On 11/03/2016 05:11 PM, Ewan D. Milne wrote:
On Wed, 2016-11-02 at 22:27 +0100, Hannes Reinecke wrote:
On 11/02/2016 04:44 PM, Ewan D. Milne wrote:
In this and other places the patch changes the code from submitting the
INQUIRY/EVPD command for the page it wants, to calling scsi_get_vpd_page().
scsi_get_vpd_page() will return -EINVAL if the device did not report in
VPD page 0 that the requested page is in the list of supported pages.
Did you actually verify that the RDAC returns all of these VPD pages in
its VPD page 0 list?
I mean, I know it's supposed to, but these are old devices.
Errm.
Old? Don't tell that to the E-Series folk; you'll never again get
something sponsored :-)
No, seriously: RDAC mode is continued to be supported even on the latest
models, and all firmware revisions I've got support VPD page 0x0.
And incidentally, this is the very same method we're using for basically
all tools accessing VPD page 0x83, be it in the kernel or something like
sg3_utils.
_Not_ asking for page 0x0 lead to crashes on several older devices.
You're right, I was curious, so I resurrected our old RDAC, and it does
in fact report the necessary supported VPD pages:
# sg_vpd -H /dev/sg4
Supported VPD pages VPD page:
00 20 00 00 12 00 80 83 85 86 87 b0 b1 c0 c1 c2 c3 ...............
10 c4 c8 c9 ca d0 e0 ......
My concern was that older devices might not do this, I think there have
been some arrays that had hidden commands/pages.
The original mpp_rdac (ie the multipath software from
Engenio/LSI/NetApp) was using the 'C?' pages to identify the array, so
they most certainly are present on all arrays.
And to properly (and safely :-) identify you have to query page 0 to
figure out if those pages are supported.
Hence we're safe here.
I could ask NetApp, though, for a final confirmation if required.
Cheers,
Hannes
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
