Re: Odd behavior of a "SAS-2" backplane with SGPIO commands

Douglas Gilbert <dgilbert@xxxxxxxxxxxx> · Thu, 03 Oct 2013 20:07:14 -0400

On 13-10-03 03:11 PM, Pasi Kärkkäinen wrote:
On Mon, Sep 02, 2013 at 05:44:35PM +0300, Pasi Kärkkäinen wrote:
On Sun, Sep 01, 2013 at 08:13:02PM +0300, Pasi Kärkkäinen wrote:
On Fri, Aug 30, 2013 at 05:38:04PM -0400, Rich wrote:
    Apparently, only about 4 months.
    P17 firmware is out.
    Tested on a card which was demonstrating the incorrect behavior before, of
    model 9201-16i.
    It does appear to contain the fixes for this problem.

This is great news, finally!
Thanks for testing and reporting!

Hmm, I checked the P17 firmware changelog PDFs for LSI 9211-8i, 9207-8i and 9201-16i,
and I can't see anything about this issue being fixed..

But if you say it's now fixed, then I guess it means LSI doesn't list every change in the changelog..

I just had some time today to test the new P17 IT firmware with LSI 9211-8i SAS HBAs.

It seems also the "LOCATE" function of LSI sas2ircu tool now actually works and makes
the red/failure leds blink on the Supermicro passive (direct-attach) backplane! Finally..

Btw do you guys have any scripts/tools to control the LEDs on these passive backplanes?
smp_read_gpio/smp_write_gpio isn't the most userfriendly interface ;) Should I try ledmon/ledctl?

There isn't much information about the SMP READ_GPIO and
WRITE_GPIO (and extended variants) in the older SAS
standards and now they have been pushed out of the recent
SAS (SPL-3) drafts. SFF-8485 ain't much help in that
respect either. If anyone has information (not covered by
an NDA) about how these are used, please send it to me.

As for blinking the LED on a SAS disk: in the sdparm package,
in the 'scripts' directory, try the sas_disk_blink script.

Doug Gilbert

I can probably hack something myself too.. but there are many other tasks on the todo-list before this :)

Thanks,

-- Pasi

    - Rich

    On Mon, Apr 22, 2013 at 8:39 PM, Pasi Kärkkäinen <[1]pasik@xxxxxx> wrote:

      On Mon, Apr 22, 2013 at 08:28:58PM -0400, Rich wrote:
      >    I'm told that the fix won't actually be in until P17, because it
      involved
      >    touching a large number of codepaths, but that it will be in P17.
      >

      That's unfortunate.. so another 5-6 months.

      Thanks for the info!

      -- Pasi

      >    - Rich
      >
      >    On Mon, Apr 22, 2013 at 8:19 PM, Pasi KÃ*rkkÃ*inen
      <[1][2]pasik@xxxxxx>
      >    wrote:
      >
      >      On Wed, Dec 19, 2012 at 09:41:42PM +0200, Pasi KÃ*rkkÃ*inen
      wrote:
      >      > On Wed, Dec 19, 2012 at 02:35:02PM -0500, Rich wrote:
      >      > > Â  Â Nope, I'm wrong.
      >      >
      >      > :(
      >      >
      >      > > Â  Â I flashed P15 on a machine, and the same behavior
      persisted,
      >      right up to
      >      > > Â  Â and including the "groups of 3 devices light up when I
      do my
      >      own
      >      > > Â  Â smp_write_gpio".
      >      > > Â  Â Whereas if I flash the experimental blob I was handed by
      >      support to try
      >      > > Â  Â and confirm they had resolved the issue, it does the
      correct
      >      thing.
      >      > > Â  Â I wonder why this change didn't make it into P15.
      >      >
      >      > Hmm.. if you have open contact channel with LSI support please
      ask
      >      them..
      >      > it'd be nice to get the fix for wider audience..
      >      >
      >
      >      It seems LSI P16 firmware has been released.. in the FW changelog
      I can
      >      see at least this:
      >
      >      "Disable SGPIO Group ID support in Channel NVDATA XML's. This
      allows the
      >      use of manufacturing page 7 or default slot mapping setting for
      direct
      >      connected drive slots."
      >
      >      -- Pasi
      >
      >      >
      >      > > Â  Â - Rich
      >      > >
      >      > > Â  Â On Wed, Dec 19, 2012 at 2:00 PM, Rich
      >      <[1][2][3]rercola@xxxxxxxxxxx> wrote:
      >      > >
      >      > > Â  Â  Â From the LSI P15 F/W release notes:
      >      > > Â  Â  Â SCGCQ00342805 (DFCT) Ã* - SlotStatus updates to SES
      managed
      >      Enclosure may
      >      > > Â  Â  Â update incorrect slots
      >      > > Â  Â  Â "Modified FW to use SES diag page 0Ah mapping only if
      SMP
      >      Discover
      >      > > Â  Â  Â DeviceSlotNum is used for Encl PhyÃ* Slot
      enumeration."
      >      > > Â  Â  Â SCGCQ00345867 (DFCT) Ã* - Channel NVDATA internal Phy
      >      reverse setting for
      >      > > Â  Â  Â 9207-4i4e 9207-8i 9217-4i4eÃ* 9217-8i and 9201-16i
      >      > > Â  Â  Â "ISSUE DESC:Ã* Internal PHY orders are reversed for
      the
      >      channel boards
      >      > > Â  Â  Â above."
      >      > > Â  Â  Â So I would submit this is likely fixed by this FW
      rev.
      >      > > Â  Â  Â - Rich
      >      > >
      >      > > Â  Â  Â On Fri, Dec 7, 2012 at 8:46 AM, Pasi
      KÃ*â*¬rkkÃ*â*¬inen
      >      <[2][3][4]pasik@xxxxxx>
      >      > > Â  Â  Â wrote:
      >      > >
      >      > > Â  Â  Â  Â On Thu, Nov 01, 2012 at 11:55:25AM -0400, Rich
      wrote:
      >      > > Â  Â  Â  Â > On Sun, Oct 21, 2012 at 8:46 AM, Pasi
      >      KÃ*â*¬rkkÃ*â*¬inen <[3][4][5]pasik@xxxxxx>
      >      > > Â  Â  Â  Â wrote:
      >      > > Â  Â  Â  Â > > On Mon, Sep 10, 2012 at 07:13:15PM +0300, Pasi
      >      KÃ*â*¬rkkÃ*â*¬inen wrote:
      >      > > Â  Â  Â  Â > >> On Mon, Sep 10, 2012 at 12:07:45PM -0400,
      Rich
      >      wrote:
      >      > > Â  Â  Â  Â > >> > On Mon, Sep 10, 2012 at 12:04 PM, Pasi
      >      KÃ*â*¬rkkÃ*â*¬inen
      >      > > Â  Â  Â  Â <[4][5][6]pasik@xxxxxx> wrote:
      >      > > Â  Â  Â  Â > >> > > On Mon, Sep 10, 2012 at 06:01:42PM +0200,
      >      Emmanuel Florac
      >      > > Â  Â  Â  Â wrote:
      >      > > Â  Â  Â  Â > >> > >> Le Mon, 10 Sep 2012 16:47:11 +0300
      >      > > Â  Â  Â  Â > >> > >> Pasi KÃ*â*¬rkkÃ*â*¬inen
      <[5][6][7]pasik@xxxxxx>
      >      Ã*Â©crivait:
      >      > > Â  Â  Â  Â > >> > >>
      >      > > Â  Â  Â  Â > >> > >> > Any replies from Supermicro/LSI ?
      >      > > Â  Â  Â  Â > >> > >> >
      >      > > Â  Â  Â  Â > >> > >>
      >      > > Â  Â  Â  Â > >> > >> Only loosely related, but Supermicro
      replaced
      >      recently my
      >      > > Â  Â  Â  Â 846E26 (dual
      >      > > Â  Â  Â  Â > >> > >> expander backplane) with 846E16 (single
      >      expander).
      >      > > Â  Â  Â  Â Apparently they
      >      > > Â  Â  Â  Â > >> > >> gave up getting the E26 to work properly
      or
      >      something: LSI
      >      > > Â  Â  Â  Â expander
      >      > > Â  Â  Â  Â > >> > >> firmware problem.
      >      > > Â  Â  Â  Â > >> > >>
      >      > > Â  Â  Â  Â > >> > >> In another (very large scale, high end)
      setup,
      >      many
      >      > > Â  Â  Â  Â different 60 slots
      >      > > Â  Â  Â  Â > >> > >> 5 LSI expanders chassis had a general
      failure
      >      of the 5th
      >      > > Â  Â  Â  Â drawer.
      >      > > Â  Â  Â  Â > >> > >> Another LSI SAS-2 expander firmware
      problem.
      >      > > Â  Â  Â  Â > >> > >>
      >      > > Â  Â  Â  Â > >> > >> I could start a rant about the evil of
      >      proprietary firmware,
      >      > > Â  Â  Â  Â etc. You
      >      > > Â  Â  Â  Â > >> > >> get my meaning :)
      >      > > Â  Â  Â  Â > >> > >>
      >      > > Â  Â  Â  Â > >> > >
      >      > > Â  Â  Â  Â > >> > > Yeah, this is a good example why we're
      trying to
      >      get the LEDs
      >      > > Â  Â  Â  Â working with
      >      > > Â  Â  Â  Â > >> > > direct attach (non-expander) backplanes
      :)
      >      > > Â  Â  Â  Â > >> > >
      >      > > Â  Â  Â  Â > >> > > -- Pasi
      >      > > Â  Â  Â  Â > >> > >
      >      > > Â  Â  Â  Â > >> >
      >      > > Â  Â  Â  Â > >> > I don't have anything useful for people,
      other
      >      than that they
      >      > > Â  Â  Â  Â have
      >      > > Â  Â  Â  Â > >> > shown me an HBA firmware that fixes the LED
      >      problem but has
      >      > > Â  Â  Â  Â other
      >      > > Â  Â  Â  Â > >> > problems they're still debugging.
      >      > > Â  Â  Â  Â > >> >
      >      > > Â  Â  Â  Â > >> > So there does exist code for this firmware
      which
      >      will fix this
      >      > > Â  Â  Â  Â problem.
      >      > > Â  Â  Â  Â > >> >
      >      > > Â  Â  Â  Â > >>
      >      > > Â  Â  Â  Â > >> That's good to know! Let's hope the fix ends
      up in
      >      an official
      >      > > Â  Â  Â  Â > >> LSI HBA firmware update in not-so-distant
      future.
      >      > > Â  Â  Â  Â > >>
      >      > > Â  Â  Â  Â > >
      >      > > Â  Â  Â  Â > > Btw what was the version number of the
      LED-fixed LSI
      >      firmware?
      >      > > Â  Â  Â  Â > > I'm just wondering in which firmware series
      the fix
      >      might end up
      >      > > Â  Â  Â  Â being included..
      >      > > Â  Â  Â  Â > >
      >      > > Â  Â  Â  Â > > Also: Any updates on this?
      >      > > Â  Â  Â  Â > >
      >      > > Â  Â  Â  Â > > Thanks,
      >      > > Â  Â  Â  Â > >
      >      > > Â  Â  Â  Â > > -- Pasi
      >      > > Â  Â  Â  Â > >
      >      > > Â  Â  Â  Â >
      >      > > Â  Â  Â  Â > The fixed FW I had demonstrated for me was still
      >      14.0.0.0 - I've
      >      > > Â  Â  Â  Â been
      >      > > Â  Â  Â  Â > told it'll be rolled into the FW release
      currently on
      >      the site and
      >      > > Â  Â  Â  Â > posted at some point, but that they have no ETA
      for
      >      that.
      >      > > Â  Â  Â  Â >
      >      > >
      >      > > Â  Â  Â  Â I can now see P15 firmware available on LSI's
      website for
      >      9211-8i
      >      > > Â  Â  Â  Â (SAS2008 based HBA).
      >      > > Â  Â  Â  Â Sadly it's missing the releasenotes/changelogs for
      both
      >      the LSI bios
      >      > > Â  Â  Â  Â and firmware..
      >      > > Â  Â  Â  Â -- Pasi
      >      > >
      >      > > References
      >      > >
      >      > > Â  Â Visible links
      >      > > Â  Â 1. mailto:[7][8]rercola@xxxxxxxxxxx
      >      > > Â  Â 2. mailto:[8][9]pasik@xxxxxx
      >      > > Â  Â 3. mailto:[9][10]pasik@xxxxxx
      >      > > Â  Â 4. mailto:[10][11]pasik@xxxxxx
      >      > > Â  Â 5. mailto:[11][12]pasik@xxxxxx
      >      > --
      >      > To unsubscribe from this list: send the line "unsubscribe
      linux-scsi"
      >      in
      >      > the body of a message to [12][13]majordomo@xxxxxxxxxxxxxxx
      >      > More majordomo info at
      >      Â [13][14]http://vger.kernel.org/majordomo-info.html
      >
      > References
      >
      >    Visible links
      >    1. mailto:[15]pasik@xxxxxx
      >    2. mailto:[16]rercola@xxxxxxxxxxx
      >    3. mailto:[17]pasik@xxxxxx
      >    4. mailto:[18]pasik@xxxxxx
      >    5. mailto:[19]pasik@xxxxxx
      >    6. mailto:[20]pasik@xxxxxx
      >    7. mailto:[21]rercola@xxxxxxxxxxx
      >    8. mailto:[22]pasik@xxxxxx
      >    9. mailto:[23]pasik@xxxxxx
      >   10. mailto:[24]pasik@xxxxxx
      >   11. mailto:[25]pasik@xxxxxx
      >   12. mailto:[26]majordomo@xxxxxxxxxxxxxxx
      >   13. [27]http://vger.kernel.org/majordomo-info.html

References

    Visible links
    1. mailto:pasik@xxxxxx
    2. mailto:pasik@xxxxxx
    3. mailto:rercola@xxxxxxxxxxx
    4. mailto:pasik@xxxxxx
    5. mailto:pasik@xxxxxx
    6. mailto:pasik@xxxxxx
    7. mailto:pasik@xxxxxx
    8. mailto:rercola@xxxxxxxxxxx
    9. mailto:pasik@xxxxxx
   10. mailto:pasik@xxxxxx
   11. mailto:pasik@xxxxxx
   12. mailto:pasik@xxxxxx
   13. mailto:majordomo@xxxxxxxxxxxxxxx
   14. http://vger.kernel.org/majordomo-info.html
   15. mailto:pasik@xxxxxx
   16. mailto:rercola@xxxxxxxxxxx
   17. mailto:pasik@xxxxxx
   18. mailto:pasik@xxxxxx
   19. mailto:pasik@xxxxxx
   20. mailto:pasik@xxxxxx
   21. mailto:rercola@xxxxxxxxxxx
   22. mailto:pasik@xxxxxx
   23. mailto:pasik@xxxxxx
   24. mailto:pasik@xxxxxx
   25. mailto:pasik@xxxxxx
   26. mailto:majordomo@xxxxxxxxxxxxxxx
   27. http://vger.kernel.org/majordomo-info.html
--
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
--
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
--
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

--
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