Regression in OVMF + RMRR series

Regression in OVMF + RMRR series

[Wei Liu]

I found this in OSSTest

http://logs.test-lab.xenproject.org/osstest/logs/59910/test-amd64-amd64-xl-qemuu-ovmf-amd64/serial-fiano1.log

Jul 25 17:48:51.468985 (d1) HVM Loader
Jul 25 17:48:51.517028 (d1) Detected Xen v4.6-unstable
Jul 25 17:48:51.517056 (d1) Xenbus rings @0xfeffc000, event channel 1
Jul 25 17:48:51.525060 (d1) System requested OVMF
Jul 25 17:48:51.525097 (d1) CPU speed is 1800 MHz
Jul 25 17:48:51.525131 (d1) Relocating guest memory for lowmem MMIO space disabled
Jul 25 17:48:51.533055 (XEN) irq.c:276: Dom1 PCI link 0 changed 0 -> 5
Jul 25 17:48:51.541041 (d1) PCI-ISA link 0 routed to IRQ5
Jul 25 17:48:51.541070 (XEN) irq.c:276: Dom1 PCI link 1 changed 0 -> 10
Jul 25 17:48:51.549023 (d1) PCI-ISA link 1 routed to IRQ10
Jul 25 17:48:51.549051 (XEN) irq.c:276: Dom1 PCI link 2 changed 0 -> 11
Jul 25 17:48:51.557040 (d1) PCI-ISA link 2 routed to IRQ11
Jul 25 17:48:51.557067 (XEN) irq.c:276: Dom1 PCI link 3 changed 0 -> 5
Jul 25 17:48:51.557093 (d1) PCI-ISA link 3 routed to IRQ5
Jul 25 17:48:51.565085 (d1) pci dev 01:2 INTD->IRQ5
Jul 25 17:48:51.565111 (d1) pci dev 01:3 INTA->IRQ10
Jul 25 17:48:51.573018 (d1) pci dev 02:0 INTA->IRQ11
Jul 25 17:48:51.573050 (d1) pci dev 04:0 INTA->IRQ5
Jul 25 17:48:51.573080 (d1) No RAM in high memory; setting high_mem resource base to 100000000
Jul 25 17:48:51.589031 (d1) pci dev 03:0 bar 10 size 002000000: 0f0000008
Jul 25 17:48:51.597018 (d1) pci dev 02:0 bar 14 size 001000000: 0f2000008
Jul 25 17:48:51.597053 (d1) pci dev 04:0 bar 30 size 000040000: 0f3000000
Jul 25 17:48:51.605026 (d1) pci dev 03:0 bar 30 size 000010000: 0f3040000
Jul 25 17:48:51.605063 (d1) pci dev 03:0 bar 14 size 000001000: 0f3050000
Jul 25 17:48:51.613035 (d1) pci dev 02:0 bar 10 size 000000100: 00000c001
Jul 25 17:48:51.613103 (d1) pci dev 04:0 bar 10 size 000000100: 00000c101
Jul 25 17:48:51.621075 (d1) pci dev 04:0 bar 14 size 000000100: 0f3051000
Jul 25 17:48:51.621109 (d1) pci dev 01:2 bar 20 size 000000020: 00000c201
Jul 25 17:48:51.629023 (d1) pci dev 01:1 bar 20 size 000000010: 00000c221
Jul 25 17:48:51.637041 (d1) Multiprocessor initialisation:
Jul 25 17:48:51.637086 (d1)  - CPU0 ... 46-bit phys ... fixed MTRRs ... var MTRRs [1/8] ... done.
Jul 25 17:48:51.645025 (d1)  - CPU1 ... 46-bit phys ... fixed MTRRs ... var MTRRs [1/8] ... done.
Jul 25 17:48:51.653020 (d1) Testing HVM environment:
Jul 25 17:48:51.653052 (d1)  - REP INSB across page boundaries ... passed
Jul 25 17:48:51.661025 (d1)  - GS base MSRs and SWAPGS ... passed
Jul 25 17:48:51.661058 (d1) Passed 2 of 2 tests
Jul 25 17:48:51.661084 (d1) Writing SMBIOS tables ...
Jul 25 17:48:51.669048 (d1) Loading OVMF ...
Jul 25 17:48:51.669078 (XEN) d1v0 Over-allocation for domain 1: 196865 > 196864
Jul 25 17:48:51.677064 (XEN) memory.c:155:d1v0 Could not allocate order=0 extent: id=1 memflags=0 (0 of 1)
Jul 25 17:48:51.677111 (d1) Loading ACPI ...
Jul 25 17:48:51.685061 (d1) vm86 TSS at fc012d00
Jul 25 17:48:51.685107 (d1) BIOS map:
Jul 25 17:48:51.685145 (d1)  ffe00000-ffffffff: Main BIOS
Jul 25 17:48:51.693030 (d1) *** HVMLoader bug at e820.c:262
Jul 25 17:48:51.693064 (d1) *** HVMLoader crashed.

Git blame shows that the change that crashes hvmloader was part of the
RMRR series.

Tiejun, could you please fix this please?  It should be easy to reproduce.

Guest configuration file is at:

http://logs.test-lab.xenproject.org/osstest/logs/59910/test-amd64-amd64-xl-qemuu-ovmf-amd64/fiano1--debianhvm.guest.osstest.cfg

Wei.

Re: Regression in OVMF + RMRR series

[Wei Liu]

I forgot to mention:  you need to use --enable-ovmf in ./configure to
enable OVMF support.

Wei.

Re: Regression in OVMF + RMRR series

[Chen, Tiejun]

> Jul 25 17:48:51.685107 (d1) BIOS map:
> Jul 25 17:48:51.685145 (d1)  ffe00000-ffffffff: Main BIOS
> Jul 25 17:48:51.693030 (d1) *** HVMLoader bug at e820.c:262
> Jul 25 17:48:51.693064 (d1) *** HVMLoader crashed.
>
> Git blame shows that the change that crashes hvmloader was part of the
> RMRR series.
>
> Tiejun, could you please fix this please?  It should be easy to reproduce.
>

This is really my responsibility and I just sent out one patch to fix this,

tools/hvmloader: sync memory map[]

Please take a review.

Thanks
Tiejun