hi, Ersek:
Thank you very much!
this patch is good for me!
i have finished compiling OvmfPkg.
best wishes,
-----邮件原件-----
发件人: Laszlo Ersek [mailto:ler...@redhat.com]
发送时间: 2012年9月5日 16:12
收件人: nicolas
抄送: edk2-devel@lists.sourceforge.net
主题: Re: [edk2] OvmfPkg compiled error
On 09/05/12 08:12, nicolas wrote:
> hi, experts:
> OvmfPkg compiled error:
> ... ...
> Architecture(s) = IA32 X64
> Build target = DEBUG
> Toolchain = VS2008
> Active Platform = e:\edkii\OvmfPkg\OvmfPkgIa32X64.dsc
> Flash Image Definition = e:\edkii\OvmfPkg\OvmfPkgIa32X64.fdf
> Processing meta-data .
> Build...
> e:\edkii\OvmfPkg\OvmfPkgIa32X64.dsc(306): error F001: Pcd
> (gPcAtChipsetPkgTokenS
> paceGuid.Pcd8259LegacyModeEdgeLevel) defined in DSC is not declared in
> DEC files
>
> but, i found Pcd8259LegacyModeEdgeLevel had been declared in
> PcAtChipsetPkg.dec .
>
> best wishes
I can reproduce this with gcc44 and
ACTIVE_PLATFORM = OvmfPkg/OvmfPkgIa32X64.dsc
TARGET_ARCH = IA32
in "Conf/target.txt".
This use of Pcd8259LegacyModeEdgeLevel was introduced in svn rev 13628
("OvmfPkg: generate full MADT dynamically, synchronize contents with
qemu"). The DSC should know about the PCD via
"OvmfPkg/AcpiPlatformDxe/AcpiPlatformDxe.inf" -->
"PcAtChipsetPkg/PcAtChipsetPkg.dec".
I notice the block
#
# ACPI Support
#
MdeModulePkg/Universal/Acpi/AcpiTableDxe/AcpiTableDxe.inf
OvmfPkg/AcpiPlatformDxe/AcpiPlatformDxe.inf
OvmfPkg/AcpiTables/AcpiTables.inf
is present in all three of
- OvmfPkg/OvmfPkgIa32X64.dsc,
- OvmfPkg/OvmfPkgIa32.dsc,
- OvmfPkg/OvmfPkgX64.dsc,
but the containing Component section is different:
- OvmfPkg/OvmfPkgIa32X64.dsc: [Components.X64]
- OvmfPkg/OvmfPkgIa32.dsc: [Components]
- OvmfPkg/OvmfPkgX64.dsc: [Components]
In "OvmfPkg/OvmfPkgIa32X64.dsc" the sections [Components.IA32] and
[Components.X64] have many more differences between each other than just
"ACPI Support", so I think this (ie. svn rev 13384) is not an oversight.
"OvmfPkg/build.sh" doesn't seem to support OvmfPkgIa32X64.dsc at all.
The "-a" option is accepted on the command line (--> PROCESSOR
variable), sanitized,
case $PROCESSOR in
IA32)
Processor=Ia32
QEMU_COMMAND=qemu
;;
X64)
Processor=X64
QEMU_COMMAND=qemu-system-x86_64
;;
*)
echo Unsupported processor architecture: $PROCESSOR
echo Only IA32 or X64 is supported
exit 1
;;
esac
and then passed on to the "build" BaseTools utility (for which"-a"
overrides TARGET_ARCH). The variable named "Processor" (case is
important) can only take the values Ia32 / X64 (see above), and then it
is used in
if [ -z "$PLATFORMFILE" ]; then
PLATFORMFILE=$WORKSPACE/OvmfPkg/OvmfPkg$Processor.dsc
fi
Hmm, I may be wrong about OvmfPkgIa32X64.dsc being supported... "-p"
sets PLATFORMFILE directly.
... I can see something similar in svn rev 13350:
Fix build break in OvmfPkgIa32X64.dsc where
gEfiSecurityPkgTokenSpaceGuid was no longer found.
It is declared in SecurityPkg.dec, but with the latest build tools,
since this package builds multiple architectures (IA32 & X64) and
there are no IA32 modules used from the SecurityPkg, it is now
necessary to place the gEfiSecurityPkgTokenSpaceGuid PCD's in a
[PcdsFixedAtBuild.X64] section.
So [Components.X64] is actually OK, and I made a mistake in svn rev
13628. The attached patch turns the error you encounter into a different
(much later) one:
Generating DXEFV FV
GenFds.py...
: error F003: Module MdeModulePkg/Universal/PCD/Dxe/Pcd.inf appears
under ARCH ['X64'] in platform
/home/lacos/src/upstream/edk2-git-svn/OvmfPkg/OvmfPkgIa32X64.dsc, but
current deduced ARCH is set(['IA32']), so NO build output could be
put into flash.
make: *** [fds] Error 1
This error is similar to when I change TARGET_ARCH to X64, but keep
ACTIVE_PLATFORM = OvmfPkg/OvmfPkgIa32X64.dsc:
Generating MAINFV FV
GenFds.py...
: error F003: Module MdeModulePkg/Universal/PCD/Pei/Pcd.inf appears
under ARCH ['IA32'] in platform
/home/lacos/src/upstream/edk2-git-svn/OvmfPkg/OvmfPkgIa32X64.dsc, but
current deduced ARCH is set(['X64']), so NO build output could be put
into flash.
make: *** [fds] Error 1
Therefore I think the attached patch could be an improvement, but I'll
let Jordan decide.
Nicolas, does the patch fix the problem for you?
Thanks & sorry!
Laszlo------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
edk2-devel mailing list
edk2-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/edk2-devel
