Please consider this change for inclusion in the 2.4 kernel.
This change is required to support the new MSA30 storage enclosure.
If you do a SCSI inquiry to a SATA disk bad things happen. This patch prevents
the inquiry from going to SATA disks.
cciss_scsi.c | 2 ++
1 files changed, 2 insertions(+)
------------------------------------------------------------------------------
diff -burN lx2425.orig/drivers/block/cciss_scsi.c lx2425/drivers/block/cciss_scsi.c
--- lx2425.orig/drivers/block/cciss_scsi.c 2003-11-28 12:26:19.000000000 -0600
+++ lx2425/drivers/block/cciss_scsi.c 2004-03-04 10:21:33.000000000 -0600
@@ -589,6 +589,8 @@
for(i=0; i<num_luns; i++) {
/* Execute an inquiry to figure the device type */
+ /* Skip over masked devices */
+ if (ld_buff->LUN[i][3] & 0xC0) continue;
memset(inq_buff, 0, sizeof(InquiryData_struct));
memcpy(scsi3addr, ld_buff->LUN[i], 8); /* ugly... */
return_code = sendcmd(CISS_INQUIRY, cntl_num, inq_buff,
On Thu, 2004-03-25 at 16:36, [email protected] wrote:
> Please consider this change for inclusion in the 2.4 kernel.
do you have a matching 2.6 patch ?
[email protected] wrote:
> Please consider this change for inclusion in the 2.4 kernel.
>
> This change is required to support the new MSA30 storage enclosure.
> If you do a SCSI inquiry to a SATA disk bad things happen. This patch prevents
> the inquiry from going to SATA disks.
I 'ack' both of those patches, but am still curious: wouldn't you want
to either (a) simulate an inquiry page via ATA's identify device or (b)
allow userspace to issue identify device?
Jeff
Yep, that would the right thing to do. But right now management wants this :(
-----Original Message-----
From: Jeff Garzik [mailto:[email protected]]
Sent: Thursday, March 25, 2004 10:14 AM
To: Miller, Mike (OS Dev)
Cc: [email protected]; [email protected]
Subject: Re: cciss updates [1 of 2]
[email protected] wrote:
> Please consider this change for inclusion in the 2.4 kernel.
>
> This change is required to support the new MSA30 storage enclosure.
> If you do a SCSI inquiry to a SATA disk bad things happen. This patch prevents
> the inquiry from going to SATA disks.
I 'ack' both of those patches, but am still curious: wouldn't you want
to either (a) simulate an inquiry page via ATA's identify device or (b)
allow userspace to issue identify device?
Jeff
Miller, Mike (OS Dev) wrote:
> Yep, that would the right thing to do. But right now management wants this :(
Cool... well I see no reason to delay these 2 patches, as what you
posted is an intermediate step towards The Right Thing. And Linux
kernel development is all about showing the "evolution" of the code, not
just the end result.
Jeff