Eu tenho um servidor Xen 4.1 com um Debian Wheezy 7.4 Dom0 (hmpyxen01). Ele tem uma placa PCI SATA que estou passando para um Ubuntu Precise 12.04 DomU (hmvmtmp01), instalado via xen-tools que neste caso é apenas um wrapper para debootstrap então o DomU é muito mínimo (ferramentas como lshw e lspci precisam ser instaladas manualmente após inicializar o DomU).
O passthrough em si parece funcionar como o controlador de armazenamento usa o módulo pciback no Dom0:
root@hmpyxen01:~# lspci -k
...snip...
00:1f.2 SATA controller: Intel Corporation 82801JI (ICH10 Family) SATA AHCI Controller
Subsystem: ASUSTeK Computer Inc. P5Q Deluxe Motherboard
Kernel driver in use: pciback
...snip...
e aparece no DomU:
root@hmvmtmp01:~# lspci -k
00:00.2 SATA controller: Intel Corporation 82801JI (ICH10 Family) SATA AHCI Controller
Subsystem: ASUSTeK Computer Inc. P5Q Deluxe Motherboard
Kernel driver in use: ahci
root@hmvmtmp01:~# lshw -class storage
*-storage
description: SATA controller
product: 82801JI (ICH10 Family) SATA AHCI Controller
vendor: Intel Corporation
physical id: 0.2
bus info: pci@0000:00:00.2
version: 00
width: 32 bits
clock: 66MHz
capabilities: storage msi pm ahci_1.0 bus_master cap_list
configuration: driver=ahci latency=0
resources: irq:29 ioport:9c00(size=8) ioport:9880(size=4) ioport:9800(size=8) ioport:9480(size=4) ioport:9400(size=32) memory:f9cfc000-f9cfc7ff
Infelizmente, nenhum dos discos anexados parece ser carregado no DomU, apenas as unidades Xen raiz / swap configuradas pelo hipervisor.
root@hmvmtmp01:~# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
xvda2 202:2 0 10G 0 disk /
xvda1 202:1 0 1G 0 disk [SWAP]
root@hmvmtmp01:~# lshw -class disk
root@hmvmtmp01:~#
Parece que o dmesg encontra o controlador & discos, mas algo falha:
root@hmvmtmp01:~# dmesg
...snip...
[ 4.192202] pcifront pci-0: Installing PCI frontend
[ 4.192263] pcifront pci-0: Creating PCI Frontend Bus 0000:00
[ 4.192636] pci 0000:00:00.2: [8086:3a22] type 0 class 0x000106
[ 4.192810] pci 0000:00:00.2: reg 10: [io 0x9c00-0x9c07]
[ 4.192886] pci 0000:00:00.2: reg 14: [io 0x9880-0x9883]
[ 4.192959] pci 0000:00:00.2: reg 18: [io 0x9800-0x9807]
[ 4.193033] pci 0000:00:00.2: reg 1c: [io 0x9480-0x9483]
[ 4.193132] pci 0000:00:00.2: reg 20: [io 0x9400-0x941f]
[ 4.193198] pci 0000:00:00.2: reg 24: [mem 0xf9cfc000-0xf9cfc7ff]
[ 4.193718] pcifront pci-0: New device on 0000:00:00.02 found.
[ 4.195283] pcifront pci-0: claiming resource 0000:00:00.2/0
[ 4.195286] pcifront pci-0: claiming resource 0000:00:00.2/1
[ 4.195287] pcifront pci-0: claiming resource 0000:00:00.2/2
[ 4.195289] pcifront pci-0: claiming resource 0000:00:00.2/3
[ 4.195290] pcifront pci-0: claiming resource 0000:00:00.2/4
[ 4.195292] pcifront pci-0: claiming resource 0000:00:00.2/5
[ 4.195325] ahci 0000:00:00.2: version 3.0
[ 4.195352] ahci 0000:00:00.2: enabling device (0000 -> 0003)
[ 4.195421] ahci 0000:00:00.2: Xen PCI mapped GSI20 to IRQ28
[ 4.195650] ahci: SSS flag set, parallel bus scan disabled
[ 4.195702] ahci 0000:00:00.2: AHCI 0001.0200 32 slots 6 ports 3 Gbps 0x3f impl SATA mode
[ 4.195705] ahci 0000:00:00.2: flags: 64bit ncq sntf stag pm led clo pmp pio slum part ccc ems sxs
[ 4.195752] ahci 0000:00:00.2: setting latency timer to 64
[ 4.235098] scsi0 : ahci
[ 4.236153] scsi1 : ahci
[ 4.238341] scsi2 : ahci
[ 4.241175] scsi3 : ahci
[ 4.242971] scsi4 : ahci
[ 4.245287] scsi5 : ahci
[ 4.245331] ata1: SATA max UDMA/133 abar m2048@0xf9cfc000 port 0xf9cfc100 irq 29
[ 4.245333] ata2: SATA max UDMA/133 abar m2048@0xf9cfc000 port 0xf9cfc180 irq 29
[ 4.245336] ata3: SATA max UDMA/133 abar m2048@0xf9cfc000 port 0xf9cfc200 irq 29
[ 4.245338] ata4: SATA max UDMA/133 abar m2048@0xf9cfc000 port 0xf9cfc280 irq 29
[ 4.245340] ata5: SATA max UDMA/133 abar m2048@0xf9cfc000 port 0xf9cfc300 irq 29
[ 4.245342] ata6: SATA max UDMA/133 abar m2048@0xf9cfc000 port 0xf9cfc380 irq 29
[ 6.427298] init: failsafe main process (429) killed by TERM signal
[ 6.460121] init: rc-sysinit main process (550) killed by TERM signal
[ 14.244071] ata1: softreset failed (1st FIS failed)
[ 14.480058] eth0: no IPv6 routers present
[ 24.244072] ata1: softreset failed (1st FIS failed)
[ 59.244072] ata1: softreset failed (1st FIS failed)
[ 59.244087] ata1: limiting SATA link speed to 1.5 Gbps
[ 64.244070] ata1: softreset failed (1st FIS failed)
[ 64.244082] ata1: reset failed, giving up
[ 74.244070] ata2: softreset failed (1st FIS failed)
[ 84.244071] ata2: softreset failed (1st FIS failed)
[ 119.244072] ata2: softreset failed (1st FIS failed)
[ 119.244084] ata2: limiting SATA link speed to 1.5 Gbps
[ 124.244071] ata2: softreset failed (1st FIS failed)
[ 124.244083] ata2: reset failed, giving up
[ 124.372116] init: udevtrigger post-stop process (335) terminated with status 1
[ 124.390542] init: udev-fallback-graphics main process (674) terminated with status 1
[ 124.407645] init: plymouth-splash main process (682) terminated with status 1
[ 134.244071] ata3: softreset failed (1st FIS failed)
[ 144.244071] ata3: softreset failed (1st FIS failed)
[ 179.244070] ata3: softreset failed (1st FIS failed)
[ 179.244083] ata3: limiting SATA link speed to 1.5 Gbps
[ 184.244071] ata3: softreset failed (1st FIS failed)
[ 184.244083] ata3: reset failed, giving up
[ 194.244071] ata4: softreset failed (1st FIS failed)
[ 204.244071] ata4: softreset failed (1st FIS failed)
[ 239.244071] ata4: softreset failed (1st FIS failed)
[ 239.244083] ata4: limiting SATA link speed to 1.5 Gbps
[ 244.244070] ata4: softreset failed (1st FIS failed)
[ 244.244082] ata4: reset failed, giving up
[ 244.564074] ata5: SATA link down (SStatus 0 SControl 300)
[ 244.884073] ata6: SATA link down (SStatus 0 SControl 300)
Como eu mencionei antes, tenho a sensação de que é algo a ver com o método de instalação do xen-tools e o DomU pode estar perdendo algo simples, mas não sei o quê. ..
Qualquer ajuda é muito apreciada.
EDIT: Informação adicional RE hotswap
Então, reiniciei a máquina DomU sem as unidades conectadas e as inseri em um sistema já em execução. Aqui está o formulário de saída que:
[ 305.424655] ata3: exception Emask 0x10 SAct 0x0 SErr 0x4040000 action 0xe frozen
[ 305.424671] ata3: irq_stat 0x00000040, connection status changed
[ 305.424681] ata3: SError: { CommWake DevExch }
[ 305.424688] ata3: hard resetting link
[ 308.353316] ata4: exception Emask 0x10 SAct 0x0 SErr 0x4040000 action 0xe frozen
[ 308.353328] ata4: irq_stat 0x00000040, connection status changed
[ 308.353337] ata4: SError: { CommWake DevExch }
[ 308.353346] ata4: hard resetting link
[ 310.651627] ata1: exception Emask 0x10 SAct 0x0 SErr 0x4040000 action 0xe frozen
[ 310.651639] ata1: irq_stat 0x00000040, connection status changed
[ 310.651648] ata1: SError: { CommWake DevExch }
[ 310.651661] ata1: hard resetting link
[ 313.120770] ata2: exception Emask 0x10 SAct 0x0 SErr 0x4040000 action 0xe frozen
[ 313.120783] ata2: irq_stat 0x00000040, connection status changed
[ 313.120791] ata2: SError: { CommWake DevExch }
[ 313.120804] ata2: hard resetting link
[ 315.428073] ata3: softreset failed (1st FIS failed)
[ 315.428083] ata3: hard resetting link
[ 318.356070] ata4: softreset failed (1st FIS failed)
[ 318.356080] ata4: hard resetting link
As unidades definitivamente funcionam como sem xen-passthrough habilitado, as unidades são todas acessíveis & legível a partir do Dom0.
Uau, 3 dias puxando meu cabelo ...
Não sei se é porque o chipset da minha placa-mãe não suporta VT-d (apenas VT-x), mas na DomU eu tenho que adicionar um iommu=soft simples aos parâmetros de inicialização do kernel em /boot/grub/menu.lst e resolveu tudo !!
São sempre as coisas simples no final.