Re: [win-pv-devel] xenbus fails to load on Win7 x64

['Marek Marczykowski-Górecki' <marmarek@xxxxxxxxxxxxxxxxxxxxxx>]

On Mon, Jul 09, 2018 at 08:17:04AM +0000, Paul Durrant wrote:
> > -----Original Message-----
> > From: 'Marek Marczykowski-Górecki'
> > [mailto:marmarek@xxxxxxxxxxxxxxxxxxxxxx]
> > Sent: 08 July 2018 01:25
> > To: Paul Durrant <Paul.Durrant@xxxxxxxxxx>
> > Cc: win-pv-devel@xxxxxxxxxxxxxxxxxxxx
> > Subject: Re: [win-pv-devel] xenbus fails to load on Win7 x64
> > 
> > On Sat, Jul 07, 2018 at 03:51:47PM +0200, 'Marek Marczykowski-Górecki'
> > wrote:
> > > On Sat, Jul 07, 2018 at 11:23:40AM +0000, Paul Durrant wrote:
> > > > > -----Original Message-----
> > > > > From: win-pv-devel [mailto:win-pv-devel-
> > bounces@xxxxxxxxxxxxxxxxxxxx] On
> > > > > Behalf Of Marek Marczykowski-Górecki
> > > > > Sent: 06 July 2018 23:53
> > > > > To: win-pv-devel@xxxxxxxxxxxxxxxxxxxx
> > > > > Subject: [win-pv-devel] xenbus fails to load on Win7 x64
> > > > >
> > > > > Hi,
> > > > >
> > > > > I'm trying to install latest build of pv drivers on Windows 7 SP1 Pro
> > (64bit).
> > > > > It's fresh install + all windows updates applied + testsigning 
> > > > > enabled.
> > > > > After getting my VM into unbootable state few times, I've backed to
> > > > > small steps - installed only xenbus. After that, device manager shows
> > > > > "Xen PV Bus" device as not working (code 39), and I think relevant
> > > > > messages from event viewer are:
> > > > >
> > > > > Kernel-PnP: The driver \Driver\xenbus failed to load for the device
> > > > >
> > PCI\VEN_5853&DEV_0001&SUBSYS_00015853&REV_01\3&267a616a&0&10.
> > > > >
> > > > > Application Popup :\SystemRoot\system32\DRIVERS\xenbus.sys
> > cannot find
> > > > > ntoskrnl.exe KeInitializeSpinLock
> > > > >
> > > > > I've tried it built myself and automatic build from jenkins. Both have
> > > > > the same result.
> > > > >
> > > > > Any ideas?
> > > >
> > > > Hi Marek,
> > > >
> > > > Which builds of the drivers are you trying? The master branch?
> > >
> > > Yes, the master branch.
> > >
> > > > The toolchain used by the public build VM is currently VS2015/WDK10 and
> > IIRC that does not actually support building for Windows 7.
> > >
> > > According to docs[1] it should work...
> > > Maybe I need to adjust some build options? Adding Windows7
> > configuration
> > > to targets.props? Anyway, I'd expect some clearer error about version
> > > incompatibility, than some unresolved symbol...
> > 
> > Ok, that was it. I've added Windows 7 configuration. Then a little
> > #ifdef as MmAllocateContiguousNodeMemory requires Windows 8+ (but
> > MmAllocateContiguousMemory does not). It builds just fine using
> > vs2017/WDK10.
> > 
> > As for build.py - right now it choose windows version based on visual
> > studio version. Wouldn't it be better to have it as an argument?
> > For my build I use slightly modified build.py - the same for all
> > drivers, with more parameters (driver name, target architecture, windows
> > version etc). SDV is broken with it, but that's offtopic here (I've run
> > into some weird problem there).
> > 
> > > > If you need to run master branch drivers on Windows 7 you can probably
> > graft the 8.2 vs2013 scripts onto the master branch code and get something
> > to build, but the intention is that master (or 9.0 as they will be when 
> > they are
> > released) drivers will only support Window 8.1 onwards as this allows us to
> > use newer kernel and storport/ndis features.
> > >
> > > So, are 9.0 drivers going to drop Windows 7 support? It is still in
> > > extended support (up to 2020)...
> > > My selfish reason is that we still have XPDM driver which don't work
> > > with any newer windows than 7.
> > 
> > Would you accept a patch adding (restoring?) Windows 7 build target?
> > I'll send the patches anyway in a moment, feel free to ignore them if
> > you don't want to keep support for Windows 7. Also, there are some
> > minor fixes, mostly independent of windows 7 build.
> 
> Marek,
> 
>   The intention, which is why master went from 8.x to 9.x, is that we 
> baseline on Windows 8. The reason for this is that Microsoft made some fairly 
> significant changes in the kernel interfaces, and if we remain backwards 
> compatible with Windows 7 we cannot use them (or certainly not without 
> jumping through hoops). Is there anything that is missing in 8.2.1 that you 
> particularly need to use? I would not be against releasing and 8.3 version if 
> need be.

Mainly changes in xeniface that Rafał made (exposing APIs to userspace).
Plus of course various bug fixes. 8.3 would work for me too. As you can
see in those patches, right now Windows 7 require just one #ifdef in all
those drivers (I haven't checked xencons, xenhid, and xenvkbd).

-- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?

Attachment: signature.asc
Description: PGP signature

_______________________________________________
win-pv-devel mailing list
win-pv-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/win-pv-devel