Re: PCI and VGA Passthrough regressions on Xen 4.4.1 vs 4.3.2

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

On Mon, Oct 06, 2014 at 03:55:36AM -0300, Zir Blazer wrote:
> There is a regression in both PCI and VGA Passthrough in Xen 4.4.1 when compared to 4.3.2. Due to the added complexity of VGA Passthrough, I was suggested to focus instead in my PCI Passthrough issue, which involves the integrated Sound Card of my Motherboard. While it passes to the DomU and works, it produces robotic, distorted and lagged noise everytime it reproduces sound. The Video Card is an absolute no-go. On the previous Xen version, everything else being equal, both works properly.I have been trying to gather as much information as possible of my issue according to Xen Wiki articles http://wiki.xen.org/wiki/Debugging_Xen and http://wiki.xen.org/wiki/Reporting_Bugs_against_Xen so this E-Mail comes attached with tons of logs.
> 
> 

Thank you for the detailed bug-report.

Interestingly enough I have been using the same exact hardware, thought

> HARDWARE
> Processor: Intel Xeon E3-1245V3 (Haswell) - http://ark.intel.com/products/75462/Intel-Xeon-Processor-E3-1245-v3-8M-Cache-3_40-GHz

I seem to see 8 CPUs while your serial output say it has 4?

Did you disable hyper-threading?

> Motherboard: Supermicro X10SAT (Chipset C226, BIOS R2.0) - http://www.supermicro.com/products/motherboard/Xeon/C220/X10SAT.cfm
> Sound Card: Integrated Realtek ALC1150
> Video Card: Haswell IGP, and a Radeon 5770 (Which is the GPU I always pass)

Anyhow, could you when booting under Xen 4.3 (and Xen 4.4) also use
iommu=verbose,debug

You should get something like this:

XEN) Enabling APIC mode:  Flat.  Using 1 I/O APICs
(XEN) ACPI: HPET id: 0x8086a701 base: 0xfed00000
(XEN) [VT-D]dmar.c:778: Host address width 39
(XEN) [VT-D]dmar.c:792: found ACPI_DMAR_DRHD:
(XEN) [VT-D]dmar.c:472:   dmaru->address = fed90000
(XEN) [VT-D]iommu.c:1158: drhd->address = fed90000 iommu->reg = ffff82c3ffd54000
(XEN) [VT-D]iommu.c:1160: cap = c0000020660462 ecap = f0101a
(XEN) [VT-D]dmar.c:383:  endpoint: 0000:00:02.0
(XEN) [VT-D]dmar.c:792: found ACPI_DMAR_DRHD:
(XEN) [VT-D]dmar.c:472:   dmaru->address = fed91000
(XEN) [VT-D]iommu.c:1158: drhd->address = fed91000 iommu->reg = ffff82c3ffd53000
(XEN) [VT-D]iommu.c:1160: cap = d2008c20660462 ecap = f010da
(XEN) [VT-D]dmar.c:397:  IOAPIC: 0000:f0:1f.0
(XEN) [VT-D]dmar.c:361:  MSI HPET: 0000:f0:0f.0
(XEN) [VT-D]dmar.c:486:   flags: INCLUDE_ALL
(XEN) [VT-D]dmar.c:797: found ACPI_DMAR_RMRR:
(XEN) [VT-D]dmar.c:383:  endpoint: 0000:00:1d.0
(XEN) [VT-D]dmar.c:383:  endpoint: 0000:00:1a.0
(XEN) [VT-D]dmar.c:383:  endpoint: 0000:00:14.0
(XEN) [VT-D]dmar.c:666:   RMRR region: base_addr d9eab000 end_address d9eb9fff
(XEN) [VT-D]dmar.c:797: found ACPI_DMAR_RMRR:
(XEN) [VT-D]dmar.c:383:  endpoint: 0000:00:02.0
(XEN) [VT-D]dmar.c:666:   RMRR region: base_addr dc000000 end_address de1fffff
(XEN) Xen ERST support is initialized.

And I only do PCI passthrough of NICs (which has been working fine
for me in Xen 4.3 - hadn't yet attempted Xen 4.4).

> September 2014: Xen 4.3.2, BIOS/UEFI Boot with Syslinux/Gummiboot, Arch Linux ISO dated Oct 2014, Linux Kernel 3.16 with default config options or custom 3.17 RC4
> Both Sound Card and Video Card works, including using my previous DomU, in either BIOS mode with either Kernel or UEFI with 3.17. The minimal DomU works as expected with only the Sound Card. The only shortcoming I found running UEFI, is that using "reboot" command in console, sometimes freezes in the last line when it says it is ready to reboot, other times it does work. Seems prone to freeze when I just turned it on and want to reboot, but can reboot consistently on prolonged sessions.
> 
> September 2014: Xen 4.4.1, BIOS/UEFI Boot with Syslinux/Gummiboot, Arch Linux ISO dated Oct 2014, Linux Kernel 3.16 with default config options or custom 3.17 RC4
> Sound Card passthrough to the minimal DomU (Which I created for testing passthrough in 4.4.1, before I downgraded to 4.3.2 to test again) does not work, robotic noise and interference as described. VGA Passthrough is a total no-go, but I suppose than fixing whatever regression annoys the Sound Card will fix the Video Card too. After tons of testing with different things (BIOS or UEFI, Compiler flags, Kernel version, Kernel config), I conclude that this is broken at least for my setup, as not even once I made it work.

Gotcha. And there were no BIOS updates in between using Xen 4.3 or Xen 4.4? I recall that BIOS v1.0 on those
boards would spit errors.