libata SATL: WWN for device id VPD page
PATCH 1/1 libata SATL: WWN for device id VPD page
libata implements a SCSI to ATA Translation Layer and should comply with the SAT (ANSI INCITS 431-2007) and SAT-2 (ANSI INCITS 465-2010).
Both standards require that if an ATA device supplies a World Wide Name (WWN) then it should appear in the corresponding SCSI device identification VPD page.
Changelog:
- add NAA descriptor to the device identification VPD page for the addressed logical unit. ATA's WWN is in NAA(5) format.
The attached patch is against lk 2.6.38 .
drivers/ata/libata-scsi.c_vpd83a renamed from drivers/ata/libata-scsi.c +19 -2
@@ -2028,12 +2028,29 @@ static unsigned int ata_scsiop_inq_80(st static unsigned int ata_scsiop_inq_83(struct ata_scsi_args *args, u8 *rbuf) { const int sat_model_serial_desc_len = 68; - int num; + int num, k; + unsigned char wwn_arr[8]; rbuf[1] = 0x83; /* this page code */ num = 4; - /* piv=0, assoc=lu, code_set=ACSII, designator=vendor */ + /* WWN starts at word 108 in IDENTIFY response */ + ata_id_string(args->id, wwn_arr, 108, 8); + /* SPL(SAS-2.1) checks for zeros rather than checking supported bit */ + for (k = 0; (k < 8) && (0 == wwn_arr[k]); ++k) + ; + if (k < 8) { /* found non-zero WWN */ + /* piv=0, assoc=lu, code_set=binary, designator=naa */ + rbuf[num + 0] = 1; + rbuf[num + 1] = 3; + rbuf[num + 2] = 0; + rbuf[num + 3] = 8; + num += 4; + memcpy(rbuf + num, wwn_arr, 8); + num += 8; + } + + /* piv=0, assoc=lu, code_set=ASCII, designator=vendor */ rbuf[num + 0] = 2; rbuf[num + 3] = ATA_ID_SERNO_LEN; num += 4;
Quoted 13 lines · Douglas Gilbert
libata implements a SCSI to ATA Translation Layer and should comply with the SAT (ANSI INCITS 431-2007) and SAT-2 (ANSI INCITS 465-2010).
Both standards require that if an ATA device supplies a World Wide Name (WWN) then it should appear in the corresponding SCSI device identification VPD page.
Changelog:
- add NAA descriptor to the device identification VPD page for the addressed logical unit. ATA's WWN is in NAA(5) format.
Quoted 4 lines
The attached patch is against lk 2.6.38 .
Signed-off-by Douglas Gilbert <[email protected]>
I should have checked 2.6.39-rc9 prior to posting. Almost the same patch as I proposed is already there.
Doug Gilbert