XPost: linux.debian.bugs.dist   
   From: iprodha.gab@gmail.com   
      
   Package: linux   
   Version: 5.10.0-36-amd64   
   Severity: important   
   Tags: amd64 bullseye disk-performance ahci regression   
      
   Problem:   
   On AMD A320 chipset systems with AHCI SATA controller, the block I/O layer   
   incorrectly displays multiple schedulers as active simultaneously:   
   "[mq-deadline] [bfq] none". This causes severe system slowdown and desktop   
   unresponsiveness during disk I/O operations.   
      
   How to reproduce:   
   1. Boot Debian 11 (bullseye) on AMD A320 chipset motherboard (ASRock   
   A320M-HDV)   
   2. Connect SATA HDD (TOSHIBA DT01ACA0 7200 RPM)   
   3. Check current scheduler: cat /sys/block/sda/queue/scheduler   
   4. Observe output: mq-deadline [bfq] none  (note two active schedulers in   
   brackets)   
   5. Perform large file copy or disk-intensive operation   
   6. Desktop environment (MATE) becomes unresponsive, system stalls   
      
   Diagnosis:   
   - System: Debian GNU/Linux 11 (bullseye), Kernel: 5.10.0-36-amd64, x86_64   
   - Motherboard: ASRock A320M-HDV (AMD Promontory A320 chipset)   
   - Disk: TOSHIBA DT01ACA0 (HDD, rotational=1, 7200 RPM SATA)   
   - Driver: ahci (multiple AHCI hosts detected: 9 instances)   
   - Scheduler state: mq-deadline [bfq] none  (bfq active but mq-deadline also   
   shown as active)   
   - Expected: Single scheduler in brackets indicating active one   
   - Current bug: Two schedulers shown with brackets [mq-deadline] [bfq]   
   - Physical system (not virtual machine)   
      
   Workaround:   
   Temporary fix:   
   # echo "bfq" > /sys/block/sda/queue/scheduler   
      
   Permanent fix via udev rule:   
   ACTION=="add|change", SUBSYSTEM=="block", ATTR{queue/rotational}=="1",   
   ATTR{queue/scheduler}="bfq"   
      
   Expected behavior:   
   According to kernel documentation and block layer design:   
   1. Only ONE scheduler should be active at a time   
   2. The active scheduler should be shown in brackets: [bfq]   
   3. Available schedulers should be listed after: [bfq] mq-deadline none   
   4. Current display "mq-dedeadline [bfq] none" is inconsistent   
      
   Impact:   
   - Severe desktop usability regression (MATE environment freezes during I/O)   
   - System becomes unresponsive for tens of seconds during file operations   
   - Affects user experience significantly on HDD-based systems   
   - Problem appears to be in scheduler initialization/display logic   
      
   Additional notes:   
   - Problem persists across reboots   
   - Not VM-related (bare metal installation)   
   - Multiple AHCI hosts detected (9), may be related to initialization order   
   - Manual scheduler selection works correctly once set   
      
   Suggested fix locations:   
   1. block/elevator.c - scheduler selection/display logic   
   2. drivers/ata/libata-core.c - ATA device initialization   
   3. /lib/udev/rules.d/60-block-scheduler.rules - default scheduler assignment   
      
   System info for reference:   
   - Debian: 11 (bullseye)   
   - Kernel: 5.10.0-36-amd64   
   - Architecture: x86_64   
   - Disk: TOSHIBA DT01ACA0 (rotational=1)   
   - Driver: ahci (9 instances)   
   - Desktop: MATE   
   - Motherboard: ASRock A320M-HDV (AMD A320 chipset)   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)