Re: XEN 4.12, redirect tape drive to domu

[Éliás Tamás <et@xxxxxxx>]

Hi!

Thank you for the responses. At the moment when I'm using PCI 
passthorugh I receive the follwoing error:

[   17.592149] scsi 4:0:6:0: Sequential-Access IBM      DDS Gen5 
A1B0 PQ: 0 ANSI: 3
[   17.607179] scsi target4:0:6: Beginning Domain Validation
[   17.612790] scsi target4:0:6: wide asynchronous
[   17.616625] scsi target4:0:6: FAST-20 WIDE SCSI 40.0 MB/s ST (50 ns, 
offset 32)
[   17.619668] scsi target4:0:6: Domain Validation skipping write tests
[   17.619670] scsi target4:0:6: Ending Domain Validation
[   19.677562] scsi 4:0:6:0: Attached scsi generic sg9 type 1
[   19.683814] st: Version 20160209, fixed bufsize 32768, s/g segs 256
[   19.685623] st 4:0:6:0: Attached scsi tape st0
[   19.685626] st 4:0:6:0: st0: try direct i/o: yes (alignment 4 B)
[19057.309731] st 4:0:6:0: Power-on or device reset occurred
[19121.837625] st 4:0:6:0: [st0] Block limits 1 - 16777215 bytes.
[20183.205282] INFO: task mt:12738 blocked for more than 120 seconds.
[20183.205408]       Not tainted 5.4.0-48-generic #52~18.04.1-Ubuntu
[20183.205516] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" 
disables this message.
[20183.205644] mt              D    0 12738  12736 0x00000000
[20183.205649] Call Trace:
[20183.205668]  __schedule+0x293/0x720
[20183.205672]  schedule+0x33/0xa0
[20183.205675]  schedule_timeout+0x1d3/0x320
[20183.205684]  ? __blk_mq_delay_run_hw_queue+0x107/0x160
[20183.205687]  ? blk_mq_run_hw_queue+0x56/0x100
[20183.205690]  wait_for_completion+0xba/0x140
[20183.205696]  ? wake_up_q+0x80/0x80
[20183.205704]  st_do_scsi.constprop.29+0x37f/0x420 [st]
[20183.205707]  st_int_ioctl+0x28f/0xf80 [st]
[20183.205711]  ? st_int_ioctl+0x28f/0xf80 [st]
[20183.205714]  st_ioctl+0x6b4/0x1770 [st]
[20183.205721]  ? _copy_to_user+0x26/0x30
[20183.205728]  ? cp_new_stat+0x152/0x180
[20183.205732]  do_vfs_ioctl+0xa9/0x640
[20183.205735]  ? __do_sys_newfstat+0x44/0x70
[20183.205738]  ksys_ioctl+0x75/0x80
[20183.205740]  __x64_sys_ioctl+0x1a/0x20
[20183.205747]  do_syscall_64+0x57/0x190
[20183.205749]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[20183.205753] RIP: 0033:0x7fc44fdca6d7
[20183.205777] Code: Bad RIP value.
[20183.205779] RSP: 002b:00007ffc346c4f78 EFLAGS: 00000246 ORIG_RAX: 
0000000000000010
[20183.205781] RAX: ffffffffffffffda RBX: 00007ffc346c5802 RCX: 
00007fc44fdca6d7
[20183.205783] RDX: 00007ffc346c4f80 RSI: 0000000040086d01 RDI: 
0000000000000003
[20183.205784] RBP: 00007ffc346c5802 R08: 0000000000000000 R09: 
0000000000000000
[20183.205785] R10: 0000000000000000 R11: 0000000000000246 R12: 
00005578d705fcf0
[20183.205787] R13: 00007ffc346c50a0 R14: 0000000000000000 R15: 
0000000000000000
[20185.674286] st 4:0:6:0: [st0] Sense Key : Medium Error [current]
[20185.674291] st 4:0:6:0: [st0] Add. Sense: Excessive write errors

If I reboot the system, tape drive works from dom0, but I get the above 
message from domu. I guess there should be some kind of fine-tuning 
available for pci passthrough?
I admint the pci device is an old one, it does not even supports 
function level reset:
29:01.0 SCSI storage controller: Adaptec AIC-7892A U160/m (rev 02)
        Subsystem: Adaptec 29160 Ultra160 SCSI Controller
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- 
ParErr+ Stepping- SERR+ FastB2B- DisINTx-
        Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=medium 
>TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 64 (10000ns min, 6250ns max), Cache Line Size: 64 bytes
        Interrupt: pin A routed to IRQ 16
        BIST result: 00
        Region 0: I/O ports at 1000 [disabled] [size=256]
        Region 1: Memory at 97900000 (64-bit, non-prefetchable) [size=4K]
        Expansion ROM at 97920000 [disabled] [size=128K]
        Capabilities: [dc] Power Management version 2
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA 
PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
        Kernel driver in use: aic7xxx
        Kernel modules: aic7xxx


These io-specific functions are supported by my server:


(XEN) I/O virtualisation disabled
(XEN) ENABLING IO-APIC IRQs
(XEN)  -> Using new ACK method
(XEN) Allocated console ring of 16 KiB.
(XEN) VMX: Supported advanced features:
(XEN)  - APIC MMIO access virtualisation
(XEN)  - APIC TPR shadow
(XEN)  - Extended Page Tables (EPT)
(XEN)  - Virtual-Processor Identifiers (VPID)
(XEN)  - Virtual NMI
(XEN)  - MSR direct-access bitmap
(XEN) HVM: ASIDs enabled.
(XEN) VMX: Disabling executable EPT superpages due to CVE-2018-12207
(XEN) HVM: VMX enabled
(XEN) HVM: Hardware Assisted Paging (HAP) detected
(XEN) HVM: HAP page sizes: 4kB, 2MB

[    0.000000] DMI: IBM IBM System x3400 M3 Server -[7379KTG]-/69Y4356, 
BIOS -[Y4E159AUS-1.14]- 05/06/2013



2020. 09. 28. 10:36 keltezéssel, Niels Dettenbach írta:
> Am Freitag, 25. September 2020, 15:20:53 CEST schrieb Éliás Tamás:
> > My question is, should it be possible to passthrough this tape drive, or
> > the complete scsi card sitting on the pci (due to it's only aim is to
> > provide access to the tape), somehow? Are there any specialities?
> hmm,
> if just one DomU should gain access to that PCI device, PCI passthrough may
> work, but if you want to use it from different DomU things may get
> complicated, because only one DomU at a time should get access to it. If xen
> allows to attach tape scsci devices (never tried it) this may be a way to
> instead.
>
> If you want to use it i.e., for a daily/nightly backup of all DomU, PCI
> hotplug and PCI attach (or storage device attach/detach iif it works with
> tape) may be way to "connect" -> backup -> "disconnect" to the volumes of
> each DomU.
>
> I wrote a xenbackup tool (shell scripts) for Dom0 side backup of LVM based
> DomU storages (with snapshots) - by default for backup-manager, but it may be
> a base to adapt it to your "batch" application with PCI passthrough and some
> ssh command to do the backup "from within" the DomUs if you need that.
>
> https://www.syndicat.com/open_source/projekte/xenbackup/
> https://github.com/nielsd/xen-host-allbackup
>
> you would have to adapt the snaps_backup_mount / snaps_backup_umount scripts
>
> DomUs/storages could be configured in:
> /etc/xenback_vols
>
>
>
> hth,
>
>
> niels.
--

Éliás Tamás
Thomas Elias

ETIT[nwpro] KFT, Ügyvezető-Hálózatbiztonsági specialista
ETIT[nwpro] Ltd, General Manager-Network security specialist

Tel. HU: +36/30-497-1626
OpenPGP pubkey: http://etit.hu/doc/et-pub.asc

Okleveles mérnök-informatikus (MSC)
Master of Science in Information Technology (MSC)
Licenced Penetration Tester (TM15-047)

Kapcsolat: http://etit.hu/index.php/hu/kapcsolat
Jogi nyilatkozat: http://etit.hu/disclaimer-email-hu.txt
Contact: http://etit.hu/index.php/en/contact
Disclaimer: http://etit.hu/disclaimer-email-en.txt