Re: [PATCH v2 1/5] ata: libata: rename ata_dev_is_zac()

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Incomplete refactoring of ZBC checks leaves SCSI emulation in an inconsistent state for Device Managed drives.
--

commit 8a69ac9e4f68da535d403caef5279dfdb19419f2
Author: Damien Le Moal <[email protected]>

ata: libata: rename ata_dev_is_zac()

The helper function ata_dev_is_zac() is renamed to ata_dev_is_zoned() to
better reflect its checks. Existing callers and an open-coded check are
updated to use the renamed helper.

> diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
> index d54ec1631e9a6..8811bfc373697 100644
> --- a/drivers/ata/libata-scsi.c
> +++ b/drivers/ata/libata-scsi.c

[ ... ]

> @@ -3656,8 +3656,7 @@ static unsigned int ata_scsi_report_supported_opcodes(struct ata_device *dev,
>  		break;
>  	case ZBC_IN:
>  	case ZBC_OUT:
> -		if (ata_id_zoned_cap(dev->id) ||
> -		    dev->class == ATA_DEV_ZAC)
> +		if (ata_dev_is_zoned(dev))

[Severity: High]
Should this refactoring also be applied to ata_scsiop_inq_std() and
ata_scsiop_read_cap() to prevent an inconsistent state for Device Managed
drives?

They still use the identical open-coded check:

drivers/ata/libata-scsi.c:ata_scsiop_inq_std() {
        ...
        if (ata_id_zoned_cap(dev->id) || dev->class == ATA_DEV_ZAC)
                memcpy(rbuf + 58, versions_zbc, sizeof(versions_zbc));
        ...
}

drivers/ata/libata-scsi.c:ata_scsiop_read_cap() {
        ...
        if (ata_id_zoned_cap(dev->id) || dev->class == ATA_DEV_ZAC)
                rbuf[12] = (1 << 4); /* RC_BASIS */
        ...
}

By only replacing the check here in ata_scsi_report_supported_opcodes(),
Device Managed zoned drives (ata_id_zoned_cap == 0x02) will correctly no
longer report support for ZBC commands, but will still incorrectly claim ZBC
capabilities in standard INQUIRY and set the ZBC-specific RC_BASIS bit in
READ CAPACITY (16).

>  			supported = 3;
>  		break;
>  	case SECURITY_PROTOCOL_IN:

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1