME>/<CD Number> --data-folder __volume_label__ -s sort.txt -b ip.bin
</pre>

==== Old way ====
# If the game contains more than track03.bin, combine the 2 binfiles into 1:
<pre>
cat track03.bin dummy.bin track05.bin > <BINFILE>
</pre>
# Convert .BIN Files to .ISO Files first with bin2iso.exe
<pre>
bin2iso.exe <BINFILE> <ISOFILE>
</pre>
# Extract Data from .ISO files
<pre>
extract.exe <ISOFILE>
</pre>
If extracted 1ST_READ.BIN has filesize of 0, LBA of track05 is wrong.

=== Downsampling ===

==== AFS Files ====

https://sizious.com/download/dreamcast/

AFS files contain multiple ADX files combined into one AFS file.

compact AFS with:

<pre>
AFSCompactor.exe
</pre>

for Alone in the Dark - The New Nightmare .XA files are AFS files, rename first with:

<pre>
rename .XA .afs *.XA
</pre>

then compact and rename back to .XA:

<pre>
rename .afs .XA *.afs
</pre>

AFSCompactor.exe will rename files inside .afs to file0001 etc... AITD4 doesnt seem to care.

extract AFS with:

<pre>
afs_extract.exe *.afs|*.XA
</pre>
or
<pre>
afsextract.exe -sc afsfile|XAfile
</pre>
or
<pre>
afsutils_v2.exe | rename .XA files to .afs beforehand.
</pre>

==== SFD Movies ====

http://www.dc-swat.ru/page/download/

Demux video to .m1v and audio to .wav first

Example (play around with bitrates):

<pre>
for f in <SFD Files>; do ffmpeg -i $f -vcodec mpeg1video -b:v 2600k -minrate 2600k -maxrate 3600k -y $f.m1v -acodec adpcm_ima_wav -y $f.wav ; done
</pre>

Open all .wavs in single Audacity project and mass export again, 16bit and 44100 Hz.

Start Dreamcast Movie Converter and remux/reencode to .SFD. Dont use batch processing, videos get truncated!

==== 4XM Movies ====

http://samples.mplayerhq.hu/game-formats/4xm/

Convert to AVI first with ffmpeg

<pre>
for i in ../rips/unmodified/aitd4/cd2/AITD4_GD_2/MOVIES/*.4XM; do ffmpeg -i "$i" -vf scale=320x160 -y "${i%.*}.avi"; done
</pre>

Downsample with 4xm_comp.exe

Speed : 30 fps
Data Rate : 110000
Resize Width 320 Height 160

=== Create ISO Images ===

For the 1st Session, chosen files must not exceed total of 69120000 bytes!

==== 1st Session ====

===== Create first ISOFile =====

<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <CHOSEN FILE> <CHOSEN FILE> <CHOSEN FILE>...
</pre>

Fill first ISOFile with zeros so it takes up exactly 68808703 bytes = LBA 45000

<pre>
dd if=/dev/zero of=<OUTPUT1.ISO> bs=1 count=1 seek=68808703
</pre>

===== Burn 1st session =====

<pre>
cdrecord -v -multi -tao -xa <OUTPUT1.ISO>
</pre>

Check MSINFO value & TOC

<pre>
cdrecord -msinfo

0,45000
</pre>

<pre>
cdrecord -toc

first: 1 last 1
track: 1 lba: 0 ( 0) 00:02:00 adr: 1 control: 4 mode: -1
track:lout lba: 33600 ( 134400) 07:30:00 adr: 1 control: 4 mode: -1
</pre>

To include subdirectories into the root directory of the CD, use
<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <SUBDIRECTORY>/=<PATH_OF_SUBDIRECTORY>
</pre>
Example:
<pre>
mkisofs -no-allow-lowercase -graft-points -l -duplicates-once -V AITD4_GD_2 -o aitd4_d2_1.iso MOVIES/=../rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES/
</pre>

==== 2nd Session ====

===== Hack 1ST_READ.BIN & IP.BIN =====

Get MSINFO Value

<pre>
cdrecord -msinfo
</pre>

<pre>
binhack

Enter name of binary: 1ST_READ.BIN
Enter name of bootsector: ip.bin
Enter msinfo value: <MSINFO Value>
File 1ST_READ.BIN successfully hacked.
Bincon detected, OS flag set to 0.
File ip.bin successfully created.
</pre>

===== Insert custom Logo =====

<pre>
logoinsert darkrealm.mr IP.BIN
</pre>

===== Burn 2nd Session =====

<pre>
mkisofs -sort sort.txt -no-allow-lowercase -l -duplicates-once -V <DISCNAME> -G <IP.BIN> -M <OUTPUT1.ISO> -C <MSINFO VALUE> -o <OUTPUT2.ISO> <GAME_DATA_DIRECTORY>
cdrecord -v -tao -xa <OUTPUT2.ISO>
</pre>

== Testing ==

* Burn whole stuff to a CD-RW
* Generate CDI from CD-RW
** Launch DiscJuggler from emulated Windows Machine and connect via iSCSI
* Test CDI with ReiCast oder ReDream

== Create CDI Images from ISO ==

* Create ISO first

<pre>
mkisofs -l -duplicates-once -graft-points -V <DISCNAME> -G <IP.BIN> -o <GAME.ISO> <GAME DIRECTORY>
</pre>

* Create DATA/DATA CDI

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

* or create AUDIO/DATA CDI, with 2nd Session starting at LBA 11702
** Note: 1ST_READ.BIN must eventually be binhacked to LBA 11702

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

Multicast

2026-04-18T20:20:57Z

Chris: /* VXLAN Interface an Bridge attachen */

== Multicast üver VPN ==

=== Multicast über Wireguard (Level 3 / Routing) ===

Voraussetzungen:

* Multicast muss für das WG Interface aktiviert sein
* Über Kommanzozeile:

<pre>
ip l set WG multicast on
</pre>

* Über systemd-networkd:

WG.network
<pre>
[Link]
Multicast=true
</pre>

* Multicast-IP-Adressen müssen bei den AllowedIPs dabei sein:

WG.netdev
<pre>
AllowedIPs=224.0.0.0/4,ff00::/8
</pre>

Achtung: Wird im .netdev-File etwas geändert muss das Interface erst gelöscht und wieder erstellt werden damit die Änderungen greifen:

<pre>
networkctl delete WG
networkctl reload
</pre>

=== mDNS ===

Über Avahi können mDNS-Abfragen "reflektiert" (in andere Subnetze weitergeleitet) werden. Wenn es sich beim dem Interface um ein Point-to-Point Interface handelt (Wireguard-Interfaces sind Point-to-Point Interfaces) muss dies auch noch erlaubt werden.

/etc/avahi/avahi-daemon.conf:

<pre>
[server]
allow-point-to-point=yes # Bei Point-to-Point Interfaces nötig

[reflector]
enable-reflector=yes
</pre>

Netzwerk A möchte die mDNS-Geräte in Netzwerk B sehen:

reflector auf Host B aktivieren.

Netzwerk B möchte die mDNS-Geräte in Netzwerk A sehen:

reflector auf Host A aktivieren.

Damit die Geräte auch auf den Hosts hinter dem Gateway gesehen werden können muss auf dem jeweiligen Gateway auch der Reflector aktiviert werden.

=== SSDP / UPnP / DLNA ===

SSDP hat mit mDNS nichts zu tun, kann also auch nicht über den avahi-daemon reflector weitergeleitet werden. Es muss ein Multicast-Router verwendet werden, hier: smcroute

/etc/smcroute.conf:

<pre>
phyint WG enable
phyint BRIDGE_LAN enable

# SSDP
mgroup from BRIDGE_LAN group 239.255.255.250
mroute from BRIDGE_LAN group 239.255.255.250 to WG
mgroup from WG group 239.255.255.250
mroute from WG group 239.255.255.250 to BRIDGE_LAN
</pre>

Netzwerk A möchte die UPnP-Geräte in Netzwerk B sehen:

smcroute auf Host A starten.

Netzwerk B möchte die UPnP-Geräte in Netzwerk A sehen:

smcroute auf Host B starten.

=== Geräte hinter dem Router ===

Damit die Geräte hinter dem Router von den Multicast-Paketen erreicht werden können:

<pre>
ip r add 192.168.1.0/24 dev WG
</pre>

<pre>
ip r add 192.168.178.0/24 dev WG
</pre>

IPv4 Forwarding muss aktiviert sein, sowie forwarding in the Firewall muss akzeptiert werden!

=== Wireguard Debugging aktivieren ===

<pre>
echo module wireguard +p > /sys/kernel/debug/dynamic_debug/control
</pre>

=== Multicast über VXLAN (Level 2 / Bridging) ===

==== VXLAN Interface auf beiden Hosts erstellen ====

Auf Host A:

<pre>
ip link add vxlan0 type vxlan id 42 dev WG remote 172.31.0.2 dstport 4789
ip link set vxlan0 up
</pre>

Auf Host B:

<pre>
ip link add vxlan0 type vxlan id 42 dev WG remote 172.31.0.1 dstport 4789
ip link set vxlan0 up
</pre>

==== VXLAN Interface an Bridge attachen ====

Auf Host B:

<pre>
ip link add br-vxlan type bridge
ip link set br-vxlan up
ip link set vxlan0 master br-vxlan
</pre>

Auf Host A:

Dies attached das vxlan0-Interface an die BRIDGE_LAN-Bridge. Wenn jetzt ein Rechner auf Site B eine DHCP-Anfrage auf dem br-vxlan-Interface stellt bekommt das Interface eine IP aus dem Adressbereich von Site A, und zwar auf dem Interface wo das vxlan0-Interface attached ist, also BRIDGE_LAN.

<pre>
ip link set vxlan0 master BRIDGE_LAN
</pre>

Die MTU evtl. noch ein wenig heruntersetzen:

Auf Host A und B:

<pre>
ip link set vxlan0 mtu 1280
</pre>

Multicast

2026-04-18T19:26:23Z

Chris: /* Multicast über VXLAN (Level 2 / Bridging) */

== Multicast üver VPN ==

=== Multicast über Wireguard (Level 3 / Routing) ===

Voraussetzungen:

* Multicast muss für das WG Interface aktiviert sein
* Über Kommanzozeile:

<pre>
ip l set WG multicast on
</pre>

* Über systemd-networkd:

WG.network
<pre>
[Link]
Multicast=true
</pre>

* Multicast-IP-Adressen müssen bei den AllowedIPs dabei sein:

WG.netdev
<pre>
AllowedIPs=224.0.0.0/4,ff00::/8
</pre>

Achtung: Wird im .netdev-File etwas geändert muss das Interface erst gelöscht und wieder erstellt werden damit die Änderungen greifen:

<pre>
networkctl delete WG
networkctl reload
</pre>

=== mDNS ===

Über Avahi können mDNS-Abfragen "reflektiert" (in andere Subnetze weitergeleitet) werden. Wenn es sich beim dem Interface um ein Point-to-Point Interface handelt (Wireguard-Interfaces sind Point-to-Point Interfaces) muss dies auch noch erlaubt werden.

/etc/avahi/avahi-daemon.conf:

<pre>
[server]
allow-point-to-point=yes # Bei Point-to-Point Interfaces nötig

[reflector]
enable-reflector=yes
</pre>

Netzwerk A möchte die mDNS-Geräte in Netzwerk B sehen:

reflector auf Host B aktivieren.

Netzwerk B möchte die mDNS-Geräte in Netzwerk A sehen:

reflector auf Host A aktivieren.

Damit die Geräte auch auf den Hosts hinter dem Gateway gesehen werden können muss auf dem jeweiligen Gateway auch der Reflector aktiviert werden.

=== SSDP / UPnP / DLNA ===

SSDP hat mit mDNS nichts zu tun, kann also auch nicht über den avahi-daemon reflector weitergeleitet werden. Es muss ein Multicast-Router verwendet werden, hier: smcroute

/etc/smcroute.conf:

<pre>
phyint WG enable
phyint BRIDGE_LAN enable

# SSDP
mgroup from BRIDGE_LAN group 239.255.255.250
mroute from BRIDGE_LAN group 239.255.255.250 to WG
mgroup from WG group 239.255.255.250
mroute from WG group 239.255.255.250 to BRIDGE_LAN
</pre>

Netzwerk A möchte die UPnP-Geräte in Netzwerk B sehen:

smcroute auf Host A starten.

Netzwerk B möchte die UPnP-Geräte in Netzwerk A sehen:

smcroute auf Host B starten.

=== Geräte hinter dem Router ===

Damit die Geräte hinter dem Router von den Multicast-Paketen erreicht werden können:

<pre>
ip r add 192.168.1.0/24 dev WG
</pre>

<pre>
ip r add 192.168.178.0/24 dev WG
</pre>

IPv4 Forwarding muss aktiviert sein, sowie forwarding in the Firewall muss akzeptiert werden!

=== Wireguard Debugging aktivieren ===

<pre>
echo module wireguard +p > /sys/kernel/debug/dynamic_debug/control
</pre>

=== Multicast über VXLAN (Level 2 / Bridging) ===

==== VXLAN Interface auf beiden Hosts erstellen ====

Auf Host A:

<pre>
ip link add vxlan0 type vxlan id 42 dev WG remote 172.31.0.2 dstport 4789
ip link set vxlan0 up
</pre>

Auf Host B:

<pre>
ip link add vxlan0 type vxlan id 42 dev WG remote 172.31.0.1 dstport 4789
ip link set vxlan0 up
</pre>

==== VXLAN Interface an Bridge attachen ====

Auf Host B:

<pre>
ip link add br-vxlan type bridge
ip link set br-vxlan up
ip link set vxlan0 master br-vxlan
</pre>

Auf Host A:

Dies attached das vxlan0-Interface an die BRIDGE_LAN-Bridge. Wenn jetzt ein Rechner auf Site B eine DHCP-Anfrage auf dem br-vxlan-Interface stellt bekommt das Interface eine IP aus dem Adressbereich von Site A, und zwar auf dem Interface wo das vxlan0-Interface attached ist, also BRIDGE_LAN.

<pre>
ip link set vxlan0 master BRIDGE_LAN
</pre>

Die MTU noch ein wenig heruntersetzen:

Auf Host A und B:

<pre>
ip link set vxlan0 mtu 1280
</pre>

Multicast

2026-04-18T19:18:37Z

Chris:

Funktionierende Toolchains

2026-01-13T23:46:20Z

Chris:

== KallistiOS Toolchain ==

Voraussetzungen:

* Wenn die Toolchain innerhalb eines CIFS-Mounts installiert wird müssen '''mfsymlinks''' in den Mountoptions aktiviert sein!
* 2026-01-14: Nur die Crosscompiler auf shodan bauen, KOS und KOS-Ports auf cortexreaver

=== Crosscompiler bauen ===
<pre>
mkdir /mnt/storage/development/dreamcast/toolchains
cd /mnt/storage/development/dreamcast/toolchains
git clone https://github.com/KallistiOS/KallistiOS.git kos
cd kos/utils/dc-chain/
cp -av Makefile.default.cfg Makefile.cfg
</pre>

==== Makefile.cfg bearbeiten ====

<pre>
# Sega Dreamcast Toolchains Maker (dc-chain)
# This file is part of KallistiOS.

#########################
### TOOLCHAIN PROFILE ###
#########################

# Choose a toolchain profile from the following available options:
# No longer supported upstream:
# - 9.3.0-legacy: Former 'stable' option, based on GCC 9.3.0 and Newlib 3.3.0.
# - 9.5.0-winxp: Most recent versions of tools which run on Windows XP.
# - 10.5.0: Last release in the GCC 10 series, released 2023-07-07.
# - 11.5.0: Last release in the GCC 11 series, released 2024-07-19.
# Supported upstream:
# - 12.4.0: Latest release in the GCC 12 series, released 2024-06-20.
# - stable: Tested stable; based on GCC 13.2.0, released 2023-07-27.
# - 13.3.0: Latest release in the GCC 13 series, released 2024-05-21.
# - 14.2.0: Latest release in the GCC 14 series, released 2024-08-01.
# Development versions:
# - 13.3.1-dev Bleeding edge GCC 13 series from git.
# - 14.2.1-dev Bleeding edge GCC 14 series from git.
# - 15.0.0-dev Bleeding edge GCC 15 series from git.
# - gccrs-dev: GCC fork for development of the GCCRS Rust compiler.
# - rustc-dev: GCC fork for development of the libgccjit rustc GCC codegen.
# If unsure, select stable. See README.md for more detailed descriptions.
toolchain_profile=stable

########################
### DOWNLOAD OPTIONS ###
########################

### Download protocol (http|https|ftp)
# Specify the protocol you want to use for downloading package files.
download_protocol=https

### Force downloader (curl|wget)
# Specify here if you'd prefer to use 'wget' or 'curl'. If neither is specified,
# a web downloader tool will be auto-detected in the following order: cURL, Wget
# You must have either Wget or cURL installed to use dc-chain.
#force_downloader=wget

### Specify GNU mirror override
# The default mirror for GNU sources is 'ftpmirror.gnu.org'
# This setting overrides the default mirror with a preferred mirror.
#gnu_mirror=mirrors.kernel.org

#####################
### BUILD OPTIONS ###
#####################

### SH toolchain install path
# Specify the directory where the SH toolchain will be installed. This setting
# must match the KOS_CC_BASE setting in your KOS "environ.sh" configuration.
sh_toolchain_path=/media/chris/development/dreamcast/toolchains/sh-elf

### ARM toolchain install path
# Specify the directory where the ARM toolchain will be installed. This setting
# must match the DC_ARM_BASE setting in your KOS "environ.sh" configuration.
arm_toolchain_path=/media/chris/development/dreamcast/toolchains/arm-eabi

### Make jobs (n|<empty>)
# Set this value to the number of parallel jobs you want to run with make.
# For fastest toolchain building, set to the number of CPU threads available.
# Using multiple jobs may cause issues in certain environments and may be
# automatically disabled. If you encounter errors building your toolchain,
# reduce the number of jobs to 1 to avoid issues and ease troubleshooting.
makejobs=16

### Verbose (1|0)
# Choose whether to actively display compilation messages on the screen.
# Messages are saved to the build log files regardless of this setting.
verbose=1

### Erase (1|0)
# Erase build directories as toolchain components are installed to save space.
erase=1

### Install toolchain debug symbols (1|0)
# Choose whether to keep the debugging symbols for the toolchain.
# This is only useful if you wish to debug the toolchain itself.
#toolchain_debug=1

########################
### LANGUAGE OPTIONS ###
########################

### Enable C++
# Builds C++ support, including the C++ compiler and its standard library. The
# vast majority of language features are supported, with C++23 and early C++26
# support included. KallistiOS provides several examples with C++ support, so it
# is enabled by default. Adding C++ support requires extra disk space and
# compilation time, so you may disable it here if you do not plan on using C++.
enable_cpp=1

### Enable Objective-C
# Builds Objective-C support. Currently, this builds the Obj-C compiler and
# runtime, but does not build a standard library. The GNUStep Base Library is
# currently in active development for future inclusion in the kos-ports library.
# KallistiOS provides examples with Obj-C support, so it is enabled by default.
enable_objc=1

### Enable Objective C++
# Builds Objective C++ support. This hybrid language allows both C++ and Obj-C
# methods to be called from both contexts. Enabled by default as both the C++
# and Obj-C languages are already enabled by default.
enable_objcpp=1

### Enable D
# Builds D support. This will build the D compiler, but does not build the
# Phobos Runtime Library. D support may only be enabled on POSIX platforms which
# have its external dependencies provided through a host DMD, GDC, or LDC
# compiler installation.
#enable_d=1

### Enable Ada
# Builds Ada support. This will build the GNAT Ada compiler and tools but does
# not build the libada runtime library. In order for this build to succeed, the
# host's GCC version must match the target Dreamcast version, plus the same
# version of GNAT and its tools must be preinstalled (and in your path).
#enable_ada=1

### Enable Rust
# Builds the work-in-progress GCCRS Rust compiler frontend for GCC. Requires a
# development version of GCC. Use with the gccrs-dev toolchain profile for the
# most recent updates to the GCCRS compiler. This option is not to be confused
# with using this toolchain as a backend for rustc; see libgccjit below.
#enable_rust=1

### Enable libgccjit
# Enables the libgccjit embeddable GCC library. This is most commonly enabled to
# allow using this toolchain as a backend for the official rustc compiler. Use
# with the rustc-dev toolchain profile if this is your use case.
#enable_libgccjit=1

###################
### GCC OPTIONS ###
###################

### Floating point precision support (m4-single-only|m4-single|m4)
# Build support for various SH4 floating-point operation ABIs. KallistiOS only
# officially supports single-precision-only mode (m4-single-only); however,
# experimental support for single-precision-default mode (m4-single) has been
# added to allow for the use of full 64-bit doubles. You may also include
# double-precision-default (m4) which is untested and unsupported, or you may
# choose to specify only m4-single-only to save disk space.
precision_modes=m4-single-only,m4-single

### Default floating point mode (m4|m4-single|m4-single-only)
# Choose the default floating point precision ABI used when GCC is invoked. This
# may be overridden by passing -m4-single-only, -m4-single, or -m4 to GCC.
# KOS currently only officially supports m4-single-only, so it is the default.
default_precision=m4-single-only

### GCC threading model (single|kos)
# KallistiOS patches to GCC provide a 'kos' thread model, which should be used.
# If you want to disable threading support for C++, Objective-C, and so forth,
# you can set this option to 'single'.
thread_model=kos

### Automatic patching for KallistiOS (1|0)
# Uncomment this option if you want to disable applying KallistiOS patches to
# toolchain source files before building. This will disable usage of the 'kos'
# thread model. Only do this if you understand what you are doing.
#use_kos_patches=0

### Disable GCC Native Language Support (1|0)
# By default, NLS allows GCC to output diagnostics in non-English languages.
# Uncomment this option to disable NLS and force GCC to output in English.
#disable_nls=1

######################
### NEWLIB OPTIONS ###
######################

### Automatic patching for Newlib (1|0)
# Uncomment this option if you want to disable the automatic patching of Newlib
# needed by KallistiOS. This will keep the generated toolchain completely raw.
# This will also disable the 'kos' thread model. Only do this if you understand
# what you are doing.
#auto_fixup_sh4_newlib=0

### C99 format specifier support (1|0)
# Define this to build Newlib with additional support for the C99 format
# specifiers, used by printf and friends. These include support for size_t,
# ptrdiff_t, intmax_t, and sized integral types.
newlib_c99_formats=1

### Multibyte character set support (1|0)
# Define this to build Newlib with additional multibyte support. This enables
# three special locales: "C-JIS", "C-SJIS", and "C-EUCJP". The multibyte
# support extends to the mb and wc functions in stdlib as well as format
# characters for the printf and scanf family of routines.
#newlib_multibyte=1

### iconv() character encoding conversions support (encoding list)
# Define a list here to enable support for the iconv() function and <iconv.h>
# header file. The given comma separated list defines for which encoding types
# to include bidirectional conversion support. For the full list of available
# encodings, see the Newlib configuration documentation.
#newlib_iconv_encodings=us_ascii,utf8,utf16,ucs_2_internal,ucs_4_internal

### Optimize Newlib for space (1|0)
# Uncomment this option to optimize for space when building Newlib. This will
# build Newlib with compiler flags which favor smaller code sizes over faster
# performance.
#newlib_opt_space=1

###################
### C++ OPTIONS ###
###################

### Timezone database support (1|0|path)
# Uncomment this option to enable building support for C++'s std::chrono::tzdb
# into your C++ standard library by using a combination of the system's local
# timezone DB and one dynamically fetched from the "IANA Time Zone Database."
# Without support enabled, only the "UTC" and "GMT" timezones will be defined.
# You can optionally provide the path to a directory with a custom "tzdata.zi"
# database file. NOTE: Enabling this will result in larger C++ binaries!
#libstdcxx_tzdb=1

#######################
### WINDOWS OPTIONS ###
#######################

### MinGW/MSYS
# Standalone binaries (1|0)
# Uncomment this option if you want static binaries that are standalone and
# require no dependencies. When this option is used, binaries can be run outside
# the MinGW/MSYS environment. This is NOT recommended; only do this if you know
# what you are doing.
#standalone_binary=1

### Force installation of BFD for SH (1|0)
# Uncomment this option if you want to force installation of 'libbfd' for the SH
# toolchain. This is required for MinGW/MSYS and can't be disabled in this
# scenario. This option is here mainly if you want to force the installation of
# 'libbfd' under other environments, but this won't be necessary in most cases,
# as 'libelf' is used almost everywhere. Please note, 'libbfd' is not portable
# if you built it on another environment. Only use this flag if you understand
# what you are doing.
#sh_force_libbfd_installation=1
</pre>

Make durchführen:
<pre>
PATH=/usr/bin make all
</pre>

==== KOS bauen ====

eviron.sh:
<pre>
# KallistiOS Environment Settings
#
# This is a sample script for configuring and customizing your
# KOS build environment. Modify it to suit your setup. Several
# settings may be enabled optionally or are provided with
# alternative values.
#
# This script is typically sourced in your current shell environment
# (probably by .bashrc, .bash_profile, or something similar), so that
# the KOS environment is set up automatically for each shell session.
#

# Build Architecture
#
# Set the major architecture you'll be building for.
# The only option here is "dreamcast" as of KOS 2.0.0.
#
export KOS_ARCH="dreamcast"

# Build Sub-Architecture
#
# Defines the sub architecture your configuration
# is targeting or uses an existing value that
# can be set externally via your IDE.
#
# Valid values:
# "pristine" - Dreamcast console or HKT-0120 devkit (default)
# "naomi" - NAOMI or NAOMI 2 arcade board
#
if [ -z "${KOS_SUBARCH}" ] ; then
export KOS_SUBARCH="pristine"
else
export KOS_SUBARCH
fi

# KOS Root Path
#
# Specifies the path to the KOS root directory
#
export KOS_BASE="/media/chris/development/dreamcast/toolchains/kos"

# KOS-Ports Path
#
# Specifies the path to the KOS-ports directory
#
export KOS_PORTS="${KOS_BASE}/../kos-ports"

# SH Compiler Prefixes
#
# Specifies the path to and prefix for the main SH
# GCC toolchain used to target the Dreamcast's SH4 CPU.
#
export KOS_CC_BASE="/media/chris/development/dreamcast/toolchains/sh-elf"
export KOS_CC_PREFIX="sh-elf"

# ARM Compiler Prefixes
#
# Specifies the path to and prefix for the additional ARM
# GCC toolchain used to target the Dreamcast's AICA SPU.
#
export DC_ARM_BASE="/media/chris/development/dreamcast/toolchains/arm-eabi"
export DC_ARM_PREFIX="arm-eabi"

# External Dreamcast Tools Path
#
# Specifies the path where Dreamcast tools that are not part of KOS are to be
# installed. This includes, for example, dc-tool-ip, dc-tool-serial, and the
# mrbc bytecode compiler. This directory, along with SH and ARM compiler
# toolchains, will be added to your PATH environment variable.
#
export DC_TOOLS_BASE="/media/chris/development/dreamcast/tools"

# CMake Toolchain Path
#
# Specifies the path to the toolchain file used to target
# KOS with the "cmake" build utility.
#
export KOS_CMAKE_TOOLCHAIN="${KOS_BASE}/utils/cmake/dreamcast.toolchain.cmake"

# Genromfs Utility Path
#
# Specifies the path to the utility which is used by KOS
# to create romdisk filesystem images.
#
export KOS_GENROMFS="${KOS_BASE}/utils/genromfs/genromfs"

# Make Utility
#
# Configures the tool to be used as the main "make" utility
# for building GNU Makefiles. On a platform such as BSD,
# the default can be changed to "gmake," for the GNU
# implementation.
#
export KOS_MAKE="make"
#export KOS_MAKE="gmake"

# Loader Utility
#
# Specifies the loader to be used with the "make run" targets
# in the KOS examples. Defaults to using a preconfigured version
# of dc-tool. Use one of the other options for a manual dc-tool-ip
# or dc-tool-serial configuration, remembering to change the values
# for the Dreamcast's IP address or the serial port interface.
#
#export KOS_LOADER="dc-tool -x"
export KOS_LOADER="dc-tool-ip -t dreamcast -x"
#export KOS_LOADER="dc-tool-ser -t /dev/ttyS0 -x"

# Default Compiler Flags
#
# Resets build flags. You can also initialize them to some preset
# default values here if you wish.
#
export KOS_INC_PATHS=""
export KOS_CFLAGS=""
export KOS_CPPFLAGS=""
export KOS_LDFLAGS=""
export KOS_AFLAGS=""
export DC_ARM_LDFLAGS=""

# Debug Builds
#
# Controls whether to disable additional debugging checks and assertions,
# such as for parameter validation or internal errors. Uncomment this if
# you do not wish to compile with this additional logic enabled.
#
#export KOS_CFLAGS="${KOS_CFLAGS} -DNDEBUG"

# Optimization Level
#
# Controls the baseline optimization level to use when building.
# Typically this is -Og (debugging), -O0, -O1, -O2, or -O3.
# NOTE: For our target, -O4 is a valid optimization level that has
# been seen to have some performance impact as well.
#
export KOS_CFLAGS="${KOS_CFLAGS} -O2"

# Additional Optimizations
#
# Uncomment this to enable what has been found empirically to be
# the optimal set of additional flags for release build performance
# on the current stable toolchain. NOTE: Certain KOS-ports and examples
# do not work properly with "-flto=auto"!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -freorder-blocks-algorithm=simple -flto=auto"

# Frame Pointers
#
# Controls whether frame pointers are emitted or not. Disabled by
# default. Enable them if you plan to use GDB for debugging.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fomit-frame-pointer"
#export KOS_CFLAGS="${KOS_CFLAGS} -fno-omit-frame-pointer -DFRAME_POINTERS"

# Stack Protector
#
# Controls whether GCC emits extra code to check for buffer overflows or stack
# smashing, which can be very useful for debugging. -fstack-protector only
# covers vulnerable objects, while -fstack-protector-strong provides medium
# coverage, and -fstack-protector-all provides full coverage. You may also
# override the default stack excepton handler by providing your own
# implementation of "void __stack_chk_fail(void)."
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fstack-protector-all"

# GCC Builtin Functions
#
# Comment out this line to enable GCC to use its own builtin implementations of
# certain standard library functions. Under certain conditions, this can allow
# compiler-optimized implementations to replace standard function invocations.
# The downside of this is that it COULD interfere with Newlib or KOS implementations
# of these functions, and it has not been tested thoroughly to ensure compatibility.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fno-builtin"

# Fast Math Instructions
#
# Uncomment this line to enable the optimized fast-math instructions (FSSRA,
# FSCA, and FSQRT) for calculating sin/cos, inverse square root, and square roots.
# These can result in substantial performance gains for these kinds of operations;
# however, they do so at the price of accuracy and are not IEEE compliant.
# NOTE: Enabling this option will also override -fno-builtin!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fbuiltin -ffast-math -ffp-contract=fast"

# SH4 Floating-Point Arithmetic Precision
#
# KallistiOS supports both the single-precision-default ABI (m4-single) and the
# single-precision-only ABI (m4-single-only). When using m4-single, the SH4 will
# be in single-precision mode upon function entry but will switch to double-
# precision mode if 64-bit doubles are used. When using m4-single-only, the SH4
# will always be in single-precision mode and 64-bit doubles will be truncated to
# 32-bit floats. Historically, m4-single-only was used in both official and
# homebrew Dreamcast software, but m4-single is the default as of KOS 2.2.0 to
# increase compatibility with newer libraries which require 64-bit doubles.
#
# WARNING: When adjusting this setting, make sure all software, including
# kos-ports and linked external libraries, are rebuilt using the same
# floating-point precision ABI setting!
#
export KOS_SH4_PRECISION="-m4-single"

# Use LRA (Local Register Allocator) Pass
#
# Uncomment this line to use the modern Local Register Allocator pass during
# code generation instead of the default older reload pass. This option is
# known to be unstable or less performant for SH at this time, but will likely
# become mandatory in future versions of GCC, so feel free to help us test.
# Only enable this setting if you understand what you are doing!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -mlra"

# Shared Compiler Configuration
#
# Include sub architecture-independent configuration file for shared
# environment settings. If you want to configure additional compiler
# options or see where other build flags are set, look at this file.
#
. ${KOS_BASE}/environ_base.sh
</pre>

<pre>
. environ.sh
make
</pre>

=== KOS-ports bauen ===

<pre>
git clone --recursive https://github.com/KallistiOS/kos-ports kos-ports
kos-ports/utils/build-all.sh
</pre>

Miniupnpd

2025-12-03T20:15:46Z

Chris:

== miniupnpd compilen ==

* git clone --recurse-submodules https://github.com/miniupnp/miniupnp.git
* ./configure --ipv6 --leasefile --portinuse --firewall=nftables --systemd
* make all
* sudo make install
* config anpassen:

/etc/miniupnpd/miniupnpd.conf:

<pre>
# miniupnpd.conf: MiniUPnP IGD & PCP/NAT-PMP daemon configuration file
# More information at https://miniupnp.tuxfamily.org or http://miniupnp.free.fr

# CAUTION: mixing up WAN and LAN interfaces may introduce security risks!
# Be sure to assign the correct interfaces to LAN and WAN and consider
# implementing UPnP permission rules at the bottom of this configuration file

# WAN network interface
ext_ifname=WAN

# if the WAN network interface for IPv6 is different than for IPv4,
# set ext_ifname6
#ext_ifname6=eth2

# If the WAN interface has several IP addresses, you
# can specify the one to use below.
# Setting ext_ip is also useful in double NAT setup, you can declare here
# the public IP address.
#ext_ip=192.168.3.251
#ext_ip=146.52.237.239

# The WAN interface must have a public IP address. Otherwise it is behind NAT
# and port forwarding is impossible. In some cases WAN interface can be
# behind unrestricted full-cone NAT 1:1 when all incoming traffic is NAT-ed and
# routed to WAN interfaces without any filtering. In this cases miniupnpd
# needs to know the public IP address and it can be learnt by asking external
# server via STUN protocol. Following option enable retrieving external
# public IP address from STUN server and detection of NAT type. You need
# to specify also external STUN server in stun_host option below.
# This option is disabled by default.
ext_perform_stun=yes
#ext_perform_stun=allow-filtered

# Specify STUN server, either hostname or IP address
# Some public STUN servers:
# stunserver.stunprotocol.org
# stun.sipgate.net
# stun.xten.com
# stun.l.google.com (on non standard port 19302)
#ext_stun_host=stunserver.stunprotocol.org
ext_stun_host=stun.xten.com
# Specify STUN UDP port, by default it is standard port 3478.
#ext_stun_port=3478

# LAN network interfaces IPs / networks
# There can be multiple listening IPs for SSDP traffic, in that case
# use multiple 'listening_ip=...' lines, one for each network interface.
# It can be IP address or network interface name (ie. "eth0")
# It is mandatory to use the network interface name in order to enable IPv6
# HTTP is available on all interfaces.
# When MULTIPLE_EXTERNAL_IP is enabled, the external IP
# address associated with the subnet follows. For example:
# listening_ip=192.168.0.1/24 88.22.44.13
# When MULTIPLE_EXTERNAL_IP is disabled, you can list associated network
# interfaces (for bridges)
# listening_ip=bridge0 em0 wlan0
#listening_ip=192.168.0.1/24
#listening_ip=10.5.0.0/16
listening_ip=BRIDGE_LAN

# Port for HTTP (descriptions and SOAP) traffic. Set to 0 for autoselect.
#http_port=0

# Port for HTTPS. Set to 0 for autoselect (default)
#https_port=0

# Path to the UNIX socket used to communicate with MiniSSDPd
# If running, MiniSSDPd will manage M-SEARCH answering.
# default is /var/run/minissdpd.sock
#minissdpdsocket=/var/run/minissdpd.sock

# Disable IPv6 (default is no : IPv6 enabled if enabled at build time)
#ipv6_disable=yes

# Enable NAT-PMP and PCP support (default is no)
enable_pcp_pmp=yes

# Enable UPNP support (default is yes)
enable_upnp=yes

# PCP
# Configure the minimum and maximum lifetime of a port mapping in seconds
# 120s and 86400s (24h) are suggested values from PCP-base
#min_lifetime=120
#max_lifetime=86400
# allow THIRD_PARTY Option for MAP and PEER Opcodes (default is no)
#pcp_allow_thirdparty=yes

# table names for netfilter nft. Default is "filter" for both
#upnp_table_name=filter
#upnp_nat_table_name=filter
# chain names for netfilter and netfilter nft
# netfilter : default are MINIUPNPD, MINIUPNPD, MINIUPNPD-POSTROUTING
# netfilter nft : default are miniupnpd, prerouting_miniupnpd, postrouting_miniupnpd
#upnp_forward_chain=forward
#upnp_nat_chain=nat
#upnp_nat_postrouting_chain=postrouting
upnp_nftables_family_split=yes

# Lease file location
#lease_file=/var/log/upnp.leases

# To enable the next few runtime options, see compile time
# ENABLE_MANUFACTURER_INFO_CONFIGURATION (config.h)

# Name of this service, default is "`uname -s` router"
#friendly_name=MiniUPnPd router

# Manufacturer name, default is "`uname -s`"
#manufacturer_name=Manufacturer corp

# Manufacturer URL, default is URL of OS vendor
#manufacturer_url=https://miniupnp.tuxfamily.org/

# Model name, default is "`uname -s` router"
#model_name=Router Model

# Model description, default is "`uname -s` router"
#model_description=Very Secure Router - Model

# Model URL, default is URL of OS vendor
#model_url=https://miniupnp.tuxfamily.org/

# Bitrates reported by daemon in bits per second
# by default miniupnpd tries to get WAN interface speed
#bitrate_up=1000000
#bitrate_down=10000000

# Secure Mode, UPnP clients can only add mappings to their own IP
# Enabled by default
#secure_mode=no

# Default presentation URL is HTTP address on port 80
# If set to an empty string, no presentationURL element will appear
# in the XML description of the device, which prevents MS Windows
# from displaying an icon in the "Network Connections" panel.
#presentation_url=http://www.mylan/index.php

# Report system uptime instead of daemon uptime
system_uptime=yes

# Notify interval in seconds. default is 900 seconds.
# As advised in the standard, announcement have a lifetime double of this value.
notify_interval=900

# Unused rules cleaning.
# never remove any rule before this threshold for the number
# of redirections is exceeded. default to 20
#clean_ruleset_threshold=10
# Clean process work interval in seconds. default to 0 (disabled).
# a 600 seconds (10 minutes) interval makes sense
clean_ruleset_interval=600

# Log packets in pf (default is no)
#packet_log=yes

# Anchor name in pf (default is miniupnpd)
#anchor=miniupnpd

# ALTQ queue in pf
# Filter rules must be used for this to be used.
# compile with PF_ENABLE_FILTER_RULES (see config.h file)
#queue=queue_name1

# Tag name in pf
#tag=tag_name1

# Make filter rules in pf quick or not. default is yes
# active when compiled with PF_ENABLE_FILTER_RULES (see config.h file)
#quickrules=no

# UUID, generate your own UUID with "make genuuid"
uuid=00000000-0000-0000-0000-000000000000

# Daemon's serial and model number when reporting to clients
# (in XML description)
#serial=12345678
#model_number=1

# If compiled with IGD_V2 defined, force reporting IGDv1 in rootDesc (default
# is no)
#force_igd_desc_v1=no

# UPnP permission rules (also enforced for NAT-PMP and PCP) for IPv4
# (allow|deny) (external port range) IP/mask (internal port range) (optional regex filter)
# A port range is <min port>-<max port> or <port> if there is only
# one port in the range.
# IP/mask format must be nnn.nnn.nnn.nnn/nn
# Regex support must be enabled at build time : ./configure --regex
# It is advised to only allow redirection of ports >= 1024
# and end the rule set with "deny 0-65535 0.0.0.0/0 0-65535"
# The following default ruleset allows specific LAN side IP addresses
# to request only ephemeral ports. It is recommended that users
# modify the IP ranges to match their own internal networks, and
# also consider implementing network-specific restrictions
# CAUTION: failure to enforce any rules may permit insecure requests to be made!
allow 1024-65535 192.168.1.0/24 1024-65535
# disallow requests whose description string matches the given regex
# deny 1024-65535 192.168.1.0/24 1024-65535 "My evil app ver [[:digit:]]*"
deny 0-65535 0.0.0.0/0 0-65535
</pre>

* miniupnpd_functions.sh-Script nach /etc/miniupnpd/ kopieren (fehlt bei nftables-Variante im Makefile):
** sudo cp netfilter_nft/scripts/miniupnpd_functions.sh /etc/miniupnpd/
* evtl. editieren: af=inet in af=ip ändern
* systemctl daemon-reload
* systemctl start miniupnpd

Damit STUN funktioniert folgende iptables/nft Regel einfügen:

* nftables:
** add rule ip filter input meta l4proto udp @th,96,32 0x2112A442 counter accept

* iptables:
** iptables -A INPUT -i WAN -p udp -m u32 --u32 "32=0x2112A442" -j ACCEPT

Miniupnpd

2025-12-03T20:00:55Z

Chris:

Miniupnpd

2025-12-03T19:01:21Z

Chris:

== miniupnpd compilen ==

* git clone --recurse-submodules https://github.com/miniupnp/miniupnp.git
* ./configure --ipv6 --leasefile --portinuse --firewall=nftables --systemd
* make all
* sudo make install
* config anpassen:

/etc/miniupnpd/miniupnpd.conf:

<pre>
# miniupnpd.conf: MiniUPnP IGD & PCP/NAT-PMP daemon configuration file
# More information at https://miniupnp.tuxfamily.org or http://miniupnp.free.fr

# CAUTION: mixing up WAN and LAN interfaces may introduce security risks!
# Be sure to assign the correct interfaces to LAN and WAN and consider
# implementing UPnP permission rules at the bottom of this configuration file

# WAN network interface
ext_ifname=WAN

# if the WAN network interface for IPv6 is different than for IPv4,
# set ext_ifname6
#ext_ifname6=eth2

# If the WAN interface has several IP addresses, you
# can specify the one to use below.
# Setting ext_ip is also useful in double NAT setup, you can declare here
# the public IP address.
#ext_ip=192.168.3.251
#ext_ip=146.52.237.239

# The WAN interface must have a public IP address. Otherwise it is behind NAT
# and port forwarding is impossible. In some cases WAN interface can be
# behind unrestricted full-cone NAT 1:1 when all incoming traffic is NAT-ed and
# routed to WAN interfaces without any filtering. In this cases miniupnpd
# needs to know the public IP address and it can be learnt by asking external
# server via STUN protocol. Following option enable retrieving external
# public IP address from STUN server and detection of NAT type. You need
# to specify also external STUN server in stun_host option below.
# This option is disabled by default.
ext_perform_stun=yes
#ext_perform_stun=allow-filtered

# Specify STUN server, either hostname or IP address
# Some public STUN servers:
# stunserver.stunprotocol.org
# stun.sipgate.net
# stun.xten.com
# stun.l.google.com (on non standard port 19302)
#ext_stun_host=stunserver.stunprotocol.org
ext_stun_host=stun.xten.com
# Specify STUN UDP port, by default it is standard port 3478.
#ext_stun_port=3478

# LAN network interfaces IPs / networks
# There can be multiple listening IPs for SSDP traffic, in that case
# use multiple 'listening_ip=...' lines, one for each network interface.
# It can be IP address or network interface name (ie. "eth0")
# It is mandatory to use the network interface name in order to enable IPv6
# HTTP is available on all interfaces.
# When MULTIPLE_EXTERNAL_IP is enabled, the external IP
# address associated with the subnet follows. For example:
# listening_ip=192.168.0.1/24 88.22.44.13
# When MULTIPLE_EXTERNAL_IP is disabled, you can list associated network
# interfaces (for bridges)
# listening_ip=bridge0 em0 wlan0
#listening_ip=192.168.0.1/24
#listening_ip=10.5.0.0/16
listening_ip=BRIDGE_LAN

# Port for HTTP (descriptions and SOAP) traffic. Set to 0 for autoselect.
#http_port=0

# Port for HTTPS. Set to 0 for autoselect (default)
#https_port=0

# Path to the UNIX socket used to communicate with MiniSSDPd
# If running, MiniSSDPd will manage M-SEARCH answering.
# default is /var/run/minissdpd.sock
#minissdpdsocket=/var/run/minissdpd.sock

# Disable IPv6 (default is no : IPv6 enabled if enabled at build time)
#ipv6_disable=yes

# Enable NAT-PMP and PCP support (default is no)
enable_pcp_pmp=yes

# Enable UPNP support (default is yes)
enable_upnp=yes

# PCP
# Configure the minimum and maximum lifetime of a port mapping in seconds
# 120s and 86400s (24h) are suggested values from PCP-base
#min_lifetime=120
#max_lifetime=86400
# allow THIRD_PARTY Option for MAP and PEER Opcodes (default is no)
#pcp_allow_thirdparty=yes

# table names for netfilter nft. Default is "filter" for both
#upnp_table_name=filter
#upnp_nat_table_name=filter
# chain names for netfilter and netfilter nft
# netfilter : default are MINIUPNPD, MINIUPNPD, MINIUPNPD-POSTROUTING
# netfilter nft : default are miniupnpd, prerouting_miniupnpd, postrouting_miniupnpd
#upnp_forward_chain=forward
#upnp_nat_chain=nat
#upnp_nat_postrouting_chain=postrouting
upnp_nftables_family_split=yes

# Lease file location
#lease_file=/var/log/upnp.leases

# To enable the next few runtime options, see compile time
# ENABLE_MANUFACTURER_INFO_CONFIGURATION (config.h)

# Name of this service, default is "`uname -s` router"
#friendly_name=MiniUPnPd router

# Manufacturer name, default is "`uname -s`"
#manufacturer_name=Manufacturer corp

# Manufacturer URL, default is URL of OS vendor
#manufacturer_url=https://miniupnp.tuxfamily.org/

# Model name, default is "`uname -s` router"
#model_name=Router Model

# Model description, default is "`uname -s` router"
#model_description=Very Secure Router - Model

# Model URL, default is URL of OS vendor
#model_url=https://miniupnp.tuxfamily.org/

# Bitrates reported by daemon in bits per second
# by default miniupnpd tries to get WAN interface speed
#bitrate_up=1000000
#bitrate_down=10000000

# Secure Mode, UPnP clients can only add mappings to their own IP
# Enabled by default
#secure_mode=no

# Default presentation URL is HTTP address on port 80
# If set to an empty string, no presentationURL element will appear
# in the XML description of the device, which prevents MS Windows
# from displaying an icon in the "Network Connections" panel.
#presentation_url=http://www.mylan/index.php

# Report system uptime instead of daemon uptime
system_uptime=yes

# Notify interval in seconds. default is 900 seconds.
# As advised in the standard, announcement have a lifetime double of this value.
notify_interval=900

# Unused rules cleaning.
# never remove any rule before this threshold for the number
# of redirections is exceeded. default to 20
#clean_ruleset_threshold=10
# Clean process work interval in seconds. default to 0 (disabled).
# a 600 seconds (10 minutes) interval makes sense
clean_ruleset_interval=600

# Log packets in pf (default is no)
#packet_log=yes

# Anchor name in pf (default is miniupnpd)
#anchor=miniupnpd

# ALTQ queue in pf
# Filter rules must be used for this to be used.
# compile with PF_ENABLE_FILTER_RULES (see config.h file)
#queue=queue_name1

# Tag name in pf
#tag=tag_name1

# Make filter rules in pf quick or not. default is yes
# active when compiled with PF_ENABLE_FILTER_RULES (see config.h file)
#quickrules=no

# UUID, generate your own UUID with "make genuuid"
uuid=00000000-0000-0000-0000-000000000000

# Daemon's serial and model number when reporting to clients
# (in XML description)
#serial=12345678
#model_number=1

# If compiled with IGD_V2 defined, force reporting IGDv1 in rootDesc (default
# is no)
#force_igd_desc_v1=no

# UPnP permission rules (also enforced for NAT-PMP and PCP) for IPv4
# (allow|deny) (external port range) IP/mask (internal port range) (optional regex filter)
# A port range is <min port>-<max port> or <port> if there is only
# one port in the range.
# IP/mask format must be nnn.nnn.nnn.nnn/nn
# Regex support must be enabled at build time : ./configure --regex
# It is advised to only allow redirection of ports >= 1024
# and end the rule set with "deny 0-65535 0.0.0.0/0 0-65535"
# The following default ruleset allows specific LAN side IP addresses
# to request only ephemeral ports. It is recommended that users
# modify the IP ranges to match their own internal networks, and
# also consider implementing network-specific restrictions
# CAUTION: failure to enforce any rules may permit insecure requests to be made!
allow 1024-65535 192.168.1.0/24 1024-65535
# disallow requests whose description string matches the given regex
# deny 1024-65535 192.168.1.0/24 1024-65535 "My evil app ver [[:digit:]]*"
deny 0-65535 0.0.0.0/0 0-65535
</pre>

* miniupnpd_functions.sh-Script nach /etc/miniupnpd/ kopieren (fehlt bei nftables-Variante im Makefile):
** sudo cp netfilter_nft/scripts/miniupnpd_functions.sh /etc/miniupnpd/
* systemctl daemon-reload
* systemctl start miniupnpd

Miniupnpd

2025-12-03T18:52:40Z

Chris:

Miniupnpd

2025-12-03T18:41:32Z

Chris: Die Seite wurde neu angelegt: „/etc/miniupnpd/miniupnpd.conf: <pre> # miniupnpd.conf: MiniUPnP IGD & PCP/NAT-PMP daemon configuration file # More information at https://miniupnp.tuxfamily.org or http://miniupnp.free.fr # CAUTION: mixing up WAN and LAN interfaces may introduce security risks! # Be sure to assign the correct interfaces to LAN and WAN and consider # implementing UPnP permission rules at the bottom of this configuration file # WAN network interface ext_ifname=WAN # if…“

/etc/miniupnpd/miniupnpd.conf:

<pre>
# miniupnpd.conf: MiniUPnP IGD & PCP/NAT-PMP daemon configuration file
# More information at https://miniupnp.tuxfamily.org or http://miniupnp.free.fr

# CAUTION: mixing up WAN and LAN interfaces may introduce security risks!
# Be sure to assign the correct interfaces to LAN and WAN and consider
# implementing UPnP permission rules at the bottom of this configuration file

# WAN network interface
ext_ifname=WAN

# if the WAN network interface for IPv6 is different than for IPv4,
# set ext_ifname6
#ext_ifname6=eth2

# If the WAN interface has several IP addresses, you
# can specify the one to use below.
# Setting ext_ip is also useful in double NAT setup, you can declare here
# the public IP address.
#ext_ip=192.168.3.251
#ext_ip=146.52.237.239

# The WAN interface must have a public IP address. Otherwise it is behind NAT
# and port forwarding is impossible. In some cases WAN interface can be
# behind unrestricted full-cone NAT 1:1 when all incoming traffic is NAT-ed and
# routed to WAN interfaces without any filtering. In this cases miniupnpd
# needs to know the public IP address and it can be learnt by asking external
# server via STUN protocol. Following option enable retrieving external
# public IP address from STUN server and detection of NAT type. You need
# to specify also external STUN server in stun_host option below.
# This option is disabled by default.
ext_perform_stun=yes
#ext_perform_stun=allow-filtered

# Specify STUN server, either hostname or IP address
# Some public STUN servers:
# stunserver.stunprotocol.org
# stun.sipgate.net
# stun.xten.com
# stun.l.google.com (on non standard port 19302)
#ext_stun_host=stunserver.stunprotocol.org
ext_stun_host=stun.xten.com
# Specify STUN UDP port, by default it is standard port 3478.
#ext_stun_port=3478

# LAN network interfaces IPs / networks
# There can be multiple listening IPs for SSDP traffic, in that case
# use multiple 'listening_ip=...' lines, one for each network interface.
# It can be IP address or network interface name (ie. "eth0")
# It is mandatory to use the network interface name in order to enable IPv6
# HTTP is available on all interfaces.
# When MULTIPLE_EXTERNAL_IP is enabled, the external IP
# address associated with the subnet follows. For example:
# listening_ip=192.168.0.1/24 88.22.44.13
# When MULTIPLE_EXTERNAL_IP is disabled, you can list associated network
# interfaces (for bridges)
# listening_ip=bridge0 em0 wlan0
#listening_ip=192.168.0.1/24
#listening_ip=10.5.0.0/16
listening_ip=BRIDGE_LAN

# Port for HTTP (descriptions and SOAP) traffic. Set to 0 for autoselect.
#http_port=0

# Port for HTTPS. Set to 0 for autoselect (default)
#https_port=0

# Path to the UNIX socket used to communicate with MiniSSDPd
# If running, MiniSSDPd will manage M-SEARCH answering.
# default is /var/run/minissdpd.sock
#minissdpdsocket=/var/run/minissdpd.sock

# Disable IPv6 (default is no : IPv6 enabled if enabled at build time)
#ipv6_disable=yes

# Enable NAT-PMP and PCP support (default is no)
enable_pcp_pmp=yes

# Enable UPNP support (default is yes)
enable_upnp=yes

# PCP
# Configure the minimum and maximum lifetime of a port mapping in seconds
# 120s and 86400s (24h) are suggested values from PCP-base
#min_lifetime=120
#max_lifetime=86400
# allow THIRD_PARTY Option for MAP and PEER Opcodes (default is no)
#pcp_allow_thirdparty=yes

# table names for netfilter nft. Default is "filter" for both
#upnp_table_name=filter
#upnp_nat_table_name=filter
# chain names for netfilter and netfilter nft
# netfilter : default are MINIUPNPD, MINIUPNPD, MINIUPNPD-POSTROUTING
# netfilter nft : default are miniupnpd, prerouting_miniupnpd, postrouting_miniupnpd
#upnp_forward_chain=forward
#upnp_nat_chain=nat
#upnp_nat_postrouting_chain=postrouting
upnp_nftables_family_split=yes

# Lease file location
#lease_file=/var/log/upnp.leases

# To enable the next few runtime options, see compile time
# ENABLE_MANUFACTURER_INFO_CONFIGURATION (config.h)

# Name of this service, default is "`uname -s` router"
#friendly_name=MiniUPnPd router

# Manufacturer name, default is "`uname -s`"
#manufacturer_name=Manufacturer corp

# Manufacturer URL, default is URL of OS vendor
#manufacturer_url=https://miniupnp.tuxfamily.org/

# Model name, default is "`uname -s` router"
#model_name=Router Model

# Model description, default is "`uname -s` router"
#model_description=Very Secure Router - Model

# Model URL, default is URL of OS vendor
#model_url=https://miniupnp.tuxfamily.org/

# Bitrates reported by daemon in bits per second
# by default miniupnpd tries to get WAN interface speed
#bitrate_up=1000000
#bitrate_down=10000000

# Secure Mode, UPnP clients can only add mappings to their own IP
# Enabled by default
#secure_mode=no

# Default presentation URL is HTTP address on port 80
# If set to an empty string, no presentationURL element will appear
# in the XML description of the device, which prevents MS Windows
# from displaying an icon in the "Network Connections" panel.
#presentation_url=http://www.mylan/index.php

# Report system uptime instead of daemon uptime
system_uptime=yes

# Notify interval in seconds. default is 900 seconds.
# As advised in the standard, announcement have a lifetime double of this value.
notify_interval=900

# Unused rules cleaning.
# never remove any rule before this threshold for the number
# of redirections is exceeded. default to 20
#clean_ruleset_threshold=10
# Clean process work interval in seconds. default to 0 (disabled).
# a 600 seconds (10 minutes) interval makes sense
clean_ruleset_interval=600

# Log packets in pf (default is no)
#packet_log=yes

# Anchor name in pf (default is miniupnpd)
#anchor=miniupnpd

# ALTQ queue in pf
# Filter rules must be used for this to be used.
# compile with PF_ENABLE_FILTER_RULES (see config.h file)
#queue=queue_name1

# Tag name in pf
#tag=tag_name1

# Make filter rules in pf quick or not. default is yes
# active when compiled with PF_ENABLE_FILTER_RULES (see config.h file)
#quickrules=no

# UUID, generate your own UUID with "make genuuid"
uuid=00000000-0000-0000-0000-000000000000

# Daemon's serial and model number when reporting to clients
# (in XML description)
#serial=12345678
#model_number=1

# If compiled with IGD_V2 defined, force reporting IGDv1 in rootDesc (default
# is no)
#force_igd_desc_v1=no

# UPnP permission rules (also enforced for NAT-PMP and PCP) for IPv4
# (allow|deny) (external port range) IP/mask (internal port range) (optional regex filter)
# A port range is <min port>-<max port> or <port> if there is only
# one port in the range.
# IP/mask format must be nnn.nnn.nnn.nnn/nn
# Regex support must be enabled at build time : ./configure --regex
# It is advised to only allow redirection of ports >= 1024
# and end the rule set with "deny 0-65535 0.0.0.0/0 0-65535"
# The following default ruleset allows specific LAN side IP addresses
# to request only ephemeral ports. It is recommended that users
# modify the IP ranges to match their own internal networks, and
# also consider implementing network-specific restrictions
# CAUTION: failure to enforce any rules may permit insecure requests to be made!
allow 1024-65535 192.168.1.0/24 1024-65535
# disallow requests whose description string matches the given regex
# deny 1024-65535 192.168.1.0/24 1024-65535 "My evil app ver [[:digit:]]*"
deny 0-65535 0.0.0.0/0 0-65535
</pre>

Kerberos

2025-09-27T00:48:20Z

Chris: /* Tests */

= MIT-KRB HowTo =

== KDC konfigurieren und KDC Datenbank erstellen ==

# /var/lib/krb5kdc/kdc.conf erstellen:

<pre>
[kdcdefaults]
kdc_ports = 89
kdc_tcp_ports = 89
kadmind_port = 750
kpasswd_port = 466

[realms]
INTERN.NEOREALM.DYNDNS.ORG = {
database_name = /var/lib/krb5kdc/principal
acl_file = /var/lib/krb5kdc/kadm5.acl
key_stash_file = /var/lib/krb5kdc/.k5.INTERN.NEOREALM.DYNDNS.ORG
kadmind_port = 750
kpasswd_port = 466
max_life = 10h 0m 0s
max_renewable_life = 7d 0h 0m 0s
#default_principal_flags = +preauth
}

[logging]
kdc = CONSOLE
kdc = SYSLOG:INFO:DAEMON
kdc = FILE:/var/log/kdc.log
admin_server = FILE:/var/log/kadmin.log
</pre>

# Datenbank erstellen:

<pre>
kdb5_util create -r <REALMNAME> -s
</pre>

# kadmin Prinzipal erstellen:

<pre>
kadmin.local
addprinc admin/admin@INTERN.NEOREALM.DYNDNS.ORG
</pre>

<pre>
*/admin@INTERN.DARKREALM.DYNDNS.ORG *
</pre>

== DNS einrichten ==

Folgende RR's in die Zone eintragen
<pre>
_kerberos.intern.darkrealm.dyndns.org TXT "INTERN.DARKREALM.DYNDNS.ORG"
_kerberos.intern.darkrealm.dyndns.org SRV 1 0 89 shodan.intern.darkrealm.dyndns.org.
_kerberos-adm.intern.darkrealm.dyndns.org SRV 1 0 777 shodan.intern.darkrealm.dyndns.org. (wird noch? nicht ausgewertet)
_kerberos.intern.darkrealm.dyndns.org SRV 1 0 89 shodan.intern.darkrealm.dyndns.org.
_kpasswd.intern.darkrealm.dyndns.org SRV 1 0 464 shodan.intern.darkrealm.dyndns.org.
</pre>
Das reicht um Rechner ohne /etc/krb5.conf auf den default realm hinzuweisen. Weil _kerberos-adm zurzeit noch nicht ausgewertet wird muss man folgendes in der /etc/krb5.conf eintragen:

<pre>
[realms]
INTERN.DARKREALM.DYNDNS.ORG = {
admin_server = shodan.intern.darkrealm.dyndns.org:777
}
</pre>

== 2. NFS mit sec=krb5p ==

Mount : mount shodan:/home /home -o sec=krb5p

Server:
rpc.svcgssd braucht keytab /etc/krb5.keytab

rpc.svcgssd braucht explizit nfs/<fqdn> principal (erstmal egal welcher FQDN, aber der client kann nicht connecten wenn fqdn vom server nicht drinsteht)

Also in die keytab:

<pre>
nfs/shodan.intern.darkrealm.dyndns.org
</pre>

Client:
Client sucht laut man rpc.gssd folgende Einträge in der keytab: (scheinbar geht es hier nur um machine credentials, egal wie der principal heisst)
<pre>
<HOSTNAME>$@<REALM>
root/<hostname>@<REALM>
nfs/<hostname>@<REALM>
host/<hostname>@<REALM>
root/<anyname>@<REALM>
nfs/<anyname>@<REALM>
host/<anyname>@<REALM>
</pre>
Also in die keytab:

<pre>
host/rasmatazz.intern.darkrealm.dyndns.org
</pre>

KVNO's müssen auf jedem Rechner gleich sein !

Keytab auf dem Server :
<pre>
slot KVNO Principal
---- ---- ---------------------------------------------------------------------
1 2 nfs/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
2 2 nfs/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

Keytab auf dem Client :
<pre>
slot KVNO Principal
---- ---- ---------------------------------------------------------------------
1 2 host/rasmatazz.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
2 2 host/rasmatazz.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

== 3. SSH mit kerberos (-K) ==

In der Server-Keytab reicht :
<pre>
host/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

Aber in der /etc/krb5.conf auf dem Server muss stehen :
<pre>
<libdefaults>
ignore_acceptor_hostname = true
</pre>
ansonsten meldet der SSH-Server daß kein keytab-entry für host/shodan@ vorhanden ist.

== 4. samba mit kerberos ==

<pre>
kinit chris@HERRMANN.AD
sudo mount //shodan.intern.darkrealm.dyndns.org/temporary /mnt/tmp -o vers=3.02,sec=krb5i,cruid=chris,user=chris,cifsacl
</pre>

== Crossrealm AD <-> MIT-KRB5 ==

Funktioniert momentan nur mit einem Windows Server als Domain Controller!

=== Auf dem MIT-KRB5 KDC ===

* Mit kadmin 2 Principals hinzufügen (bei beiden gleiches Passwort setzen):
** addprinc krbtgt/INTERN.DARKREALM.DYNDNS.ORG@TESTDOMAIN.AD
** krbtgt/TESTDOMAIN.AD@INTERN.DARKREALM.DYNDNS.ORG

=== Auf dem Windows-Domaincontroller ===

* Active Directory-Domänen und -Vertrauensstellungen öffnen
* Neue Vertrauensstellung
* Name: INTERN.DARKREALM.DYNDNS.ORG
* Vertrauenstyp: Bereichsvertrauensstellung
* Nicht transitiv oder Transitiv, ist egal
* Bidirektional
* Vertrauensstellungskennwort eingeben (gleiches wie auf der MIT-KRB5 Seite)

* Zuständigen KDC für INTERN.DARKREALM.DYNDNS.ORG festlegen:
** ksetup /addkdc INTERN.DARKREALM.DYNDNS.ORG shodan.intern.darkrealm.dyndns.org
* Rechnernamen zu dem Realm zuordnen:
** ksetup /addhosttorealmmap shodan.intern.darkrealm.dyndns.org INTERN.DARKREALM.DYNDNS.ORG

=== Tests ===

* Ticket aus dem INTERN.DARKREALM.DYNDNS.ORG-Realm heraus vom TESTDOMAIN.AD-Realm holen

<pre>
kinit chris@INTERN.DARKREALM.DYNDNS.ORG
kvno HTTP/testdomain.ad@TESTDOMAIN.AD
</pre>

==== Fehler ====

Bekommt man einen Fehler "kvno: Generic error (see e-text) while getting credentials for HTTP/vm-winserver2025.testdomain.ad@TESTDOMAIN.AD" o.ä., dann entweder:

* modprinc +no_auth_data_required krbtgt/TESTDOMAIN.AD@INTERN.DARKREALM.DYNDNS.ORG

oder in /var/lib/krb5kdc/kdc.conf im [realms]-Abschnitt für den INTERN.DARKREALM.DYNDNS.ORG-Realms folgende Zeile einfügen:

'''disable_pac = true'''

Quelle: https://bugzilla.redhat.com/show_bug.cgi?id=2016312

=== Beispiel-LDIF eines Trusts ===

<pre>
dn: CN=INTERN.DARKREALM.DYNDNS.ORG,CN=System,DC=testdomain,DC=ad
objectClass: top
objectClass: leaf
objectClass: trustedDomain
cn: INTERN.DARKREALM.DYNDNS.ORG
distinguishedName: CN=INTERN.DARKREALM.DYNDNS.ORG,CN=System,DC=testdomain,DC=a
d
instanceType: 4
whenCreated: 20250926000103.0Z
whenChanged: 20250926000103.0Z
uSNCreated: 16453
uSNChanged: 16454
showInAdvancedViewOnly: TRUE
name: INTERN.DARKREALM.DYNDNS.ORG
objectGUID:: ojk5Xjs+Z0ieoCAx6wEjsA==
trustDirection: 3
trustPartner: INTERN.DARKREALM.DYNDNS.ORG
trustPosixOffset: 0
trustType: 3
trustAttributes: 0
flatName: INTERN.DARKREALM.DYNDNS.ORG
objectCategory: CN=Trusted-Domain,CN=Schema,CN=Configuration,DC=testdomain,DC=
ad
isCriticalSystemObject: TRUE
dSCorePropagationData: 16010101000000.0Z
trustAuthIncoming:
trustAuthOutgoing:
</pre>

<pre>
dn: CN=INTERN.DARKREALM.DYNDNS.ORG$,CN=Users,DC=testdomain,DC=ad
objectClass: top
objectClass: person
objectClass: organizationalPerson
objectClass: user
cn: INTERN.DARKREALM.DYNDNS.ORG$
distinguishedName: CN=INTERN.DARKREALM.DYNDNS.ORG$,CN=Users,DC=testdomain,DC=a
d
instanceType: 4
whenCreated: 20250926005246.0Z
whenChanged: 20250926005247.0Z
uSNCreated: 24597
uSNChanged: 24599
name: INTERN.DARKREALM.DYNDNS.ORG$
objectGUID:: HPPriJ7qRU+1bDWJWV9Yvw==
userAccountControl: 2082
badPwdCount: 0
codePage: 0
countryCode: 0
badPasswordTime: 0
lastLogoff: 0
lastLogon: 0
pwdLastSet: 0
primaryGroupID: 529
objectSid:: AQUAAAAAAAUVAAAAjZWqvXpflmUzfrM8UgQAAA==
accountExpires: 9223372036854775807
logonCount: 0
sAMAccountName: INTERN.DARKREALM.DYNDNS.ORG$
sAMAccountType: 805306370
objectCategory: CN=Person,CN=Schema,CN=Configuration,DC=testdomain,DC=ad
isCriticalSystemObject: TRUE
dSCorePropagationData: 16010101000000.0Z
</pre>

== Squid Crossrealm ==

Kein Plan wie und warum es funktioniert. Ist auch noch nicht wirklich crossrealm.

In /etc/squid/squid.conf folgende Zeile einfügen:

<pre>
auth_param negotiate program /usr/libexec/squid/negotiate_kerberos_auth -k /etc/squid.smb.keytab -i
</pre>

<pre>
msktutil -c -s HTTP/shodan.intern.darkrealm.dyndns.org -h shodan.intern.darkrealm.dyndns.org -k squid.smb.keytab --computer-name squid-http --upn HTTP/shodan.intern.darkrealm.dyndns.org --verbose --realm AD.INTERN.DARKREALM.DYNDNS.ORG --no-reverse-lookups --use-service-account
</pre>

Dann erstellte Keytab nach /etc/squid.smb.keytab kopieren und squid neustarten.

Kerberos

2025-09-27T00:45:17Z

Chris: /* Auf dem Windows-Domaincontroller */

= MIT-KRB HowTo =

== KDC konfigurieren und KDC Datenbank erstellen ==

# /var/lib/krb5kdc/kdc.conf erstellen:

<pre>
[kdcdefaults]
kdc_ports = 89
kdc_tcp_ports = 89
kadmind_port = 750
kpasswd_port = 466

[realms]
INTERN.NEOREALM.DYNDNS.ORG = {
database_name = /var/lib/krb5kdc/principal
acl_file = /var/lib/krb5kdc/kadm5.acl
key_stash_file = /var/lib/krb5kdc/.k5.INTERN.NEOREALM.DYNDNS.ORG
kadmind_port = 750
kpasswd_port = 466
max_life = 10h 0m 0s
max_renewable_life = 7d 0h 0m 0s
#default_principal_flags = +preauth
}

[logging]
kdc = CONSOLE
kdc = SYSLOG:INFO:DAEMON
kdc = FILE:/var/log/kdc.log
admin_server = FILE:/var/log/kadmin.log
</pre>

# Datenbank erstellen:

<pre>
kdb5_util create -r <REALMNAME> -s
</pre>

# kadmin Prinzipal erstellen:

<pre>
kadmin.local
addprinc admin/admin@INTERN.NEOREALM.DYNDNS.ORG
</pre>

<pre>
*/admin@INTERN.DARKREALM.DYNDNS.ORG *
</pre>

== DNS einrichten ==

Folgende RR's in die Zone eintragen
<pre>
_kerberos.intern.darkrealm.dyndns.org TXT "INTERN.DARKREALM.DYNDNS.ORG"
_kerberos.intern.darkrealm.dyndns.org SRV 1 0 89 shodan.intern.darkrealm.dyndns.org.
_kerberos-adm.intern.darkrealm.dyndns.org SRV 1 0 777 shodan.intern.darkrealm.dyndns.org. (wird noch? nicht ausgewertet)
_kerberos.intern.darkrealm.dyndns.org SRV 1 0 89 shodan.intern.darkrealm.dyndns.org.
_kpasswd.intern.darkrealm.dyndns.org SRV 1 0 464 shodan.intern.darkrealm.dyndns.org.
</pre>
Das reicht um Rechner ohne /etc/krb5.conf auf den default realm hinzuweisen. Weil _kerberos-adm zurzeit noch nicht ausgewertet wird muss man folgendes in der /etc/krb5.conf eintragen:

<pre>
[realms]
INTERN.DARKREALM.DYNDNS.ORG = {
admin_server = shodan.intern.darkrealm.dyndns.org:777
}
</pre>

== 2. NFS mit sec=krb5p ==

Mount : mount shodan:/home /home -o sec=krb5p

Server:
rpc.svcgssd braucht keytab /etc/krb5.keytab

rpc.svcgssd braucht explizit nfs/<fqdn> principal (erstmal egal welcher FQDN, aber der client kann nicht connecten wenn fqdn vom server nicht drinsteht)

Also in die keytab:

<pre>
nfs/shodan.intern.darkrealm.dyndns.org
</pre>

Client:
Client sucht laut man rpc.gssd folgende Einträge in der keytab: (scheinbar geht es hier nur um machine credentials, egal wie der principal heisst)
<pre>
<HOSTNAME>$@<REALM>
root/<hostname>@<REALM>
nfs/<hostname>@<REALM>
host/<hostname>@<REALM>
root/<anyname>@<REALM>
nfs/<anyname>@<REALM>
host/<anyname>@<REALM>
</pre>
Also in die keytab:

<pre>
host/rasmatazz.intern.darkrealm.dyndns.org
</pre>

KVNO's müssen auf jedem Rechner gleich sein !

Keytab auf dem Server :
<pre>
slot KVNO Principal
---- ---- ---------------------------------------------------------------------
1 2 nfs/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
2 2 nfs/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

Keytab auf dem Client :
<pre>
slot KVNO Principal
---- ---- ---------------------------------------------------------------------
1 2 host/rasmatazz.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
2 2 host/rasmatazz.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

== 3. SSH mit kerberos (-K) ==

In der Server-Keytab reicht :
<pre>
host/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

Aber in der /etc/krb5.conf auf dem Server muss stehen :
<pre>
<libdefaults>
ignore_acceptor_hostname = true
</pre>
ansonsten meldet der SSH-Server daß kein keytab-entry für host/shodan@ vorhanden ist.

== 4. samba mit kerberos ==

<pre>
kinit chris@HERRMANN.AD
sudo mount //shodan.intern.darkrealm.dyndns.org/temporary /mnt/tmp -o vers=3.02,sec=krb5i,cruid=chris,user=chris,cifsacl
</pre>

== Crossrealm AD <-> MIT-KRB5 ==

Funktioniert momentan nur mit einem Windows Server als Domain Controller!

=== Auf dem MIT-KRB5 KDC ===

* Mit kadmin 2 Principals hinzufügen (bei beiden gleiches Passwort setzen):
** addprinc krbtgt/INTERN.DARKREALM.DYNDNS.ORG@TESTDOMAIN.AD
** krbtgt/TESTDOMAIN.AD@INTERN.DARKREALM.DYNDNS.ORG

=== Auf dem Windows-Domaincontroller ===

* Active Directory-Domänen und -Vertrauensstellungen öffnen
* Neue Vertrauensstellung
* Name: INTERN.DARKREALM.DYNDNS.ORG
* Vertrauenstyp: Bereichsvertrauensstellung
* Nicht transitiv oder Transitiv, ist egal
* Bidirektional
* Vertrauensstellungskennwort eingeben (gleiches wie auf der MIT-KRB5 Seite)

* Zuständigen KDC für INTERN.DARKREALM.DYNDNS.ORG festlegen:
** ksetup /addkdc INTERN.DARKREALM.DYNDNS.ORG shodan.intern.darkrealm.dyndns.org
* Rechnernamen zu dem Realm zuordnen:
** ksetup /addhosttorealmmap shodan.intern.darkrealm.dyndns.org INTERN.DARKREALM.DYNDNS.ORG

=== Tests ===

* Ticket aus dem INTERN.DARKREALM.DYNDNS.ORG-Realm heraus vom TESTDOMAIN.AD-Realm holen

<pre>
kinit chris@INTERN.DARKREALM.DYNDNS.ORG
kvno HTTP/testdomain.ad@TESTDOMAIN.AD
</pre>

=== Beispiel-LDIF eines Trusts ===

<pre>
dn: CN=INTERN.DARKREALM.DYNDNS.ORG,CN=System,DC=testdomain,DC=ad
objectClass: top
objectClass: leaf
objectClass: trustedDomain
cn: INTERN.DARKREALM.DYNDNS.ORG
distinguishedName: CN=INTERN.DARKREALM.DYNDNS.ORG,CN=System,DC=testdomain,DC=a
d
instanceType: 4
whenCreated: 20250926000103.0Z
whenChanged: 20250926000103.0Z
uSNCreated: 16453
uSNChanged: 16454
showInAdvancedViewOnly: TRUE
name: INTERN.DARKREALM.DYNDNS.ORG
objectGUID:: ojk5Xjs+Z0ieoCAx6wEjsA==
trustDirection: 3
trustPartner: INTERN.DARKREALM.DYNDNS.ORG
trustPosixOffset: 0
trustType: 3
trustAttributes: 0
flatName: INTERN.DARKREALM.DYNDNS.ORG
objectCategory: CN=Trusted-Domain,CN=Schema,CN=Configuration,DC=testdomain,DC=
ad
isCriticalSystemObject: TRUE
dSCorePropagationData: 16010101000000.0Z
trustAuthIncoming:
trustAuthOutgoing:
</pre>

<pre>
dn: CN=INTERN.DARKREALM.DYNDNS.ORG$,CN=Users,DC=testdomain,DC=ad
objectClass: top
objectClass: person
objectClass: organizationalPerson
objectClass: user
cn: INTERN.DARKREALM.DYNDNS.ORG$
distinguishedName: CN=INTERN.DARKREALM.DYNDNS.ORG$,CN=Users,DC=testdomain,DC=a
d
instanceType: 4
whenCreated: 20250926005246.0Z
whenChanged: 20250926005247.0Z
uSNCreated: 24597
uSNChanged: 24599
name: INTERN.DARKREALM.DYNDNS.ORG$
objectGUID:: HPPriJ7qRU+1bDWJWV9Yvw==
userAccountControl: 2082
badPwdCount: 0
codePage: 0
countryCode: 0
badPasswordTime: 0
lastLogoff: 0
lastLogon: 0
pwdLastSet: 0
primaryGroupID: 529
objectSid:: AQUAAAAAAAUVAAAAjZWqvXpflmUzfrM8UgQAAA==
accountExpires: 9223372036854775807
logonCount: 0
sAMAccountName: INTERN.DARKREALM.DYNDNS.ORG$
sAMAccountType: 805306370
objectCategory: CN=Person,CN=Schema,CN=Configuration,DC=testdomain,DC=ad
isCriticalSystemObject: TRUE
dSCorePropagationData: 16010101000000.0Z
</pre>

== Squid Crossrealm ==

Kein Plan wie und warum es funktioniert. Ist auch noch nicht wirklich crossrealm.

In /etc/squid/squid.conf folgende Zeile einfügen:

<pre>
auth_param negotiate program /usr/libexec/squid/negotiate_kerberos_auth -k /etc/squid.smb.keytab -i
</pre>

<pre>
msktutil -c -s HTTP/shodan.intern.darkrealm.dyndns.org -h shodan.intern.darkrealm.dyndns.org -k squid.smb.keytab --computer-name squid-http --upn HTTP/shodan.intern.darkrealm.dyndns.org --verbose --realm AD.INTERN.DARKREALM.DYNDNS.ORG --no-reverse-lookups --use-service-account
</pre>

Dann erstellte Keytab nach /etc/squid.smb.keytab kopieren und squid neustarten.

Kerberos

2025-09-26T23:29:46Z

Chris: /* Crossrealm AD MIT-KRB5 */

= MIT-KRB HowTo =

== KDC konfigurieren und KDC Datenbank erstellen ==

# /var/lib/krb5kdc/kdc.conf erstellen:

<pre>
[kdcdefaults]
kdc_ports = 89
kdc_tcp_ports = 89
kadmind_port = 750
kpasswd_port = 466

[realms]
INTERN.NEOREALM.DYNDNS.ORG = {
database_name = /var/lib/krb5kdc/principal
acl_file = /var/lib/krb5kdc/kadm5.acl
key_stash_file = /var/lib/krb5kdc/.k5.INTERN.NEOREALM.DYNDNS.ORG
kadmind_port = 750
kpasswd_port = 466
max_life = 10h 0m 0s
max_renewable_life = 7d 0h 0m 0s
#default_principal_flags = +preauth
}

[logging]
kdc = CONSOLE
kdc = SYSLOG:INFO:DAEMON
kdc = FILE:/var/log/kdc.log
admin_server = FILE:/var/log/kadmin.log
</pre>

# Datenbank erstellen:

<pre>
kdb5_util create -r <REALMNAME> -s
</pre>

# kadmin Prinzipal erstellen:

<pre>
kadmin.local
addprinc admin/admin@INTERN.NEOREALM.DYNDNS.ORG
</pre>

<pre>
*/admin@INTERN.DARKREALM.DYNDNS.ORG *
</pre>

== DNS einrichten ==

Folgende RR's in die Zone eintragen
<pre>
_kerberos.intern.darkrealm.dyndns.org TXT "INTERN.DARKREALM.DYNDNS.ORG"
_kerberos.intern.darkrealm.dyndns.org SRV 1 0 89 shodan.intern.darkrealm.dyndns.org.
_kerberos-adm.intern.darkrealm.dyndns.org SRV 1 0 777 shodan.intern.darkrealm.dyndns.org. (wird noch? nicht ausgewertet)
_kerberos.intern.darkrealm.dyndns.org SRV 1 0 89 shodan.intern.darkrealm.dyndns.org.
_kpasswd.intern.darkrealm.dyndns.org SRV 1 0 464 shodan.intern.darkrealm.dyndns.org.
</pre>
Das reicht um Rechner ohne /etc/krb5.conf auf den default realm hinzuweisen. Weil _kerberos-adm zurzeit noch nicht ausgewertet wird muss man folgendes in der /etc/krb5.conf eintragen:

<pre>
[realms]
INTERN.DARKREALM.DYNDNS.ORG = {
admin_server = shodan.intern.darkrealm.dyndns.org:777
}
</pre>

== 2. NFS mit sec=krb5p ==

Mount : mount shodan:/home /home -o sec=krb5p

Server:
rpc.svcgssd braucht keytab /etc/krb5.keytab

rpc.svcgssd braucht explizit nfs/<fqdn> principal (erstmal egal welcher FQDN, aber der client kann nicht connecten wenn fqdn vom server nicht drinsteht)

Also in die keytab:

<pre>
nfs/shodan.intern.darkrealm.dyndns.org
</pre>

Client:
Client sucht laut man rpc.gssd folgende Einträge in der keytab: (scheinbar geht es hier nur um machine credentials, egal wie der principal heisst)
<pre>
<HOSTNAME>$@<REALM>
root/<hostname>@<REALM>
nfs/<hostname>@<REALM>
host/<hostname>@<REALM>
root/<anyname>@<REALM>
nfs/<anyname>@<REALM>
host/<anyname>@<REALM>
</pre>
Also in die keytab:

<pre>
host/rasmatazz.intern.darkrealm.dyndns.org
</pre>

KVNO's müssen auf jedem Rechner gleich sein !

Keytab auf dem Server :
<pre>
slot KVNO Principal
---- ---- ---------------------------------------------------------------------
1 2 nfs/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
2 2 nfs/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

Keytab auf dem Client :
<pre>
slot KVNO Principal
---- ---- ---------------------------------------------------------------------
1 2 host/rasmatazz.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
2 2 host/rasmatazz.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

== 3. SSH mit kerberos (-K) ==

In der Server-Keytab reicht :
<pre>
host/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

Aber in der /etc/krb5.conf auf dem Server muss stehen :
<pre>
<libdefaults>
ignore_acceptor_hostname = true
</pre>
ansonsten meldet der SSH-Server daß kein keytab-entry für host/shodan@ vorhanden ist.

== 4. samba mit kerberos ==

<pre>
kinit chris@HERRMANN.AD
sudo mount //shodan.intern.darkrealm.dyndns.org/temporary /mnt/tmp -o vers=3.02,sec=krb5i,cruid=chris,user=chris,cifsacl
</pre>

== Crossrealm AD <-> MIT-KRB5 ==

Funktioniert momentan nur mit einem Windows Server als Domain Controller!

=== Auf dem MIT-KRB5 KDC ===

* Mit kadmin 2 Principals hinzufügen (bei beiden gleiches Passwort setzen):
** addprinc krbtgt/INTERN.DARKREALM.DYNDNS.ORG@TESTDOMAIN.AD
** krbtgt/TESTDOMAIN.AD@INTERN.DARKREALM.DYNDNS.ORG

=== Auf dem Windows-Domaincontroller ===

* Active Directory-Domänen und -Vertrauensstellungen öffnen
* Neue Vertrauensstellung
* Name: INTERN.DARKREALM.DYNDNS.ORG
* Vertrauenstyp: Bereichsvertrauensstellung
* Nicht transitiv oder Transitiv, ist egal
* Bidirektional
* Vertrauensstellungskennwort eingeben (gleiches wie auf der MIT-KRB5 Seite)

* Zuständigen KDC für INTERN.DARKREALM.DYNDNS.ORG festlegen:
** ksetup /addkdc INTERN.DARKREALM.DYNDNS.ORG shodan.intern.darkrealm.dyndns.org
* Rechnernamen zu dem Realm zuordnen:
** ksetup /addhosttorealmmap shodan.intern.darkrealm.dyndns.org INTERN.DARKREALM.DYNDNS.ORG

Achtung:

=== Tests ===

* Ticket aus dem INTERN.DARKREALM.DYNDNS.ORG-Realm heraus vom TESTDOMAIN.AD-Realm holen

<pre>
kinit chris@INTERN.DARKREALM.DYNDNS.ORG
kvno HTTP/testdomain.ad@TESTDOMAIN.AD
</pre>

=== Beispiel-LDIF eines Trusts ===

<pre>
dn: CN=INTERN.DARKREALM.DYNDNS.ORG,CN=System,DC=testdomain,DC=ad
objectClass: top
objectClass: leaf
objectClass: trustedDomain
cn: INTERN.DARKREALM.DYNDNS.ORG
distinguishedName: CN=INTERN.DARKREALM.DYNDNS.ORG,CN=System,DC=testdomain,DC=a
d
instanceType: 4
whenCreated: 20250926000103.0Z
whenChanged: 20250926000103.0Z
uSNCreated: 16453
uSNChanged: 16454
showInAdvancedViewOnly: TRUE
name: INTERN.DARKREALM.DYNDNS.ORG
objectGUID:: ojk5Xjs+Z0ieoCAx6wEjsA==
trustDirection: 3
trustPartner: INTERN.DARKREALM.DYNDNS.ORG
trustPosixOffset: 0
trustType: 3
trustAttributes: 0
flatName: INTERN.DARKREALM.DYNDNS.ORG
objectCategory: CN=Trusted-Domain,CN=Schema,CN=Configuration,DC=testdomain,DC=
ad
isCriticalSystemObject: TRUE
dSCorePropagationData: 16010101000000.0Z
trustAuthIncoming:
trustAuthOutgoing:
</pre>

<pre>
dn: CN=INTERN.DARKREALM.DYNDNS.ORG$,CN=Users,DC=testdomain,DC=ad
objectClass: top
objectClass: person
objectClass: organizationalPerson
objectClass: user
cn: INTERN.DARKREALM.DYNDNS.ORG$
distinguishedName: CN=INTERN.DARKREALM.DYNDNS.ORG$,CN=Users,DC=testdomain,DC=a
d
instanceType: 4
whenCreated: 20250926005246.0Z
whenChanged: 20250926005247.0Z
uSNCreated: 24597
uSNChanged: 24599
name: INTERN.DARKREALM.DYNDNS.ORG$
objectGUID:: HPPriJ7qRU+1bDWJWV9Yvw==
userAccountControl: 2082
badPwdCount: 0
codePage: 0
countryCode: 0
badPasswordTime: 0
lastLogoff: 0
lastLogon: 0
pwdLastSet: 0
primaryGroupID: 529
objectSid:: AQUAAAAAAAUVAAAAjZWqvXpflmUzfrM8UgQAAA==
accountExpires: 9223372036854775807
logonCount: 0
sAMAccountName: INTERN.DARKREALM.DYNDNS.ORG$
sAMAccountType: 805306370
objectCategory: CN=Person,CN=Schema,CN=Configuration,DC=testdomain,DC=ad
isCriticalSystemObject: TRUE
dSCorePropagationData: 16010101000000.0Z
</pre>

== Squid Crossrealm ==

Kein Plan wie und warum es funktioniert. Ist auch noch nicht wirklich crossrealm.

In /etc/squid/squid.conf folgende Zeile einfügen:

<pre>
auth_param negotiate program /usr/libexec/squid/negotiate_kerberos_auth -k /etc/squid.smb.keytab -i
</pre>

<pre>
msktutil -c -s HTTP/shodan.intern.darkrealm.dyndns.org -h shodan.intern.darkrealm.dyndns.org -k squid.smb.keytab --computer-name squid-http --upn HTTP/shodan.intern.darkrealm.dyndns.org --verbose --realm AD.INTERN.DARKREALM.DYNDNS.ORG --no-reverse-lookups --use-service-account
</pre>

Dann erstellte Keytab nach /etc/squid.smb.keytab kopieren und squid neustarten.

Kerberos

2025-09-26T01:23:02Z

Chris: /* Crossrealm AD MIT-KRB5 */

= MIT-KRB HowTo =

== KDC konfigurieren und KDC Datenbank erstellen ==

# /var/lib/krb5kdc/kdc.conf erstellen:

<pre>
[kdcdefaults]
kdc_ports = 89
kdc_tcp_ports = 89
kadmind_port = 750
kpasswd_port = 466

[realms]
INTERN.NEOREALM.DYNDNS.ORG = {
database_name = /var/lib/krb5kdc/principal
acl_file = /var/lib/krb5kdc/kadm5.acl
key_stash_file = /var/lib/krb5kdc/.k5.INTERN.NEOREALM.DYNDNS.ORG
kadmind_port = 750
kpasswd_port = 466
max_life = 10h 0m 0s
max_renewable_life = 7d 0h 0m 0s
#default_principal_flags = +preauth
}

[logging]
kdc = CONSOLE
kdc = SYSLOG:INFO:DAEMON
kdc = FILE:/var/log/kdc.log
admin_server = FILE:/var/log/kadmin.log
</pre>

# Datenbank erstellen:

<pre>
kdb5_util create -r <REALMNAME> -s
</pre>

# kadmin Prinzipal erstellen:

<pre>
kadmin.local
addprinc admin/admin@INTERN.NEOREALM.DYNDNS.ORG
</pre>

<pre>
*/admin@INTERN.DARKREALM.DYNDNS.ORG *
</pre>

== DNS einrichten ==

Folgende RR's in die Zone eintragen
<pre>
_kerberos.intern.darkrealm.dyndns.org TXT "INTERN.DARKREALM.DYNDNS.ORG"
_kerberos.intern.darkrealm.dyndns.org SRV 1 0 89 shodan.intern.darkrealm.dyndns.org.
_kerberos-adm.intern.darkrealm.dyndns.org SRV 1 0 777 shodan.intern.darkrealm.dyndns.org. (wird noch? nicht ausgewertet)
_kerberos.intern.darkrealm.dyndns.org SRV 1 0 89 shodan.intern.darkrealm.dyndns.org.
_kpasswd.intern.darkrealm.dyndns.org SRV 1 0 464 shodan.intern.darkrealm.dyndns.org.
</pre>
Das reicht um Rechner ohne /etc/krb5.conf auf den default realm hinzuweisen. Weil _kerberos-adm zurzeit noch nicht ausgewertet wird muss man folgendes in der /etc/krb5.conf eintragen:

<pre>
[realms]
INTERN.DARKREALM.DYNDNS.ORG = {
admin_server = shodan.intern.darkrealm.dyndns.org:777
}
</pre>

== 2. NFS mit sec=krb5p ==

Mount : mount shodan:/home /home -o sec=krb5p

Server:
rpc.svcgssd braucht keytab /etc/krb5.keytab

rpc.svcgssd braucht explizit nfs/<fqdn> principal (erstmal egal welcher FQDN, aber der client kann nicht connecten wenn fqdn vom server nicht drinsteht)

Also in die keytab:

<pre>
nfs/shodan.intern.darkrealm.dyndns.org
</pre>

Client:
Client sucht laut man rpc.gssd folgende Einträge in der keytab: (scheinbar geht es hier nur um machine credentials, egal wie der principal heisst)
<pre>
<HOSTNAME>$@<REALM>
root/<hostname>@<REALM>
nfs/<hostname>@<REALM>
host/<hostname>@<REALM>
root/<anyname>@<REALM>
nfs/<anyname>@<REALM>
host/<anyname>@<REALM>
</pre>
Also in die keytab:

<pre>
host/rasmatazz.intern.darkrealm.dyndns.org
</pre>

KVNO's müssen auf jedem Rechner gleich sein !

Keytab auf dem Server :
<pre>
slot KVNO Principal
---- ---- ---------------------------------------------------------------------
1 2 nfs/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
2 2 nfs/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

Keytab auf dem Client :
<pre>
slot KVNO Principal
---- ---- ---------------------------------------------------------------------
1 2 host/rasmatazz.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
2 2 host/rasmatazz.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

== 3. SSH mit kerberos (-K) ==

In der Server-Keytab reicht :
<pre>
host/shodan.intern.darkrealm.dyndns.org@INTERN.DARKREALM.DYNDNS.ORG
</pre>

Aber in der /etc/krb5.conf auf dem Server muss stehen :
<pre>
<libdefaults>
ignore_acceptor_hostname = true
</pre>
ansonsten meldet der SSH-Server daß kein keytab-entry für host/shodan@ vorhanden ist.

== 4. samba mit kerberos ==

<pre>
kinit chris@HERRMANN.AD
sudo mount //shodan.intern.darkrealm.dyndns.org/temporary /mnt/tmp -o vers=3.02,sec=krb5i,cruid=chris,user=chris,cifsacl
</pre>

== Crossrealm AD <-> MIT-KRB5 ==

Funktioniert momentan nur mit einem Windows Server als Domain Controller!

=== Auf dem MIT-KRB5 KDC ===

* Mit kadmin 2 Principals hinzufügen (bei beiden gleiches Passwort setzen):
** addprinc krbtgt/INTERN.DARKREALM.DYNDNS.ORG@TESTDOMAIN.AD
** krbtgt/TESTDOMAIN.AD@INTERN.DARKREALM.DYNDNS.ORG

=== Auf dem Windows-Domaincontroller ===

* Active Directory-Domänen und -Vertrauensstellungen öffnen
* Neue Vertrauensstellung
* Name: INTERN.DARKREALM.DYNDNS.ORG
* Vertrauenstyp: Bereichsvertrauensstellung
* Nicht transitiv oder Transitiv, ist egal
* Bidirektional
* Vertrauensstellungskennwort eingeben (gleiches wie auf der MIT-KRB5 Seite)

* Zuständigen KDC für INTERN.DARKREALM.DYNDNS.ORG festlegen:
** ksetup /addkdc INTERN.DARKREALM.DYNDNS.ORG shodan.intern.darkrealm.dyndns.org
* Rechnernamen zu dem Realm zuordnen:
** ksetup /addhosttorealmmap shodan.intern.darkrealm.dyndns.org INTERN.DARKREALM.DYNDNS.ORG

=== Beispiel-LDIF eines Trusts ===

<pre>
dn: CN=INTERN.DARKREALM.DYNDNS.ORG,CN=System,DC=testdomain,DC=ad
objectClass: top
objectClass: leaf
objectClass: trustedDomain
cn: INTERN.DARKREALM.DYNDNS.ORG
distinguishedName: CN=INTERN.DARKREALM.DYNDNS.ORG,CN=System,DC=testdomain,DC=a
d
instanceType: 4
whenCreated: 20250926000103.0Z
whenChanged: 20250926000103.0Z
uSNCreated: 16453
uSNChanged: 16454
showInAdvancedViewOnly: TRUE
name: INTERN.DARKREALM.DYNDNS.ORG
objectGUID:: ojk5Xjs+Z0ieoCAx6wEjsA==
trustDirection: 3
trustPartner: INTERN.DARKREALM.DYNDNS.ORG
trustPosixOffset: 0
trustType: 3
trustAttributes: 0
flatName: INTERN.DARKREALM.DYNDNS.ORG
objectCategory: CN=Trusted-Domain,CN=Schema,CN=Configuration,DC=testdomain,DC=
ad
isCriticalSystemObject: TRUE
dSCorePropagationData: 16010101000000.0Z
trustAuthIncoming:
trustAuthOutgoing:
</pre>

<pre>
dn: CN=INTERN.DARKREALM.DYNDNS.ORG$,CN=Users,DC=testdomain,DC=ad
objectClass: top
objectClass: person
objectClass: organizationalPerson
objectClass: user
cn: INTERN.DARKREALM.DYNDNS.ORG$
distinguishedName: CN=INTERN.DARKREALM.DYNDNS.ORG$,CN=Users,DC=testdomain,DC=a
d
instanceType: 4
whenCreated: 20250926005246.0Z
whenChanged: 20250926005247.0Z
uSNCreated: 24597
uSNChanged: 24599
name: INTERN.DARKREALM.DYNDNS.ORG$
objectGUID:: HPPriJ7qRU+1bDWJWV9Yvw==
userAccountControl: 2082
badPwdCount: 0
codePage: 0
countryCode: 0
badPasswordTime: 0
lastLogoff: 0
lastLogon: 0
pwdLastSet: 0
primaryGroupID: 529
objectSid:: AQUAAAAAAAUVAAAAjZWqvXpflmUzfrM8UgQAAA==
accountExpires: 9223372036854775807
logonCount: 0
sAMAccountName: INTERN.DARKREALM.DYNDNS.ORG$
sAMAccountType: 805306370
objectCategory: CN=Person,CN=Schema,CN=Configuration,DC=testdomain,DC=ad
isCriticalSystemObject: TRUE
dSCorePropagationData: 16010101000000.0Z
</pre>

== Squid Crossrealm ==

Kein Plan wie und warum es funktioniert. Ist auch noch nicht wirklich crossrealm.

In /etc/squid/squid.conf folgende Zeile einfügen:

<pre>
auth_param negotiate program /usr/libexec/squid/negotiate_kerberos_auth -k /etc/squid.smb.keytab -i
</pre>

<pre>
msktutil -c -s HTTP/shodan.intern.darkrealm.dyndns.org -h shodan.intern.darkrealm.dyndns.org -k squid.smb.keytab --computer-name squid-http --upn HTTP/shodan.intern.darkrealm.dyndns.org --verbose --realm AD.INTERN.DARKREALM.DYNDNS.ORG --no-reverse-lookups --use-service-account
</pre>

Dann erstellte Keytab nach /etc/squid.smb.keytab kopieren und squid neustarten.

Kerberos

2025-09-26T00:07:02Z

Chris: /* Auf dem Windows-Domaincontroller */

Kerberos

2025-09-26T00:04:07Z

Chris:

Steam Games Directories

2025-09-06T13:56:23Z

Chris: /* Resident Evil 3 */

== Resident Evil 0 HD Remaster ==

%SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\339340\remote

Verknüpfung erstellen :

<pre>
mklink /D 339340 "c:\Users\chris\Saved Games\Resident Evil 0"
</pre>

== Resident Evil HD Remaster ==

Default : %SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\304240\remote

Verknüpfung erstellen:

<pre>
mklink /D 304240 "c:\Users\chris\Saved Games\Resident Evil"
</pre>

== Resident Evil 2 ==

%SystemDrive%\Users\Public\Documents\Steam\CODEX\883710

Verknüpfung erstellen :

<pre>
mklink /d 883710 "c:\Users\chris\Saved Games\Resident Evil 2"
</pre>

== Resident Evil 3 ==

%SystemDrive%\Users\Public\Documents\Steam\CODEX\254700

Verknüpfung erstellen :

Windows:

<pre>
mklink /d 254700 "c:\Users\chris\Saved Games\Resident Evil 3"
</pre>

Linux:

<pre>
ln -s /media/chris/Saved\ Games/Resident\ Evil\ 3/ ~/.wine/drive_c/users/Public/Documents/Steam/CODEX/254700
</pre>

== Resident Evil 4 ==

Default : C:\ProgramData\Steam\%UserName%\%Appid%\

Verknüpfung erstellen:

<pre>
mklink /d Delacroix "c:\Users\chris\Saved Games\Resident Evil 4"
</pre>

== Resident Evil 5 Gold ==

%SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\21690\remote

Verknüpfung erstellen :

<pre>
mklink /D 21690 "c:\Users\chris\Saved Games\Resident Evil 5"
</pre>

== Resident Evil 6 ==

Savegames müssen beim Importieren auf neuen Rechnern / Maschinen hexeditiert werden, sonst werden sie als korrupt gemeldet.

* Neue Speicherdaten erstellen
* Neu erstellte savegame.bin mit hexeditor öffnen
* Offset 16-23 (Hex 10-17) kopieren
* alte savegame.bin mit hexeditor öffnen
* An Offset 16-23 (Hex 10-17) Werte von neu erstellter savegame.bin einfügen

Version 1.0.6 (Ali123)

c:\Resident Evil 6\Profile

Verknüpfung erstellen :

<pre>
mklink /D Profile "c:\Users\chris\Saved Games\Resident Evil 6"
</pre>

Version 1.0.6 (RLD)

C:\ProgramData\Steam\Delacroix

<pre>
mklink /D 221040 "C:\Users\chris\Saved Games\Resident Evil 6"
</pre>

== Resident Evil 7 ==

Default : %DEFAULT_CPY_DIR% = %USERNAME%\Documents\CPY_SAVES\<PlayerName>\418370

SavePath in c:\Resident Evil 7\cpy.ini ändern.

== Resident Evil 7 Gold ==

%SystemDrive%\Users\Public\Documents\Steam\CODEX\418370

Verknüpfung erstellen :

<pre>
mklink /D 418370 "c:\Users\chris\Saved Games\Resident Evil 7"
</pre>

== Resident Evil Revelations ==

Windows
<pre>
C:\users\chris\Local Settings\Application Data\FLT\Steam\222480\remote
</pre>

Linux
<pre>
~/.wine/drive_c/users/chris/AppData/Local/FLT/Steam/222480/remote
</pre>

== Resident Evil Revelations 2 ==

%SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\287290\remote

Verknüpfung erstellen:

<pre>
mklink /D 287290 "c:\Users\chris\Saved Games\Resident Evil Revelations 2"
ln -s ~/WINE/Saved\ Games/Resident\ Evil\ Revelations\ 2 287290
</pre>

Mit Folder Redirections alles von Linux aus:

<pre>
ln -s ../AppData/Roaming/Steam/CODEX/287290 Resident\ Evil\ Revelations\ 2
</pre>

Linux

<pre>
~/.wine/drive_c/users/chris/AppData/Roaming/Steam/CODEX/
</pre>

Steam Games Directories

2025-09-06T13:55:59Z

Chris: /* Resident Evil 3 */

== Resident Evil 0 HD Remaster ==

%SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\339340\remote

Verknüpfung erstellen :

<pre>
mklink /D 339340 "c:\Users\chris\Saved Games\Resident Evil 0"
</pre>

== Resident Evil HD Remaster ==

Default : %SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\304240\remote

Verknüpfung erstellen:

<pre>
mklink /D 304240 "c:\Users\chris\Saved Games\Resident Evil"
</pre>

== Resident Evil 2 ==

%SystemDrive%\Users\Public\Documents\Steam\CODEX\883710

Verknüpfung erstellen :

<pre>
mklink /d 883710 "c:\Users\chris\Saved Games\Resident Evil 2"
</pre>

== Resident Evil 3 ==

%SystemDrive%\Users\Public\Documents\Steam\CODEX\254700

Verknüpfung erstellen :

Windows:

<pre>
mklink /d 254700 "c:\Users\chris\Saved Games\Resident Evil 2"
</pre>

Linux:

<pre>
ln -s /media/chris/Saved\ Games/Resident\ Evil\ 3/ ~/.wine/drive_c/users/Public/Documents/Steam/CODEX/254700
</pre>

== Resident Evil 4 ==

Default : C:\ProgramData\Steam\%UserName%\%Appid%\

Verknüpfung erstellen:

<pre>
mklink /d Delacroix "c:\Users\chris\Saved Games\Resident Evil 4"
</pre>

== Resident Evil 5 Gold ==

%SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\21690\remote

Verknüpfung erstellen :

<pre>
mklink /D 21690 "c:\Users\chris\Saved Games\Resident Evil 5"
</pre>

== Resident Evil 6 ==

Savegames müssen beim Importieren auf neuen Rechnern / Maschinen hexeditiert werden, sonst werden sie als korrupt gemeldet.

* Neue Speicherdaten erstellen
* Neu erstellte savegame.bin mit hexeditor öffnen
* Offset 16-23 (Hex 10-17) kopieren
* alte savegame.bin mit hexeditor öffnen
* An Offset 16-23 (Hex 10-17) Werte von neu erstellter savegame.bin einfügen

Version 1.0.6 (Ali123)

c:\Resident Evil 6\Profile

Verknüpfung erstellen :

<pre>
mklink /D Profile "c:\Users\chris\Saved Games\Resident Evil 6"
</pre>

Version 1.0.6 (RLD)

C:\ProgramData\Steam\Delacroix

<pre>
mklink /D 221040 "C:\Users\chris\Saved Games\Resident Evil 6"
</pre>

== Resident Evil 7 ==

Default : %DEFAULT_CPY_DIR% = %USERNAME%\Documents\CPY_SAVES\<PlayerName>\418370

SavePath in c:\Resident Evil 7\cpy.ini ändern.

== Resident Evil 7 Gold ==

%SystemDrive%\Users\Public\Documents\Steam\CODEX\418370

Verknüpfung erstellen :

<pre>
mklink /D 418370 "c:\Users\chris\Saved Games\Resident Evil 7"
</pre>

== Resident Evil Revelations ==

Windows
<pre>
C:\users\chris\Local Settings\Application Data\FLT\Steam\222480\remote
</pre>

Linux
<pre>
~/.wine/drive_c/users/chris/AppData/Local/FLT/Steam/222480/remote
</pre>

== Resident Evil Revelations 2 ==

%SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\287290\remote

Verknüpfung erstellen:

<pre>
mklink /D 287290 "c:\Users\chris\Saved Games\Resident Evil Revelations 2"
ln -s ~/WINE/Saved\ Games/Resident\ Evil\ Revelations\ 2 287290
</pre>

Mit Folder Redirections alles von Linux aus:

<pre>
ln -s ../AppData/Roaming/Steam/CODEX/287290 Resident\ Evil\ Revelations\ 2
</pre>

Linux

<pre>
~/.wine/drive_c/users/chris/AppData/Roaming/Steam/CODEX/
</pre>

Steam Games Directories

2025-09-06T13:52:41Z

Chris: /* Resident Evil 2 */

== Resident Evil 0 HD Remaster ==

%SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\339340\remote

Verknüpfung erstellen :

<pre>
mklink /D 339340 "c:\Users\chris\Saved Games\Resident Evil 0"
</pre>

== Resident Evil HD Remaster ==

Default : %SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\304240\remote

Verknüpfung erstellen:

<pre>
mklink /D 304240 "c:\Users\chris\Saved Games\Resident Evil"
</pre>

== Resident Evil 2 ==

%SystemDrive%\Users\Public\Documents\Steam\CODEX\883710

Verknüpfung erstellen :

<pre>
mklink /d 883710 "c:\Users\chris\Saved Games\Resident Evil 2"
</pre>

== Resident Evil 3 ==

%SystemDrive%\Users\Public\Documents\Steam\CODEX\254700

Verknüpfung erstellen :

<pre>
mklink /d 254700 "c:\Users\chris\Saved Games\Resident Evil 2"
</pre>

== Resident Evil 4 ==

Default : C:\ProgramData\Steam\%UserName%\%Appid%\

Verknüpfung erstellen:

<pre>
mklink /d Delacroix "c:\Users\chris\Saved Games\Resident Evil 4"
</pre>

== Resident Evil 5 Gold ==

%SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\21690\remote

Verknüpfung erstellen :

<pre>
mklink /D 21690 "c:\Users\chris\Saved Games\Resident Evil 5"
</pre>

== Resident Evil 6 ==

Savegames müssen beim Importieren auf neuen Rechnern / Maschinen hexeditiert werden, sonst werden sie als korrupt gemeldet.

* Neue Speicherdaten erstellen
* Neu erstellte savegame.bin mit hexeditor öffnen
* Offset 16-23 (Hex 10-17) kopieren
* alte savegame.bin mit hexeditor öffnen
* An Offset 16-23 (Hex 10-17) Werte von neu erstellter savegame.bin einfügen

Version 1.0.6 (Ali123)

c:\Resident Evil 6\Profile

Verknüpfung erstellen :

<pre>
mklink /D Profile "c:\Users\chris\Saved Games\Resident Evil 6"
</pre>

Version 1.0.6 (RLD)

C:\ProgramData\Steam\Delacroix

<pre>
mklink /D 221040 "C:\Users\chris\Saved Games\Resident Evil 6"
</pre>

== Resident Evil 7 ==

Default : %DEFAULT_CPY_DIR% = %USERNAME%\Documents\CPY_SAVES\<PlayerName>\418370

SavePath in c:\Resident Evil 7\cpy.ini ändern.

== Resident Evil 7 Gold ==

%SystemDrive%\Users\Public\Documents\Steam\CODEX\418370

Verknüpfung erstellen :

<pre>
mklink /D 418370 "c:\Users\chris\Saved Games\Resident Evil 7"
</pre>

== Resident Evil Revelations ==

Windows
<pre>
C:\users\chris\Local Settings\Application Data\FLT\Steam\222480\remote
</pre>

Linux
<pre>
~/.wine/drive_c/users/chris/AppData/Local/FLT/Steam/222480/remote
</pre>

== Resident Evil Revelations 2 ==

%SystemDrive%\Users\%USERNAME%\AppData\Roaming\Steam\CODEX\287290\remote

Verknüpfung erstellen:

<pre>
mklink /D 287290 "c:\Users\chris\Saved Games\Resident Evil Revelations 2"
ln -s ~/WINE/Saved\ Games/Resident\ Evil\ Revelations\ 2 287290
</pre>

Mit Folder Redirections alles von Linux aus:

<pre>
ln -s ../AppData/Roaming/Steam/CODEX/287290 Resident\ Evil\ Revelations\ 2
</pre>

Linux

<pre>
~/.wine/drive_c/users/chris/AppData/Roaming/Steam/CODEX/
</pre>

Visual Pinball

2025-08-28T00:03:04Z

Chris: /* BGFX Backend */

== Linux Settings ==

=== BGFX Backend ===

<pre>
PlayfieldDisplay = 2
BGSet = 1
ViewDTRotation = 90
ViewCabRotation = 90

; Nudge
LRAxis = 0
LRAxisFlip = 0
UDAxis = 0
UDAxisFlip = 0
PBWEnabled = 0
PBWNormalMount = 0
PBWDefaultLayout = 1
PBWRotationCB = 0
PBWRotationvalue = 0
PBWAccelGainX = 0
PBWAccelGainY = 0
PBWAccelMaxX = 0
PBWAccelMaxY = 0
EnableNudgeFilter = 0
EnableLegacyNudge = 0
LegacyNudgeStrength = 0
AccelVelocityInput = 0

; Tilt plumb
TiltSensCB = 0
TiltSensValue = 0
TiltInertia = 0

; Plunger
PlungerAxis = 0
ReversePlungerAxis = 0
DeadZone = 0
PlungerRetract = 0

; Joystick input mappings
JoyLFlipKey = 10
JoyRFlipKey = 11
JoyStagedLFlipKey = 0
JoyStagedRFlipKey = 0
JoyPlungerKey = 1
JoyAddCreditKey = 3
JoyAddCredit2Key = 0
JoyLMagnaSave = 0
JoyRMagnaSave = 0
JoyStartGameKey = 7
JoyExitGameKey = 0
JoyFrameCount = 0
JoyVolumeUp = 0
JoyVolumeDown = 0
JoyLTiltKey = 0
JoyCTiltKey = 0
JoyRTiltKey = 0
JoyMechTiltKey = 0
JoyDebugKey = 0
JoyDebuggerKey = 0
JoyCustom1 =
JoyCustom2 =
JoyCustom3 =
JoyCustom4 =
JoyPMBuyIn =
JoyPMCoin3 =
JoyPMCoin4 =
JoyPMCoinDoor = 5
JoyPMCancel = 14
JoyPMDown = 13
JoyPMUp = 12
JoyPMEnter = 15
JoyLockbarKey = 0
JoyTableRecenterKey = 0
JoyTableUpKey = 0
JoyTableDownKey = 0
JoyPauseKey = 0
JoyTweakKey = 0

[ScoreView]
ScoreViewDisplay =
ScoreViewFullScreen = 0
; Output mode
; 0 - Disabled
; 1 - Embedded in another output
; 2 - Native system window (maybe a window, exclusive fullscreen, VR headset,...)
ScoreViewOutput = 2
ScoreViewWndX = 1920
ScoreViewWndY = 0
ScoreViewWidth = 1600
ScoreViewHeight = 900
; ScoreView Priority (Highest wins)
Priority.B2SLegacyDMD =
Priority.ScoreView =
</pre>

Visual Pinball

2025-08-27T22:59:06Z

Chris: /* BGFX Backend */

Visual Pinball

2025-08-27T22:14:51Z

Chris: /* BGFX Backend */

== Linux Settings ==

=== BGFX Backend ===

<pre>
PlayfieldDisplay = 2
BGSet = 1
ViewDTRotation = 90
ViewCabRotation = 90
</pre>

Visual Pinball

2025-08-27T22:14:25Z

Chris: Die Seite wurde neu angelegt: „== Linux Settings == === BGFX Backend === PlayfieldDisplay = 2 BGSet = 1 ViewDTRotation = 90 ViewCabRotation = 90“

== Linux Settings ==

=== BGFX Backend ===

PlayfieldDisplay = 2
BGSet = 1
ViewDTRotation = 90
ViewCabRotation = 90

Paperless-ngx

2025-08-17T22:36:47Z

Chris:

== Installation ==

* sudo lxc-create -n paperless-ngx -B zfs -t download -- --dist debian --release bookworm --arch amd64

* dpkg-reconfigure tzdata
* dpkg-reconfigure locales
* apt install python3 python3-pip python3-dev imagemagick fonts-liberation gnupg libpq-dev default-libmysqlclient-dev pkg-config libmagic-dev mime-support libzbar0 poppler-utils apt-transport-https man-db vim bash-completion openssh-server unpaper ghostscript icc-profiles-free qpdf liblept5 libxml2 pngquant zlib1g tesseract-ocr redis tesseract-ocr-deu git locate curl wget python3.11-venv emacs-nox autoconf libtool libleptonica-dev curl wget vim sudo mariadb-server
* adduser paperless --system --home /opt/paperless-ngx --group
* curl -O -L https://github.com/paperless-ngx/paperless-ngx/releases/download/v2.1.3/paperless-ngx-v2.1.3.tar.xz
* tar -xf paperless-ngx-v2.1.3.tar.xz -C/opt/
* cd /opt/paperless-ngx
* mkdir media
* mkdir data
* mkdir consume
* chown -R paperless:paperless .
* chown paperless:paperless media
* chown paperless:paperless data
* chown paperless:paperless consume
* sudo -Hu paperless python3 -m venv venv-paperless-ngx
* sudo -Hu paperless -s
* . venv-paperless-ngx/bin/activate
* pip3 install -r requirements.txt
* cd src
* vim /opt/paperless-ngx/paperless.conf

<pre>
# Have a look at the docs for documentation.
# https://docs.paperless-ngx.com/configuration/

# Debug. Only enable this for development.

#PAPERLESS_DEBUG=false

# Required services

PAPERLESS_REDIS=redis://localhost:6379
PAPERLESS_DBENGINE=mariadb
PAPERLESS_DBHOST=shodan.intern.darkrealm.dyndns.org
PAPERLESS_DBPORT=3306
PAPERLESS_DBNAME=paperless
PAPERLESS_DBUSER=paperless
PAPERLESS_DBPASS=paperless
PAPERLESS_DBSSLMODE=DISABLED

# Paths and folders

#PAPERLESS_CONSUMPTION_DIR=../consume
#PAPERLESS_DATA_DIR=../data
#PAPERLESS_TRASH_DIR=../trash
#PAPERLESS_MEDIA_ROOT=../media
PAPERLESS_CONSUMPTION_DIR=/mnt/storage/dms/consume
PAPERLESS_DATA_DIR=/mnt/storage/dms/data
PAPERLESS_TRASH_DIR=/mnt/storage/dms/trash
PAPERLESS_MEDIA_ROOT=/mnt/storage/dms/media
PAPERLESS_STATICDIR=../static
#PAPERLESS_FILENAME_FORMAT=
PAPERLESS_FILENAME_FORMAT={{owner_username}}/{{correspondent}}/{{document_type}}/{{created_year}}/{{title}}
PAPERLESS_FILENAME_FORMAT_REMOVE_NONE=true

# Security and hosting

PAPERLESS_SECRET_KEY=Peem7AhD
PAPERLESS_URL=https://paperless.darkrealm.dyndns.org
#PAPERLESS_CSRF_TRUSTED_ORIGINS=https://example.com # can be set using PAPERLESS_URL
#PAPERLESS_ALLOWED_HOSTS=example.com,www.example.com # can be set using PAPERLESS_URL
#PAPERLESS_CORS_ALLOWED_HOSTS=https://localhost:8080,https://example.com # can be set using PAPERLESS_URL
#PAPERLESS_FORCE_SCRIPT_NAME=
#PAPERLESS_STATIC_URL=/static/
#PAPERLESS_AUTO_LOGIN_USERNAME=
#PAPERLESS_COOKIE_PREFIX=
#PAPERLESS_ENABLE_HTTP_REMOTE_USER=false

# OCR settings

PAPERLESS_OCR_LANGUAGE=eng+deu
#PAPERLESS_OCR_MODE=skip
#PAPERLESS_OCR_SKIP_ARCHIVE_FILE=never
#PAPERLESS_OCR_OUTPUT_TYPE=pdfa
#PAPERLESS_OCR_PAGES=1
#PAPERLESS_OCR_IMAGE_DPI=300
#PAPERLESS_OCR_CLEAN=clean
#PAPERLESS_OCR_DESKEW=true
#PAPERLESS_OCR_ROTATE_PAGES=true
#PAPERLESS_OCR_ROTATE_PAGES_THRESHOLD=4
#PAPERLESS_OCR_USER_ARGS={}
#PAPERLESS_CONVERT_MEMORY_LIMIT=0
#PAPERLESS_CONVERT_TMPDIR=/var/tmp/paperless

PAPERLESS_ENABLE_NLTK=true
#PAPERLESS_NLTK_DIR=../nltk_data
PAPERLESS_NLTK_DIR=/mnt/storage/dms/nltk_data
# Software tweaks

#PAPERLESS_TASK_WORKERS=1
#PAPERLESS_THREADS_PER_WORKER=1
PAPERLESS_TIME_ZONE=Europe/Berlin
#PAPERLESS_CONSUMER_POLLING=10
#PAPERLESS_CONSUMER_DELETE_DUPLICATES=false
PAPERLESS_CONSUMER_RECURSIVE=true
#PAPERLESS_CONSUMER_IGNORE_PATTERNS=[".DS_STORE/*", "._*", ".stfolder/*", ".stversions/*", ".localized/*", "desktop.ini"]
#PAPERLESS_CONSUMER_SUBDIRS_AS_TAGS=false
PAPERLESS_CONSUMER_ENABLE_BARCODES=true
#PAPERLESS_CONSUMER_BARCODE_STRING=PATCHT
#PAPERLESS_CONSUMER_BARCODE_UPSCALE=0.0
#PAPERLESS_CONSUMER_BARCODE_DPI=300
#PAPERLESS_CONSUMER_ENABLE_COLLATE_DOUBLE_SIDED=false
#PAPERLESS_CONSUMER_COLLATE_DOUBLE_SIDED_SUBDIR_NAME=double-sided
#PAPERLESS_CONSUMER_COLLATE_DOUBLE_SIDED_TIFF_SUPPORT=false
#PAPERLESS_PRE_CONSUME_SCRIPT=/path/to/an/arbitrary/script.sh
#PAPERLESS_POST_CONSUME_SCRIPT=/path/to/an/arbitrary/script.sh
#PAPERLESS_FILENAME_DATE_ORDER=YMD
#PAPERLESS_FILENAME_PARSE_TRANSFORMS=[]
#PAPERLESS_NUMBER_OF_SUGGESTED_DATES=5
#PAPERLESS_THUMBNAIL_FONT_NAME=
#PAPERLESS_IGNORE_DATES=
#PAPERLESS_ENABLE_UPDATE_CHECK=

# Tika settings

PAPERLESS_TIKA_ENABLED=true
PAPERLESS_TIKA_ENDPOINT=http://localhost:9998
PAPERLESS_TIKA_GOTENBERG_ENDPOINT=http://localhost:3000

# Binaries

#PAPERLESS_CONVERT_BINARY=/usr/bin/convert
#PAPERLESS_GS_BINARY=/usr/bin/gs
</pre>

* DB und User erstellen

* sudo -Hu paperless python3 manage.py migrate
* sudo -Hu paperless python3 manage.py createsuperuser
* cp -av ../scripts/*.service /etc/systemd/system/
* cp -av ../scripts/*.socket /etc/systemd/system/
* systemctl daemon-reload

ImageMagick Policy in /etc/ImageMagick-6/policy.xml anpassen:
<pre>
<policy domain="coder" rights="read|write" pattern="PDF" />
</pre>

JBIG2ENC compilen:
* git clone https://github.com/agl/jbig2enc
* cd jbig2enc
* ./autogen.sh
* ./configure
* make
* make install

NLTK installieren
* sudo -Hu paperless -s
* . venv-paperless-ngx/bin/activate
* python3
* import nltk
* nltk.download() (Nach /opt/paperless-ngx/nltk_data herunterladen)
* punkt, snowball_data und stopwords runterladen

Apache Tika und Gotenberg installieren (momentan über Docker)
* apt install docker-compose
* docker run -d -p 127.0.0.1:9998:9998 apache/tika
* docker run -d -p 127.0.0.1:3000:3000 gotenberg/gotenberg:8

== Update ==

* wget aktuellstes tar.xz
* systemctl stop paperless-*
* tar -xf <release.tar.xz> -C/opt/
* cd /opt/paperless-ngx
* cp -av paperless.conf.keep paperless.conf
* chown paperless:paperless . -R
* sudo -Hu paperless -s
* source /opt/paperless-ngx/venv-paperless-ngx/bin/activate
* pip install -r requirements.txt
* cd src
* python3 manage.py migrate

== Fixes ==

=== NLTK Fix ===

<pre>
sudo -Hu paperless -s
source /opt/paperless-ngx/venv-paperless-ngx/bin/activate
cd src
python3 -W ignore::RuntimeWarning -m nltk.downloader -d "/mnt/storage/dms/nltk_data" punkt_tab
</pre>

=== Logs show "possible incompatible database column" when deleting documents ===

You may see errors when deleting documents like:

<pre>
Data too long for column 'transaction_id' at row 1
</pre>

This error can occur in installations which have upgraded from a version of Paperless-ngx that used Django 4 (Paperless-ngx versions prior to v2.13.0) with a MariaDB/MySQL database. Due to the backawards-incompatible change in Django 5, the column "documents_document.transaction_id" will need to be re-created, which can be done with a one-time run of the following management command:

<pre>
sudo -Hu paperless -s
source /opt/paperless-ngx/venv-paperless-ngx/bin/activate
cd src
python3 manage.py convert_mariadb_uuid
</pre>

== Komplette Neuinstallation inkl. Komplettrestore ==

* lxc init images:debian/12 paperless-ngx
* lxc start papalass
* lxc shell papalass
* passwd

Anschliessend vorgehen wie bei [[#Installation]]

* die gesicherte ZIP-Datei des exports auf den neuen Paperless-Server kopieren und nach /temp entpacken
* sudo -Hu paperless -s
* source /opt/paperless-ngx/venv-paperless-ngx/bin/activate
* cd src
* python3 manage.py document_importer /temp/

Multicast

2025-07-21T19:00:02Z

Chris:

== Multicast über Wireguard ==

Voraussetzungen:

* Multicast muss für das WG Interface aktiviert sein
* Über Kommanzozeile:

<pre>
ip l set WG multicast on
</pre>

* Über systemd-networkd:

WG.network
<pre>
[Link]
Multicast=true
</pre>

* Multicast-IP-Adressen müssen bei den AllowedIPs dabei sein:

WG.netdev
<pre>
AllowedIPs=224.0.0.0/4,ff00::/8
</pre>

Achtung: Wird im .netdev-File etwas geändert muss das Interface erst gelöscht und wieder erstellt werden damit die Änderungen greifen:

<pre>
networkctl delete WG
networkctl reload
</pre>

=== mDNS ===

Über Avahi können mDNS-Abfragen "reflektiert" (in andere Subnetze weitergeleitet) werden. Wenn es sich beim dem Interface um ein Point-to-Point Interface handelt (Wireguard-Interfaces sind Point-to-Point Interfaces) muss dies auch noch erlaubt werden.

/etc/avahi/avahi-daemon.conf:

<pre>
[server]
allow-point-to-point=yes # Bei Point-to-Point Interfaces nötig

[reflector]
enable-reflector=yes
</pre>

Netzwerk A möchte die mDNS-Geräte in Netzwerk B sehen:

reflector auf Host B aktivieren.

Netzwerk B möchte die mDNS-Geräte in Netzwerk A sehen:

reflector auf Host A aktivieren.

Damit die Geräte auch auf den Hosts hinter dem Gateway gesehen werden können muss auf dem jeweiligen Gateway auch der Reflector aktiviert werden.

=== SSDP / UPnP / DLNA ===

SSDP hat mit mDNS nichts zu tun, kann also auch nicht über den avahi-daemon reflector weitergeleitet werden. Es muss ein Multicast-Router verwendet werden, hier: smcroute

/etc/smcroute.conf:

<pre>
phyint WG enable
phyint BRIDGE_LAN enable

# SSDP
mgroup from BRIDGE_LAN group 239.255.255.250
mroute from BRIDGE_LAN group 239.255.255.250 to WG
mgroup from WG group 239.255.255.250
mroute from WG group 239.255.255.250 to BRIDGE_LAN
</pre>

Netzwerk A möchte die UPnP-Geräte in Netzwerk B sehen:

smcroute auf Host A starten.

Netzwerk B möchte die UPnP-Geräte in Netzwerk A sehen:

smcroute auf Host B starten.

=== Geräte hinter dem Router ===

Damit die Geräte hinter dem Router von den Multicast-Paketen erreicht werden können:

<pre>
ip r add 192.168.1.0/24 dev WG
</pre>

<pre>
ip r add 192.168.178.0/24 dev WG
</pre>

IPv4 Forwarding muss aktiviert sein, sowie forwarding in the Firewall muss akzeptiert werden!

=== Wireguard Debugging aktivieren ===

<pre>
echo module wireguard +p > /sys/kernel/debug/dynamic_debug/control
</pre>

Multicast

2025-07-19T02:38:47Z

Chris: /* Geräte hinter dem Router */

Multicast

2025-07-18T23:34:13Z

Chris:

Initramfs

2025-07-15T22:28:03Z

Chris:

== Inhalt anzeigen ==

<pre>
lsinitrd <INITRAMFS>
</pre>

== Entpacken ==

<pre>
lsinitrd --unpack <INITRAMFS>
</pre>

== CPIO ==

=== Inhalt anzeigen ===

<pre>
cpio -t < <INITRAMFS>
</pre>

=== Inhalt entpacken ===

<pre>
cpio -i < <INITRAMFS>
</pre>

Unix Helper Stuff

2025-06-19T17:38:59Z

Chris:

= Little UNIX Helpers =

== Verzeichnisse nach Größe sortieren ==

<pre>
du -d 1 | sort -n
</pre>

*du -d 1 = Nur in ein tieferliegendes Verzeichnis absteigen
*sort -n = Aufsteigend nach Nummern sortieren

[http://en.wikipedia.org/wiki/Sort_%28Unix%29]

== Nur Dateien listen die NACH dem <Datum> modifiziert wurden ==

<pre>
find -newermt "YYYY-MM-DD HH:MM"
</pre>

== locate-Ergebnisse nach String durchsuchen ==
=== Achtung: Funktioniert bis jetzt nur bei Dateinamen OHNE Leerzeichen !===
<pre>
locate <Dateinamen> | xargs grep <string>
</pre>

== ACLs nur für Directories setzen ==

<pre>
find . -type d -exec setfacl -m u::rwx,u::rwx,o::rx {} \;
</pre>

* -type : Typ angeben ( f=Datei, d=Directory)
* -exec : Befehl ausführen
* {} : Platzhalter für die find-Ergebniss
* \; : Muss ESCaped werden

== Alle extended attributes aus Verzeichnisbaum entfernen ==

<pre>
getfattr -h -R * | grep "file:" | cut -d " " -f 3- | sed -e 's/.*/\"&\"/' | xargs setfattr -x user.DOSATTRIB --
</pre>

getfattr:
* -h : Symbolische links nicht dereferenzieren
* -R : Rekursiv

grep: gibt nur die Zeilen aus, die "file:" enthalten

cut:
* -d " ": Feldtrenner ist in dem Fall das Leerzeichen (" ")
* -f 3- : Alles wegschneiden bis zum dritten Feld (in dem Fall bis zum dritten Leerzeichen)

sed:
* -e :
* 's/.*/\"&\"/' : umschließt den stdout mit Anführungszeichen

xargs:

== Nach bestimmten ACLs suchen ==

<pre>
getfacl -nRs . | awk -v RS= -v ORS='\n\n' '/group:3000040/'
</pre>

== Verzeichnisse, Unterverzeichnisse und Dateien nach uppercase umwandeln ==

<pre>
find . -exec perl-rename 'y/a-z/A-Z/' {} \;
</pre>

Paperless-ngx

2025-05-26T20:55:53Z

Chris: /* Installation */

== Installation ==

* sudo lxc-create -n paperless-ngx -B zfs -t download -- --dist debian --release bookworm --arch amd64

* dpkg-reconfigure tzdata
* dpkg-reconfigure locales
* apt install python3 python3-pip python3-dev imagemagick fonts-liberation gnupg libpq-dev default-libmysqlclient-dev pkg-config libmagic-dev mime-support libzbar0 poppler-utils apt-transport-https man-db vim bash-completion openssh-server unpaper ghostscript icc-profiles-free qpdf liblept5 libxml2 pngquant zlib1g tesseract-ocr redis tesseract-ocr-deu git locate curl wget python3.11-venv emacs-nox autoconf libtool libleptonica-dev curl wget vim sudo mariadb-server
* adduser paperless --system --home /opt/paperless-ngx --group
* curl -O -L https://github.com/paperless-ngx/paperless-ngx/releases/download/v2.1.3/paperless-ngx-v2.1.3.tar.xz
* tar -xf paperless-ngx-v2.1.3.tar.xz -C/opt/
* cd /opt/paperless-ngx
* mkdir media
* mkdir data
* mkdir consume
* chown -R paperless:paperless .
* chown paperless:paperless media
* chown paperless:paperless data
* chown paperless:paperless consume
* sudo -Hu paperless python3 -m venv venv-paperless-ngx
* sudo -Hu paperless -s
* . venv-paperless-ngx/bin/activate
* pip3 install -r requirements.txt
* cd src
* vim /opt/paperless-ngx/paperless.conf

<pre>
# Have a look at the docs for documentation.
# https://docs.paperless-ngx.com/configuration/

# Debug. Only enable this for development.

#PAPERLESS_DEBUG=false

# Required services

PAPERLESS_REDIS=redis://localhost:6379
PAPERLESS_DBENGINE=mariadb
PAPERLESS_DBHOST=shodan.intern.darkrealm.dyndns.org
PAPERLESS_DBPORT=3306
PAPERLESS_DBNAME=paperless
PAPERLESS_DBUSER=paperless
PAPERLESS_DBPASS=paperless
PAPERLESS_DBSSLMODE=DISABLED

# Paths and folders

#PAPERLESS_CONSUMPTION_DIR=../consume
#PAPERLESS_DATA_DIR=../data
#PAPERLESS_TRASH_DIR=../trash
#PAPERLESS_MEDIA_ROOT=../media
PAPERLESS_CONSUMPTION_DIR=/mnt/storage/dms/consume
PAPERLESS_DATA_DIR=/mnt/storage/dms/data
PAPERLESS_TRASH_DIR=/mnt/storage/dms/trash
PAPERLESS_MEDIA_ROOT=/mnt/storage/dms/media
PAPERLESS_STATICDIR=../static
#PAPERLESS_FILENAME_FORMAT=
PAPERLESS_FILENAME_FORMAT={{owner_username}}/{{correspondent}}/{{document_type}}/{{created_year}}/{{title}}
PAPERLESS_FILENAME_FORMAT_REMOVE_NONE=true

# Security and hosting

PAPERLESS_SECRET_KEY=Peem7AhD
PAPERLESS_URL=https://paperless.darkrealm.dyndns.org
#PAPERLESS_CSRF_TRUSTED_ORIGINS=https://example.com # can be set using PAPERLESS_URL
#PAPERLESS_ALLOWED_HOSTS=example.com,www.example.com # can be set using PAPERLESS_URL
#PAPERLESS_CORS_ALLOWED_HOSTS=https://localhost:8080,https://example.com # can be set using PAPERLESS_URL
#PAPERLESS_FORCE_SCRIPT_NAME=
#PAPERLESS_STATIC_URL=/static/
#PAPERLESS_AUTO_LOGIN_USERNAME=
#PAPERLESS_COOKIE_PREFIX=
#PAPERLESS_ENABLE_HTTP_REMOTE_USER=false

# OCR settings

PAPERLESS_OCR_LANGUAGE=eng+deu
#PAPERLESS_OCR_MODE=skip
#PAPERLESS_OCR_SKIP_ARCHIVE_FILE=never
#PAPERLESS_OCR_OUTPUT_TYPE=pdfa
#PAPERLESS_OCR_PAGES=1
#PAPERLESS_OCR_IMAGE_DPI=300
#PAPERLESS_OCR_CLEAN=clean
#PAPERLESS_OCR_DESKEW=true
#PAPERLESS_OCR_ROTATE_PAGES=true
#PAPERLESS_OCR_ROTATE_PAGES_THRESHOLD=4
#PAPERLESS_OCR_USER_ARGS={}
#PAPERLESS_CONVERT_MEMORY_LIMIT=0
#PAPERLESS_CONVERT_TMPDIR=/var/tmp/paperless

PAPERLESS_ENABLE_NLTK=true
#PAPERLESS_NLTK_DIR=../nltk_data
PAPERLESS_NLTK_DIR=/mnt/storage/dms/nltk_data
# Software tweaks

#PAPERLESS_TASK_WORKERS=1
#PAPERLESS_THREADS_PER_WORKER=1
PAPERLESS_TIME_ZONE=Europe/Berlin
#PAPERLESS_CONSUMER_POLLING=10
#PAPERLESS_CONSUMER_DELETE_DUPLICATES=false
PAPERLESS_CONSUMER_RECURSIVE=true
#PAPERLESS_CONSUMER_IGNORE_PATTERNS=[".DS_STORE/*", "._*", ".stfolder/*", ".stversions/*", ".localized/*", "desktop.ini"]
#PAPERLESS_CONSUMER_SUBDIRS_AS_TAGS=false
PAPERLESS_CONSUMER_ENABLE_BARCODES=true
#PAPERLESS_CONSUMER_BARCODE_STRING=PATCHT
#PAPERLESS_CONSUMER_BARCODE_UPSCALE=0.0
#PAPERLESS_CONSUMER_BARCODE_DPI=300
#PAPERLESS_CONSUMER_ENABLE_COLLATE_DOUBLE_SIDED=false
#PAPERLESS_CONSUMER_COLLATE_DOUBLE_SIDED_SUBDIR_NAME=double-sided
#PAPERLESS_CONSUMER_COLLATE_DOUBLE_SIDED_TIFF_SUPPORT=false
#PAPERLESS_PRE_CONSUME_SCRIPT=/path/to/an/arbitrary/script.sh
#PAPERLESS_POST_CONSUME_SCRIPT=/path/to/an/arbitrary/script.sh
#PAPERLESS_FILENAME_DATE_ORDER=YMD
#PAPERLESS_FILENAME_PARSE_TRANSFORMS=[]
#PAPERLESS_NUMBER_OF_SUGGESTED_DATES=5
#PAPERLESS_THUMBNAIL_FONT_NAME=
#PAPERLESS_IGNORE_DATES=
#PAPERLESS_ENABLE_UPDATE_CHECK=

# Tika settings

PAPERLESS_TIKA_ENABLED=true
PAPERLESS_TIKA_ENDPOINT=http://localhost:9998
PAPERLESS_TIKA_GOTENBERG_ENDPOINT=http://localhost:3000

# Binaries

#PAPERLESS_CONVERT_BINARY=/usr/bin/convert
#PAPERLESS_GS_BINARY=/usr/bin/gs
</pre>

* DB und User erstellen

* sudo -Hu paperless python3 manage.py migrate
* sudo -Hu paperless python3 manage.py createsuperuser
* cp -av ../scripts/*.service /etc/systemd/system/
* cp -av ../scripts/*.socket /etc/systemd/system/
* systemctl daemon-reload

ImageMagick Policy in /etc/ImageMagick-6/policy.xml anpassen:
<pre>
<policy domain="coder" rights="read|write" pattern="PDF" />
</pre>

JBIG2ENC compilen:
* git clone https://github.com/agl/jbig2enc
* cd jbig2enc
* ./autogen.sh
* ./configure
* make
* make install

NLTK installieren
* sudo -Hu paperless -s
* . venv-paperless-ngx/bin/activate
* python3
* import nltk
* nltk.download() (Nach /opt/paperless-ngx/nltk_data herunterladen)
* punkt, snowball_data und stopwords runterladen

Apache Tika und Gotenberg installieren (momentan über Docker)
* apt install docker-compose
* docker run -d -p 127.0.0.1:9998:9998 apache/tika
* docker run -d -p 127.0.0.1:3000:3000 gotenberg/gotenberg:8

== Update ==

* wget aktuellstes tar.xf
* systemctl stop paperless-*
* tar -xf <release.tar.xf> -C/opt/
* cd /opt/paperless-ngx
* cp -av paperless.conf.keep paperless.conf
* chown paperless:paperless . -R
* sudo -Hu paperless -s
* source /opt/paperless-ngx/venv-paperless-ngx/bin/activate
* pip install -r requirements.txt
* cd src
* python3 manage.py migrate

== Fixes ==

=== NLTK Fix ===

<pre>
sudo -Hu paperless -s
source /opt/paperless-ngx/venv-paperless-ngx/bin/activate
cd src
python3 -W ignore::RuntimeWarning -m nltk.downloader -d "/mnt/storage/dms/nltk_data" punkt_tab
</pre>

=== Logs show "possible incompatible database column" when deleting documents ===

You may see errors when deleting documents like:

<pre>
Data too long for column 'transaction_id' at row 1
</pre>

This error can occur in installations which have upgraded from a version of Paperless-ngx that used Django 4 (Paperless-ngx versions prior to v2.13.0) with a MariaDB/MySQL database. Due to the backawards-incompatible change in Django 5, the column "documents_document.transaction_id" will need to be re-created, which can be done with a one-time run of the following management command:

<pre>
sudo -Hu paperless -s
source /opt/paperless-ngx/venv-paperless-ngx/bin/activate
cd src
python3 manage.py convert_mariadb_uuid
</pre>

== Komplette Neuinstallation inkl. Komplettrestore ==

* lxc init images:debian/12 paperless-ngx
* lxc start papalass
* lxc shell papalass
* passwd

Anschliessend vorgehen wie bei [[#Installation]]

* die gesicherte ZIP-Datei des exports auf den neuen Paperless-Server kopieren und nach /temp entpacken
* sudo -Hu paperless -s
* source /opt/paperless-ngx/venv-paperless-ngx/bin/activate
* cd src
* python3 manage.py document_importer /temp/

Tomb Raider Times Exclusive Level DC Conversion

2025-05-14T19:26:12Z

Chris:

== Files replaced ==

* script.dat
* data/cutseq.pak

== Files inserted ==

* english.dat
* data/office.tr4
* data/times.tr4

== Dummy Texture Files ==

* data/office.tex
* data/times.tex

Tomb Raider Times Exclusive Level DC Conversion

2025-05-14T19:22:00Z

Chris: /* Files inserted */

== Files replaced ==

* script.dat
* data/cutseq.pak

== Files inserted ==

* english.dat
* data/office.tr4
* data/times.tr4

Tomb Raider Times Exclusive Level DC Conversion

2025-05-14T19:20:22Z

Chris:

== Files replaced ==

* script.dat
* data/cutseq.pak

== Files inserted ==

* english.dat

Tomb Raider Times Exclusive Level DC Conversion

2025-05-14T19:20:05Z

Chris: Die Seite wurde neu angelegt: „== Files replaced == script.dat data/cutseq.pak == Files inserted == english.dat“

== Files replaced ==

script.dat
data/cutseq.pak

== Files inserted ==

english.dat

Dreamcast

2025-05-14T18:29:21Z

Chris: /* Create CDI Images from ISO */

== CDR Sizes ==

80min/700MB: 1st Session 674MB Maximum

90min/800MB: max 748MB when 45000 Sectors in 1st Session with dummy burn

90min/800MB: max ~735MB when 45000 Sectors in 1st Session

== Access Files on CDI Images ==

# Extract the seperate Track first

<pre>
cdirip <CDIFILE>.cdi
</pre>

# In case of 2nd Data Track, fix the LBA first

<pre>
isofix <TRACK02>.iso <LBA Value reported by cdirip>
</pre>

== What binhack does ==

Changes LBA Value at offset 1461904 ( 8 Bytes before CD001 string) from 6EB0 (45166) to LBA Value+166 specified during binhack.

Adds Region Codes for JAP,EUR and USA and strings for that.
Adds some bootcode, probably SH4 ASM to make the GDROM accept CDRs.

== What cdda does ==

Changes some bytes, probably SH4 ASM to enable playback of CDDA tracks.

== What dahack does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to 9600 (150). Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== What hack3 does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to LBA Value+150 specified during dahack. Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== Another Copy Protection ==

Searching for CDE4 036A (or CDE4 436A) and replacing with 0900 0900 is another Copy Protection defeat.

CDE4 436A -> 0900 0900
1032 0D8B -> 0800 0D8B

== LBA 45000 Trick ==

69120000 bytes in first session will result in LBA 45000.

== Fill 1st Session ISO File to 69120000 bytes with zeros ==

6912000 bytes = 33750 2048-byte-sectors

CORRECTION : Must be 33598 Sectors !!! = 68808703 bytes

1st Session Leadout = 11250 Sectors = 23040000 Bytes

2nd Session will start at Sector 45000

When using extract.exe, add 150 Sectors to LBA reported in gdinfo_FG.exe track05

<pre>
dd if=/dev/zero of=<ISOFILE> bs=1 count=1 seek=69119999
</pre>

== Creating & Burning DATA/DATA ISOs ==

== Merging track03 & track05 into one file ==

Create dummy file first
Calculate 1185760800 - <SIZE OF track03.bin + SIZE OF track05.bin>

<pre>
dd if=/dev/zero bs=1 count=0 seek=<RESULT> of=dummy.bin
cat track03.bin dummy.bin track05.bin > combined.bin
</pre>

This also helps when extract.exe only outputs garbage out of track03.iso along with track05.iso. This way only the combined.iso needs to be extracted.

== Burning DATA/DATA CDIs ==

<pre>
cdirip <CDIFile> <Destination Directory>
cdrecord dev=/dev/sr0 -v -multi -tao -xa tdata01.iso
cdrecord dev=/dev/sr0 -v -tao -xa tdata02.iso
</pre>

== Full rip process from GD-ROM to CD-ROM ==

=== Get GDROM Data from Dreamcast with http-ack ===

https://www.gotwalls.com

https://github.com/sega-dreamcast/httpd-ack

# Boot HTTP-ack from CD
# Swap HTTP-ack CD with Original GD-ROm
# Connect via Browser to http://192.168.1.30
# Download all files starting with track03.bin
# Download disc.gdi

=== Extract Data from Tracks ===

==== Use gditools ====

https://sourceforge.net/projects/dcisotools/

Either use GUI or gditools.py

<pre>
python2.7 ./gditools.py
Usage: ./gditools.py -i input_gdi [options]

-h, --help Display this help
-l, --list List all files in the filesystem and exit
-o [outdir] Output directory. Default: gdi folder
-s [filename] Create a sorttxt file with custom name
(It uses *data-folder* as prefix)
-b [ipname] Dump the ip.bin with custom name
-e [filename] Dump a single file from the filesystem
--extract-all Dump all the files in the *data-folder*
--data-folder [name] *data-folder* subfolder. Default: data
(__volume_label__ --> Use ISO9660 volume label)
--sort-spacer [num] Sorttxt entries are sperated by num
--silent Minimal verbosity mode
[no option] Display gdi infos if not silent

gditools.py by FamilyGuy, http://sourceforge.net/p/dcisotools/
Licensed under GPLv3, see licences folder.

iso9660.py by Barney Gale, http://github.com/barneygale
Licensed under a BSD-based license, see licences folder.
</pre>

<pre>
python2.7 ./gditools.py -i <disc.gdi> --extract-all -o /mnt/development/dreamcast/rips/orgs/<GAME_NAME>/<CD Number> --data-folder __volume_label__ -s sort.txt -b ip.bin
</pre>

==== Old way ====
# If the game contains more than track03.bin, combine the 2 binfiles into 1:
<pre>
cat track03.bin dummy.bin track05.bin > <BINFILE>
</pre>
# Convert .BIN Files to .ISO Files first with bin2iso.exe
<pre>
bin2iso.exe <BINFILE> <ISOFILE>
</pre>
# Extract Data from .ISO files
<pre>
extract.exe <ISOFILE>
</pre>
If extracted 1ST_READ.BIN has filesize of 0, LBA of track05 is wrong.

=== Downsampling ===

==== AFS Files ====

https://sizious.com/download/dreamcast/

AFS files contain multiple ADX files combined into one AFS file.

compact AFS with:

<pre>
AFSCompactor.exe
</pre>

for Alone in the Dark - The New Nightmare .XA files are AFS files, rename first with:

<pre>
rename .XA .afs *.XA
</pre>

then compact and rename back to .XA:

<pre>
rename .afs .XA *.afs
</pre>

AFSCompactor.exe will rename files inside .afs to file0001 etc... AITD4 doesnt seem to care.

extract AFS with:

<pre>
afs_extract.exe *.afs|*.XA
</pre>
or
<pre>
afsextract.exe -sc afsfile|XAfile
</pre>
or
<pre>
afsutils_v2.exe | rename .XA files to .afs beforehand.
</pre>

==== SFD Movies ====

http://www.dc-swat.ru/page/download/

Demux video to .m1v and audio to .wav first

Example (play around with bitrates):

<pre>
for f in <SFD Files>; do ffmpeg -i $f -vcodec mpeg1video -b:v 2600k -minrate 2600k -maxrate 3600k -y $f.m1v -acodec adpcm_ima_wav -y $f.wav ; done
</pre>

Open all .wavs in single Audacity project and mass export again, 16bit and 44100 Hz.

Start Dreamcast Movie Converter and remux/reencode to .SFD. Dont use batch processing, videos get truncated!

==== 4XM Movies ====

http://samples.mplayerhq.hu/game-formats/4xm/

Convert to AVI first with ffmpeg

<pre>
for i in ../rips/unmodified/aitd4/cd2/AITD4_GD_2/MOVIES/*.4XM; do ffmpeg -i "$i" -vf scale=320x160 -y "${i%.*}.avi"; done
</pre>

Downsample with 4xm_comp.exe

Speed : 30 fps
Data Rate : 110000
Resize Width 320 Height 160

=== Create ISO Images ===

For the 1st Session, chosen files must not exceed total of 69120000 bytes!

==== 1st Session ====

===== Create first ISOFile =====

<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <CHOSEN FILE> <CHOSEN FILE> <CHOSEN FILE>...
</pre>

Fill first ISOFile with zeros so it takes up exactly 68808703 bytes = LBA 45000

<pre>
dd if=/dev/zero of=<OUTPUT1.ISO> bs=1 count=1 seek=68808703
</pre>

===== Burn 1st session =====

<pre>
cdrecord -v -multi -tao -xa <OUTPUT1.ISO>
</pre>

Check MSINFO value & TOC

<pre>
cdrecord -msinfo

0,45000
</pre>

<pre>
cdrecord -toc

first: 1 last 1
track: 1 lba: 0 ( 0) 00:02:00 adr: 1 control: 4 mode: -1
track:lout lba: 33600 ( 134400) 07:30:00 adr: 1 control: 4 mode: -1
</pre>

To include subdirectories into the root directory of the CD, use
<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <SUBDIRECTORY>/=<PATH_OF_SUBDIRECTORY>
</pre>
Example:
<pre>
mkisofs -no-allow-lowercase -graft-points -l -duplicates-once -V AITD4_GD_2 -o aitd4_d2_1.iso MOVIES/=../rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES/
</pre>

==== 2nd Session ====

===== Hack 1ST_READ.BIN & IP.BIN =====

Get MSINFO Value

<pre>
cdrecord -msinfo
</pre>

<pre>
binhack

Enter name of binary: 1ST_READ.BIN
Enter name of bootsector: ip.bin
Enter msinfo value: <MSINFO Value>
File 1ST_READ.BIN successfully hacked.
Bincon detected, OS flag set to 0.
File ip.bin successfully created.
</pre>

===== Insert custom Logo =====

<pre>
logoinsert darkrealm.mr IP.BIN
</pre>

===== Burn 2nd Session =====

<pre>
mkisofs -no-allow-lowercase -l -duplicates-once -V <DISCNAME> -G <IP.BIN> -M <OUTPUT1.ISO> -C <MSINFO VALUE> -o <OUTPUT2.ISO> <GAME_DATA_DIRECTORY>
cdrecord -v -tao -xa <OUTPUT2.ISO>
</pre>

== Testing ==

* Burn whole stuff to a CD-RW
* Generate CDI from CD-RW
** Launch DiscJuggler from emulated Windows Machine and connect via iSCSI
* Test CDI with ReiCast oder ReDream

== Create CDI Images from ISO ==

* Create ISO first

<pre>
mkisofs -l -duplicates-once -graft-points -V <DISCNAME> -G <IP.BIN> -o <GAME.ISO> <GAME DIRECTORY>
</pre>

* Create DATA/DATA CDI

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

* or create AUDIO/DATA CDI, with 2nd Session starting at LBA 11702
** Note: 1ST_READ.BIN must eventually be binhacked to LBA 11702

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

Dreamcast

2025-05-14T18:28:55Z

Chris:

== CDR Sizes ==

80min/700MB: 1st Session 674MB Maximum

90min/800MB: max 748MB when 45000 Sectors in 1st Session with dummy burn

90min/800MB: max ~735MB when 45000 Sectors in 1st Session

== Access Files on CDI Images ==

# Extract the seperate Track first

<pre>
cdirip <CDIFILE>.cdi
</pre>

# In case of 2nd Data Track, fix the LBA first

<pre>
isofix <TRACK02>.iso <LBA Value reported by cdirip>
</pre>

== What binhack does ==

Changes LBA Value at offset 1461904 ( 8 Bytes before CD001 string) from 6EB0 (45166) to LBA Value+166 specified during binhack.

Adds Region Codes for JAP,EUR and USA and strings for that.
Adds some bootcode, probably SH4 ASM to make the GDROM accept CDRs.

== What cdda does ==

Changes some bytes, probably SH4 ASM to enable playback of CDDA tracks.

== What dahack does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to 9600 (150). Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== What hack3 does ==

Changes LBA Values at offsets 1301408,1302380 and 1461904. First 2 offsets get rewritten from 5EB0 (45150) to LBA Value+150 specified during dahack. Third offset gets rewritten from 6EB0 (45166) to LBA Value+166 specified during dahack. 3rd offset change is basically the same what binhack does.

== Another Copy Protection ==

Searching for CDE4 036A (or CDE4 436A) and replacing with 0900 0900 is another Copy Protection defeat.

CDE4 436A -> 0900 0900
1032 0D8B -> 0800 0D8B

== LBA 45000 Trick ==

69120000 bytes in first session will result in LBA 45000.

== Fill 1st Session ISO File to 69120000 bytes with zeros ==

6912000 bytes = 33750 2048-byte-sectors

CORRECTION : Must be 33598 Sectors !!! = 68808703 bytes

1st Session Leadout = 11250 Sectors = 23040000 Bytes

2nd Session will start at Sector 45000

When using extract.exe, add 150 Sectors to LBA reported in gdinfo_FG.exe track05

<pre>
dd if=/dev/zero of=<ISOFILE> bs=1 count=1 seek=69119999
</pre>

== Creating & Burning DATA/DATA ISOs ==

== Merging track03 & track05 into one file ==

Create dummy file first
Calculate 1185760800 - <SIZE OF track03.bin + SIZE OF track05.bin>

<pre>
dd if=/dev/zero bs=1 count=0 seek=<RESULT> of=dummy.bin
cat track03.bin dummy.bin track05.bin > combined.bin
</pre>

This also helps when extract.exe only outputs garbage out of track03.iso along with track05.iso. This way only the combined.iso needs to be extracted.

== Burning DATA/DATA CDIs ==

<pre>
cdirip <CDIFile> <Destination Directory>
cdrecord dev=/dev/sr0 -v -multi -tao -xa tdata01.iso
cdrecord dev=/dev/sr0 -v -tao -xa tdata02.iso
</pre>

== Full rip process from GD-ROM to CD-ROM ==

=== Get GDROM Data from Dreamcast with http-ack ===

https://www.gotwalls.com

https://github.com/sega-dreamcast/httpd-ack

# Boot HTTP-ack from CD
# Swap HTTP-ack CD with Original GD-ROm
# Connect via Browser to http://192.168.1.30
# Download all files starting with track03.bin
# Download disc.gdi

=== Extract Data from Tracks ===

==== Use gditools ====

https://sourceforge.net/projects/dcisotools/

Either use GUI or gditools.py

<pre>
python2.7 ./gditools.py
Usage: ./gditools.py -i input_gdi [options]

-h, --help Display this help
-l, --list List all files in the filesystem and exit
-o [outdir] Output directory. Default: gdi folder
-s [filename] Create a sorttxt file with custom name
(It uses *data-folder* as prefix)
-b [ipname] Dump the ip.bin with custom name
-e [filename] Dump a single file from the filesystem
--extract-all Dump all the files in the *data-folder*
--data-folder [name] *data-folder* subfolder. Default: data
(__volume_label__ --> Use ISO9660 volume label)
--sort-spacer [num] Sorttxt entries are sperated by num
--silent Minimal verbosity mode
[no option] Display gdi infos if not silent

gditools.py by FamilyGuy, http://sourceforge.net/p/dcisotools/
Licensed under GPLv3, see licences folder.

iso9660.py by Barney Gale, http://github.com/barneygale
Licensed under a BSD-based license, see licences folder.
</pre>

<pre>
python2.7 ./gditools.py -i <disc.gdi> --extract-all -o /mnt/development/dreamcast/rips/orgs/<GAME_NAME>/<CD Number> --data-folder __volume_label__ -s sort.txt -b ip.bin
</pre>

==== Old way ====
# If the game contains more than track03.bin, combine the 2 binfiles into 1:
<pre>
cat track03.bin dummy.bin track05.bin > <BINFILE>
</pre>
# Convert .BIN Files to .ISO Files first with bin2iso.exe
<pre>
bin2iso.exe <BINFILE> <ISOFILE>
</pre>
# Extract Data from .ISO files
<pre>
extract.exe <ISOFILE>
</pre>
If extracted 1ST_READ.BIN has filesize of 0, LBA of track05 is wrong.

=== Downsampling ===

==== AFS Files ====

https://sizious.com/download/dreamcast/

AFS files contain multiple ADX files combined into one AFS file.

compact AFS with:

<pre>
AFSCompactor.exe
</pre>

for Alone in the Dark - The New Nightmare .XA files are AFS files, rename first with:

<pre>
rename .XA .afs *.XA
</pre>

then compact and rename back to .XA:

<pre>
rename .afs .XA *.afs
</pre>

AFSCompactor.exe will rename files inside .afs to file0001 etc... AITD4 doesnt seem to care.

extract AFS with:

<pre>
afs_extract.exe *.afs|*.XA
</pre>
or
<pre>
afsextract.exe -sc afsfile|XAfile
</pre>
or
<pre>
afsutils_v2.exe | rename .XA files to .afs beforehand.
</pre>

==== SFD Movies ====

http://www.dc-swat.ru/page/download/

Demux video to .m1v and audio to .wav first

Example (play around with bitrates):

<pre>
for f in <SFD Files>; do ffmpeg -i $f -vcodec mpeg1video -b:v 2600k -minrate 2600k -maxrate 3600k -y $f.m1v -acodec adpcm_ima_wav -y $f.wav ; done
</pre>

Open all .wavs in single Audacity project and mass export again, 16bit and 44100 Hz.

Start Dreamcast Movie Converter and remux/reencode to .SFD. Dont use batch processing, videos get truncated!

==== 4XM Movies ====

http://samples.mplayerhq.hu/game-formats/4xm/

Convert to AVI first with ffmpeg

<pre>
for i in ../rips/unmodified/aitd4/cd2/AITD4_GD_2/MOVIES/*.4XM; do ffmpeg -i "$i" -vf scale=320x160 -y "${i%.*}.avi"; done
</pre>

Downsample with 4xm_comp.exe

Speed : 30 fps
Data Rate : 110000
Resize Width 320 Height 160

=== Create ISO Images ===

For the 1st Session, chosen files must not exceed total of 69120000 bytes!

==== 1st Session ====

===== Create first ISOFile =====

<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <CHOSEN FILE> <CHOSEN FILE> <CHOSEN FILE>...
</pre>

Fill first ISOFile with zeros so it takes up exactly 68808703 bytes = LBA 45000

<pre>
dd if=/dev/zero of=<OUTPUT1.ISO> bs=1 count=1 seek=68808703
</pre>

===== Burn 1st session =====

<pre>
cdrecord -v -multi -tao -xa <OUTPUT1.ISO>
</pre>

Check MSINFO value & TOC

<pre>
cdrecord -msinfo

0,45000
</pre>

<pre>
cdrecord -toc

first: 1 last 1
track: 1 lba: 0 ( 0) 00:02:00 adr: 1 control: 4 mode: -1
track:lout lba: 33600 ( 134400) 07:30:00 adr: 1 control: 4 mode: -1
</pre>

To include subdirectories into the root directory of the CD, use
<pre>
mkisofs -l -duplicates-once -no-allow-lowercase -graft-points -V <DISCNAME> -o <OUTPUT1.ISO> <SUBDIRECTORY>/=<PATH_OF_SUBDIRECTORY>
</pre>
Example:
<pre>
mkisofs -no-allow-lowercase -graft-points -l -duplicates-once -V AITD4_GD_2 -o aitd4_d2_1.iso MOVIES/=../rips/modified/aitd4/cd2/AITD4_GD_2/MOVIES/
</pre>

==== 2nd Session ====

===== Hack 1ST_READ.BIN & IP.BIN =====

Get MSINFO Value

<pre>
cdrecord -msinfo
</pre>

<pre>
binhack

Enter name of binary: 1ST_READ.BIN
Enter name of bootsector: ip.bin
Enter msinfo value: <MSINFO Value>
File 1ST_READ.BIN successfully hacked.
Bincon detected, OS flag set to 0.
File ip.bin successfully created.
</pre>

===== Insert custom Logo =====

<pre>
logoinsert darkrealm.mr IP.BIN
</pre>

===== Burn 2nd Session =====

<pre>
mkisofs -no-allow-lowercase -l -duplicates-once -V <DISCNAME> -G <IP.BIN> -M <OUTPUT1.ISO> -C <MSINFO VALUE> -o <OUTPUT2.ISO> <GAME_DATA_DIRECTORY>
cdrecord -v -tao -xa <OUTPUT2.ISO>
</pre>

== Testing ==

* Burn whole stuff to a CD-RW
* Generate CDI from CD-RW
** Launch DiscJuggler from emulated Windows Machine and connect via iSCSI
* Test CDI with ReiCast oder ReDream

== Create CDI Images from ISO ==

* Create ISO first

<pre>
mkisofs -l -duplicates-once -graft-points -V <DISCNAME> -G <IP.BIN> -o <GAME.ISO> <GAME DIRECTORY>
</pre>

* Create DATA/DATA CDI

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

* or create AUDIO/DATA CDI, with 2nd Session starting at LBA 11702
** Note: 1ST_READ.BIN must eventually be binacked to LBA 11702

<pre>
wine cdi4dc.exe <GAME.ISO> /media/chris/development/dreamcast/backups/CDI/tests/<GAME.CDI> -d
</pre>

Funktionierende Toolchains

2025-04-18T22:17:58Z

Chris:

== KallistiOS Toolchain ==

Voraussetzungen:

* Wenn die Toolchain innerhalb eines CIFS-Mounts installiert wird müssen '''mfsymlinks''' in den Mountoptions aktiviert sein!

=== Crosscompiler bauen ===
<pre>
mkdir /media/chris/development/dreamcast/toolchains
cd /media/chris/development/dreamcast/toolchains
git clone https://github.com/KallistiOS/KallistiOS.git kos
cd kos/utils/dc-chain/
cp -av Makefile.default.cfg Makefile.cfg
</pre>

==== Makefile.cfg bearbeiten ====

<pre>
# Sega Dreamcast Toolchains Maker (dc-chain)
# This file is part of KallistiOS.

#########################
### TOOLCHAIN PROFILE ###
#########################

# Choose a toolchain profile from the following available options:
# No longer supported upstream:
# - 9.3.0-legacy: Former 'stable' option, based on GCC 9.3.0 and Newlib 3.3.0.
# - 9.5.0-winxp: Most recent versions of tools which run on Windows XP.
# - 10.5.0: Last release in the GCC 10 series, released 2023-07-07.
# - 11.5.0: Last release in the GCC 11 series, released 2024-07-19.
# Supported upstream:
# - 12.4.0: Latest release in the GCC 12 series, released 2024-06-20.
# - stable: Tested stable; based on GCC 13.2.0, released 2023-07-27.
# - 13.3.0: Latest release in the GCC 13 series, released 2024-05-21.
# - 14.2.0: Latest release in the GCC 14 series, released 2024-08-01.
# Development versions:
# - 13.3.1-dev Bleeding edge GCC 13 series from git.
# - 14.2.1-dev Bleeding edge GCC 14 series from git.
# - 15.0.0-dev Bleeding edge GCC 15 series from git.
# - gccrs-dev: GCC fork for development of the GCCRS Rust compiler.
# - rustc-dev: GCC fork for development of the libgccjit rustc GCC codegen.
# If unsure, select stable. See README.md for more detailed descriptions.
toolchain_profile=stable

########################
### DOWNLOAD OPTIONS ###
########################

### Download protocol (http|https|ftp)
# Specify the protocol you want to use for downloading package files.
download_protocol=https

### Force downloader (curl|wget)
# Specify here if you'd prefer to use 'wget' or 'curl'. If neither is specified,
# a web downloader tool will be auto-detected in the following order: cURL, Wget
# You must have either Wget or cURL installed to use dc-chain.
#force_downloader=wget

### Specify GNU mirror override
# The default mirror for GNU sources is 'ftpmirror.gnu.org'
# This setting overrides the default mirror with a preferred mirror.
#gnu_mirror=mirrors.kernel.org

#####################
### BUILD OPTIONS ###
#####################

### SH toolchain install path
# Specify the directory where the SH toolchain will be installed. This setting
# must match the KOS_CC_BASE setting in your KOS "environ.sh" configuration.
sh_toolchain_path=/media/chris/development/dreamcast/toolchains/sh-elf

### ARM toolchain install path
# Specify the directory where the ARM toolchain will be installed. This setting
# must match the DC_ARM_BASE setting in your KOS "environ.sh" configuration.
arm_toolchain_path=/media/chris/development/dreamcast/toolchains/arm-eabi

### Make jobs (n|<empty>)
# Set this value to the number of parallel jobs you want to run with make.
# For fastest toolchain building, set to the number of CPU threads available.
# Using multiple jobs may cause issues in certain environments and may be
# automatically disabled. If you encounter errors building your toolchain,
# reduce the number of jobs to 1 to avoid issues and ease troubleshooting.
makejobs=16

### Verbose (1|0)
# Choose whether to actively display compilation messages on the screen.
# Messages are saved to the build log files regardless of this setting.
verbose=1

### Erase (1|0)
# Erase build directories as toolchain components are installed to save space.
erase=1

### Install toolchain debug symbols (1|0)
# Choose whether to keep the debugging symbols for the toolchain.
# This is only useful if you wish to debug the toolchain itself.
#toolchain_debug=1

########################
### LANGUAGE OPTIONS ###
########################

### Enable C++
# Builds C++ support, including the C++ compiler and its standard library. The
# vast majority of language features are supported, with C++23 and early C++26
# support included. KallistiOS provides several examples with C++ support, so it
# is enabled by default. Adding C++ support requires extra disk space and
# compilation time, so you may disable it here if you do not plan on using C++.
enable_cpp=1

### Enable Objective-C
# Builds Objective-C support. Currently, this builds the Obj-C compiler and
# runtime, but does not build a standard library. The GNUStep Base Library is
# currently in active development for future inclusion in the kos-ports library.
# KallistiOS provides examples with Obj-C support, so it is enabled by default.
enable_objc=1

### Enable Objective C++
# Builds Objective C++ support. This hybrid language allows both C++ and Obj-C
# methods to be called from both contexts. Enabled by default as both the C++
# and Obj-C languages are already enabled by default.
enable_objcpp=1

### Enable D
# Builds D support. This will build the D compiler, but does not build the
# Phobos Runtime Library. D support may only be enabled on POSIX platforms which
# have its external dependencies provided through a host DMD, GDC, or LDC
# compiler installation.
#enable_d=1

### Enable Ada
# Builds Ada support. This will build the GNAT Ada compiler and tools but does
# not build the libada runtime library. In order for this build to succeed, the
# host's GCC version must match the target Dreamcast version, plus the same
# version of GNAT and its tools must be preinstalled (and in your path).
#enable_ada=1

### Enable Rust
# Builds the work-in-progress GCCRS Rust compiler frontend for GCC. Requires a
# development version of GCC. Use with the gccrs-dev toolchain profile for the
# most recent updates to the GCCRS compiler. This option is not to be confused
# with using this toolchain as a backend for rustc; see libgccjit below.
#enable_rust=1

### Enable libgccjit
# Enables the libgccjit embeddable GCC library. This is most commonly enabled to
# allow using this toolchain as a backend for the official rustc compiler. Use
# with the rustc-dev toolchain profile if this is your use case.
#enable_libgccjit=1

###################
### GCC OPTIONS ###
###################

### Floating point precision support (m4-single-only|m4-single|m4)
# Build support for various SH4 floating-point operation ABIs. KallistiOS only
# officially supports single-precision-only mode (m4-single-only); however,
# experimental support for single-precision-default mode (m4-single) has been
# added to allow for the use of full 64-bit doubles. You may also include
# double-precision-default (m4) which is untested and unsupported, or you may
# choose to specify only m4-single-only to save disk space.
precision_modes=m4-single-only,m4-single

### Default floating point mode (m4|m4-single|m4-single-only)
# Choose the default floating point precision ABI used when GCC is invoked. This
# may be overridden by passing -m4-single-only, -m4-single, or -m4 to GCC.
# KOS currently only officially supports m4-single-only, so it is the default.
default_precision=m4-single-only

### GCC threading model (single|kos)
# KallistiOS patches to GCC provide a 'kos' thread model, which should be used.
# If you want to disable threading support for C++, Objective-C, and so forth,
# you can set this option to 'single'.
thread_model=kos

### Automatic patching for KallistiOS (1|0)
# Uncomment this option if you want to disable applying KallistiOS patches to
# toolchain source files before building. This will disable usage of the 'kos'
# thread model. Only do this if you understand what you are doing.
#use_kos_patches=0

### Disable GCC Native Language Support (1|0)
# By default, NLS allows GCC to output diagnostics in non-English languages.
# Uncomment this option to disable NLS and force GCC to output in English.
#disable_nls=1

######################
### NEWLIB OPTIONS ###
######################

### Automatic patching for Newlib (1|0)
# Uncomment this option if you want to disable the automatic patching of Newlib
# needed by KallistiOS. This will keep the generated toolchain completely raw.
# This will also disable the 'kos' thread model. Only do this if you understand
# what you are doing.
#auto_fixup_sh4_newlib=0

### C99 format specifier support (1|0)
# Define this to build Newlib with additional support for the C99 format
# specifiers, used by printf and friends. These include support for size_t,
# ptrdiff_t, intmax_t, and sized integral types.
newlib_c99_formats=1

### Multibyte character set support (1|0)
# Define this to build Newlib with additional multibyte support. This enables
# three special locales: "C-JIS", "C-SJIS", and "C-EUCJP". The multibyte
# support extends to the mb and wc functions in stdlib as well as format
# characters for the printf and scanf family of routines.
#newlib_multibyte=1

### iconv() character encoding conversions support (encoding list)
# Define a list here to enable support for the iconv() function and <iconv.h>
# header file. The given comma separated list defines for which encoding types
# to include bidirectional conversion support. For the full list of available
# encodings, see the Newlib configuration documentation.
#newlib_iconv_encodings=us_ascii,utf8,utf16,ucs_2_internal,ucs_4_internal

### Optimize Newlib for space (1|0)
# Uncomment this option to optimize for space when building Newlib. This will
# build Newlib with compiler flags which favor smaller code sizes over faster
# performance.
#newlib_opt_space=1

###################
### C++ OPTIONS ###
###################

### Timezone database support (1|0|path)
# Uncomment this option to enable building support for C++'s std::chrono::tzdb
# into your C++ standard library by using a combination of the system's local
# timezone DB and one dynamically fetched from the "IANA Time Zone Database."
# Without support enabled, only the "UTC" and "GMT" timezones will be defined.
# You can optionally provide the path to a directory with a custom "tzdata.zi"
# database file. NOTE: Enabling this will result in larger C++ binaries!
#libstdcxx_tzdb=1

#######################
### WINDOWS OPTIONS ###
#######################

### MinGW/MSYS
# Standalone binaries (1|0)
# Uncomment this option if you want static binaries that are standalone and
# require no dependencies. When this option is used, binaries can be run outside
# the MinGW/MSYS environment. This is NOT recommended; only do this if you know
# what you are doing.
#standalone_binary=1

### Force installation of BFD for SH (1|0)
# Uncomment this option if you want to force installation of 'libbfd' for the SH
# toolchain. This is required for MinGW/MSYS and can't be disabled in this
# scenario. This option is here mainly if you want to force the installation of
# 'libbfd' under other environments, but this won't be necessary in most cases,
# as 'libelf' is used almost everywhere. Please note, 'libbfd' is not portable
# if you built it on another environment. Only use this flag if you understand
# what you are doing.
#sh_force_libbfd_installation=1
</pre>

Make durchführen:
<pre>
PATH=/usr/bin make all
</pre>

==== KOS bauen ====

eviron.sh:
<pre>
# KallistiOS Environment Settings
#
# This is a sample script for configuring and customizing your
# KOS build environment. Modify it to suit your setup. Several
# settings may be enabled optionally or are provided with
# alternative values.
#
# This script is typically sourced in your current shell environment
# (probably by .bashrc, .bash_profile, or something similar), so that
# the KOS environment is set up automatically for each shell session.
#

# Build Architecture
#
# Set the major architecture you'll be building for.
# The only option here is "dreamcast" as of KOS 2.0.0.
#
export KOS_ARCH="dreamcast"

# Build Sub-Architecture
#
# Defines the sub architecture your configuration
# is targeting or uses an existing value that
# can be set externally via your IDE.
#
# Valid values:
# "pristine" - Dreamcast console or HKT-0120 devkit (default)
# "naomi" - NAOMI or NAOMI 2 arcade board
#
if [ -z "${KOS_SUBARCH}" ] ; then
export KOS_SUBARCH="pristine"
else
export KOS_SUBARCH
fi

# KOS Root Path
#
# Specifies the path to the KOS root directory
#
export KOS_BASE="/media/chris/development/dreamcast/toolchains/kos"

# KOS-Ports Path
#
# Specifies the path to the KOS-ports directory
#
export KOS_PORTS="${KOS_BASE}/../kos-ports"

# SH Compiler Prefixes
#
# Specifies the path to and prefix for the main SH
# GCC toolchain used to target the Dreamcast's SH4 CPU.
#
export KOS_CC_BASE="/media/chris/development/dreamcast/toolchains/sh-elf"
export KOS_CC_PREFIX="sh-elf"

# ARM Compiler Prefixes
#
# Specifies the path to and prefix for the additional ARM
# GCC toolchain used to target the Dreamcast's AICA SPU.
#
export DC_ARM_BASE="/media/chris/development/dreamcast/toolchains/arm-eabi"
export DC_ARM_PREFIX="arm-eabi"

# External Dreamcast Tools Path
#
# Specifies the path where Dreamcast tools that are not part of KOS are to be
# installed. This includes, for example, dc-tool-ip, dc-tool-serial, and the
# mrbc bytecode compiler. This directory, along with SH and ARM compiler
# toolchains, will be added to your PATH environment variable.
#
export DC_TOOLS_BASE="/media/chris/development/dreamcast/tools"

# CMake Toolchain Path
#
# Specifies the path to the toolchain file used to target
# KOS with the "cmake" build utility.
#
export KOS_CMAKE_TOOLCHAIN="${KOS_BASE}/utils/cmake/dreamcast.toolchain.cmake"

# Genromfs Utility Path
#
# Specifies the path to the utility which is used by KOS
# to create romdisk filesystem images.
#
export KOS_GENROMFS="${KOS_BASE}/utils/genromfs/genromfs"

# Make Utility
#
# Configures the tool to be used as the main "make" utility
# for building GNU Makefiles. On a platform such as BSD,
# the default can be changed to "gmake," for the GNU
# implementation.
#
export KOS_MAKE="make"
#export KOS_MAKE="gmake"

# Loader Utility
#
# Specifies the loader to be used with the "make run" targets
# in the KOS examples. Defaults to using a preconfigured version
# of dc-tool. Use one of the other options for a manual dc-tool-ip
# or dc-tool-serial configuration, remembering to change the values
# for the Dreamcast's IP address or the serial port interface.
#
#export KOS_LOADER="dc-tool -x"
export KOS_LOADER="dc-tool-ip -t dreamcast -x"
#export KOS_LOADER="dc-tool-ser -t /dev/ttyS0 -x"

# Default Compiler Flags
#
# Resets build flags. You can also initialize them to some preset
# default values here if you wish.
#
export KOS_INC_PATHS=""
export KOS_CFLAGS=""
export KOS_CPPFLAGS=""
export KOS_LDFLAGS=""
export KOS_AFLAGS=""
export DC_ARM_LDFLAGS=""

# Debug Builds
#
# Controls whether to disable additional debugging checks and assertions,
# such as for parameter validation or internal errors. Uncomment this if
# you do not wish to compile with this additional logic enabled.
#
#export KOS_CFLAGS="${KOS_CFLAGS} -DNDEBUG"

# Optimization Level
#
# Controls the baseline optimization level to use when building.
# Typically this is -Og (debugging), -O0, -O1, -O2, or -O3.
# NOTE: For our target, -O4 is a valid optimization level that has
# been seen to have some performance impact as well.
#
export KOS_CFLAGS="${KOS_CFLAGS} -O2"

# Additional Optimizations
#
# Uncomment this to enable what has been found empirically to be
# the optimal set of additional flags for release build performance
# on the current stable toolchain. NOTE: Certain KOS-ports and examples
# do not work properly with "-flto=auto"!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -freorder-blocks-algorithm=simple -flto=auto"

# Frame Pointers
#
# Controls whether frame pointers are emitted or not. Disabled by
# default. Enable them if you plan to use GDB for debugging.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fomit-frame-pointer"
#export KOS_CFLAGS="${KOS_CFLAGS} -fno-omit-frame-pointer -DFRAME_POINTERS"

# Stack Protector
#
# Controls whether GCC emits extra code to check for buffer overflows or stack
# smashing, which can be very useful for debugging. -fstack-protector only
# covers vulnerable objects, while -fstack-protector-strong provides medium
# coverage, and -fstack-protector-all provides full coverage. You may also
# override the default stack excepton handler by providing your own
# implementation of "void __stack_chk_fail(void)."
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fstack-protector-all"

# GCC Builtin Functions
#
# Comment out this line to enable GCC to use its own builtin implementations of
# certain standard library functions. Under certain conditions, this can allow
# compiler-optimized implementations to replace standard function invocations.
# The downside of this is that it COULD interfere with Newlib or KOS implementations
# of these functions, and it has not been tested thoroughly to ensure compatibility.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fno-builtin"

# Fast Math Instructions
#
# Uncomment this line to enable the optimized fast-math instructions (FSSRA,
# FSCA, and FSQRT) for calculating sin/cos, inverse square root, and square roots.
# These can result in substantial performance gains for these kinds of operations;
# however, they do so at the price of accuracy and are not IEEE compliant.
# NOTE: Enabling this option will also override -fno-builtin!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fbuiltin -ffast-math -ffp-contract=fast"

# SH4 Floating-Point Arithmetic Precision
#
# KallistiOS supports both the single-precision-default ABI (m4-single) and the
# single-precision-only ABI (m4-single-only). When using m4-single, the SH4 will
# be in single-precision mode upon function entry but will switch to double-
# precision mode if 64-bit doubles are used. When using m4-single-only, the SH4
# will always be in single-precision mode and 64-bit doubles will be truncated to
# 32-bit floats. Historically, m4-single-only was used in both official and
# homebrew Dreamcast software, but m4-single is the default as of KOS 2.2.0 to
# increase compatibility with newer libraries which require 64-bit doubles.
#
# WARNING: When adjusting this setting, make sure all software, including
# kos-ports and linked external libraries, are rebuilt using the same
# floating-point precision ABI setting!
#
export KOS_SH4_PRECISION="-m4-single"

# Use LRA (Local Register Allocator) Pass
#
# Uncomment this line to use the modern Local Register Allocator pass during
# code generation instead of the default older reload pass. This option is
# known to be unstable or less performant for SH at this time, but will likely
# become mandatory in future versions of GCC, so feel free to help us test.
# Only enable this setting if you understand what you are doing!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -mlra"

# Shared Compiler Configuration
#
# Include sub architecture-independent configuration file for shared
# environment settings. If you want to configure additional compiler
# options or see where other build flags are set, look at this file.
#
. ${KOS_BASE}/environ_base.sh
</pre>

<pre>
. environ.sh
make
</pre>

=== KOS-ports bauen ===

<pre>
git clone --recursive https://github.com/KallistiOS/kos-ports kos-ports
kos-ports/utils/build-all.sh
</pre>

Funktionierende Toolchains

2025-04-18T22:17:16Z

Chris: /* KallistiOS Toolchain */

== KallistiOS Toolchain ==

Voraussetzungen:

* Wenn die Toolchain innerhalb eines CIFS-Mounts installiert wird müssen *mfsymlinks in den Mountoptions aktiviert sein!

=== Crosscompiler bauen ===
<pre>
mkdir /media/chris/development/dreamcast/toolchains
cd /media/chris/development/dreamcast/toolchains
git clone https://github.com/KallistiOS/KallistiOS.git kos
cd kos/utils/dc-chain/
cp -av Makefile.default.cfg Makefile.cfg
</pre>

==== Makefile.cfg bearbeiten ====

<pre>
# Sega Dreamcast Toolchains Maker (dc-chain)
# This file is part of KallistiOS.

#########################
### TOOLCHAIN PROFILE ###
#########################

# Choose a toolchain profile from the following available options:
# No longer supported upstream:
# - 9.3.0-legacy: Former 'stable' option, based on GCC 9.3.0 and Newlib 3.3.0.
# - 9.5.0-winxp: Most recent versions of tools which run on Windows XP.
# - 10.5.0: Last release in the GCC 10 series, released 2023-07-07.
# - 11.5.0: Last release in the GCC 11 series, released 2024-07-19.
# Supported upstream:
# - 12.4.0: Latest release in the GCC 12 series, released 2024-06-20.
# - stable: Tested stable; based on GCC 13.2.0, released 2023-07-27.
# - 13.3.0: Latest release in the GCC 13 series, released 2024-05-21.
# - 14.2.0: Latest release in the GCC 14 series, released 2024-08-01.
# Development versions:
# - 13.3.1-dev Bleeding edge GCC 13 series from git.
# - 14.2.1-dev Bleeding edge GCC 14 series from git.
# - 15.0.0-dev Bleeding edge GCC 15 series from git.
# - gccrs-dev: GCC fork for development of the GCCRS Rust compiler.
# - rustc-dev: GCC fork for development of the libgccjit rustc GCC codegen.
# If unsure, select stable. See README.md for more detailed descriptions.
toolchain_profile=stable

########################
### DOWNLOAD OPTIONS ###
########################

### Download protocol (http|https|ftp)
# Specify the protocol you want to use for downloading package files.
download_protocol=https

### Force downloader (curl|wget)
# Specify here if you'd prefer to use 'wget' or 'curl'. If neither is specified,
# a web downloader tool will be auto-detected in the following order: cURL, Wget
# You must have either Wget or cURL installed to use dc-chain.
#force_downloader=wget

### Specify GNU mirror override
# The default mirror for GNU sources is 'ftpmirror.gnu.org'
# This setting overrides the default mirror with a preferred mirror.
#gnu_mirror=mirrors.kernel.org

#####################
### BUILD OPTIONS ###
#####################

### SH toolchain install path
# Specify the directory where the SH toolchain will be installed. This setting
# must match the KOS_CC_BASE setting in your KOS "environ.sh" configuration.
sh_toolchain_path=/media/chris/development/dreamcast/toolchains/sh-elf

### ARM toolchain install path
# Specify the directory where the ARM toolchain will be installed. This setting
# must match the DC_ARM_BASE setting in your KOS "environ.sh" configuration.
arm_toolchain_path=/media/chris/development/dreamcast/toolchains/arm-eabi

### Make jobs (n|<empty>)
# Set this value to the number of parallel jobs you want to run with make.
# For fastest toolchain building, set to the number of CPU threads available.
# Using multiple jobs may cause issues in certain environments and may be
# automatically disabled. If you encounter errors building your toolchain,
# reduce the number of jobs to 1 to avoid issues and ease troubleshooting.
makejobs=16

### Verbose (1|0)
# Choose whether to actively display compilation messages on the screen.
# Messages are saved to the build log files regardless of this setting.
verbose=1

### Erase (1|0)
# Erase build directories as toolchain components are installed to save space.
erase=1

### Install toolchain debug symbols (1|0)
# Choose whether to keep the debugging symbols for the toolchain.
# This is only useful if you wish to debug the toolchain itself.
#toolchain_debug=1

########################
### LANGUAGE OPTIONS ###
########################

### Enable C++
# Builds C++ support, including the C++ compiler and its standard library. The
# vast majority of language features are supported, with C++23 and early C++26
# support included. KallistiOS provides several examples with C++ support, so it
# is enabled by default. Adding C++ support requires extra disk space and
# compilation time, so you may disable it here if you do not plan on using C++.
enable_cpp=1

### Enable Objective-C
# Builds Objective-C support. Currently, this builds the Obj-C compiler and
# runtime, but does not build a standard library. The GNUStep Base Library is
# currently in active development for future inclusion in the kos-ports library.
# KallistiOS provides examples with Obj-C support, so it is enabled by default.
enable_objc=1

### Enable Objective C++
# Builds Objective C++ support. This hybrid language allows both C++ and Obj-C
# methods to be called from both contexts. Enabled by default as both the C++
# and Obj-C languages are already enabled by default.
enable_objcpp=1

### Enable D
# Builds D support. This will build the D compiler, but does not build the
# Phobos Runtime Library. D support may only be enabled on POSIX platforms which
# have its external dependencies provided through a host DMD, GDC, or LDC
# compiler installation.
#enable_d=1

### Enable Ada
# Builds Ada support. This will build the GNAT Ada compiler and tools but does
# not build the libada runtime library. In order for this build to succeed, the
# host's GCC version must match the target Dreamcast version, plus the same
# version of GNAT and its tools must be preinstalled (and in your path).
#enable_ada=1

### Enable Rust
# Builds the work-in-progress GCCRS Rust compiler frontend for GCC. Requires a
# development version of GCC. Use with the gccrs-dev toolchain profile for the
# most recent updates to the GCCRS compiler. This option is not to be confused
# with using this toolchain as a backend for rustc; see libgccjit below.
#enable_rust=1

### Enable libgccjit
# Enables the libgccjit embeddable GCC library. This is most commonly enabled to
# allow using this toolchain as a backend for the official rustc compiler. Use
# with the rustc-dev toolchain profile if this is your use case.
#enable_libgccjit=1

###################
### GCC OPTIONS ###
###################

### Floating point precision support (m4-single-only|m4-single|m4)
# Build support for various SH4 floating-point operation ABIs. KallistiOS only
# officially supports single-precision-only mode (m4-single-only); however,
# experimental support for single-precision-default mode (m4-single) has been
# added to allow for the use of full 64-bit doubles. You may also include
# double-precision-default (m4) which is untested and unsupported, or you may
# choose to specify only m4-single-only to save disk space.
precision_modes=m4-single-only,m4-single

### Default floating point mode (m4|m4-single|m4-single-only)
# Choose the default floating point precision ABI used when GCC is invoked. This
# may be overridden by passing -m4-single-only, -m4-single, or -m4 to GCC.
# KOS currently only officially supports m4-single-only, so it is the default.
default_precision=m4-single-only

### GCC threading model (single|kos)
# KallistiOS patches to GCC provide a 'kos' thread model, which should be used.
# If you want to disable threading support for C++, Objective-C, and so forth,
# you can set this option to 'single'.
thread_model=kos

### Automatic patching for KallistiOS (1|0)
# Uncomment this option if you want to disable applying KallistiOS patches to
# toolchain source files before building. This will disable usage of the 'kos'
# thread model. Only do this if you understand what you are doing.
#use_kos_patches=0

### Disable GCC Native Language Support (1|0)
# By default, NLS allows GCC to output diagnostics in non-English languages.
# Uncomment this option to disable NLS and force GCC to output in English.
#disable_nls=1

######################
### NEWLIB OPTIONS ###
######################

### Automatic patching for Newlib (1|0)
# Uncomment this option if you want to disable the automatic patching of Newlib
# needed by KallistiOS. This will keep the generated toolchain completely raw.
# This will also disable the 'kos' thread model. Only do this if you understand
# what you are doing.
#auto_fixup_sh4_newlib=0

### C99 format specifier support (1|0)
# Define this to build Newlib with additional support for the C99 format
# specifiers, used by printf and friends. These include support for size_t,
# ptrdiff_t, intmax_t, and sized integral types.
newlib_c99_formats=1

### Multibyte character set support (1|0)
# Define this to build Newlib with additional multibyte support. This enables
# three special locales: "C-JIS", "C-SJIS", and "C-EUCJP". The multibyte
# support extends to the mb and wc functions in stdlib as well as format
# characters for the printf and scanf family of routines.
#newlib_multibyte=1

### iconv() character encoding conversions support (encoding list)
# Define a list here to enable support for the iconv() function and <iconv.h>
# header file. The given comma separated list defines for which encoding types
# to include bidirectional conversion support. For the full list of available
# encodings, see the Newlib configuration documentation.
#newlib_iconv_encodings=us_ascii,utf8,utf16,ucs_2_internal,ucs_4_internal

### Optimize Newlib for space (1|0)
# Uncomment this option to optimize for space when building Newlib. This will
# build Newlib with compiler flags which favor smaller code sizes over faster
# performance.
#newlib_opt_space=1

###################
### C++ OPTIONS ###
###################

### Timezone database support (1|0|path)
# Uncomment this option to enable building support for C++'s std::chrono::tzdb
# into your C++ standard library by using a combination of the system's local
# timezone DB and one dynamically fetched from the "IANA Time Zone Database."
# Without support enabled, only the "UTC" and "GMT" timezones will be defined.
# You can optionally provide the path to a directory with a custom "tzdata.zi"
# database file. NOTE: Enabling this will result in larger C++ binaries!
#libstdcxx_tzdb=1

#######################
### WINDOWS OPTIONS ###
#######################

### MinGW/MSYS
# Standalone binaries (1|0)
# Uncomment this option if you want static binaries that are standalone and
# require no dependencies. When this option is used, binaries can be run outside
# the MinGW/MSYS environment. This is NOT recommended; only do this if you know
# what you are doing.
#standalone_binary=1

### Force installation of BFD for SH (1|0)
# Uncomment this option if you want to force installation of 'libbfd' for the SH
# toolchain. This is required for MinGW/MSYS and can't be disabled in this
# scenario. This option is here mainly if you want to force the installation of
# 'libbfd' under other environments, but this won't be necessary in most cases,
# as 'libelf' is used almost everywhere. Please note, 'libbfd' is not portable
# if you built it on another environment. Only use this flag if you understand
# what you are doing.
#sh_force_libbfd_installation=1
</pre>

Make durchführen:
<pre>
PATH=/usr/bin make all
</pre>

==== KOS bauen ====

eviron.sh:
<pre>
# KallistiOS Environment Settings
#
# This is a sample script for configuring and customizing your
# KOS build environment. Modify it to suit your setup. Several
# settings may be enabled optionally or are provided with
# alternative values.
#
# This script is typically sourced in your current shell environment
# (probably by .bashrc, .bash_profile, or something similar), so that
# the KOS environment is set up automatically for each shell session.
#

# Build Architecture
#
# Set the major architecture you'll be building for.
# The only option here is "dreamcast" as of KOS 2.0.0.
#
export KOS_ARCH="dreamcast"

# Build Sub-Architecture
#
# Defines the sub architecture your configuration
# is targeting or uses an existing value that
# can be set externally via your IDE.
#
# Valid values:
# "pristine" - Dreamcast console or HKT-0120 devkit (default)
# "naomi" - NAOMI or NAOMI 2 arcade board
#
if [ -z "${KOS_SUBARCH}" ] ; then
export KOS_SUBARCH="pristine"
else
export KOS_SUBARCH
fi

# KOS Root Path
#
# Specifies the path to the KOS root directory
#
export KOS_BASE="/media/chris/development/dreamcast/toolchains/kos"

# KOS-Ports Path
#
# Specifies the path to the KOS-ports directory
#
export KOS_PORTS="${KOS_BASE}/../kos-ports"

# SH Compiler Prefixes
#
# Specifies the path to and prefix for the main SH
# GCC toolchain used to target the Dreamcast's SH4 CPU.
#
export KOS_CC_BASE="/media/chris/development/dreamcast/toolchains/sh-elf"
export KOS_CC_PREFIX="sh-elf"

# ARM Compiler Prefixes
#
# Specifies the path to and prefix for the additional ARM
# GCC toolchain used to target the Dreamcast's AICA SPU.
#
export DC_ARM_BASE="/media/chris/development/dreamcast/toolchains/arm-eabi"
export DC_ARM_PREFIX="arm-eabi"

# External Dreamcast Tools Path
#
# Specifies the path where Dreamcast tools that are not part of KOS are to be
# installed. This includes, for example, dc-tool-ip, dc-tool-serial, and the
# mrbc bytecode compiler. This directory, along with SH and ARM compiler
# toolchains, will be added to your PATH environment variable.
#
export DC_TOOLS_BASE="/media/chris/development/dreamcast/tools"

# CMake Toolchain Path
#
# Specifies the path to the toolchain file used to target
# KOS with the "cmake" build utility.
#
export KOS_CMAKE_TOOLCHAIN="${KOS_BASE}/utils/cmake/dreamcast.toolchain.cmake"

# Genromfs Utility Path
#
# Specifies the path to the utility which is used by KOS
# to create romdisk filesystem images.
#
export KOS_GENROMFS="${KOS_BASE}/utils/genromfs/genromfs"

# Make Utility
#
# Configures the tool to be used as the main "make" utility
# for building GNU Makefiles. On a platform such as BSD,
# the default can be changed to "gmake," for the GNU
# implementation.
#
export KOS_MAKE="make"
#export KOS_MAKE="gmake"

# Loader Utility
#
# Specifies the loader to be used with the "make run" targets
# in the KOS examples. Defaults to using a preconfigured version
# of dc-tool. Use one of the other options for a manual dc-tool-ip
# or dc-tool-serial configuration, remembering to change the values
# for the Dreamcast's IP address or the serial port interface.
#
#export KOS_LOADER="dc-tool -x"
export KOS_LOADER="dc-tool-ip -t dreamcast -x"
#export KOS_LOADER="dc-tool-ser -t /dev/ttyS0 -x"

# Default Compiler Flags
#
# Resets build flags. You can also initialize them to some preset
# default values here if you wish.
#
export KOS_INC_PATHS=""
export KOS_CFLAGS=""
export KOS_CPPFLAGS=""
export KOS_LDFLAGS=""
export KOS_AFLAGS=""
export DC_ARM_LDFLAGS=""

# Debug Builds
#
# Controls whether to disable additional debugging checks and assertions,
# such as for parameter validation or internal errors. Uncomment this if
# you do not wish to compile with this additional logic enabled.
#
#export KOS_CFLAGS="${KOS_CFLAGS} -DNDEBUG"

# Optimization Level
#
# Controls the baseline optimization level to use when building.
# Typically this is -Og (debugging), -O0, -O1, -O2, or -O3.
# NOTE: For our target, -O4 is a valid optimization level that has
# been seen to have some performance impact as well.
#
export KOS_CFLAGS="${KOS_CFLAGS} -O2"

# Additional Optimizations
#
# Uncomment this to enable what has been found empirically to be
# the optimal set of additional flags for release build performance
# on the current stable toolchain. NOTE: Certain KOS-ports and examples
# do not work properly with "-flto=auto"!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -freorder-blocks-algorithm=simple -flto=auto"

# Frame Pointers
#
# Controls whether frame pointers are emitted or not. Disabled by
# default. Enable them if you plan to use GDB for debugging.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fomit-frame-pointer"
#export KOS_CFLAGS="${KOS_CFLAGS} -fno-omit-frame-pointer -DFRAME_POINTERS"

# Stack Protector
#
# Controls whether GCC emits extra code to check for buffer overflows or stack
# smashing, which can be very useful for debugging. -fstack-protector only
# covers vulnerable objects, while -fstack-protector-strong provides medium
# coverage, and -fstack-protector-all provides full coverage. You may also
# override the default stack excepton handler by providing your own
# implementation of "void __stack_chk_fail(void)."
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fstack-protector-all"

# GCC Builtin Functions
#
# Comment out this line to enable GCC to use its own builtin implementations of
# certain standard library functions. Under certain conditions, this can allow
# compiler-optimized implementations to replace standard function invocations.
# The downside of this is that it COULD interfere with Newlib or KOS implementations
# of these functions, and it has not been tested thoroughly to ensure compatibility.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fno-builtin"

# Fast Math Instructions
#
# Uncomment this line to enable the optimized fast-math instructions (FSSRA,
# FSCA, and FSQRT) for calculating sin/cos, inverse square root, and square roots.
# These can result in substantial performance gains for these kinds of operations;
# however, they do so at the price of accuracy and are not IEEE compliant.
# NOTE: Enabling this option will also override -fno-builtin!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fbuiltin -ffast-math -ffp-contract=fast"

# SH4 Floating-Point Arithmetic Precision
#
# KallistiOS supports both the single-precision-default ABI (m4-single) and the
# single-precision-only ABI (m4-single-only). When using m4-single, the SH4 will
# be in single-precision mode upon function entry but will switch to double-
# precision mode if 64-bit doubles are used. When using m4-single-only, the SH4
# will always be in single-precision mode and 64-bit doubles will be truncated to
# 32-bit floats. Historically, m4-single-only was used in both official and
# homebrew Dreamcast software, but m4-single is the default as of KOS 2.2.0 to
# increase compatibility with newer libraries which require 64-bit doubles.
#
# WARNING: When adjusting this setting, make sure all software, including
# kos-ports and linked external libraries, are rebuilt using the same
# floating-point precision ABI setting!
#
export KOS_SH4_PRECISION="-m4-single"

# Use LRA (Local Register Allocator) Pass
#
# Uncomment this line to use the modern Local Register Allocator pass during
# code generation instead of the default older reload pass. This option is
# known to be unstable or less performant for SH at this time, but will likely
# become mandatory in future versions of GCC, so feel free to help us test.
# Only enable this setting if you understand what you are doing!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -mlra"

# Shared Compiler Configuration
#
# Include sub architecture-independent configuration file for shared
# environment settings. If you want to configure additional compiler
# options or see where other build flags are set, look at this file.
#
. ${KOS_BASE}/environ_base.sh
</pre>

<pre>
. environ.sh
make
</pre>

=== KOS-ports bauen ===

<pre>
git clone --recursive https://github.com/KallistiOS/kos-ports kos-ports
kos-ports/utils/build-all.sh
</pre>

Funktionierende Toolchains

2025-04-18T22:16:59Z

Chris:

== KallistiOS Toolchain ==

Voraussetzungen:

* Wenn die Toolchain innerhalb eines CIFS-Mounts installiert wird müssen *mfsymlinks* in den Mountoptions aktiviert sein!

=== Crosscompiler bauen ===
<pre>
mkdir /media/chris/development/dreamcast/toolchains
cd /media/chris/development/dreamcast/toolchains
git clone https://github.com/KallistiOS/KallistiOS.git kos
cd kos/utils/dc-chain/
cp -av Makefile.default.cfg Makefile.cfg
</pre>

==== Makefile.cfg bearbeiten ====

<pre>
# Sega Dreamcast Toolchains Maker (dc-chain)
# This file is part of KallistiOS.

#########################
### TOOLCHAIN PROFILE ###
#########################

# Choose a toolchain profile from the following available options:
# No longer supported upstream:
# - 9.3.0-legacy: Former 'stable' option, based on GCC 9.3.0 and Newlib 3.3.0.
# - 9.5.0-winxp: Most recent versions of tools which run on Windows XP.
# - 10.5.0: Last release in the GCC 10 series, released 2023-07-07.
# - 11.5.0: Last release in the GCC 11 series, released 2024-07-19.
# Supported upstream:
# - 12.4.0: Latest release in the GCC 12 series, released 2024-06-20.
# - stable: Tested stable; based on GCC 13.2.0, released 2023-07-27.
# - 13.3.0: Latest release in the GCC 13 series, released 2024-05-21.
# - 14.2.0: Latest release in the GCC 14 series, released 2024-08-01.
# Development versions:
# - 13.3.1-dev Bleeding edge GCC 13 series from git.
# - 14.2.1-dev Bleeding edge GCC 14 series from git.
# - 15.0.0-dev Bleeding edge GCC 15 series from git.
# - gccrs-dev: GCC fork for development of the GCCRS Rust compiler.
# - rustc-dev: GCC fork for development of the libgccjit rustc GCC codegen.
# If unsure, select stable. See README.md for more detailed descriptions.
toolchain_profile=stable

########################
### DOWNLOAD OPTIONS ###
########################

### Download protocol (http|https|ftp)
# Specify the protocol you want to use for downloading package files.
download_protocol=https

### Force downloader (curl|wget)
# Specify here if you'd prefer to use 'wget' or 'curl'. If neither is specified,
# a web downloader tool will be auto-detected in the following order: cURL, Wget
# You must have either Wget or cURL installed to use dc-chain.
#force_downloader=wget

### Specify GNU mirror override
# The default mirror for GNU sources is 'ftpmirror.gnu.org'
# This setting overrides the default mirror with a preferred mirror.
#gnu_mirror=mirrors.kernel.org

#####################
### BUILD OPTIONS ###
#####################

### SH toolchain install path
# Specify the directory where the SH toolchain will be installed. This setting
# must match the KOS_CC_BASE setting in your KOS "environ.sh" configuration.
sh_toolchain_path=/media/chris/development/dreamcast/toolchains/sh-elf

### ARM toolchain install path
# Specify the directory where the ARM toolchain will be installed. This setting
# must match the DC_ARM_BASE setting in your KOS "environ.sh" configuration.
arm_toolchain_path=/media/chris/development/dreamcast/toolchains/arm-eabi

### Make jobs (n|<empty>)
# Set this value to the number of parallel jobs you want to run with make.
# For fastest toolchain building, set to the number of CPU threads available.
# Using multiple jobs may cause issues in certain environments and may be
# automatically disabled. If you encounter errors building your toolchain,
# reduce the number of jobs to 1 to avoid issues and ease troubleshooting.
makejobs=16

### Verbose (1|0)
# Choose whether to actively display compilation messages on the screen.
# Messages are saved to the build log files regardless of this setting.
verbose=1

### Erase (1|0)
# Erase build directories as toolchain components are installed to save space.
erase=1

### Install toolchain debug symbols (1|0)
# Choose whether to keep the debugging symbols for the toolchain.
# This is only useful if you wish to debug the toolchain itself.
#toolchain_debug=1

########################
### LANGUAGE OPTIONS ###
########################

### Enable C++
# Builds C++ support, including the C++ compiler and its standard library. The
# vast majority of language features are supported, with C++23 and early C++26
# support included. KallistiOS provides several examples with C++ support, so it
# is enabled by default. Adding C++ support requires extra disk space and
# compilation time, so you may disable it here if you do not plan on using C++.
enable_cpp=1

### Enable Objective-C
# Builds Objective-C support. Currently, this builds the Obj-C compiler and
# runtime, but does not build a standard library. The GNUStep Base Library is
# currently in active development for future inclusion in the kos-ports library.
# KallistiOS provides examples with Obj-C support, so it is enabled by default.
enable_objc=1

### Enable Objective C++
# Builds Objective C++ support. This hybrid language allows both C++ and Obj-C
# methods to be called from both contexts. Enabled by default as both the C++
# and Obj-C languages are already enabled by default.
enable_objcpp=1

### Enable D
# Builds D support. This will build the D compiler, but does not build the
# Phobos Runtime Library. D support may only be enabled on POSIX platforms which
# have its external dependencies provided through a host DMD, GDC, or LDC
# compiler installation.
#enable_d=1

### Enable Ada
# Builds Ada support. This will build the GNAT Ada compiler and tools but does
# not build the libada runtime library. In order for this build to succeed, the
# host's GCC version must match the target Dreamcast version, plus the same
# version of GNAT and its tools must be preinstalled (and in your path).
#enable_ada=1

### Enable Rust
# Builds the work-in-progress GCCRS Rust compiler frontend for GCC. Requires a
# development version of GCC. Use with the gccrs-dev toolchain profile for the
# most recent updates to the GCCRS compiler. This option is not to be confused
# with using this toolchain as a backend for rustc; see libgccjit below.
#enable_rust=1

### Enable libgccjit
# Enables the libgccjit embeddable GCC library. This is most commonly enabled to
# allow using this toolchain as a backend for the official rustc compiler. Use
# with the rustc-dev toolchain profile if this is your use case.
#enable_libgccjit=1

###################
### GCC OPTIONS ###
###################

### Floating point precision support (m4-single-only|m4-single|m4)
# Build support for various SH4 floating-point operation ABIs. KallistiOS only
# officially supports single-precision-only mode (m4-single-only); however,
# experimental support for single-precision-default mode (m4-single) has been
# added to allow for the use of full 64-bit doubles. You may also include
# double-precision-default (m4) which is untested and unsupported, or you may
# choose to specify only m4-single-only to save disk space.
precision_modes=m4-single-only,m4-single

### Default floating point mode (m4|m4-single|m4-single-only)
# Choose the default floating point precision ABI used when GCC is invoked. This
# may be overridden by passing -m4-single-only, -m4-single, or -m4 to GCC.
# KOS currently only officially supports m4-single-only, so it is the default.
default_precision=m4-single-only

### GCC threading model (single|kos)
# KallistiOS patches to GCC provide a 'kos' thread model, which should be used.
# If you want to disable threading support for C++, Objective-C, and so forth,
# you can set this option to 'single'.
thread_model=kos

### Automatic patching for KallistiOS (1|0)
# Uncomment this option if you want to disable applying KallistiOS patches to
# toolchain source files before building. This will disable usage of the 'kos'
# thread model. Only do this if you understand what you are doing.
#use_kos_patches=0

### Disable GCC Native Language Support (1|0)
# By default, NLS allows GCC to output diagnostics in non-English languages.
# Uncomment this option to disable NLS and force GCC to output in English.
#disable_nls=1

######################
### NEWLIB OPTIONS ###
######################

### Automatic patching for Newlib (1|0)
# Uncomment this option if you want to disable the automatic patching of Newlib
# needed by KallistiOS. This will keep the generated toolchain completely raw.
# This will also disable the 'kos' thread model. Only do this if you understand
# what you are doing.
#auto_fixup_sh4_newlib=0

### C99 format specifier support (1|0)
# Define this to build Newlib with additional support for the C99 format
# specifiers, used by printf and friends. These include support for size_t,
# ptrdiff_t, intmax_t, and sized integral types.
newlib_c99_formats=1

### Multibyte character set support (1|0)
# Define this to build Newlib with additional multibyte support. This enables
# three special locales: "C-JIS", "C-SJIS", and "C-EUCJP". The multibyte
# support extends to the mb and wc functions in stdlib as well as format
# characters for the printf and scanf family of routines.
#newlib_multibyte=1

### iconv() character encoding conversions support (encoding list)
# Define a list here to enable support for the iconv() function and <iconv.h>
# header file. The given comma separated list defines for which encoding types
# to include bidirectional conversion support. For the full list of available
# encodings, see the Newlib configuration documentation.
#newlib_iconv_encodings=us_ascii,utf8,utf16,ucs_2_internal,ucs_4_internal

### Optimize Newlib for space (1|0)
# Uncomment this option to optimize for space when building Newlib. This will
# build Newlib with compiler flags which favor smaller code sizes over faster
# performance.
#newlib_opt_space=1

###################
### C++ OPTIONS ###
###################

### Timezone database support (1|0|path)
# Uncomment this option to enable building support for C++'s std::chrono::tzdb
# into your C++ standard library by using a combination of the system's local
# timezone DB and one dynamically fetched from the "IANA Time Zone Database."
# Without support enabled, only the "UTC" and "GMT" timezones will be defined.
# You can optionally provide the path to a directory with a custom "tzdata.zi"
# database file. NOTE: Enabling this will result in larger C++ binaries!
#libstdcxx_tzdb=1

#######################
### WINDOWS OPTIONS ###
#######################

### MinGW/MSYS
# Standalone binaries (1|0)
# Uncomment this option if you want static binaries that are standalone and
# require no dependencies. When this option is used, binaries can be run outside
# the MinGW/MSYS environment. This is NOT recommended; only do this if you know
# what you are doing.
#standalone_binary=1

### Force installation of BFD for SH (1|0)
# Uncomment this option if you want to force installation of 'libbfd' for the SH
# toolchain. This is required for MinGW/MSYS and can't be disabled in this
# scenario. This option is here mainly if you want to force the installation of
# 'libbfd' under other environments, but this won't be necessary in most cases,
# as 'libelf' is used almost everywhere. Please note, 'libbfd' is not portable
# if you built it on another environment. Only use this flag if you understand
# what you are doing.
#sh_force_libbfd_installation=1
</pre>

Make durchführen:
<pre>
PATH=/usr/bin make all
</pre>

==== KOS bauen ====

eviron.sh:
<pre>
# KallistiOS Environment Settings
#
# This is a sample script for configuring and customizing your
# KOS build environment. Modify it to suit your setup. Several
# settings may be enabled optionally or are provided with
# alternative values.
#
# This script is typically sourced in your current shell environment
# (probably by .bashrc, .bash_profile, or something similar), so that
# the KOS environment is set up automatically for each shell session.
#

# Build Architecture
#
# Set the major architecture you'll be building for.
# The only option here is "dreamcast" as of KOS 2.0.0.
#
export KOS_ARCH="dreamcast"

# Build Sub-Architecture
#
# Defines the sub architecture your configuration
# is targeting or uses an existing value that
# can be set externally via your IDE.
#
# Valid values:
# "pristine" - Dreamcast console or HKT-0120 devkit (default)
# "naomi" - NAOMI or NAOMI 2 arcade board
#
if [ -z "${KOS_SUBARCH}" ] ; then
export KOS_SUBARCH="pristine"
else
export KOS_SUBARCH
fi

# KOS Root Path
#
# Specifies the path to the KOS root directory
#
export KOS_BASE="/media/chris/development/dreamcast/toolchains/kos"

# KOS-Ports Path
#
# Specifies the path to the KOS-ports directory
#
export KOS_PORTS="${KOS_BASE}/../kos-ports"

# SH Compiler Prefixes
#
# Specifies the path to and prefix for the main SH
# GCC toolchain used to target the Dreamcast's SH4 CPU.
#
export KOS_CC_BASE="/media/chris/development/dreamcast/toolchains/sh-elf"
export KOS_CC_PREFIX="sh-elf"

# ARM Compiler Prefixes
#
# Specifies the path to and prefix for the additional ARM
# GCC toolchain used to target the Dreamcast's AICA SPU.
#
export DC_ARM_BASE="/media/chris/development/dreamcast/toolchains/arm-eabi"
export DC_ARM_PREFIX="arm-eabi"

# External Dreamcast Tools Path
#
# Specifies the path where Dreamcast tools that are not part of KOS are to be
# installed. This includes, for example, dc-tool-ip, dc-tool-serial, and the
# mrbc bytecode compiler. This directory, along with SH and ARM compiler
# toolchains, will be added to your PATH environment variable.
#
export DC_TOOLS_BASE="/media/chris/development/dreamcast/tools"

# CMake Toolchain Path
#
# Specifies the path to the toolchain file used to target
# KOS with the "cmake" build utility.
#
export KOS_CMAKE_TOOLCHAIN="${KOS_BASE}/utils/cmake/dreamcast.toolchain.cmake"

# Genromfs Utility Path
#
# Specifies the path to the utility which is used by KOS
# to create romdisk filesystem images.
#
export KOS_GENROMFS="${KOS_BASE}/utils/genromfs/genromfs"

# Make Utility
#
# Configures the tool to be used as the main "make" utility
# for building GNU Makefiles. On a platform such as BSD,
# the default can be changed to "gmake," for the GNU
# implementation.
#
export KOS_MAKE="make"
#export KOS_MAKE="gmake"

# Loader Utility
#
# Specifies the loader to be used with the "make run" targets
# in the KOS examples. Defaults to using a preconfigured version
# of dc-tool. Use one of the other options for a manual dc-tool-ip
# or dc-tool-serial configuration, remembering to change the values
# for the Dreamcast's IP address or the serial port interface.
#
#export KOS_LOADER="dc-tool -x"
export KOS_LOADER="dc-tool-ip -t dreamcast -x"
#export KOS_LOADER="dc-tool-ser -t /dev/ttyS0 -x"

# Default Compiler Flags
#
# Resets build flags. You can also initialize them to some preset
# default values here if you wish.
#
export KOS_INC_PATHS=""
export KOS_CFLAGS=""
export KOS_CPPFLAGS=""
export KOS_LDFLAGS=""
export KOS_AFLAGS=""
export DC_ARM_LDFLAGS=""

# Debug Builds
#
# Controls whether to disable additional debugging checks and assertions,
# such as for parameter validation or internal errors. Uncomment this if
# you do not wish to compile with this additional logic enabled.
#
#export KOS_CFLAGS="${KOS_CFLAGS} -DNDEBUG"

# Optimization Level
#
# Controls the baseline optimization level to use when building.
# Typically this is -Og (debugging), -O0, -O1, -O2, or -O3.
# NOTE: For our target, -O4 is a valid optimization level that has
# been seen to have some performance impact as well.
#
export KOS_CFLAGS="${KOS_CFLAGS} -O2"

# Additional Optimizations
#
# Uncomment this to enable what has been found empirically to be
# the optimal set of additional flags for release build performance
# on the current stable toolchain. NOTE: Certain KOS-ports and examples
# do not work properly with "-flto=auto"!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -freorder-blocks-algorithm=simple -flto=auto"

# Frame Pointers
#
# Controls whether frame pointers are emitted or not. Disabled by
# default. Enable them if you plan to use GDB for debugging.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fomit-frame-pointer"
#export KOS_CFLAGS="${KOS_CFLAGS} -fno-omit-frame-pointer -DFRAME_POINTERS"

# Stack Protector
#
# Controls whether GCC emits extra code to check for buffer overflows or stack
# smashing, which can be very useful for debugging. -fstack-protector only
# covers vulnerable objects, while -fstack-protector-strong provides medium
# coverage, and -fstack-protector-all provides full coverage. You may also
# override the default stack excepton handler by providing your own
# implementation of "void __stack_chk_fail(void)."
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fstack-protector-all"

# GCC Builtin Functions
#
# Comment out this line to enable GCC to use its own builtin implementations of
# certain standard library functions. Under certain conditions, this can allow
# compiler-optimized implementations to replace standard function invocations.
# The downside of this is that it COULD interfere with Newlib or KOS implementations
# of these functions, and it has not been tested thoroughly to ensure compatibility.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fno-builtin"

# Fast Math Instructions
#
# Uncomment this line to enable the optimized fast-math instructions (FSSRA,
# FSCA, and FSQRT) for calculating sin/cos, inverse square root, and square roots.
# These can result in substantial performance gains for these kinds of operations;
# however, they do so at the price of accuracy and are not IEEE compliant.
# NOTE: Enabling this option will also override -fno-builtin!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fbuiltin -ffast-math -ffp-contract=fast"

# SH4 Floating-Point Arithmetic Precision
#
# KallistiOS supports both the single-precision-default ABI (m4-single) and the
# single-precision-only ABI (m4-single-only). When using m4-single, the SH4 will
# be in single-precision mode upon function entry but will switch to double-
# precision mode if 64-bit doubles are used. When using m4-single-only, the SH4
# will always be in single-precision mode and 64-bit doubles will be truncated to
# 32-bit floats. Historically, m4-single-only was used in both official and
# homebrew Dreamcast software, but m4-single is the default as of KOS 2.2.0 to
# increase compatibility with newer libraries which require 64-bit doubles.
#
# WARNING: When adjusting this setting, make sure all software, including
# kos-ports and linked external libraries, are rebuilt using the same
# floating-point precision ABI setting!
#
export KOS_SH4_PRECISION="-m4-single"

# Use LRA (Local Register Allocator) Pass
#
# Uncomment this line to use the modern Local Register Allocator pass during
# code generation instead of the default older reload pass. This option is
# known to be unstable or less performant for SH at this time, but will likely
# become mandatory in future versions of GCC, so feel free to help us test.
# Only enable this setting if you understand what you are doing!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -mlra"

# Shared Compiler Configuration
#
# Include sub architecture-independent configuration file for shared
# environment settings. If you want to configure additional compiler
# options or see where other build flags are set, look at this file.
#
. ${KOS_BASE}/environ_base.sh
</pre>

<pre>
. environ.sh
make
</pre>

=== KOS-ports bauen ===

<pre>
git clone --recursive https://github.com/KallistiOS/kos-ports kos-ports
kos-ports/utils/build-all.sh
</pre>

Funktionierende Toolchains

2025-04-18T22:15:05Z

Chris:

== KallistiOS Toolchain ==

=== Crosscompiler bauen ===
<pre>
mkdir /media/chris/development/dreamcast/toolchains
cd /media/chris/development/dreamcast/toolchains
git clone https://github.com/KallistiOS/KallistiOS.git kos
cd kos/utils/dc-chain/
cp -av Makefile.default.cfg Makefile.cfg
</pre>

==== Makefile.cfg bearbeiten ====

<pre>
# Sega Dreamcast Toolchains Maker (dc-chain)
# This file is part of KallistiOS.

#########################
### TOOLCHAIN PROFILE ###
#########################

# Choose a toolchain profile from the following available options:
# No longer supported upstream:
# - 9.3.0-legacy: Former 'stable' option, based on GCC 9.3.0 and Newlib 3.3.0.
# - 9.5.0-winxp: Most recent versions of tools which run on Windows XP.
# - 10.5.0: Last release in the GCC 10 series, released 2023-07-07.
# - 11.5.0: Last release in the GCC 11 series, released 2024-07-19.
# Supported upstream:
# - 12.4.0: Latest release in the GCC 12 series, released 2024-06-20.
# - stable: Tested stable; based on GCC 13.2.0, released 2023-07-27.
# - 13.3.0: Latest release in the GCC 13 series, released 2024-05-21.
# - 14.2.0: Latest release in the GCC 14 series, released 2024-08-01.
# Development versions:
# - 13.3.1-dev Bleeding edge GCC 13 series from git.
# - 14.2.1-dev Bleeding edge GCC 14 series from git.
# - 15.0.0-dev Bleeding edge GCC 15 series from git.
# - gccrs-dev: GCC fork for development of the GCCRS Rust compiler.
# - rustc-dev: GCC fork for development of the libgccjit rustc GCC codegen.
# If unsure, select stable. See README.md for more detailed descriptions.
toolchain_profile=stable

########################
### DOWNLOAD OPTIONS ###
########################

### Download protocol (http|https|ftp)
# Specify the protocol you want to use for downloading package files.
download_protocol=https

### Force downloader (curl|wget)
# Specify here if you'd prefer to use 'wget' or 'curl'. If neither is specified,
# a web downloader tool will be auto-detected in the following order: cURL, Wget
# You must have either Wget or cURL installed to use dc-chain.
#force_downloader=wget

### Specify GNU mirror override
# The default mirror for GNU sources is 'ftpmirror.gnu.org'
# This setting overrides the default mirror with a preferred mirror.
#gnu_mirror=mirrors.kernel.org

#####################
### BUILD OPTIONS ###
#####################

### SH toolchain install path
# Specify the directory where the SH toolchain will be installed. This setting
# must match the KOS_CC_BASE setting in your KOS "environ.sh" configuration.
sh_toolchain_path=/media/chris/development/dreamcast/toolchains/sh-elf

### ARM toolchain install path
# Specify the directory where the ARM toolchain will be installed. This setting
# must match the DC_ARM_BASE setting in your KOS "environ.sh" configuration.
arm_toolchain_path=/media/chris/development/dreamcast/toolchains/arm-eabi

### Make jobs (n|<empty>)
# Set this value to the number of parallel jobs you want to run with make.
# For fastest toolchain building, set to the number of CPU threads available.
# Using multiple jobs may cause issues in certain environments and may be
# automatically disabled. If you encounter errors building your toolchain,
# reduce the number of jobs to 1 to avoid issues and ease troubleshooting.
makejobs=16

### Verbose (1|0)
# Choose whether to actively display compilation messages on the screen.
# Messages are saved to the build log files regardless of this setting.
verbose=1

### Erase (1|0)
# Erase build directories as toolchain components are installed to save space.
erase=1

### Install toolchain debug symbols (1|0)
# Choose whether to keep the debugging symbols for the toolchain.
# This is only useful if you wish to debug the toolchain itself.
#toolchain_debug=1

########################
### LANGUAGE OPTIONS ###
########################

### Enable C++
# Builds C++ support, including the C++ compiler and its standard library. The
# vast majority of language features are supported, with C++23 and early C++26
# support included. KallistiOS provides several examples with C++ support, so it
# is enabled by default. Adding C++ support requires extra disk space and
# compilation time, so you may disable it here if you do not plan on using C++.
enable_cpp=1

### Enable Objective-C
# Builds Objective-C support. Currently, this builds the Obj-C compiler and
# runtime, but does not build a standard library. The GNUStep Base Library is
# currently in active development for future inclusion in the kos-ports library.
# KallistiOS provides examples with Obj-C support, so it is enabled by default.
enable_objc=1

### Enable Objective C++
# Builds Objective C++ support. This hybrid language allows both C++ and Obj-C
# methods to be called from both contexts. Enabled by default as both the C++
# and Obj-C languages are already enabled by default.
enable_objcpp=1

### Enable D
# Builds D support. This will build the D compiler, but does not build the
# Phobos Runtime Library. D support may only be enabled on POSIX platforms which
# have its external dependencies provided through a host DMD, GDC, or LDC
# compiler installation.
#enable_d=1

### Enable Ada
# Builds Ada support. This will build the GNAT Ada compiler and tools but does
# not build the libada runtime library. In order for this build to succeed, the
# host's GCC version must match the target Dreamcast version, plus the same
# version of GNAT and its tools must be preinstalled (and in your path).
#enable_ada=1

### Enable Rust
# Builds the work-in-progress GCCRS Rust compiler frontend for GCC. Requires a
# development version of GCC. Use with the gccrs-dev toolchain profile for the
# most recent updates to the GCCRS compiler. This option is not to be confused
# with using this toolchain as a backend for rustc; see libgccjit below.
#enable_rust=1

### Enable libgccjit
# Enables the libgccjit embeddable GCC library. This is most commonly enabled to
# allow using this toolchain as a backend for the official rustc compiler. Use
# with the rustc-dev toolchain profile if this is your use case.
#enable_libgccjit=1

###################
### GCC OPTIONS ###
###################

### Floating point precision support (m4-single-only|m4-single|m4)
# Build support for various SH4 floating-point operation ABIs. KallistiOS only
# officially supports single-precision-only mode (m4-single-only); however,
# experimental support for single-precision-default mode (m4-single) has been
# added to allow for the use of full 64-bit doubles. You may also include
# double-precision-default (m4) which is untested and unsupported, or you may
# choose to specify only m4-single-only to save disk space.
precision_modes=m4-single-only,m4-single

### Default floating point mode (m4|m4-single|m4-single-only)
# Choose the default floating point precision ABI used when GCC is invoked. This
# may be overridden by passing -m4-single-only, -m4-single, or -m4 to GCC.
# KOS currently only officially supports m4-single-only, so it is the default.
default_precision=m4-single-only

### GCC threading model (single|kos)
# KallistiOS patches to GCC provide a 'kos' thread model, which should be used.
# If you want to disable threading support for C++, Objective-C, and so forth,
# you can set this option to 'single'.
thread_model=kos

### Automatic patching for KallistiOS (1|0)
# Uncomment this option if you want to disable applying KallistiOS patches to
# toolchain source files before building. This will disable usage of the 'kos'
# thread model. Only do this if you understand what you are doing.
#use_kos_patches=0

### Disable GCC Native Language Support (1|0)
# By default, NLS allows GCC to output diagnostics in non-English languages.
# Uncomment this option to disable NLS and force GCC to output in English.
#disable_nls=1

######################
### NEWLIB OPTIONS ###
######################

### Automatic patching for Newlib (1|0)
# Uncomment this option if you want to disable the automatic patching of Newlib
# needed by KallistiOS. This will keep the generated toolchain completely raw.
# This will also disable the 'kos' thread model. Only do this if you understand
# what you are doing.
#auto_fixup_sh4_newlib=0

### C99 format specifier support (1|0)
# Define this to build Newlib with additional support for the C99 format
# specifiers, used by printf and friends. These include support for size_t,
# ptrdiff_t, intmax_t, and sized integral types.
newlib_c99_formats=1

### Multibyte character set support (1|0)
# Define this to build Newlib with additional multibyte support. This enables
# three special locales: "C-JIS", "C-SJIS", and "C-EUCJP". The multibyte
# support extends to the mb and wc functions in stdlib as well as format
# characters for the printf and scanf family of routines.
#newlib_multibyte=1

### iconv() character encoding conversions support (encoding list)
# Define a list here to enable support for the iconv() function and <iconv.h>
# header file. The given comma separated list defines for which encoding types
# to include bidirectional conversion support. For the full list of available
# encodings, see the Newlib configuration documentation.
#newlib_iconv_encodings=us_ascii,utf8,utf16,ucs_2_internal,ucs_4_internal

### Optimize Newlib for space (1|0)
# Uncomment this option to optimize for space when building Newlib. This will
# build Newlib with compiler flags which favor smaller code sizes over faster
# performance.
#newlib_opt_space=1

###################
### C++ OPTIONS ###
###################

### Timezone database support (1|0|path)
# Uncomment this option to enable building support for C++'s std::chrono::tzdb
# into your C++ standard library by using a combination of the system's local
# timezone DB and one dynamically fetched from the "IANA Time Zone Database."
# Without support enabled, only the "UTC" and "GMT" timezones will be defined.
# You can optionally provide the path to a directory with a custom "tzdata.zi"
# database file. NOTE: Enabling this will result in larger C++ binaries!
#libstdcxx_tzdb=1

#######################
### WINDOWS OPTIONS ###
#######################

### MinGW/MSYS
# Standalone binaries (1|0)
# Uncomment this option if you want static binaries that are standalone and
# require no dependencies. When this option is used, binaries can be run outside
# the MinGW/MSYS environment. This is NOT recommended; only do this if you know
# what you are doing.
#standalone_binary=1

### Force installation of BFD for SH (1|0)
# Uncomment this option if you want to force installation of 'libbfd' for the SH
# toolchain. This is required for MinGW/MSYS and can't be disabled in this
# scenario. This option is here mainly if you want to force the installation of
# 'libbfd' under other environments, but this won't be necessary in most cases,
# as 'libelf' is used almost everywhere. Please note, 'libbfd' is not portable
# if you built it on another environment. Only use this flag if you understand
# what you are doing.
#sh_force_libbfd_installation=1
</pre>

Make durchführen:
<pre>
PATH=/usr/bin make all
</pre>

==== KOS bauen ====

eviron.sh:
<pre>
# KallistiOS Environment Settings
#
# This is a sample script for configuring and customizing your
# KOS build environment. Modify it to suit your setup. Several
# settings may be enabled optionally or are provided with
# alternative values.
#
# This script is typically sourced in your current shell environment
# (probably by .bashrc, .bash_profile, or something similar), so that
# the KOS environment is set up automatically for each shell session.
#

# Build Architecture
#
# Set the major architecture you'll be building for.
# The only option here is "dreamcast" as of KOS 2.0.0.
#
export KOS_ARCH="dreamcast"

# Build Sub-Architecture
#
# Defines the sub architecture your configuration
# is targeting or uses an existing value that
# can be set externally via your IDE.
#
# Valid values:
# "pristine" - Dreamcast console or HKT-0120 devkit (default)
# "naomi" - NAOMI or NAOMI 2 arcade board
#
if [ -z "${KOS_SUBARCH}" ] ; then
export KOS_SUBARCH="pristine"
else
export KOS_SUBARCH
fi

# KOS Root Path
#
# Specifies the path to the KOS root directory
#
export KOS_BASE="/media/chris/development/dreamcast/toolchains/kos"

# KOS-Ports Path
#
# Specifies the path to the KOS-ports directory
#
export KOS_PORTS="${KOS_BASE}/../kos-ports"

# SH Compiler Prefixes
#
# Specifies the path to and prefix for the main SH
# GCC toolchain used to target the Dreamcast's SH4 CPU.
#
export KOS_CC_BASE="/media/chris/development/dreamcast/toolchains/sh-elf"
export KOS_CC_PREFIX="sh-elf"

# ARM Compiler Prefixes
#
# Specifies the path to and prefix for the additional ARM
# GCC toolchain used to target the Dreamcast's AICA SPU.
#
export DC_ARM_BASE="/media/chris/development/dreamcast/toolchains/arm-eabi"
export DC_ARM_PREFIX="arm-eabi"

# External Dreamcast Tools Path
#
# Specifies the path where Dreamcast tools that are not part of KOS are to be
# installed. This includes, for example, dc-tool-ip, dc-tool-serial, and the
# mrbc bytecode compiler. This directory, along with SH and ARM compiler
# toolchains, will be added to your PATH environment variable.
#
export DC_TOOLS_BASE="/media/chris/development/dreamcast/tools"

# CMake Toolchain Path
#
# Specifies the path to the toolchain file used to target
# KOS with the "cmake" build utility.
#
export KOS_CMAKE_TOOLCHAIN="${KOS_BASE}/utils/cmake/dreamcast.toolchain.cmake"

# Genromfs Utility Path
#
# Specifies the path to the utility which is used by KOS
# to create romdisk filesystem images.
#
export KOS_GENROMFS="${KOS_BASE}/utils/genromfs/genromfs"

# Make Utility
#
# Configures the tool to be used as the main "make" utility
# for building GNU Makefiles. On a platform such as BSD,
# the default can be changed to "gmake," for the GNU
# implementation.
#
export KOS_MAKE="make"
#export KOS_MAKE="gmake"

# Loader Utility
#
# Specifies the loader to be used with the "make run" targets
# in the KOS examples. Defaults to using a preconfigured version
# of dc-tool. Use one of the other options for a manual dc-tool-ip
# or dc-tool-serial configuration, remembering to change the values
# for the Dreamcast's IP address or the serial port interface.
#
#export KOS_LOADER="dc-tool -x"
export KOS_LOADER="dc-tool-ip -t dreamcast -x"
#export KOS_LOADER="dc-tool-ser -t /dev/ttyS0 -x"

# Default Compiler Flags
#
# Resets build flags. You can also initialize them to some preset
# default values here if you wish.
#
export KOS_INC_PATHS=""
export KOS_CFLAGS=""
export KOS_CPPFLAGS=""
export KOS_LDFLAGS=""
export KOS_AFLAGS=""
export DC_ARM_LDFLAGS=""

# Debug Builds
#
# Controls whether to disable additional debugging checks and assertions,
# such as for parameter validation or internal errors. Uncomment this if
# you do not wish to compile with this additional logic enabled.
#
#export KOS_CFLAGS="${KOS_CFLAGS} -DNDEBUG"

# Optimization Level
#
# Controls the baseline optimization level to use when building.
# Typically this is -Og (debugging), -O0, -O1, -O2, or -O3.
# NOTE: For our target, -O4 is a valid optimization level that has
# been seen to have some performance impact as well.
#
export KOS_CFLAGS="${KOS_CFLAGS} -O2"

# Additional Optimizations
#
# Uncomment this to enable what has been found empirically to be
# the optimal set of additional flags for release build performance
# on the current stable toolchain. NOTE: Certain KOS-ports and examples
# do not work properly with "-flto=auto"!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -freorder-blocks-algorithm=simple -flto=auto"

# Frame Pointers
#
# Controls whether frame pointers are emitted or not. Disabled by
# default. Enable them if you plan to use GDB for debugging.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fomit-frame-pointer"
#export KOS_CFLAGS="${KOS_CFLAGS} -fno-omit-frame-pointer -DFRAME_POINTERS"

# Stack Protector
#
# Controls whether GCC emits extra code to check for buffer overflows or stack
# smashing, which can be very useful for debugging. -fstack-protector only
# covers vulnerable objects, while -fstack-protector-strong provides medium
# coverage, and -fstack-protector-all provides full coverage. You may also
# override the default stack excepton handler by providing your own
# implementation of "void __stack_chk_fail(void)."
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fstack-protector-all"

# GCC Builtin Functions
#
# Comment out this line to enable GCC to use its own builtin implementations of
# certain standard library functions. Under certain conditions, this can allow
# compiler-optimized implementations to replace standard function invocations.
# The downside of this is that it COULD interfere with Newlib or KOS implementations
# of these functions, and it has not been tested thoroughly to ensure compatibility.
#
export KOS_CFLAGS="${KOS_CFLAGS} -fno-builtin"

# Fast Math Instructions
#
# Uncomment this line to enable the optimized fast-math instructions (FSSRA,
# FSCA, and FSQRT) for calculating sin/cos, inverse square root, and square roots.
# These can result in substantial performance gains for these kinds of operations;
# however, they do so at the price of accuracy and are not IEEE compliant.
# NOTE: Enabling this option will also override -fno-builtin!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -fbuiltin -ffast-math -ffp-contract=fast"

# SH4 Floating-Point Arithmetic Precision
#
# KallistiOS supports both the single-precision-default ABI (m4-single) and the
# single-precision-only ABI (m4-single-only). When using m4-single, the SH4 will
# be in single-precision mode upon function entry but will switch to double-
# precision mode if 64-bit doubles are used. When using m4-single-only, the SH4
# will always be in single-precision mode and 64-bit doubles will be truncated to
# 32-bit floats. Historically, m4-single-only was used in both official and
# homebrew Dreamcast software, but m4-single is the default as of KOS 2.2.0 to
# increase compatibility with newer libraries which require 64-bit doubles.
#
# WARNING: When adjusting this setting, make sure all software, including
# kos-ports and linked external libraries, are rebuilt using the same
# floating-point precision ABI setting!
#
export KOS_SH4_PRECISION="-m4-single"

# Use LRA (Local Register Allocator) Pass
#
# Uncomment this line to use the modern Local Register Allocator pass during
# code generation instead of the default older reload pass. This option is
# known to be unstable or less performant for SH at this time, but will likely
# become mandatory in future versions of GCC, so feel free to help us test.
# Only enable this setting if you understand what you are doing!
#
#export KOS_CFLAGS="${KOS_CFLAGS} -mlra"

# Shared Compiler Configuration
#
# Include sub architecture-independent configuration file for shared
# environment settings. If you want to configure additional compiler
# options or see where other build flags are set, look at this file.
#
. ${KOS_BASE}/environ_base.sh
</pre>

<pre>
. environ.sh
make
</pre>

=== KOS-ports bauen ===

<pre>
git clone --recursive https://github.com/KallistiOS/kos-ports kos-ports
kos-ports/utils/build-all.sh
</pre>