Re: [PATCH v5 01/40] scsi: define all additional sense codes and their qualifiers

[Bart Van Assche <bart.vanassche@gmail.com>]

On 9/6/26 8:51 PM, Damien Le Moal wrote:
> On 9/7/26 12:14, sashiko-bot@kernel.org wrote:
>> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
>> - [Low] The newly introduced `scsi_sense_code` enum injects hundreds of generic, un-prefixed identifiers into the global namespace.
>> --
>>
>> commit 483fa6eeda05c51714847f3283212cc3d768cf57
>> Author: Damien Le Moal <dlemoal@kernel.org>
>>
>> scsi: define all additional sense codes and their qualifiers
>>
>> Introduce include/scsi/scsi_sense.h to define all existing additional
>> sense codes and their combinations with additional sense code qualifiers.
>> The sense key macro definitions in include/scsi/scsi_proto.h are also
>> moved into this new file, and scsi_sense.h is included in scsi_proto.h.
>>
>>> diff --git a/include/scsi/scsi_sense.h b/include/scsi/scsi_sense.h
>>> new file mode 100644
>>> index 0000000000000..e668ac1c2302c
>>> --- /dev/null
>>> +++ b/include/scsi/scsi_sense.h
>> [ ... ]
>>> +enum scsi_sense_code {
>>> +	ASC(NO_ADDITIONAL_SENSE_INFORMATION,			0x00),
>>> +	FILEMARK_DETECTED					= 0x0001,
>> [ ... ]
>>> +	ASC(WARNING,						0x0B),
>>
>> [Severity: Low]
>> Does defining scsi_sense_code this way inject generic, un-prefixed
>> identifiers like WARNING and WRITE_ERROR into the global namespace?
>>
>> Could this cause symbol collisions since scsi_sense.h is widely included
>> through scsi_proto.h?
> 
> Martin, James,
> 
> This is a valid concern which I had when writing scsi_sense.h. But so far,
> building with allyesconfig/allmodconfig did not great any issue. But I can
> change these names if you think it is better to avoid issues going forward.

Hi Damien,

My opinion is that introducing a new un-prefixed global identifier with
a name like WARNING is completely unacceptable because it clashes with
an existing global identifier. Please fix.

Thanks,

Bart.