Geoffroy Van Cutsem
toggle quoted message
Show quoted text
From: acrn-users@... <acrn-users@...>
On Behalf Of Steve Osselton
Sent: Friday, January 10, 2020 1:27 PM
To: acrn-users@...
Subject: Re: [acrn-users] Problem with Industry Scenario on a NUC7i7DNH
Thanks Geoffroy. Odd that the same version of the compiler works with Clear Linux ?
Thanks Steve, I can confirm that I see the problem too using Ubuntu 19.10 (gcc 9.2.1). Ubuntu 19.04 compiles ACRN
correctly (uses gcc 8.3.0). I have filed an issue to track this:
https://github.com/projectacrn/acrn-hypervisor/issues/4358.
Geoffroy
Hi Geoffroy,
Build from a clean checkout on 19.10 (not 19.04) with:
make all BOARD=nuc7i7dnb SCENARIO=industry RELEASE=0
We use a gcc 9.2.1 in Clear Linux to build ACRN these days, I’ll see if I can spin up an Ubuntu 19.04 machine to
reproduce the failure.
Geoffroy
Hi Liu,
Followed exactly these instructions, except on Ubuntu 19.10. To me looks like a gcc 9 or config issue as clearly
doesn't like ‘-mindirect-branch’ and ‘-fcf-protection’ used together. Am building a Clear Linux VM now ....
Hi Steve,
We don’t met this issue with Ubuntu 18.04.
Do you install following package? As mentioned on:
https://projectacrn.github.io/1.5/getting-started/building-from-source.html
BR.
Fuzhong
From:
acrn-users@... <acrn-users@...>
On Behalf Of Steve Osselton
Sent: Friday, January 10, 2020 5:33 PM
To: acrn-users@...
Subject: Re: [acrn-users] Problem with Industry Scenario on a NUC7i7DNH
Hi Yin,
Will validate whether suggested quick fix works. Looks like not happy building with gcc 9 (am using Ubuntu 19.10):
cc -Iinclude -Iinclude/lib -Iinclude/lib/crypto -Iinclude/common -Iinclude/arch/x86 -Iinclude/arch/x86/boot -Iinclude/arch/x86/guest -Iinclude/arch/x86/lib -Iinclude/debug
-Iinclude/public -Iinclude/dm -Iinclude/hw -Iboot/include -Iboot/include/guest -I/home/steve/git/acrn-hypervisor/build/hypervisor/include -Iarch/x86/configs/nuc7i7dnb -Iscenarios/industry -I. -c -include /home/steve/git/acrn-hypervisor/build/hypervisor/include/config.h
-Wall -W -ffunction-sections -fdata-sections -fshort-wchar -ffreestanding -fsigned-char -m64 -mno-mmx -mno-sse -mno-sse2 -mno-80387 -mno-fp-ret-in-387 -mno-red-zone -mpopcnt -nostdinc -nostdlib -fno-common -Werror -O2 -fpie -fstack-protector-strong -DSTACK_PROTECTOR
-mindirect-branch=thunk-extern -mindirect-branch-register -DCONFIG_RETPOLINE -DHV_DEBUG -DPROFILING_ON -fno-omit-frame-pointer -gdwarf-2 pre_build/static_checks.c -o /home/steve/git/acrn-hypervisor/build/hypervisor/pre_build/static_checks.o -MMD -MT /home/steve/git/acrn-hypervisor/build/hypervisor/pre_build/static_checks.o
In file included from pre_build/static_checks.c:7:
include/lib/util.h: In function ‘mem_aligned_check’:
include/lib/util.h:30:1: error: ‘-mindirect-branch’ and ‘-fcf-protection’ are not compatible
Should I submit this as a bug report? Will try building on Clear Linux ...
Hi Steve,
On 2020/1/10
下午5:12, Steve Osselton wrote:
> Hi Yin,
>
> What is the preferred way? Is there any chance of having the default
> value increased as I guess it would be nice
> to support Samsung SSDs "out of the box", without having to rebuild.
The default way is change it through "make menuconfig" command and
rebuild the ACRN. Changing Kconfig is kind of hack here.
For default value, we need trade off like whether there are many such
devices.
Could you please try the hack first so we could know whether this is
the root cause of your issue? Thanks.
Regards
Yin, Fengwei
>
> Cheers Steve.
>
> On Fri, 10 Jan 2020 at 09:06, Yin, Fengwei <fengwei.yin@...
> <mailto:fengwei.yin@...>> wrote:
>
> Hi Steve,
>
> On 2020/1/10
下午5:02, Steve Osselton wrote:
> > Hi Yin,
> >
> > I have a Samsung PCI SSD installed. Looking at it's capabilities
> I get:
> >
> > Capabilities: [b0] MSI-X: Enable+ Count=33 Masked-
> >
> > So looks like this is the problem. It it easy to up the number of
> > supported MSIX tables?
> The easiest way to try is change the file:
> hypervisor/arch/x86/Kconfig
> which has:
> config MAX_MSIX_TABLE_NUM
> int "Maximum number of MSI-X tables per device"
> range 1 2048
> default 16 ---> 64
>
> You could try it as quick check. It's not the preferred way...
>
>
> Regards
> Yin, Fengwei
>
> >
> > Cheers Steve.
> >
> > On Fri, 10 Jan 2020 at 05:00, Yin, Fengwei <fengwei.yin@...
> <mailto:fengwei.yin@...>
> > <mailto:fengwei.yin@... <mailto:fengwei.yin@...>>> wrote:
> >
> > Hi Osselton,
> > It looks like there is one PCI device in your system which
> has more than
> > 16 MSIX table.
> >
> > Can you boot to native linux and dump the MSI table of each
> PCI devices
> > on your board?
> >
> >
> > Regards
> > Yin, Fengwei
> >
> > On 2020/1/9
下午6:30, Steve Osselton wrote:
> > > Hi Liz,
> > >
> > > Was just following the instructions for example industry
> scenario
> > for
> > > the 1.4 release, didn't build anything myself.
> > > Have tried again from scratch using latest updates for 1.5:
> > >
> > > 1 - Installed clear-32030-live-server.iso onto hard disk.
> > > 2 - Boot:
> > >
> > > # wget
> > >
> >
> https://raw.githubusercontent.com/projectacrn/acrn-hypervisor/master/doc/getting-started/acrn_quick_setup.sh
> > > # chmod +x acrn_quick_setup.sh
> > > # ./acrn_quick_setup.sh -s 32030 -d -i
> > > # efibootmgr -c -l '\EFI\acrn\acrn.efi' -d /dev/sda -p 1
> -L ACRN -u
> > > "uart=port@0x3f8"
> > >
> > > 3 - Reboot:
> > >
> > > Console output:
> > >
> > > ACRN Hypervisor
> > > calibrate_tsc, tsc_khz=2100000
> > > [11187656us][cpu=0][(null)][sev=2][seq=1]:HV version
> > > 1.5-unstable-2020-01-05 11:33:11- DBG (daily tag:) build by
> > mockbuild,
> > > start time 11171478us
> > > [11201829us][cpu=0][(null)][sev=2][seq=2]:API version 1.0
> > > [11207570us][cpu=0][(null)][sev=2][seq=3]:Detect
> processor: Intel(R)
> > > Core(TM) i7-8650U CPU @ 1.90GHz
> > > [11218392us][cpu=0][(null)][sev=2][seq=4]:Assertion failed
> in file
> > > hw/pci.c,line 448 : fatal error
> > >
> > > Still failing ...
> > >
> > > Cheers Steve.
> > >
> > >
> > > On Thu, 9 Jan 2020 at 02:40, Liu, Fuzhong
> <fuzhong.liu@... <mailto:fuzhong.liu@...>
> > <mailto:fuzhong.liu@... <mailto:fuzhong.liu@...>>
> > > <mailto:fuzhong.liu@...
> <mailto:fuzhong.liu@...> <mailto:fuzhong.liu@...
> <mailto:fuzhong.liu@...>>>> wrote:
> > >
> > > Hi Steve____
> > >
> > > Seem like, you are using very old codebase, can you
> try with v1.4
> > > codebase first?____
> > >
> > > If not; please share the commit id.____
> > >
> > > Thanks!____
> > >
> > > __ __
> > >
> > > BR.____
> > >
> > > Fuzhong____
> > >
> > > __ __
> > >
> > > *From:*
acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>
> > > <mailto:acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>>
> > > <acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>
> > > <mailto:acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>>> *On Behalf Of *Steve
> > Osselton
> > > *Sent:* Wednesday, January 8, 2020 11:34 PM
> > > *To:*
acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>
> > > <mailto:acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>>
> > > *Subject:* Re: [acrn-users] Problem with Industry
> Scenario on a
> > > NUC7i7DNH____
> > >
> > > __ __
> > >
> > > Hi Liu,____
> > >
> > > __ __
> > >
> > > On inspection was running an old BIOS version, have
> upgraded to
> > > latest version (DNi70067) and disabled hyper
> threading.____
> > >
> > > Things are improved in that I now get diagnostics from
> the serial
> > > port:____
> > >
> > > __ __
> > >
> > > ACRN Hypervisor
> > > calibrate_tsc, tsc_khz=2100000
> > > [11041475us][cpu=0][sev=2][seq=1]:HV version
> > 1.4-unstable-2019-11-18
> > > 09:13:55- DBG (daily tag:) build by mockbuild, start time
> > 11028016us
> > > [11054921us][cpu=0][sev=2][seq=2]:API version 1.0
> > > [11059914us][cpu=0][sev=2][seq=3]:Detect processor:
> Intel(R)
> > > Core(TM) i7-8650U CPU @ 1.90GHz
> > > [11069846us][cpu=0][sev=2][seq=4]:Assertion failed in file
> > > hw/pci.c,line 250 : fatal error____
> > >
> > > __ __
> > >
> > > So still failing.____
> > >
> > > __ __
> > >
> > > Also as a comment, the BIOS settings descriptions, are
> badly
> > out of
> > > step with what you see from the latest visual BIOS.____
> > >
> > > __ __
> > >
> > > Cheers Steve.____
> > >
> > > __ __
> > >
> > > On Wed, 8 Jan 2020 at 13:46, Liu, Fuzhong
> > <fuzhong.liu@... <mailto:fuzhong.liu@...>
> <mailto:fuzhong.liu@... <mailto:fuzhong.liu@...>>
> > > <mailto:fuzhong.liu@...
> <mailto:fuzhong.liu@...>
> > <mailto:fuzhong.liu@...
> <mailto:fuzhong.liu@...>>>> wrote:____
> > >
> > > Hi Steve,____
> > >
> > > Please disable Hyper thread in BIOS for
> workaround.____
> > >
> > > This issue is fixed in v1.5; will release soon.____
> > >
> > > ____
> > >
> > > BR.____
> > >
> > > Fuzhong____
> > >
> > > ____
> > >
> > > *From:*
acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>
> > > <mailto:acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>>
> > > <acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>
> > > <mailto:acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>>> *On Behalf Of *Steve
> > > Osselton
> > > *Sent:* Wednesday, January 8, 2020 8:32 PM
> > > *To:*
acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>
> > > <mailto:acrn-users@...
> <mailto:acrn-users@...>
> > <mailto:acrn-users@...
> <mailto:acrn-users@...>>>
> > > *Subject:* [acrn-users] Problem with Industry
> Scenario on a
> > > NUC7i7DNH____
> > >
> > > ____
> > >
> > > Hi,
> > >
> > > Am essentially following the deployment
> instructions for the
> > > ACRN industry scenario
> > > using a NUC7i7DNH with a serial port connected.
> > >
> > >
> https://projectacrn.github.io/latest/getting-started/rt_industry.html
> > >
> > > Flashing clear-31670-live-server.iso worked fine. Then
> > followed
> > > the "Use the pre-installed industry ACRN hypervisor"
> > > section and installed serial port support. Again
> this all
> > worked
> > > fine and looked good.
> > >
> > > I then rebooted the NUC. At HDMI console get
> "Intel NUC"
> > banner
> > > with the following at top LHS of screen:
> > >
> > > detected 8 processes, 8 enabled
> > > current on process 0
> > >
> > > And on serial port:
> > >
> > > E0S0 S0
> > >
> > > Not sure what has happened here. Am I missing
> something
> > obvious ?
> > >
> > > Thanks Steve ____
> > >
> > >
> > > ____
> > >
> > > __ __
> > >
> > > -- ____
> > >
> > > Engineering Director____
> > >
> > > IOTech Systems Ltd.____
> > >
> > > __
> > >
> > >
> > >
> > > --
> > > Engineering Director
> > > IOTech Systems Ltd.
> > >
> >
> >
> >
> >
> >
> > --
> > Engineering Director
> > IOTech Systems Ltd.
> >
>
>
>
>
>
> --
> Engineering Director
> IOTech Systems Ltd.
>
--
--
--
--
|
