Hi all,

I am experiencing a consistent BD-R burning failure with K3b/cdrecord on kernel 6.x
that used to work fine on kernel 5.x. I have reproduced it on two completely different
machines, which rules out a hardware-specific issue.

== Hardware ==

Machine 1: Samsung NP505P with HLDS BDRE BU50N BW51 optical drive (/dev/sr0)
Machine 2: Dell 7140 tablet with the same usb external optical drive

== Software ==

K3b Version:  26.4.2
KDE Version:  6.26.0
Qt Version:   6.11.1
mkisofs:      3.2a09
cdrecord:     3.2a09

== Kernels Tested ==

- Kernel 6.12.94-1-MANJARO → FAILS
- Kernel 6.18              → FAILS (on Dell 7140)
- Kernel 5.x               → WORKED fine (approximately one year ago)

== What Happens ==

K3b starts a SAO BD-R burn session and cdrecord immediately fails at sector 0
with the following critical error:

  Sense Key: 0x5 Illegal Request
  Sense Code: 0x21 Qual 0x02 (invalid address for write)

Additionally, the following warnings appear before the write attempt:

  WARNING: Phys disk size 4452354 differs from rzone size 24438784! Prerecorded disk?
  WARNING: Phys start: 25088 Phys end 4477441

Note: the disc is new and blank. The "Phys disk size" warning appears to be a known
cdrecord misinterpretation of the BD-R disc type identifier bytes ("BDR"), not a real
problem with the disc itself.

The error message "buffer underrun" printed by cdrecord at the end of the log is
misleading, it is a secondary consequence of the write failure, not the root cause.

Notably, mkisofs itself did NOT crash - it only terminated because cdrecord exited
early and closed the pipe.

== Suspected Root Cause ==

There is a known bug in cdrecord/wodim where, when the kernel minor version is >= 6,
it attempts to access obsoleted device files (/dev/scd<n>), causing the burn to fail.
This was referenced in Debian bug #833346 and a related Thomas Schmitt post on the
debian-user mailing list (November 2025). Since both kernels I tested have a minor
version >= 6 (6.12 and 6.18), this strongly correlates with my observations.

== Possible Workaround ==

Switching from cdrecord to xorriso or cdrskin (both from the libburn project)
may resolve the issue. Both tools support BD-R and BD-R DL (50GB dual layer)
media and are actively maintained.

Install xorriso:
  sudo pacman -S xorriso

Burn with xorriso as cdrecord drop-in:
  xorriso -as cdrecord -v dev=/dev/sr0 -dao your_image.iso

Or in K3b: Settings → Configure K3b → Programs → point to xorriso.

== Full cdrecord Log ==

Devices
-----------------------
hp HLDS BDRE   BU50N BW51 (/dev/sr0, CD-R, CD-RW, CD-ROM, DVD-ROM, DVD-R, DVD-RW, DVD-R DL, BD-ROM, BD-R, BD-RE, DVD+R, DVD+RW, DVD+R DL) [DVD-ROM, DVD-R Sequential, DVD-R Dual Layer Sequential, DVD-R Dual Layer Jump, DVD-RAM, DVD-RW Restricted Overwrite, DVD-RW Sequential, DVD+RW, DVD+R, DVD+R Dual Layer, CD-ROM, CD-R, CD-RW, BD-ROM, BD-R Sequential (SRM), BD-R Random (RRM), BD-RE] [SAO, TAO, RAW, SAO/R96P, SAO/R96R, RAW/R16, RAW/R96P, RAW/R96R, Restricted Overwrite, Layer Jump, Random Recording, Sequential Recording, Sequential Recording + POW] [%7]

K3b::IsoImager
-----------------------
mkisofs print size result: 24320424 (49808228352 bytes)

System
-----------------------
K3b Version: 26.4.2
KDE Version: 6.26.0
Qt Version:  6.11.1
Kernel:      6.12.94-1-MANJARO

Used versions
-----------------------
mkisofs: 3.2a09
cdrecord: 3.2a09

cdrecord
-----------------------
scsidev: '/dev/sr0'
devname: '/dev/sr0'
scsibus: -2 target: -2 lun: -2
Warning: Open by 'devname' is unintentional and not supported.
Linux sg driver version: 3.5.27
SCSI buffer size: 64512
cdrecord: Warning: Cannot read drive buffer.
cdrecord: Warning: The DMA speed test has been skipped.
Cdrecord-ProDVD-ProBD-Clone 3.02a09 (--) Copyright (C) 1995-2016 Joerg Schilling
TOC Type: 3 = CD-ROM XA mode 2
Using libscg version 'schily-0.9'.
Driveropts: 'burnfree'
atapi: 1
Device type    : Removable CD-ROM
Version        : 0
Response Format: 2
Capabilities   :
Vendor_info    : 'hp HLDS '
Identifikation : 'BDRE   BU50N    '
Revision       : 'BW51'
Device seems to be: Generic mmc2 DVD-R/DVD-RW/DVD-RAM.
Current: BD-R sequential recording
Profile: BD-RE
Profile: BD-R random recording
Profile: BD-R sequential recording (current)
Profile: BD-ROM
Profile: DVD+R/DL
Profile: DVD+R
Profile: DVD+RW
Profile: DVD-R/DL layer jump recording
Profile: DVD-R/DL sequential recording
Profile: DVD-RW sequential recording
Profile: DVD-RW restricted overwrite
Profile: DVD-RAM
Profile: DVD-R sequential recording
Profile: DVD-ROM
Profile: CD-RW
Profile: CD-R
Profile: CD-ROM
Profile: Removable Disk
Using generic SCSI-3/mmc-3 BD-R driver (mmc_bdr).
Driver flags   : NO-CD BD MMC-3 BURNFREE
Supported modes: PACKET SAO LAYER_JUMP
Drive buf size : 1900544 = 1856 KB
FIFO size      : 4194304 = 4096 KB
resid: 60
resid: 28
Track 01: data  47500 MB       
Total size:     47500 MB = 24320424 sectors
Current Secsize: 2048
WARNING: Phys disk size 4452354 differs from rzone size 24438784! Prerecorded disk?
WARNING: Phys start: 25088 Phys end 4477441
Blocks total: 24438784 Blocks current: 24438784 Blocks remaining: 118360
Reducing transfer size from 64512 to 32768 bytes.
Starting to write CD/DVD/BD at speed 0 in real SAO mode for single session.
Last chance to quit, starting real write in 3 seconds.
   2 seconds.
   1 seconds.
   0 seconds. Operation starts.
Waiting for reader process to fill input buffer ... resid: 60
input buffer ready.
Starting new track at sector: 0
Track 01:    0 of 47500 MB written.
cdrecord: Input/output error. write_g1: scsi sendcmd: no error
CDB:  2A 00 00 00 00 00 00 00 10 00
status: 0x2 (CHECK CONDITION)
Sense Bytes: 70 00 05 00 00 00 00 0A 00 00 00 00 21 02 00 00
Sense Key: 0x5 Illegal Request, Segment 0
Sense Code: 0x21 Qual 0x02 (invalid address for write) Fru 0x0
Sense flags: Blk 0 (not valid)
resid: 32768
cmd finished after 1.794s timeout 200s
cdrecord: The current problem looks like a buffer underrun.
cdrecord: It looks like 'driveropts=burnfree' does not work for this drive.
cdrecord: Please report.
cdrecord: Make sure that you are root, enable DMA and check your HW/OS set up.
write track data: error after 0 bytes
Writing  time:   36.682s (00:00:36.682)
Average write speed 302.1x.
Fixating...
operation 99% done
=== last message repeated 16 times. ===
Fixating time:   32.029s (00:00:32.029)
cdrecord: fifo had 129 puts and 2 gets.
cdrecord: fifo was 0 times empty and 1 times full, min fill was 99%.
BURN-Free was not used.

cdrecord command:
-----------------------
/usr/bin/cdrecord -v gracetime=2 dev=/dev/sr0 speed=6 -sao driveropts=burnfree -xa -tsize=24320424s -

mkisofs
-----------------------
24320424
mkisofs: Warning: Cannot add inode hints with -no-cache-inodes.
=== last message repeated 2 times. ===
Setting input-charset to 'UTF-8' from locale.
  0.00% done, estimate finish Sat Jun 27 10:22:54 2026
=== last message repeated 2 times. ===
  0.01% done, estimate finish Sat Jun 27 10:22:54 2026

mkisofs calculate size command:
-----------------------
/usr/bin/mkisofs -gui -graft-points -print-size -quiet -volid TO_REC_214 -volset  -appid K3B THE CD KREATOR (C) 1998-2018 SEBASTIAN TRUEG, MICHAL MALEK AND LESLIE ZHAI -publisher  -preparer  -sysid LINUX -volset-size 1 -volset-seqno 1 -sort /tmp/k3b.tnzxVB -rational-rock -hide-list /tmp/k3b.EyUgqE -joliet -joliet-long -hide-joliet-list /tmp/k3b.qNIZwf -no-cache-inodes -udf -full-iso9660-filenames -iso-level 3 -path-list /tmp/k3b.mUYTGS

mkisofs command:
-----------------------
/usr/bin/mkisofs -gui -graft-points -volid TO_REC_214 -volset  -appid K3B THE CD KREATOR (C) 1998-2018 SEBASTIAN TRUEG, MICHAL MALEK AND LESLIE ZHAI -publisher  -preparer  -sysid LINUX -volset-size 1 -volset-seqno 1 -sort /tmp/k3b.hlvqfl -rational-rock -hide-list /tmp/k3b.fqnxtl -joliet -joliet-long -hide-joliet-list /tmp/k3b.Oainxp -no-cache-inodes -udf -full-iso9660-filenames -iso-level 3 -path-list /tmp/k3b.hidgrN

== Questions for the Community ==

1. Can anyone else reproduce this with kernel 6.x and BD-R media?
2. Is there a cleaner fix within K3b itself, or is switching to xorriso the
   recommended long-term solution?
3. Should this be escalated to the upstream cdrecord/wodim maintainers or
   to the linux-scsi mailing list?

Thanks in advance.