invalid log directory version error

invalid log directory version error

[Karel Balej]

Hello,

ever since updating to kernel v6.17, I have been getting

	ata2.00: Invalid log directory version 0x0000

errors (one upon every boot).

I have not bisected the issue but it seems that the guilty ata_dev_err()
call was introduced in 6d4405b16d37 and by the looks of that it is more
likely just an uncovering of an existing issue rather than a new
problem.

I have not been seeing any functional regressions (which along with the
above is why I'm not marking this as a regression) but error is an error
so I'm wondering whether there is anything to fix (or maybe just tone
down the log level?).

Please let me know, I can also test patches and if necessary, try to
perform the bisect.

The disk referred to as ata2.00 above should be Samsung SSD 840, here
are all related dmesg messages:

	[    1.406544] ata2.00: SATA link up 6.0 Gbps (SStatus 133 SControl 330)
	[    1.411861] ata2.00: Model 'Samsung SSD 840 Series', rev 'DXT07B0Q', applying quirks: noncqtrim zeroaftertrim
	[    1.412014] ata2.00: ATA-9: Samsung SSD 840 Series, DXT07B0Q, max UDMA/133
	[    1.412019] ata2.00: 234441648 sectors, multi 16: LBA48 NCQ (depth 0/32)
	[    1.413802] ata2.00: Features: Dev-Sleep DIPM
	[    1.417160] ata2.00: Invalid log directory version 0x0000
	[    1.417218] ata2.00: configured for UDMA/133
	[    1.449749] ata2.00: Enabling discard_zeroes_data
	[    1.457968] ata2.00: Enabling discard_zeroes_data

Thank you and best regards,
K. B.

Re: invalid log directory version error

[Niklas Cassel]

Hello Karel,

On Wed, Oct 15, 2025 at 07:56:31PM +0200, Karel Balej wrote:
> Hello,
> 
> ever since updating to kernel v6.17, I have been getting
> 
> 	ata2.00: Invalid log directory version 0x0000
> 
> errors (one upon every boot).
> 
> I have not bisected the issue but it seems that the guilty ata_dev_err()
> call was introduced in 6d4405b16d37 and by the looks of that it is more
> likely just an uncovering of an existing issue rather than a new
> problem.
> 
> I have not been seeing any functional regressions (which along with the
> above is why I'm not marking this as a regression) but error is an error
> so I'm wondering whether there is anything to fix (or maybe just tone
> down the log level?).
> 
> Please let me know, I can also test patches and if necessary, try to
> perform the bisect.

There is a fix queued up for this already:
https://git.kernel.org/pub/scm/linux/kernel/git/libata/linux.git/commit/?h=for-6.18-fixes&id=12d724f2852d094d68dccaf5101e0ef89a971cde

Will most likely be included in v6.18-rc2.


Kind regards,
Niklas

Re: invalid log directory version error

[Karel Balej]

Niklas Cassel, 2025-10-15T19:59:55+02:00:
> Hello Karel,
>
> On Wed, Oct 15, 2025 at 07:56:31PM +0200, Karel Balej wrote:
>> Hello,
>> 
>> ever since updating to kernel v6.17, I have been getting
>> 
>> 	ata2.00: Invalid log directory version 0x0000
>> 
>> errors (one upon every boot).
>> 
>> I have not bisected the issue but it seems that the guilty ata_dev_err()
>> call was introduced in 6d4405b16d37 and by the looks of that it is more
>> likely just an uncovering of an existing issue rather than a new
>> problem.
>> 
>> I have not been seeing any functional regressions (which along with the
>> above is why I'm not marking this as a regression) but error is an error
>> so I'm wondering whether there is anything to fix (or maybe just tone
>> down the log level?).
>> 
>> Please let me know, I can also test patches and if necessary, try to
>> perform the bisect.
>
> There is a fix queued up for this already:
> https://git.kernel.org/pub/scm/linux/kernel/git/libata/linux.git/commit/?h=for-6.18-fixes&id=12d724f2852d094d68dccaf5101e0ef89a971cde
>
> Will most likely be included in v6.18-rc2.

Ah, excellent, thank you!

Best regards,
K. B.