* [Qemu-devel] The maximum limit of virtual network device
@ 2017-07-06 6:20 Wu, Jiaxin
2017-07-06 8:11 ` Daniel P. Berrange
2017-07-06 8:31 ` Laszlo Ersek
0 siblings, 2 replies; 6+ messages in thread
From: Wu, Jiaxin @ 2017-07-06 6:20 UTC (permalink / raw)
To: qemu-devel@nongnu.org; +Cc: Wu, Jiaxin
Hello experts,
We know QEMU has the capability to create the multiple network devices in one QEMU guest with the -device syntax. But I met the below failure when I'm trying to create more than 30 virtual devices with the each TAP backend:
qemu-system-x86_64: -device e1000: PCI: no slot/function available for e1000, all in use.
The corresponding QEMU command shows as following:
sudo qemu-system-x86_64 -pflash OVMF.fd -global e1000.romfile="" -netdev tap,id=hostnet0,ifname=tap0,script=no,downscript=no -device e1000,netdev=hostnet0 -netdev tap,id=hostnet1,ifname=tap1,script=no,downscript=no -device e1000,netdev=hostnet1 -netdev tap,id=hostnet2,ifname=tap2,script=no,downscript=no -device e1000,netdev=hostnet2 -netdev tap,id=hostnet3,ifname=tap3,script=no,downscript=no -device e1000,netdev=hostnet3 -netdev tap,id=hostnet4,ifname=tap4,script=no,downscript=no -device e1000,netdev=hostnet4 -netdev tap,id=hostnet5,ifname=tap5,script=no,downscript=no -device e1000,netdev=hostnet5 -netdev tap,id=hostnet6,ifname=tap6,script=no,downscript=no -device e1000,netdev=hostnet6 -netdev tap,id=hostnet7,ifname=tap7,script=no,downscript=no -device e1000,netdev=hostnet7 -netdev tap,id=hostnet8,ifname=tap8,script=no,downscript=no -device e1000,netdev=hostnet8 -netdev tap,id=hostnet9,ifname=tap9,script=no,downscript=no -device e1000,netdev=hostnet9 -netdev tap,id=hostnet10,ifname=tap10,script=no,downscript=no -device e1000,netdev=hostnet10 -netdev tap,id=hostnet11,ifname=tap11,script=no,downscript=no -device e1000,netdev=hostnet11 -netdev tap,id=hostnet12,ifname=tap12,script=no,downscript=no -device e1000,netdev=hostnet12 -netdev tap,id=hostnet13,ifname=tap13,script=no,downscript=no -device e1000,netdev=hostnet13 -netdev tap,id=hostnet14,ifname=tap14,script=no,downscript=no -device e1000,netdev=hostnet14 -netdev tap,id=hostnet15,ifname=tap15,script=no,downscript=no -device e1000,netdev=hostnet15 -netdev tap,id=hostnet16,ifname=tap16,script=no,downscript=no -device e1000,netdev=hostnet16 -netdev tap,id=hostnet17,ifname=tap17,script=no,downscript=no -device e1000,netdev=hostnet17 -netdev tap,id=hostnet18,ifname=tap18,script=no,downscript=no -device e1000,netdev=hostnet18 -netdev tap,id=hostnet19,ifname=tap19,script=no,downscript=no -device e1000,netdev=hostnet19 -netdev tap,id=hostnet20,ifname=tap20,script=no,downscript=no -device e1000,netdev=hostnet20 -netdev tap,id=hostnet21,ifname=tap21,script=no,downscript=no -device e1000,netdev=hostnet21 -netdev tap,id=hostnet22,ifname=tap22,script=no,downscript=no -device e1000,netdev=hostnet22 -netdev tap,id=hostnet23,ifname=tap23,script=no,downscript=no -device e1000,netdev=hostnet23 -netdev tap,id=hostnet24,ifname=tap24,script=no,downscript=no -device e1000,netdev=hostnet24 -netdev tap,id=hostnet25,ifname=tap25,script=no,downscript=no -device e1000,netdev=hostnet25 -netdev tap,id=hostnet26,ifname=tap26,script=no,downscript=no -device e1000,netdev=hostnet26 -netdev tap,id=hostnet27,ifname=tap27,script=no,downscript=no -device e1000,netdev=hostnet27 -netdev tap,id=hostnet28,ifname=tap28,script=no,downscript=no -device e1000,netdev=hostnet28 -netdev tap,id=hostnet29,ifname=tap29,script=no,downscript=no -device e1000,netdev=hostnet29
>From above, the max limit of virtual network device in one guest is about 29? If not, how can I avoid such failure? My use case is to create more than 150 network devices in one guest. Please provide your comments on this.
Thank you!
Best Regards.
Jiaxin
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] The maximum limit of virtual network device
2017-07-06 6:20 [Qemu-devel] The maximum limit of virtual network device Wu, Jiaxin
@ 2017-07-06 8:11 ` Daniel P. Berrange
2017-07-06 8:31 ` Laszlo Ersek
1 sibling, 0 replies; 6+ messages in thread
From: Daniel P. Berrange @ 2017-07-06 8:11 UTC (permalink / raw)
To: Wu, Jiaxin; +Cc: qemu-devel@nongnu.org
On Thu, Jul 06, 2017 at 06:20:54AM +0000, Wu, Jiaxin wrote:
> Hello experts,
>
> We know QEMU has the capability to create the multiple network devices in one QEMU guest with the -device syntax. But I met the below failure when I'm trying to create more than 30 virtual devices with the each TAP backend:
>
> qemu-system-x86_64: -device e1000: PCI: no slot/function available for e1000, all in use.
>
> The corresponding QEMU command shows as following:
[snip]
> From above, the max limit of virtual network device in one guest is about
> 29? If not, how can I avoid such failure? My use case is to create more than
> 150 network devices in one guest. Please provide your comments on this.
As the error message above shows, you have run out of PCI slots on the default
PCI bus. There are two ways to get around this. The first is to make use of
multifunction=on and specify individual PCI functions. There are 7 functions
per slot, so that would take you to 203 devices. The downside of multifunction
is that you can't hotplug individual functions. The second approach is to
just add a bunch of PCI bridges, so that you have more slots available.
Regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] The maximum limit of virtual network device
2017-07-06 6:20 [Qemu-devel] The maximum limit of virtual network device Wu, Jiaxin
2017-07-06 8:11 ` Daniel P. Berrange
@ 2017-07-06 8:31 ` Laszlo Ersek
2017-07-06 9:24 ` Marcel Apfelbaum
1 sibling, 1 reply; 6+ messages in thread
From: Laszlo Ersek @ 2017-07-06 8:31 UTC (permalink / raw)
To: Wu, Jiaxin, qemu-devel@nongnu.org; +Cc: Marcel Apfelbaum
Hi Jiaxin,
it's nice to see a question from you on qemu-devel! :)
On 07/06/17 08:20, Wu, Jiaxin wrote:
> Hello experts,
>
> We know QEMU has the capability to create the multiple network devices
> in one QEMU guest with the -device syntax. But I met the below failure
> when I'm trying to create more than 30 virtual devices with the each
> TAP backend:
>
> qemu-system-x86_64: -device e1000: PCI: no slot/function available for
> e1000, all in use.
>
> The corresponding QEMU command shows as following:
>
> sudo qemu-system-x86_64 \
> -pflash OVMF.fd \
> -global e1000.romfile="" \
> -netdev tap,id=hostnet0,ifname=tap0,script=no,downscript=no \
> -device e1000,netdev=hostnet0 \
> -netdev tap,id=hostnet1,ifname=tap1,script=no,downscript=no \
> -device e1000,netdev=hostnet1 \
> -netdev tap,id=hostnet2,ifname=tap2,script=no,downscript=no \
> -device e1000,netdev=hostnet2 \
> -netdev tap,id=hostnet3,ifname=tap3,script=no,downscript=no \
> -device e1000,netdev=hostnet3 \
> -netdev tap,id=hostnet4,ifname=tap4,script=no,downscript=no \
> -device e1000,netdev=hostnet4 \
> -netdev tap,id=hostnet5,ifname=tap5,script=no,downscript=no \
> -device e1000,netdev=hostnet5 \
> -netdev tap,id=hostnet6,ifname=tap6,script=no,downscript=no \
> -device e1000,netdev=hostnet6 \
> -netdev tap,id=hostnet7,ifname=tap7,script=no,downscript=no \
> -device e1000,netdev=hostnet7 \
> -netdev tap,id=hostnet8,ifname=tap8,script=no,downscript=no \
> -device e1000,netdev=hostnet8 \
> -netdev tap,id=hostnet9,ifname=tap9,script=no,downscript=no \
> -device e1000,netdev=hostnet9 \
> -netdev tap,id=hostnet10,ifname=tap10,script=no,downscript=no \
> -device e1000,netdev=hostnet10 \
> -netdev tap,id=hostnet11,ifname=tap11,script=no,downscript=no \
> -device e1000,netdev=hostnet11 \
> -netdev tap,id=hostnet12,ifname=tap12,script=no,downscript=no \
> -device e1000,netdev=hostnet12 \
> -netdev tap,id=hostnet13,ifname=tap13,script=no,downscript=no \
> -device e1000,netdev=hostnet13 \
> -netdev tap,id=hostnet14,ifname=tap14,script=no,downscript=no \
> -device e1000,netdev=hostnet14 \
> -netdev tap,id=hostnet15,ifname=tap15,script=no,downscript=no \
> -device e1000,netdev=hostnet15 \
> -netdev tap,id=hostnet16,ifname=tap16,script=no,downscript=no \
> -device e1000,netdev=hostnet16 \
> -netdev tap,id=hostnet17,ifname=tap17,script=no,downscript=no \
> -device e1000,netdev=hostnet17 \
> -netdev tap,id=hostnet18,ifname=tap18,script=no,downscript=no \
> -device e1000,netdev=hostnet18 \
> -netdev tap,id=hostnet19,ifname=tap19,script=no,downscript=no \
> -device e1000,netdev=hostnet19 \
> -netdev tap,id=hostnet20,ifname=tap20,script=no,downscript=no \
> -device e1000,netdev=hostnet20 \
> -netdev tap,id=hostnet21,ifname=tap21,script=no,downscript=no \
> -device e1000,netdev=hostnet21 \
> -netdev tap,id=hostnet22,ifname=tap22,script=no,downscript=no \
> -device e1000,netdev=hostnet22 \
> -netdev tap,id=hostnet23,ifname=tap23,script=no,downscript=no \
> -device e1000,netdev=hostnet23 \
> -netdev tap,id=hostnet24,ifname=tap24,script=no,downscript=no \
> -device e1000,netdev=hostnet24 \
> -netdev tap,id=hostnet25,ifname=tap25,script=no,downscript=no \
> -device e1000,netdev=hostnet25 \
> -netdev tap,id=hostnet26,ifname=tap26,script=no,downscript=no \
> -device e1000,netdev=hostnet26 \
> -netdev tap,id=hostnet27,ifname=tap27,script=no,downscript=no \
> -device e1000,netdev=hostnet27 \
> -netdev tap,id=hostnet28,ifname=tap28,script=no,downscript=no \
> -device e1000,netdev=hostnet28 \
> -netdev tap,id=hostnet29,ifname=tap29,script=no,downscript=no \
> -device e1000,netdev=hostnet29
>
> From above, the max limit of virtual network device in one guest is
> about 29? If not, how can I avoid such failure? My use case is to
> create more than 150 network devices in one guest. Please provide your
> comments on this.
You are seeing the above symptom because the above command line
instructs QEMU to do the following:
- use the i440fx machine type,
- use a single PCI bus (= the main root bridge),
- add the e1000 cards to separate slots (always using function 0) on
that bus.
Accordingly, there are three things you can do to remedy this:
- Use the Q35 machine type and work with a PCI Express hierarchy rather
than a PCI hierarchy. I'm mentioning this only for completeness,
because it won't directly help your use case. But, I certainly want to
highlight "docs/pcie.txt". Please read it sometime; it has nice
examples and makes good points.
- Use multiple PCI bridges to attach the devices. For this, several ways
are possible:
- use multiple root buses, with the pxb or pxb-pcie devices (see
"docs/pci_expander_bridge.txt" and "docs/pcie.txt")
- use multiple normal PCI bridges
- use multiple PCI Express root ports or downstream ports (but for
this, you'll likely have to use the PCI Express variant of the
e1000, namely e1000e)
- If you don't need hot-plug / hot-unplug, aggregate eights of e1000
NICs into multifunction PCI devices.
Now, I would normally recommend sticking with i440fx for simplicity.
However, each PCI bridge requires 4KB of IO space (meaning (1 + 5) * 4KB
= 24KB), and OVMF on the i440fx does not support that much (only
0x4000). So, I'll recommend Q35 for IO space purposes; OVMF on Q35
provides 0xA000 (40KB).
For scaling higher than this, a PCI Express hierarchy should be used
with PCI Express devices that require no IO space at all. However, that
setup is even more problematic *for now*; please see section "3. IO
space issues" in "docs/pcie.txt". We have open OVMF and QEMU BZs for
limiting IO space allocation to cases when it is really necessary:
https://bugzilla.redhat.com/show_bug.cgi?id=1344299
https://bugzilla.redhat.com/show_bug.cgi?id=1434740
Therefore I guess the simplest example I can give now is:
- use Q35 (for a larger IO space),
- plug a DMI-PCI bridge into the root bridge,
- plug 5 PCI bridges into the DMI-PCI bridge,
- plug 31 NICs per PCI bridge, each NIC into a separate slot.
This follows the following recommendation of "2.3 PCI only hierarchy" in
"docs/pcie.txt" (slightly rewrapped here):
> 2.3 PCI only hierarchy
> ======================
> Legacy PCI devices can be plugged into pcie.0 as Integrated Endpoints,
> but, as mentioned in section 5, doing so means the legacy PCI
> device in question will be incapable of hot-unplugging.
> Besides that use DMI-PCI Bridges (i82801b11-bridge) in combination
> with PCI-PCI Bridges (pci-bridge) to start PCI hierarchies.
>
> Prefer flat hierarchies. For most scenarios a single DMI-PCI Bridge
> (having 32 slots) and several PCI-PCI Bridges attached to it (each
> supporting also 32 slots) will support hundreds of legacy devices. The
> recommendation is to populate one PCI-PCI Bridge under the DMI-PCI
> Bridge until is full and then plug a new PCI-PCI Bridge...
Here's a command line. Please note that the OVMF boot may take quite
long with this, as the E3522X2.EFI driver from BootUtil (-D
E1000_ENABLE) binds all 150 e1000 NICs in succession! Watching the OVMF
debug log is recommended.
qemu-system-x86_64 \
\
-machine q35,vmport=off,accel=kvm \
-pflash OVMF.fd \
-global e1000.romfile="" \
-m 2048 \
-debugcon file:debug.log \
-global isa-debugcon.iobase=0x402 \
\
-netdev tap,id=hostnet0,ifname=tap0,script=no,downscript=no \
-netdev tap,id=hostnet1,ifname=tap1,script=no,downscript=no \
-netdev tap,id=hostnet2,ifname=tap2,script=no,downscript=no \
-netdev tap,id=hostnet3,ifname=tap3,script=no,downscript=no \
-netdev tap,id=hostnet4,ifname=tap4,script=no,downscript=no \
-netdev tap,id=hostnet5,ifname=tap5,script=no,downscript=no \
-netdev tap,id=hostnet6,ifname=tap6,script=no,downscript=no \
-netdev tap,id=hostnet7,ifname=tap7,script=no,downscript=no \
-netdev tap,id=hostnet8,ifname=tap8,script=no,downscript=no \
-netdev tap,id=hostnet9,ifname=tap9,script=no,downscript=no \
-netdev tap,id=hostnet10,ifname=tap10,script=no,downscript=no \
-netdev tap,id=hostnet11,ifname=tap11,script=no,downscript=no \
-netdev tap,id=hostnet12,ifname=tap12,script=no,downscript=no \
-netdev tap,id=hostnet13,ifname=tap13,script=no,downscript=no \
-netdev tap,id=hostnet14,ifname=tap14,script=no,downscript=no \
-netdev tap,id=hostnet15,ifname=tap15,script=no,downscript=no \
-netdev tap,id=hostnet16,ifname=tap16,script=no,downscript=no \
-netdev tap,id=hostnet17,ifname=tap17,script=no,downscript=no \
-netdev tap,id=hostnet18,ifname=tap18,script=no,downscript=no \
-netdev tap,id=hostnet19,ifname=tap19,script=no,downscript=no \
-netdev tap,id=hostnet20,ifname=tap20,script=no,downscript=no \
-netdev tap,id=hostnet21,ifname=tap21,script=no,downscript=no \
-netdev tap,id=hostnet22,ifname=tap22,script=no,downscript=no \
-netdev tap,id=hostnet23,ifname=tap23,script=no,downscript=no \
-netdev tap,id=hostnet24,ifname=tap24,script=no,downscript=no \
-netdev tap,id=hostnet25,ifname=tap25,script=no,downscript=no \
-netdev tap,id=hostnet26,ifname=tap26,script=no,downscript=no \
-netdev tap,id=hostnet27,ifname=tap27,script=no,downscript=no \
-netdev tap,id=hostnet28,ifname=tap28,script=no,downscript=no \
-netdev tap,id=hostnet29,ifname=tap29,script=no,downscript=no \
-netdev tap,id=hostnet30,ifname=tap30,script=no,downscript=no \
-netdev tap,id=hostnet31,ifname=tap31,script=no,downscript=no \
-netdev tap,id=hostnet32,ifname=tap32,script=no,downscript=no \
-netdev tap,id=hostnet33,ifname=tap33,script=no,downscript=no \
-netdev tap,id=hostnet34,ifname=tap34,script=no,downscript=no \
-netdev tap,id=hostnet35,ifname=tap35,script=no,downscript=no \
-netdev tap,id=hostnet36,ifname=tap36,script=no,downscript=no \
-netdev tap,id=hostnet37,ifname=tap37,script=no,downscript=no \
-netdev tap,id=hostnet38,ifname=tap38,script=no,downscript=no \
-netdev tap,id=hostnet39,ifname=tap39,script=no,downscript=no \
-netdev tap,id=hostnet40,ifname=tap40,script=no,downscript=no \
-netdev tap,id=hostnet41,ifname=tap41,script=no,downscript=no \
-netdev tap,id=hostnet42,ifname=tap42,script=no,downscript=no \
-netdev tap,id=hostnet43,ifname=tap43,script=no,downscript=no \
-netdev tap,id=hostnet44,ifname=tap44,script=no,downscript=no \
-netdev tap,id=hostnet45,ifname=tap45,script=no,downscript=no \
-netdev tap,id=hostnet46,ifname=tap46,script=no,downscript=no \
-netdev tap,id=hostnet47,ifname=tap47,script=no,downscript=no \
-netdev tap,id=hostnet48,ifname=tap48,script=no,downscript=no \
-netdev tap,id=hostnet49,ifname=tap49,script=no,downscript=no \
-netdev tap,id=hostnet50,ifname=tap50,script=no,downscript=no \
-netdev tap,id=hostnet51,ifname=tap51,script=no,downscript=no \
-netdev tap,id=hostnet52,ifname=tap52,script=no,downscript=no \
-netdev tap,id=hostnet53,ifname=tap53,script=no,downscript=no \
-netdev tap,id=hostnet54,ifname=tap54,script=no,downscript=no \
-netdev tap,id=hostnet55,ifname=tap55,script=no,downscript=no \
-netdev tap,id=hostnet56,ifname=tap56,script=no,downscript=no \
-netdev tap,id=hostnet57,ifname=tap57,script=no,downscript=no \
-netdev tap,id=hostnet58,ifname=tap58,script=no,downscript=no \
-netdev tap,id=hostnet59,ifname=tap59,script=no,downscript=no \
-netdev tap,id=hostnet60,ifname=tap60,script=no,downscript=no \
-netdev tap,id=hostnet61,ifname=tap61,script=no,downscript=no \
-netdev tap,id=hostnet62,ifname=tap62,script=no,downscript=no \
-netdev tap,id=hostnet63,ifname=tap63,script=no,downscript=no \
-netdev tap,id=hostnet64,ifname=tap64,script=no,downscript=no \
-netdev tap,id=hostnet65,ifname=tap65,script=no,downscript=no \
-netdev tap,id=hostnet66,ifname=tap66,script=no,downscript=no \
-netdev tap,id=hostnet67,ifname=tap67,script=no,downscript=no \
-netdev tap,id=hostnet68,ifname=tap68,script=no,downscript=no \
-netdev tap,id=hostnet69,ifname=tap69,script=no,downscript=no \
-netdev tap,id=hostnet70,ifname=tap70,script=no,downscript=no \
-netdev tap,id=hostnet71,ifname=tap71,script=no,downscript=no \
-netdev tap,id=hostnet72,ifname=tap72,script=no,downscript=no \
-netdev tap,id=hostnet73,ifname=tap73,script=no,downscript=no \
-netdev tap,id=hostnet74,ifname=tap74,script=no,downscript=no \
-netdev tap,id=hostnet75,ifname=tap75,script=no,downscript=no \
-netdev tap,id=hostnet76,ifname=tap76,script=no,downscript=no \
-netdev tap,id=hostnet77,ifname=tap77,script=no,downscript=no \
-netdev tap,id=hostnet78,ifname=tap78,script=no,downscript=no \
-netdev tap,id=hostnet79,ifname=tap79,script=no,downscript=no \
-netdev tap,id=hostnet80,ifname=tap80,script=no,downscript=no \
-netdev tap,id=hostnet81,ifname=tap81,script=no,downscript=no \
-netdev tap,id=hostnet82,ifname=tap82,script=no,downscript=no \
-netdev tap,id=hostnet83,ifname=tap83,script=no,downscript=no \
-netdev tap,id=hostnet84,ifname=tap84,script=no,downscript=no \
-netdev tap,id=hostnet85,ifname=tap85,script=no,downscript=no \
-netdev tap,id=hostnet86,ifname=tap86,script=no,downscript=no \
-netdev tap,id=hostnet87,ifname=tap87,script=no,downscript=no \
-netdev tap,id=hostnet88,ifname=tap88,script=no,downscript=no \
-netdev tap,id=hostnet89,ifname=tap89,script=no,downscript=no \
-netdev tap,id=hostnet90,ifname=tap90,script=no,downscript=no \
-netdev tap,id=hostnet91,ifname=tap91,script=no,downscript=no \
-netdev tap,id=hostnet92,ifname=tap92,script=no,downscript=no \
-netdev tap,id=hostnet93,ifname=tap93,script=no,downscript=no \
-netdev tap,id=hostnet94,ifname=tap94,script=no,downscript=no \
-netdev tap,id=hostnet95,ifname=tap95,script=no,downscript=no \
-netdev tap,id=hostnet96,ifname=tap96,script=no,downscript=no \
-netdev tap,id=hostnet97,ifname=tap97,script=no,downscript=no \
-netdev tap,id=hostnet98,ifname=tap98,script=no,downscript=no \
-netdev tap,id=hostnet99,ifname=tap99,script=no,downscript=no \
-netdev tap,id=hostnet100,ifname=tap100,script=no,downscript=no \
-netdev tap,id=hostnet101,ifname=tap101,script=no,downscript=no \
-netdev tap,id=hostnet102,ifname=tap102,script=no,downscript=no \
-netdev tap,id=hostnet103,ifname=tap103,script=no,downscript=no \
-netdev tap,id=hostnet104,ifname=tap104,script=no,downscript=no \
-netdev tap,id=hostnet105,ifname=tap105,script=no,downscript=no \
-netdev tap,id=hostnet106,ifname=tap106,script=no,downscript=no \
-netdev tap,id=hostnet107,ifname=tap107,script=no,downscript=no \
-netdev tap,id=hostnet108,ifname=tap108,script=no,downscript=no \
-netdev tap,id=hostnet109,ifname=tap109,script=no,downscript=no \
-netdev tap,id=hostnet110,ifname=tap110,script=no,downscript=no \
-netdev tap,id=hostnet111,ifname=tap111,script=no,downscript=no \
-netdev tap,id=hostnet112,ifname=tap112,script=no,downscript=no \
-netdev tap,id=hostnet113,ifname=tap113,script=no,downscript=no \
-netdev tap,id=hostnet114,ifname=tap114,script=no,downscript=no \
-netdev tap,id=hostnet115,ifname=tap115,script=no,downscript=no \
-netdev tap,id=hostnet116,ifname=tap116,script=no,downscript=no \
-netdev tap,id=hostnet117,ifname=tap117,script=no,downscript=no \
-netdev tap,id=hostnet118,ifname=tap118,script=no,downscript=no \
-netdev tap,id=hostnet119,ifname=tap119,script=no,downscript=no \
-netdev tap,id=hostnet120,ifname=tap120,script=no,downscript=no \
-netdev tap,id=hostnet121,ifname=tap121,script=no,downscript=no \
-netdev tap,id=hostnet122,ifname=tap122,script=no,downscript=no \
-netdev tap,id=hostnet123,ifname=tap123,script=no,downscript=no \
-netdev tap,id=hostnet124,ifname=tap124,script=no,downscript=no \
-netdev tap,id=hostnet125,ifname=tap125,script=no,downscript=no \
-netdev tap,id=hostnet126,ifname=tap126,script=no,downscript=no \
-netdev tap,id=hostnet127,ifname=tap127,script=no,downscript=no \
-netdev tap,id=hostnet128,ifname=tap128,script=no,downscript=no \
-netdev tap,id=hostnet129,ifname=tap129,script=no,downscript=no \
-netdev tap,id=hostnet130,ifname=tap130,script=no,downscript=no \
-netdev tap,id=hostnet131,ifname=tap131,script=no,downscript=no \
-netdev tap,id=hostnet132,ifname=tap132,script=no,downscript=no \
-netdev tap,id=hostnet133,ifname=tap133,script=no,downscript=no \
-netdev tap,id=hostnet134,ifname=tap134,script=no,downscript=no \
-netdev tap,id=hostnet135,ifname=tap135,script=no,downscript=no \
-netdev tap,id=hostnet136,ifname=tap136,script=no,downscript=no \
-netdev tap,id=hostnet137,ifname=tap137,script=no,downscript=no \
-netdev tap,id=hostnet138,ifname=tap138,script=no,downscript=no \
-netdev tap,id=hostnet139,ifname=tap139,script=no,downscript=no \
-netdev tap,id=hostnet140,ifname=tap140,script=no,downscript=no \
-netdev tap,id=hostnet141,ifname=tap141,script=no,downscript=no \
-netdev tap,id=hostnet142,ifname=tap142,script=no,downscript=no \
-netdev tap,id=hostnet143,ifname=tap143,script=no,downscript=no \
-netdev tap,id=hostnet144,ifname=tap144,script=no,downscript=no \
-netdev tap,id=hostnet145,ifname=tap145,script=no,downscript=no \
-netdev tap,id=hostnet146,ifname=tap146,script=no,downscript=no \
-netdev tap,id=hostnet147,ifname=tap147,script=no,downscript=no \
-netdev tap,id=hostnet148,ifname=tap148,script=no,downscript=no \
-netdev tap,id=hostnet149,ifname=tap149,script=no,downscript=no \
\
-device i82801b11-bridge,id=dmi-pci-bridge \
\
-device pci-bridge,id=bridge-1,chassis_nr=1,bus=dmi-pci-bridge \
-device pci-bridge,id=bridge-2,chassis_nr=2,bus=dmi-pci-bridge \
-device pci-bridge,id=bridge-3,chassis_nr=3,bus=dmi-pci-bridge \
-device pci-bridge,id=bridge-4,chassis_nr=4,bus=dmi-pci-bridge \
-device pci-bridge,id=bridge-5,chassis_nr=5,bus=dmi-pci-bridge \
\
-device e1000,netdev=hostnet0,bus=bridge-1,addr=0x1.0 \
-device e1000,netdev=hostnet1,bus=bridge-1,addr=0x2.0 \
-device e1000,netdev=hostnet2,bus=bridge-1,addr=0x3.0 \
-device e1000,netdev=hostnet3,bus=bridge-1,addr=0x4.0 \
-device e1000,netdev=hostnet4,bus=bridge-1,addr=0x5.0 \
-device e1000,netdev=hostnet5,bus=bridge-1,addr=0x6.0 \
-device e1000,netdev=hostnet6,bus=bridge-1,addr=0x7.0 \
-device e1000,netdev=hostnet7,bus=bridge-1,addr=0x8.0 \
-device e1000,netdev=hostnet8,bus=bridge-1,addr=0x9.0 \
-device e1000,netdev=hostnet9,bus=bridge-1,addr=0xa.0 \
-device e1000,netdev=hostnet10,bus=bridge-1,addr=0xb.0 \
-device e1000,netdev=hostnet11,bus=bridge-1,addr=0xc.0 \
-device e1000,netdev=hostnet12,bus=bridge-1,addr=0xd.0 \
-device e1000,netdev=hostnet13,bus=bridge-1,addr=0xe.0 \
-device e1000,netdev=hostnet14,bus=bridge-1,addr=0xf.0 \
-device e1000,netdev=hostnet15,bus=bridge-1,addr=0x10.0 \
-device e1000,netdev=hostnet16,bus=bridge-1,addr=0x11.0 \
-device e1000,netdev=hostnet17,bus=bridge-1,addr=0x12.0 \
-device e1000,netdev=hostnet18,bus=bridge-1,addr=0x13.0 \
-device e1000,netdev=hostnet19,bus=bridge-1,addr=0x14.0 \
-device e1000,netdev=hostnet20,bus=bridge-1,addr=0x15.0 \
-device e1000,netdev=hostnet21,bus=bridge-1,addr=0x16.0 \
-device e1000,netdev=hostnet22,bus=bridge-1,addr=0x17.0 \
-device e1000,netdev=hostnet23,bus=bridge-1,addr=0x18.0 \
-device e1000,netdev=hostnet24,bus=bridge-1,addr=0x19.0 \
-device e1000,netdev=hostnet25,bus=bridge-1,addr=0x1a.0 \
-device e1000,netdev=hostnet26,bus=bridge-1,addr=0x1b.0 \
-device e1000,netdev=hostnet27,bus=bridge-1,addr=0x1c.0 \
-device e1000,netdev=hostnet28,bus=bridge-1,addr=0x1d.0 \
-device e1000,netdev=hostnet29,bus=bridge-1,addr=0x1e.0 \
-device e1000,netdev=hostnet30,bus=bridge-1,addr=0x1f.0 \
\
-device e1000,netdev=hostnet31,bus=bridge-2,addr=0x1.0 \
-device e1000,netdev=hostnet32,bus=bridge-2,addr=0x2.0 \
-device e1000,netdev=hostnet33,bus=bridge-2,addr=0x3.0 \
-device e1000,netdev=hostnet34,bus=bridge-2,addr=0x4.0 \
-device e1000,netdev=hostnet35,bus=bridge-2,addr=0x5.0 \
-device e1000,netdev=hostnet36,bus=bridge-2,addr=0x6.0 \
-device e1000,netdev=hostnet37,bus=bridge-2,addr=0x7.0 \
-device e1000,netdev=hostnet38,bus=bridge-2,addr=0x8.0 \
-device e1000,netdev=hostnet39,bus=bridge-2,addr=0x9.0 \
-device e1000,netdev=hostnet40,bus=bridge-2,addr=0xa.0 \
-device e1000,netdev=hostnet41,bus=bridge-2,addr=0xb.0 \
-device e1000,netdev=hostnet42,bus=bridge-2,addr=0xc.0 \
-device e1000,netdev=hostnet43,bus=bridge-2,addr=0xd.0 \
-device e1000,netdev=hostnet44,bus=bridge-2,addr=0xe.0 \
-device e1000,netdev=hostnet45,bus=bridge-2,addr=0xf.0 \
-device e1000,netdev=hostnet46,bus=bridge-2,addr=0x10.0 \
-device e1000,netdev=hostnet47,bus=bridge-2,addr=0x11.0 \
-device e1000,netdev=hostnet48,bus=bridge-2,addr=0x12.0 \
-device e1000,netdev=hostnet49,bus=bridge-2,addr=0x13.0 \
-device e1000,netdev=hostnet50,bus=bridge-2,addr=0x14.0 \
-device e1000,netdev=hostnet51,bus=bridge-2,addr=0x15.0 \
-device e1000,netdev=hostnet52,bus=bridge-2,addr=0x16.0 \
-device e1000,netdev=hostnet53,bus=bridge-2,addr=0x17.0 \
-device e1000,netdev=hostnet54,bus=bridge-2,addr=0x18.0 \
-device e1000,netdev=hostnet55,bus=bridge-2,addr=0x19.0 \
-device e1000,netdev=hostnet56,bus=bridge-2,addr=0x1a.0 \
-device e1000,netdev=hostnet57,bus=bridge-2,addr=0x1b.0 \
-device e1000,netdev=hostnet58,bus=bridge-2,addr=0x1c.0 \
-device e1000,netdev=hostnet59,bus=bridge-2,addr=0x1d.0 \
-device e1000,netdev=hostnet60,bus=bridge-2,addr=0x1e.0 \
-device e1000,netdev=hostnet61,bus=bridge-2,addr=0x1f.0 \
\
-device e1000,netdev=hostnet62,bus=bridge-3,addr=0x1.0 \
-device e1000,netdev=hostnet63,bus=bridge-3,addr=0x2.0 \
-device e1000,netdev=hostnet64,bus=bridge-3,addr=0x3.0 \
-device e1000,netdev=hostnet65,bus=bridge-3,addr=0x4.0 \
-device e1000,netdev=hostnet66,bus=bridge-3,addr=0x5.0 \
-device e1000,netdev=hostnet67,bus=bridge-3,addr=0x6.0 \
-device e1000,netdev=hostnet68,bus=bridge-3,addr=0x7.0 \
-device e1000,netdev=hostnet69,bus=bridge-3,addr=0x8.0 \
-device e1000,netdev=hostnet70,bus=bridge-3,addr=0x9.0 \
-device e1000,netdev=hostnet71,bus=bridge-3,addr=0xa.0 \
-device e1000,netdev=hostnet72,bus=bridge-3,addr=0xb.0 \
-device e1000,netdev=hostnet73,bus=bridge-3,addr=0xc.0 \
-device e1000,netdev=hostnet74,bus=bridge-3,addr=0xd.0 \
-device e1000,netdev=hostnet75,bus=bridge-3,addr=0xe.0 \
-device e1000,netdev=hostnet76,bus=bridge-3,addr=0xf.0 \
-device e1000,netdev=hostnet77,bus=bridge-3,addr=0x10.0 \
-device e1000,netdev=hostnet78,bus=bridge-3,addr=0x11.0 \
-device e1000,netdev=hostnet79,bus=bridge-3,addr=0x12.0 \
-device e1000,netdev=hostnet80,bus=bridge-3,addr=0x13.0 \
-device e1000,netdev=hostnet81,bus=bridge-3,addr=0x14.0 \
-device e1000,netdev=hostnet82,bus=bridge-3,addr=0x15.0 \
-device e1000,netdev=hostnet83,bus=bridge-3,addr=0x16.0 \
-device e1000,netdev=hostnet84,bus=bridge-3,addr=0x17.0 \
-device e1000,netdev=hostnet85,bus=bridge-3,addr=0x18.0 \
-device e1000,netdev=hostnet86,bus=bridge-3,addr=0x19.0 \
-device e1000,netdev=hostnet87,bus=bridge-3,addr=0x1a.0 \
-device e1000,netdev=hostnet88,bus=bridge-3,addr=0x1b.0 \
-device e1000,netdev=hostnet89,bus=bridge-3,addr=0x1c.0 \
-device e1000,netdev=hostnet90,bus=bridge-3,addr=0x1d.0 \
-device e1000,netdev=hostnet91,bus=bridge-3,addr=0x1e.0 \
-device e1000,netdev=hostnet92,bus=bridge-3,addr=0x1f.0 \
\
-device e1000,netdev=hostnet93,bus=bridge-4,addr=0x1.0 \
-device e1000,netdev=hostnet94,bus=bridge-4,addr=0x2.0 \
-device e1000,netdev=hostnet95,bus=bridge-4,addr=0x3.0 \
-device e1000,netdev=hostnet96,bus=bridge-4,addr=0x4.0 \
-device e1000,netdev=hostnet97,bus=bridge-4,addr=0x5.0 \
-device e1000,netdev=hostnet98,bus=bridge-4,addr=0x6.0 \
-device e1000,netdev=hostnet99,bus=bridge-4,addr=0x7.0 \
-device e1000,netdev=hostnet100,bus=bridge-4,addr=0x8.0 \
-device e1000,netdev=hostnet101,bus=bridge-4,addr=0x9.0 \
-device e1000,netdev=hostnet102,bus=bridge-4,addr=0xa.0 \
-device e1000,netdev=hostnet103,bus=bridge-4,addr=0xb.0 \
-device e1000,netdev=hostnet104,bus=bridge-4,addr=0xc.0 \
-device e1000,netdev=hostnet105,bus=bridge-4,addr=0xd.0 \
-device e1000,netdev=hostnet106,bus=bridge-4,addr=0xe.0 \
-device e1000,netdev=hostnet107,bus=bridge-4,addr=0xf.0 \
-device e1000,netdev=hostnet108,bus=bridge-4,addr=0x10.0 \
-device e1000,netdev=hostnet109,bus=bridge-4,addr=0x11.0 \
-device e1000,netdev=hostnet110,bus=bridge-4,addr=0x12.0 \
-device e1000,netdev=hostnet111,bus=bridge-4,addr=0x13.0 \
-device e1000,netdev=hostnet112,bus=bridge-4,addr=0x14.0 \
-device e1000,netdev=hostnet113,bus=bridge-4,addr=0x15.0 \
-device e1000,netdev=hostnet114,bus=bridge-4,addr=0x16.0 \
-device e1000,netdev=hostnet115,bus=bridge-4,addr=0x17.0 \
-device e1000,netdev=hostnet116,bus=bridge-4,addr=0x18.0 \
-device e1000,netdev=hostnet117,bus=bridge-4,addr=0x19.0 \
-device e1000,netdev=hostnet118,bus=bridge-4,addr=0x1a.0 \
-device e1000,netdev=hostnet119,bus=bridge-4,addr=0x1b.0 \
-device e1000,netdev=hostnet120,bus=bridge-4,addr=0x1c.0 \
-device e1000,netdev=hostnet121,bus=bridge-4,addr=0x1d.0 \
-device e1000,netdev=hostnet122,bus=bridge-4,addr=0x1e.0 \
-device e1000,netdev=hostnet123,bus=bridge-4,addr=0x1f.0 \
\
-device e1000,netdev=hostnet124,bus=bridge-5,addr=0x1.0 \
-device e1000,netdev=hostnet125,bus=bridge-5,addr=0x2.0 \
-device e1000,netdev=hostnet126,bus=bridge-5,addr=0x3.0 \
-device e1000,netdev=hostnet127,bus=bridge-5,addr=0x4.0 \
-device e1000,netdev=hostnet128,bus=bridge-5,addr=0x5.0 \
-device e1000,netdev=hostnet129,bus=bridge-5,addr=0x6.0 \
-device e1000,netdev=hostnet130,bus=bridge-5,addr=0x7.0 \
-device e1000,netdev=hostnet131,bus=bridge-5,addr=0x8.0 \
-device e1000,netdev=hostnet132,bus=bridge-5,addr=0x9.0 \
-device e1000,netdev=hostnet133,bus=bridge-5,addr=0xa.0 \
-device e1000,netdev=hostnet134,bus=bridge-5,addr=0xb.0 \
-device e1000,netdev=hostnet135,bus=bridge-5,addr=0xc.0 \
-device e1000,netdev=hostnet136,bus=bridge-5,addr=0xd.0 \
-device e1000,netdev=hostnet137,bus=bridge-5,addr=0xe.0 \
-device e1000,netdev=hostnet138,bus=bridge-5,addr=0xf.0 \
-device e1000,netdev=hostnet139,bus=bridge-5,addr=0x10.0 \
-device e1000,netdev=hostnet140,bus=bridge-5,addr=0x11.0 \
-device e1000,netdev=hostnet141,bus=bridge-5,addr=0x12.0 \
-device e1000,netdev=hostnet142,bus=bridge-5,addr=0x13.0 \
-device e1000,netdev=hostnet143,bus=bridge-5,addr=0x14.0 \
-device e1000,netdev=hostnet144,bus=bridge-5,addr=0x15.0 \
-device e1000,netdev=hostnet145,bus=bridge-5,addr=0x16.0 \
-device e1000,netdev=hostnet146,bus=bridge-5,addr=0x17.0 \
-device e1000,netdev=hostnet147,bus=bridge-5,addr=0x18.0 \
-device e1000,netdev=hostnet148,bus=bridge-5,addr=0x19.0 \
-device e1000,netdev=hostnet149,bus=bridge-5,addr=0x1a.0
Thanks
Laszlo
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] The maximum limit of virtual network device
2017-07-06 8:31 ` Laszlo Ersek
@ 2017-07-06 9:24 ` Marcel Apfelbaum
2017-07-06 9:46 ` Laszlo Ersek
0 siblings, 1 reply; 6+ messages in thread
From: Marcel Apfelbaum @ 2017-07-06 9:24 UTC (permalink / raw)
To: Laszlo Ersek, Wu, Jiaxin, qemu-devel@nongnu.org,
Alexander Bezzubikov
On 06/07/2017 11:31, Laszlo Ersek wrote:
> Hi Jiaxin,
>
> it's nice to see a question from you on qemu-devel! :)
>
> On 07/06/17 08:20, Wu, Jiaxin wrote:
>> Hello experts,
>>
>> We know QEMU has the capability to create the multiple network devices
>> in one QEMU guest with the -device syntax. But I met the below failure
>> when I'm trying to create more than 30 virtual devices with the each
>> TAP backend:
>>
>> qemu-system-x86_64: -device e1000: PCI: no slot/function available for
>> e1000, all in use.
>>
>> The corresponding QEMU command shows as following:
>>
>> sudo qemu-system-x86_64 \
>> -pflash OVMF.fd \
>> -global e1000.romfile="" \
>> -netdev tap,id=hostnet0,ifname=tap0,script=no,downscript=no \
>> -device e1000,netdev=hostnet0 \
[...]
>> -netdev tap,id=hostnet29,ifname=tap29,script=no,downscript=no \
>> -device e1000,netdev=hostnet29
>>
>> From above, the max limit of virtual network device in one guest is
>> about 29? If not, how can I avoid such failure? My use case is to
>> create more than 150 network devices in one guest. Please provide your
>> comments on this.
>
> You are seeing the above symptom because the above command line
> instructs QEMU to do the following:
> - use the i440fx machine type,
> - use a single PCI bus (= the main root bridge),
> - add the e1000 cards to separate slots (always using function 0) on
> that bus.
>
> Accordingly, there are three things you can do to remedy this:
>
> - Use the Q35 machine type and work with a PCI Express hierarchy rather
> than a PCI hierarchy. I'm mentioning this only for completeness,
> because it won't directly help your use case. But, I certainly want to
> highlight "docs/pcie.txt". Please read it sometime; it has nice
> examples and makes good points.
>
> - Use multiple PCI bridges to attach the devices. For this, several ways
> are possible:
>
> - use multiple root buses, with the pxb or pxb-pcie devices (see
> "docs/pci_expander_bridge.txt" and "docs/pcie.txt")
>
> - use multiple normal PCI bridges
>
> - use multiple PCI Express root ports or downstream ports (but for
> this, you'll likely have to use the PCI Express variant of the
> e1000, namely e1000e)
>
> - If you don't need hot-plug / hot-unplug, aggregate eights of e1000
> NICs into multifunction PCI devices.
> > Now, I would normally recommend sticking with i440fx for simplicity.
> However, each PCI bridge requires 4KB of IO space (meaning (1 + 5) * 4KB
> = 24KB), and OVMF on the i440fx does not support that much (only
> 0x4000). So, I'll recommend Q35 for IO space purposes; OVMF on Q35
> provides 0xA000 (40KB).
So if we use OVMF, going for Q35 gives us actually more IO space, nice!
However recommending Q35 for IO space seems odd :)
>
> For scaling higher than this, a PCI Express hierarchy should be used
> with PCI Express devices that require no IO space at all. However, that
> setup is even more problematic *for now*; please see section "3. IO
> space issues" in "docs/pcie.txt". We have open OVMF and QEMU BZs for
> limiting IO space allocation to cases when it is really necessary:
>
> https://bugzilla.redhat.com/show_bug.cgi?id=1344299
> https://bugzilla.redhat.com/show_bug.cgi?id=1434740
>
> Therefore I guess the simplest example I can give now is:
> - use Q35 (for a larger IO space),
> - plug a DMI-PCI bridge into the root bridge,
> - plug 5 PCI bridges into the DMI-PCI bridge,
> - plug 31 NICs per PCI bridge, each NIC into a separate slot.
>
The setup looks OK to me (assuming OVMF is needed, otherwise
PC + pci-bridges will result in more devices),
I do have a little concern.
We want to deprecate the dmi-pci bridge since it does not support
hot-plug (for itself or devices behind it).
Alexandr (CCed) is a GSOC student working on a generic
pcie-pci bridge that can (eventually) be hot-plugged
into a PCIe Root Port and keeps the machine cleaner.
See:
https://lists.gnu.org/archive/html/qemu-devel/2017-06/msg05498.html
If is a "lab" project it doesn't really matter, but I wanted
to point out the direction.
Thanks,
Marcel
> This follows the following recommendation of "2.3 PCI only hierarchy" in
> "docs/pcie.txt" (slightly rewrapped here):
>
>> 2.3 PCI only hierarchy
>> ======================
>> Legacy PCI devices can be plugged into pcie.0 as Integrated Endpoints,
>> but, as mentioned in section 5, doing so means the legacy PCI
>> device in question will be incapable of hot-unplugging.
>> Besides that use DMI-PCI Bridges (i82801b11-bridge) in combination
>> with PCI-PCI Bridges (pci-bridge) to start PCI hierarchies.
>>
>> Prefer flat hierarchies. For most scenarios a single DMI-PCI Bridge
>> (having 32 slots) and several PCI-PCI Bridges attached to it (each
>> supporting also 32 slots) will support hundreds of legacy devices. The
>> recommendation is to populate one PCI-PCI Bridge under the DMI-PCI
>> Bridge until is full and then plug a new PCI-PCI Bridge...
>
> Here's a command line. Please note that the OVMF boot may take quite
> long with this, as the E3522X2.EFI driver from BootUtil (-D
> E1000_ENABLE) binds all 150 e1000 NICs in succession! Watching the OVMF
> debug log is recommended.
>
> qemu-system-x86_64 \
> \
> -machine q35,vmport=off,accel=kvm \
> -pflash OVMF.fd \
> -global e1000.romfile="" \
> -m 2048 \
> -debugcon file:debug.log \
> -global isa-debugcon.iobase=0x402 \
> \
> -netdev tap,id=hostnet0,ifname=tap0,script=no,downscript=no \
[...]
> -netdev tap,id=hostnet149,ifname=tap149,script=no,downscript=no \
> \
> -device i82801b11-bridge,id=dmi-pci-bridge \
> \
> -device pci-bridge,id=bridge-1,chassis_nr=1,bus=dmi-pci-bridge \
> -device pci-bridge,id=bridge-2,chassis_nr=2,bus=dmi-pci-bridge \
> -device pci-bridge,id=bridge-3,chassis_nr=3,bus=dmi-pci-bridge \
> -device pci-bridge,id=bridge-4,chassis_nr=4,bus=dmi-pci-bridge \
> -device pci-bridge,id=bridge-5,chassis_nr=5,bus=dmi-pci-bridge \
> \
> -device e1000,netdev=hostnet0,bus=bridge-1,addr=0x1.0 \
[...]
> -device e1000,netdev=hostnet149,bus=bridge-5,addr=0x1a.0
>
> Thanks
> Laszlo
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] The maximum limit of virtual network device
2017-07-06 9:24 ` Marcel Apfelbaum
@ 2017-07-06 9:46 ` Laszlo Ersek
2017-07-06 14:49 ` Wu, Jiaxin
0 siblings, 1 reply; 6+ messages in thread
From: Laszlo Ersek @ 2017-07-06 9:46 UTC (permalink / raw)
To: Marcel Apfelbaum, Wu, Jiaxin, qemu-devel@nongnu.org,
Alexander Bezzubikov
On 07/06/17 11:24, Marcel Apfelbaum wrote:
> On 06/07/2017 11:31, Laszlo Ersek wrote:
>> > Now, I would normally recommend sticking with i440fx for simplicity.
>> However, each PCI bridge requires 4KB of IO space (meaning (1 + 5) * 4KB
>> = 24KB), and OVMF on the i440fx does not support that much (only
>> 0x4000). So, I'll recommend Q35 for IO space purposes; OVMF on Q35
>> provides 0xA000 (40KB).
>
> So if we use OVMF, going for Q35 gives us actually more IO space, nice!
> However recommending Q35 for IO space seems odd :)
OVMF used to have only 0x4000 bytes for PCI IO aperture, since the
beginning. In <https://bugzilla.redhat.com/show_bug.cgi?id=1333238> I
investigated how much I could grow the aperture. On Q35, it was possible
to grow it to 0xA000 bytes (but even then you have to disable vmport,
which sort of sits in the middle otherwise). On i440fx, the IO ports in
use by platform devices were so badly distributed that moving beyond
0x4000 was not possible. See in particular:
https://bugzilla.redhat.com/show_bug.cgi?id=1333238#c16
https://bugzilla.redhat.com/show_bug.cgi?id=1333238#c19
>> Therefore I guess the simplest example I can give now is:
>> - use Q35 (for a larger IO space),
>> - plug a DMI-PCI bridge into the root bridge,
>> - plug 5 PCI bridges into the DMI-PCI bridge,
>> - plug 31 NICs per PCI bridge, each NIC into a separate slot.
>>
>
> The setup looks OK to me (assuming OVMF is needed, otherwise
> PC + pci-bridges will result in more devices),
OVMF is quite needed; Jiaxin is one of the edk2 networking maintainers,
and I think he's using QEMU and OVMF as a testbed for otherwise
physically-oriented UEFI development.
> I do have a little concern.
> We want to deprecate the dmi-pci bridge since it does not support
> hot-plug (for itself or devices behind it).
> Alexandr (CCed) is a GSOC student working on a generic
> pcie-pci bridge that can (eventually) be hot-plugged
> into a PCIe Root Port and keeps the machine cleaner.
Nice!
Please include an update to "docs/pcie.txt" in the scope :)
Thanks!
Laszlo
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] The maximum limit of virtual network device
2017-07-06 9:46 ` Laszlo Ersek
@ 2017-07-06 14:49 ` Wu, Jiaxin
0 siblings, 0 replies; 6+ messages in thread
From: Wu, Jiaxin @ 2017-07-06 14:49 UTC (permalink / raw)
To: Laszlo Ersek, Marcel Apfelbaum, qemu-devel@nongnu.org,
Alexander Bezzubikov
Thanks Laszlo and Marcel for the comments/suggestion. It is quite helpful against my usage case.
> >> - use Q35 (for a larger IO space),
> >> - plug a DMI-PCI bridge into the root bridge,
> >> - plug 5 PCI bridges into the DMI-PCI bridge,
> >> - plug 31 NICs per PCI bridge, each NIC into a separate slot.
Best regards!
Jiaxin
> -----Original Message-----
> From: Laszlo Ersek [mailto:lersek@redhat.com]
> Sent: Thursday, July 6, 2017 5:46 PM
> To: Marcel Apfelbaum <marcel@redhat.com>; Wu, Jiaxin
> <jiaxin.wu@intel.com>; qemu-devel@nongnu.org; Alexander Bezzubikov
> <zuban32s@gmail.com>
> Subject: Re: [Qemu-devel] The maximum limit of virtual network device
>
> On 07/06/17 11:24, Marcel Apfelbaum wrote:
> > On 06/07/2017 11:31, Laszlo Ersek wrote:
>
> >> > Now, I would normally recommend sticking with i440fx for simplicity.
> >> However, each PCI bridge requires 4KB of IO space (meaning (1 + 5) * 4KB
> >> = 24KB), and OVMF on the i440fx does not support that much (only
> >> 0x4000). So, I'll recommend Q35 for IO space purposes; OVMF on Q35
> >> provides 0xA000 (40KB).
> >
> > So if we use OVMF, going for Q35 gives us actually more IO space, nice!
> > However recommending Q35 for IO space seems odd :)
>
> OVMF used to have only 0x4000 bytes for PCI IO aperture, since the
> beginning. In <https://bugzilla.redhat.com/show_bug.cgi?id=1333238> I
> investigated how much I could grow the aperture. On Q35, it was possible
> to grow it to 0xA000 bytes (but even then you have to disable vmport,
> which sort of sits in the middle otherwise). On i440fx, the IO ports in
> use by platform devices were so badly distributed that moving beyond
> 0x4000 was not possible. See in particular:
>
> https://bugzilla.redhat.com/show_bug.cgi?id=1333238#c16
> https://bugzilla.redhat.com/show_bug.cgi?id=1333238#c19
>
> >> Therefore I guess the simplest example I can give now is:
> >> - use Q35 (for a larger IO space),
> >> - plug a DMI-PCI bridge into the root bridge,
> >> - plug 5 PCI bridges into the DMI-PCI bridge,
> >> - plug 31 NICs per PCI bridge, each NIC into a separate slot.
> >>
> >
> > The setup looks OK to me (assuming OVMF is needed, otherwise
> > PC + pci-bridges will result in more devices),
>
> OVMF is quite needed; Jiaxin is one of the edk2 networking maintainers,
> and I think he's using QEMU and OVMF as a testbed for otherwise
> physically-oriented UEFI development.
>
> > I do have a little concern.
> > We want to deprecate the dmi-pci bridge since it does not support
> > hot-plug (for itself or devices behind it).
> > Alexandr (CCed) is a GSOC student working on a generic
> > pcie-pci bridge that can (eventually) be hot-plugged
> > into a PCIe Root Port and keeps the machine cleaner.
>
> Nice!
>
> Please include an update to "docs/pcie.txt" in the scope :)
>
> Thanks!
> Laszlo
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2017-07-06 14:50 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-07-06 6:20 [Qemu-devel] The maximum limit of virtual network device Wu, Jiaxin
2017-07-06 8:11 ` Daniel P. Berrange
2017-07-06 8:31 ` Laszlo Ersek
2017-07-06 9:24 ` Marcel Apfelbaum
2017-07-06 9:46 ` Laszlo Ersek
2017-07-06 14:49 ` Wu, Jiaxin
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).