scsi: devinfo: drop "Promise"/"" entry from scsi_static_device_list
PATCH 1/1 scsi: devinfo: drop "Promise"/"" entry from scsi_static_device_list
"Promise" "" is too generic, and affects any device from this vendor, regardless of its actual bugs. Applying BLIST_SPARSELUN globally is an overbroad approach.
It was originally intended solely for "st_shasta" family of controllers (SuperTrak EX8350/8300/16350/16300/EX12350/EX4350/EX24350), as established in commit e0b2e597d5dd ([SCSI] stex: fix id mapping issue):
" -- add an entry in scsi_devindo.c to force sequential lun scan
(for st_shasta controllers)"
Removing this catch-all entry prevents unintended behavior on other Promise storage products that do not require this legacy workaround.
---
All email addresses from @tw.promise.com or @promise.com, in the git rep, no longer exist.
--- drivers/scsi/scsi_devinfo.c | 1 - 1 file changed, 1 deletion(-)
drivers/scsi/scsi_devinfo.c +0 -1
@@ -216,7 +216,6 @@ static struct { {"PIONEER", "CD-ROM DRM-604X", NULL, BLIST_FORCELUN | BLIST_SINGLELUN}, {"PIONEER", "CD-ROM DRM-624X", NULL, BLIST_FORCELUN | BLIST_SINGLELUN}, {"Promise", "VTrak E610f", NULL, BLIST_SPARSELUN | BLIST_NO_RSOC}, - {"Promise", "", NULL, BLIST_SPARSELUN}, {"QEMU", "QEMU CD-ROM", NULL, BLIST_SKIP_VPD_PAGES}, {"QNAP", "iSCSI Storage", NULL, BLIST_MAX_1024}, {"SYNOLOGY", "iSCSI Storage", NULL, BLIST_MAX_1024},
Signature · 1 line
2.54.0
Quoted 31 lines · Xose Vazquez Perez
"Promise" "" is too generic, and affects any device from this vendor, regardless of its actual bugs. Applying
BLIST_SPARSELUNglobally is an overbroad approach.It was originally intended solely for "
st_shasta" family of controllers (SuperTrak EX8350/8300/16350/16300/EX12350/EX4350/EX24350), as established in commite0b2e597d5dd([SCSI] stex: fix id mapping issue):
" -- add an entry inscsi_devindo.cto force sequential lun scan
(forst_shastacontrollers)"Removing this catch-all entry prevents unintended behavior on other Promise storage products that do not require this legacy workaround.
Cc Christoph Hellwig <[email protected]>Cc James E.J. Bottomley <[email protected]>Cc Martin K. Petersen <[email protected]>Cc SCSI-ML <[email protected]>Signed-off-by Xose Vazquez Perez <[email protected]>---
All email addresses from @tw.promise.com or @promise.com, in the git rep, no longer exist.--- drivers/scsi/scsi_devinfo.c | 1 - 1 file changed, 1 deletion(-)drivers/scsi/scsi_devinfo.c +0
-0
drivers/scsi/scsi_devinfo.c static struct { · line 221
Quoted 7 lines
drivers/scsi/scsi_devinfo.c +0
-1@@ -216,7 +216,6 @@ static struct { {"PIONEER", "CD-ROM DRM-604X", NULL, BLIST_FORCELUN | BLIST_SINGLELUN}, {"PIONEER", "CD-ROM DRM-624X", NULL, BLIST_FORCELUN | BLIST_SINGLELUN}, {"Promise", "VTrak E610f", NULL, BLIST_SPARSELUN | BLIST_NO_RSOC}, - {"Promise", "", NULL, BLIST_SPARSELUN}, {"QEMU", "QEMU CD-ROM", NULL, BLIST_SKIP_VPD_PAGES}, {"QNAP", "iSCSI Storage", NULL, BLIST_MAX_1024}, {"SYNOLOGY", "iSCSI Storage", NULL, BLIST_MAX_1024},
Please, drop this patch.
The stex driver forces a sequential LUN scan for st_shasta by failing REPORT_LUNS:
case REPORT_LUNS:
/*
* The shasta firmware does not report actual luns in the
* target, so fail the command to force sequential lun
* scan.
* Also, the console device does not support this command.
*/
if (hba->cardtype == st_shasta || id == host->max_id - 1) {
stex_invalid_field(cmd, done);
return 0;
}
break;Without the BLIST_SPARSELUN flag, this scan will truncate at the first LUN gap.
Although later firmware versions might fix this problem:
8<---
Version/Build - 1.1.0.25
Date Release - 11/2/2005
2. Compact Lun mapping during firmware initialize to avoid lun gap. [...]
Version/Build - 1.2.0.10
Date Release - 02/09/2006
3. Add REPORT LUNS support. (not enabled yet)
[...]
Version/Build - 1.3.0.20
1. Enable ENGINE_SUPPORT_REPORT_LUNS,ENGINE_SUPPORT_SCSIOP_MODE_SENSE_SELECT and ENGINE_SUPPORT_SCSIOP_INQUIRY_EVPD to support Vista SCSI compliance test.
2. To work with item 1, firmware needs to return actual transferred data length.
This requires driver update at the same time.
8<---
The driver logic remains tied to this legacy behavior. And given the lack of active maintenance and vendor support to verify firmware-level changes, it is safer to keep the current BLIST_SPARSELUN entry to prevent any regression.