#84 hdparm -I /dev/sda crashes device

[Siegfried Schnieders]

I am using Debian 10 "Buster"

# hdparm -V
hdparm v9.58
# uname -v
#1 SMP Debian 4.19.28-2 (2019-03-15)

When I issue the command

hdparm -I /dev/sda

I get

ATA device, with non-removable media
    Model Number:       2.5" SATA SSD 3SE                       
    Serial Number:      20140430AA0000000263
    Firmware Revision:  S130710 
    Transport:          Serial, ATA8-AST, SATA 1.0a, SATA II Extensions, SATA Rev 2.5, SATA Rev 2.6, SATA Rev 3.0
Standards:
    Supported: 8 7 6 5 
    Likely used: 8
Configuration:
    Logical     max current
    cylinders   16383   16383
    heads       16  16
    sectors/track   63  63
    --
    CHS current addressable sectors:    16514064
    LBA    user addressable sectors:    31277232
    LBA48  user addressable sectors:    31277232

Then after about 30 seconds I get kernel messages

[  106.761031] ata1.01: revalidatin failed (errno=-5)
[  129.285000] ata1.01: revalidatin failed (errno=-5)
[  173.316994] ata1.01: revalidatin failed (errno=-5)
SG_IO: bad/missing sense data, sb[]: [  185.648280]: print_req_error: I/O error, dev sda, sector 2048
... more Buffer I/O error and print_req_error ...
[  185.625550] Aborting journal on device sda1-8
... more Buffer I/O error and print_req_error ...
[  185.652707] Ext4-fs (sda1): previous I/O error to superblock detected
... more Buffer I/O error and print_req_error ...
[  185.652840] JDB2: Error -5 detected when updating journal superblock for sda1-8
[  185.653050] Ext4-fs error (device sda1): ext4_journal_check_start:61: Detected aborted journal 
[  185.653052] Ext4-fs (sda1): Remounting file system read-only
... more Buffer I/O error and print_req_error ...
[  185.653096] Ext4-fs (sda1): I/O error while writing superblock
... more Buffer I/O error and print_req_error ...
70 00 05 00 00 00 00 0a 00 40 50 00 21 04 00 ... 00
    Logical  Sector size:                   512 bytes
    ... more hdparm information ...
Checksum correct.

Afterwards the system is unusable. I only can halt it (with magic SysRq). Restart works fine.

The problem was reproducible with another SSD of the same type, and with the SSD in another computer.

The command

hdparm --Istdout /dev/sda | hdparm --Istdin

gives

ATA device, with non-removable media
    Model Number:       2.5" SATA SSD 3SE                       
    Serial Number:      20140430AA0000000263
    Firmware Revision:  S130710 
    Transport:          Serial, ATA8-AST, SATA 1.0a, SATA II Extensions, SATA Rev 2.5, SATA Rev 2.6, SATA Rev 3.0
Standards:
    Supported: 8 7 6 5 
    Likely used: 8
Configuration:
    Logical     max current
    cylinders   16383   16383
    heads       16  16
    sectors/track   63  63
    --
    CHS current addressable sectors:    16514064
    LBA    user addressable sectors:    31277232
    LBA48  user addressable sectors:    31277232
    Logical  Sector size:                   512 bytes

and throws an error

HDIO_DRIVE_CMD(identify) failed: Bad file descriptor

to stdout.

With hdparm version 9.51 (installed from the package from Debian 9) the command hdparm -I /dev/sda work hdparm --Istdout /dev/sda | hdparm --Istdin but have different outputs.

[Siegfried Schnieders]

I downloaded and compiled versions 9.52 - 9.57 and tested the two commands

hdparm -I /dev/sda
hdparm --Istdout /dev/sda | hdparm --Istdin 

With version 9.56 both commands give the same output (hdparm -I with an additional line /dev/sda:).

With versions 9.52 - 9.55 the "-I" calls give the same output als the 9.56 versions.
The "--Istout ... | --Istdin" versions give inplausible information:

ATA device, with non-removable media
powers-up in standby; SET FEATURES subcmd spins-up.
    Model Number:       .2"5S TA ASS DS3 E                      
    Serial Number:      02414003AA0000002036
    Firmware Revision:  1S0317 0
    Transport:          0x3f10
Standards:
    Supported: 14 13 12 11 10 9 
    Likely used: 14
Configuration:
    Logical     max current
    cylinders   65343   0
    heads       4096    0
    sectors/track   16128   0
    --
    Logical/Physical Sector size:           512 bytes
    device size with M = 1024*1024:  2107703808 MBytes
    device size with M = 1000*1000:  2210087628 MBytes (2210087 GB)
    cache/buffer size  = unknown
Capabilities:
    IORDY(may be)(cannot be disabled)
    Standby timer values: spec'd by Vendor
    R/W multiple sector transfer: Max = 128 Current = ?
    DMA: not supported
    PIO: unknown
Integrity word not set (found 0xa505, expected 0x100a5)

Same does the version 9.51 from Debian 9.

Version 9.57 behaves like the originally used version 9.58.

[Siegfried Schnieders]

The error still exists in versoin 9.62.

[Mark Lord]

Kernel (libata) bug, not hdparm.