LSI megaRAID controller: Kernel panic on virtual disk creation

[Desai, Kashyap]

Steve:

Thanks for your suggestion. We found the root cause of this issue. In our environment we had few drives which fails to respond READ CAP 16 even though it is SPC-3 complaint. 

Recently upstream kernel added code for read cap 16 predication based on SPC-3 flag from Inquiry in side "daregister(struct cam_periph *periph, void *arg)"

        /* Predict whether device may support READ CAPACITY(16). */
        if (SID_ANSI_REV(&cgd->inq_data) >= SCSI_REV_SPC3) {
                softc->flags |= DA_FLAG_CAN_RC16;
                softc->state = DA_STATE_PROBE_RC16;
        }

Above code will force OS to only send READ CAP 16 for drives which are SPC3 compliant. They never try REAC CAP 10.

Check below link. There is already discussion at upstream to debate this issue. 
http://lists.freebsd.org/pipermail/freebsd-fs/2013-January/016278.html

There may be some drives which are SPC-3 complaint but do not follow complete spec and will have failure for some SCSI commands.
So considering SCSI_REV_SPC3 or greater does not means all Drives will always respond to READ CAP16. 
We should try to fall back to READ CAP 10 if Read cap 16 fails.

Thanks, Kashyap

> -----Original Message-----
> From: owner-freebsd-scsi at freebsd.org [mailto:owner-freebsd-
> scsi at freebsd.org] On Behalf Of Steven Hartland
> Sent: Friday, May 17, 2013 7:46 PM
> To: Saxena, Sumit; freebsd-scsi at freebsd.org
> Subject: Re: LSI megaRAID controller: Kernel panic on virtual disk
> creation
> 
> That indeed doesn't look like look like an issue unrelated to driver.
> 
> Could you provide information on which version your running and what the
> exact command you where running to create the panic?
> 
> While I'm here IIRC 9286 is a 2208 chipset card, so MegaRAI based, which
> I did some significant fixes for our MFI driver in r247367 &
> r247369 so might be worth reviewing to see if any of thats needed in
> mrsas?
> 
>     Regards
>     Steve
> ----- Original Message -----
> > Hi ,
> >
> > While doing some testing on "mrsas" driver(LSIs' latest driver for
> > next generation MegaRAID controllers- which will soon be submitted to
> > upstream kernel), I am facing kernel panic  while creating Virtual
> drive(RAID0) on LSI controller-9286-e (6 Gb/s RAID controller). I am
> using latest upstream kernel.  The error message seen on panic is:
> >
> > "reboot after panic: Provider da1 lacks sectorsize"
> >
> > Below is the call stack:
> >
> > ----------------------------------------------
> > kdb_enter() at kdb_enter+0x3e/frame 0xffffff8000285920
> > vpanic() at vpanic+0x146/frame 0xffffff8000285960
> > kassert_panic() at kassert_panic+0x136/frame 0xffffff80002859d0
> > g_access() at g_access+0x3ba/frame 0xffffff8000285a60
> > g_raid_md_taste_jmicron() at g_raid_md_taste_jmicron+0x6d/frame
> > 0xffffff8000285b00
> > g_raid_taste() at g_raid_taste+0x17f/frame 0xffffff8000285b50
> > g_new_provider_event() at g_new_provider_event+0xda/frame
> > 0xffffff8000285b70
> > g_run_events() at g_run_events+0x1a7/frame 0xffffff8000285bb0
> > fork_exit() at fork_exit+0x84/frame 0xffffff8000285bf0
> > fork_trampoline() at fork_trampoline+0xe/frame 0xffffff8000285bf0
> > --- trap 0, rip = 0, rsp = 0xffffff8000285cb0, rbp = 0 ---
> > -------------------------------------------------
> >
> > "/dev/da1" is virtual drive created.
> >
> > The problem does not seems to be related to <mrsas> driver. call stack
> indicates that panic is observed  in GEOM module.
> > Has someone face such issue earlier related to GEOM module?
> 
> 
> ================================================
> This e.mail is private and confidential between Multiplay (UK) Ltd. and
> the person or entity to whom it is addressed. In the event of
> misdirection, the recipient is prohibited from using, copying, printing
> or otherwise disseminating it or any information contained in it.
> 
> In the event of misdirection, illegible or incomplete transmission
> please telephone +44 845 868 1337 or return the E.mail to
> postmaster at multiplay.co.uk.
> 
> _______________________________________________
> freebsd-scsi at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-scsi
> To unsubscribe, send any mail to "freebsd-scsi-unsubscribe at freebsd.org"