illumos gate

Jorge has re-tested his patched image, incorporating the additional patch from the attached Gerrit change and confirmed that it no longer panics, and that the passthrough devices are still working.

dtrace shows that the BAR allocation is now failing (with ENOMEM), as expected:

  8  81180                vm_map_find:entry vm_map_find(addr=ffffe000, len=1000)

              vmm`vmm_mmio_alloc+0xe2
              vmm`vm_map_mmio+0x14
              ppt`ppt_map_mmio+0x101

  8  82269             vm_mapping_gap:entry vm_mapping_gap(addr=ffffe000, size=1000)

  8  82270            vm_mapping_gap:return                 0
  8  81181               vm_map_find:return                12

The passed-through device has a successful BAR mapping at a different location (pciconf in a FreeBSD guest):

xhci0@pci0:0:8:0:       class=0x0c0330 card=0x00151912 chip=0x00151912 rev=0x02 hdr=0x00
    vendor     = 'Renesas Technology Corp.'
    device     = 'uPD720202 USB 3.0 Host Controller'
    class      = serial bus
    subclass   = USB
    bar   [10] = type Prefetchable Memory, range 64, base 0xc0004000, size 8192, enabled
    cap 01[50] = powerspec 3  supports D0 D3  current D0
    cap 05[70] = MSI supports 2 messages, 64 bit enabled with 1 message
    cap 11[90] = MSI-X supports 8 messages
                 Table in map 0x10[0x1000], PBA in map 0x10[0x1080]
    cap 10[a0] = PCI-Express 2 endpoint max data 128(128)
                 link x1(x1) speed 5.0(5.0) ASPM disabled(L0s/L1) ClockPM disabled

and dtrace also shows that this allocation was successful:

  6  81180                vm_map_find:entry vm_map_find(addr=c0004000, len=1000)
  6  82269             vm_mapping_gap:entry vm_mapping_gap(addr=c0004000, size=1000)
  6  82270            vm_mapping_gap:return                 1
  6  81181               vm_map_find:return                 0

Booted a fresh PI with CR 1383, 1411, and 1412 applied

root@fermion:~ # pciconf -lvb
hostb0@pci0:0:0:0:      class=0x060000 card=0x00000000 chip=0x12378086 rev=0x00 hdr=0x00
    vendor     = 'Intel Corporation'
    device     = '440FX - 82441FX PMC [Natoma]'
    class      = bridge
    subclass   = HOST-PCI
virtio_pci0@pci0:0:5:0: class=0x010000 card=0x00021af4 chip=0x10011af4 rev=0x00 hdr=0x00
    vendor     = 'Red Hat, Inc.'
    device     = 'Virtio block device'
    class      = mass storage
    subclass   = SCSI
    bar   [10] = type I/O Port, range 32, base 0x2000, size 128, enabled
    bar   [14] = type Memory, range 32, base 0xc0000000, size 8192, enabled
virtio_pci1@pci0:0:6:0: class=0x020000 card=0x00011af4 chip=0x10001af4 rev=0x00 hdr=0x00
    vendor     = 'Red Hat, Inc.'
    device     = 'Virtio network device'
    class      = network
    subclass   = ethernet
    bar   [10] = type I/O Port, range 32, base 0x2080, size 32, enabled
    bar   [14] = type Memory, range 32, base 0xc0002000, size 8192, enabled
xhci0@pci0:0:8:0:       class=0x0c0330 card=0x00151912 chip=0x00151912 rev=0x02 hdr=0x00
    vendor     = 'Renesas Technology Corp.'
    device     = 'uPD720202 USB 3.0 Host Controller'
    class      = serial bus
    subclass   = USB
    bar   [10] = type Prefetchable Memory, range 64, base 0xc0004000, size 8192, enabled
ixl0@pci0:0:8:1:        class=0x020000 card=0x37d215d9 chip=0x37d28086 rev=0x09 hdr=0x00
    vendor     = 'Intel Corporation'
    device     = 'Ethernet Connection X722 for 10GBASE-T'
    class      = network
    subclass   = ethernet
    bar   [10] = type Prefetchable Memory, range 64, base 0xc1000000, size 16777216, enabled
    bar   [1c] = type Prefetchable Memory, range 64, base 0xc2000000, size 32768, enabled
isab0@pci0:0:31:0:      class=0x060100 card=0x00000000 chip=0x70008086 rev=0x00 hdr=0x00
    vendor     = 'Intel Corporation'
    device     = '82371SB PIIX3 ISA [Natoma/Triton II]'
    class      = bridge
    subclass   = PCI-ISA

One of the 4 seperate USB controller is passed to a Win 10 VM (NVMe based) and is working fine, a i40e NIC + 2nd USB controller is passed to a freebsd VM

Test USB dongle showed up and was functional in both, as was the i40e nic.
Did a quick check if NVMe, vioblk, and viona were also working as expected with was the case.

This change has also been slightly less extensively tested applied directly to gate (that is, without the other commits that were in the branch tested by Jorge). Booted various VMs running OmniOS, SmartOS, Windows, FreeBSD including some with configured pass-through devices. All tests were successful.