xl pci-detach vs xm pci-detach in Xen 4.3 (one works, the other does not)

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

Something is odd. If I do the pci-detach in Xend 
(xm pci-detach latest 0000:01:00.0) the guest tells me:

[   20.866562] pcifront pci-0: Rescanning PCI Frontend Bus 0000:00
[   20.996602] pcifront pci-0: backend going away!
[   20.996824] pci_bus 0000:00: busn_res: [bus 00-ff] is released
[   20.997089] pcifront pci-0: Disconnecting PCI Frontend Buses
[   21.006171] pcifront pci-0: 22 freeing event channel 17

and I want to attach it back (xm pci-attach latest 0000:01:00.0):

# [   78.823076] pcifront pci-0: Installing PCI frontend
[   78.823332] pcifront pci-0: Creating PCI Frontend Bus 0000:00
[   78.823674] pcifront pci-0: PCI host bridge to bus 0000:00
[   78.823686] pci_bus 0000:00: root bus resource [io  0x0000-0xffff]
[   78.823697] pci_bus 0000:00: root bus resource [mem 0x00000000-0xfffffffff]
[   78.823703] pci_bus 0000:00: root bus resource [bus 00-ff]
[   78.823911] pci 0000:00:00.0: [8086:105e] type 00 class 0x020000
[   78.824234] pci 0000:00:00.0: reg 10: [mem 0xfe4a0000-0xfe4bffff]
[   78.824378] pci 0000:00:00.0: reg 14: [mem 0xfe480000-0xfe49ffff]
[   78.824489] pci 0000:00:00.0: reg 18: [io  0xe020-0xe03f]
[   78.827837] pcifront pci-0: claiming resource 0000:00:00.0/0
[   78.827845] pcifront pci-0: claiming resource 0000:00:00.0/1
[   78.827850] pcifront pci-0: claiming resource 0000:00:00.0/2
[   78.831779] e1000e: Intel(R) PRO/1000 Network Driver - 2.3.2-k
[   78.831790] e1000e: Copyright(c) 1999 - 2013 Intel Corporation.
[   78.831842] e1000e 0000:00:00.0: Disabling ASPM  L1
[   78.831937] e1000e 0000:00:00.0: enabling device (0000 -> 0002)
[   78.834698] xen_map_pirq_gsi: returning irq 34 for gsi 16
[   78.834707] e1000e 0000:00:00.0: Xen PCI mapped GSI16 to IRQ34
[   78.835330] e1000e 0000:00:00.0: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode
[   79.002430] e1000e 0000:00:00.0 eth0: (PCI Express:2.5GT/s:Width x4) 00:15:17:8f:18:a2
[   79.002441] e1000e 0000:00:00.0 eth0: Intel(R) PRO/1000 Network Connection
[   79.002521] e1000e 0000:00:00.0 eth0: MAC: 0, PHY: 4, PBA No: D50868-003
[   79.310912] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready
[   79.313442] device eth0 entered promiscuous mode
[   82.230982] e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
[   82.231133] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[   82.231225] switch: port 1(eth0) entered forwarding state
[   82.231242] switch: port 1(eth0) entered forwarding state

Doing this with xl I get:

15:42:23 # 57 :~/
> xl -f pci-detach latest 0000:01:00.0
libxl: error: libxl_pci.c:1231:do_pci_remove: xc_physdev_unmap_pirq irq=82
libxl: error: libxl_pci.c:1235:do_pci_remove: xc_domain_irq_permission irq=82


and in the guest:
# kill -1 1
# [   28.516427] switch: port 1(eth0) entered disabled state
[   28.516590] device eth0 left promiscuous mode
[   28.516599] switch: port 1(eth0) entered disabled state
[   28.547664] pcifront pci-0: Rescanning PCI Frontend Bus 0000:00
[   28.654770] pci_bus 0000:00: busn_res: [bus 00-ff] is released
[   28.654982] pcifront pci-0: 22 freeing event channel 17
[   28.655228] pcifront pci-0: failed to write error node for device/pci/0 (22 freeing event channel 17)


but this with Xen 4.3 that has at its tip:

commit 365c95f7de789e1dca03f119eab7dc61fe0f77c9
Author: Jan Beulich <jbeulich@suse.com>
Date:   Tue Jun 4 09:29:07 2013 +0200

    x86/xsave: properly check guest input to XSETBV

any thoughts? I don't know if this is a regression or not, but the
libxl errors are pointing me to the recent XSA issue - which I thought
was fixed?

Guest config:
extra="console=hvc0 debug "
kernel="/mnt/lab/latest/vmlinuz"
ramdisk="/mnt/lab/latest/initramfs.cpio.gz"
memory=1024
vcpus=2
name="latest"
on_crash="preserve"
vnc=1
vnclisten="0.0.0.0"
pci=["01:00.0"]
e820_host=1
disk=['phy:/dev/sdb2,xvda,w']