Re: [edk2-devel] Problems building some Arm platforms (__stack_chk_guard/__stack_chk_fail, _GLOBAL_OFFSET_TABLE_)

["Pedro Falcato" <pedro.falcato@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 7555 bytes --]

Hi,

So, what's your suggestion? Do you want me to add that to my package as
well?

There is obviously a huge problem with EDK2 build tools not knowing what a
damn runtime library is. I personally don't think it's appropriate for me
to say "I want my package built with the stack protector", particularly
because my package shouldn't know what a damn stack protector is, and
whoever is building it knows best, and because of that this should be dealt
with by the build system itself (and no, setting it up in a platform dsc
isn't a correct approach, since my package should be buildable standalone).
And what if someone doesn't want the stack protector? Do they just create a
new AARCH64_GCC5_NOSSP toolchain and build it with that? Do they just sed
AARCH64_GCC5 to use -fno-stack-protector? Do they just limit themselves to
toolchains without default SSP? IMO the correct way to go about things is
to have the build tools automatically insert dependencies on runtime
libraries and have a way to tag a toolchain such that you can easily
selectively enable instrumentation (such as the SSP, UBSAN, ASAN, etc) and
build them in any combination.

Anyway, </rant>. I'll defer to your experience with the EDK2 build system
for a solution.

All the best,
Pedro

On Sat, Aug 20, 2022 at 4:06 AM Rebecca Cran <rebecca@bsdio.com> wrote:

> Other platform such as Platform/Qemu/SbsaQemu/SbsaQemu.dsc have:
>
>
>   # Add support for GCC stack protector
>   NULL|MdePkg/Library/BaseStackCheckLib/BaseStackCheckLib.inf
>
>
> --
>
> Rebecca Cran
>
>
> On 8/19/22 20:34, Pedro Falcato wrote:
>
> I see the issue: We're not passing -fno-stack-protector to AARCH64 GCC
> toolchains (although arguably we should just enable support for it...). Can
> you try hacking up your tools_def to add -fno-stack-protector to
> GCC_AARCH64_CC_FLAGS? It should fix your build issues.
>
> On Sat, Aug 20, 2022 at 2:51 AM Rebecca Cran <rebecca@bsdio.com> wrote:
>
>> I'm using gcc version 9.4.0 (Ubuntu 9.4.0-1ubuntu1~20.04.1) on Ubuntu
>> 20.04.4, from the Ubuntu repo.
>>
>> I'm building it with:
>>
>>
>> export PACKAGES_PATH=$PWD/edk2:$PWD/edk2-platforms:$PWD/edk2-non-osi
>> export WORKSPACE=$PWD
>> export GCC5_AARCH64_PREFIX=aarch64-linux-gnu-
>>
>> . ./edk2/edksetup.sh
>> build -p ./Features/Ext4Pkg/Ext4Pkg.dsc -a AARCH64 -b RELEASE -t GCC5
>>
>>
>> Some of the messages are:
>>
>>
>> "aarch64-linux-gnu-gcc" -o
>> /home/bcran/src/beaglebone/Build/Ext4Pkg/RELEASE_GCC5/AARCH64/Features/Ext4Pkg/Ext4Dxe/Ext4Dxe/DEBUG/Ext4Dxe.dll
>> -Wl,--emit-relocs -nostdlib -Wl,--gc-sections -u _ModuleEntryPoint
>> -Wl,-e,_ModuleEntryPoint,-Map,/home/bcran/src/beaglebone/Build/Ext4Pkg/RELEASE_GCC5/AARCH64/Features/Ext4Pkg/Ext4Dxe/Ext4Dxe/DEBUG/Ext4Dxe.map
>> -z common-page-size=0x20 -z common-page-size=0x1000 -flto -Os
>> -L/home/bcran/src/beaglebone/edk2/ArmPkg/Library/GccLto -llto-aarch64
>> -Wl,-plugin-opt=-pass-through=-llto-aarch64 -Wno-lto-type-mismatch
>> -Wl,--start-group,@/home/bcran/src/beaglebone/Build/Ext4Pkg/RELEASE_GCC5/AARCH64/Features/Ext4Pkg/Ext4Dxe/Ext4Dxe/OUTPUT/static_library_files.lst,--end-group
>> -g -Os -fshort-wchar -fno-builtin -fno-strict-aliasing -Wall -Werror
>> -Wno-array-bounds -include AutoGen.h -fno-common -ffunction-sections
>> -fdata-sections -DSTRING_ARRAY_NAME=Ext4DxeStrings -g -Os -fshort-wchar
>> -fno-builtin -fno-strict-aliasing -Wall -Werror -Wno-array-bounds -include
>> AutoGen.h -fno-common -mlittle-endian -fno-short-enums -fverbose-asm
>> -funsigned-char -ffunction-sections -fdata-sections -Wno-address
>> -fno-asynchronous-unwind-tables -fno-unwind-tables -fno-pic -fno-pie
>> -ffixed-x18 -mcmodel=small -flto -Wno-unused-but-set-variable
>> -Wno-unused-const-variable -D DISABLE_NEW_DEPRECATED_INTERFACES
>> -Wl,--script=/home/bcran/src/beaglebone/edk2/BaseTools/Scripts/GccBase.lds
>> -Wl,--defsym=PECOFF_HEADER_SIZE=0x228 -Wno-error
>> /usr/lib/gcc-cross/aarch64-linux-gnu/9/../../../../aarch64-linux-gnu/bin/ld:
>> /tmp/Ext4Dxe.dll.sETaOX.ltrans0.ltrans.o: in function
>> `InternalAllocatePool.constprop.0':
>> /home/bcran/src/beaglebone/edk2/MdePkg/Library/UefiMemoryAllocationLib/MemoryAllocationLib.c:368:
>> undefined reference to `__stack_chk_guard'
>> /usr/lib/gcc-cross/aarch64-linux-gnu/9/../../../../aarch64-linux-gnu/bin/ld:
>> /tmp/Ext4Dxe.dll.sETaOX.ltrans0.ltrans.o: relocation
>> R_AARCH64_ADR_PREL_PG_HI21 against symbol `__stack_chk_guard' which may
>> bind externally can not be used when making a shared object; recompile with
>> -fPIC
>> /home/bcran/src/beaglebone/edk2/MdePkg/Library/UefiMemoryAllocationLib/MemoryAllocationLib.c:368:(.text.InternalAllocatePool.constprop.0+0xc):
>> dangerous relocation: unsupported relocation
>> /usr/lib/gcc-cross/aarch64-linux-gnu/9/../../../../aarch64-linux-gnu/bin/ld:
>> /home/bcran/src/beaglebone/edk2/MdePkg/Library/UefiMemoryAllocationLib/MemoryAllocationLib.c:368:
>> undefined reference to `__stack_chk_guard'
>> /usr/lib/gcc-cross/aarch64-linux-gnu/9/../../../../aarch64-linux-gnu/bin/ld:
>> /home/bcran/src/beaglebone/edk2/MdePkg/Library/UefiMemoryAllocationLib/MemoryAllocationLib.c:382:
>> undefined reference to `__stack_chk_fail'
>> /usr/lib/gcc-cross/aarch64-linux-gnu/9/../../../../aarch64-linux-gnu/bin/ld:
>> /tmp/Ext4Dxe.dll.sETaOX.ltrans0.ltrans.o: in function
>> `OrderedCollectionInsert.constprop.0':
>> /home/bcran/src/beaglebone/edk2/MdePkg/Library/BaseOrderedCollectionRedBlackTreeLib/BaseOrderedCollectionRedBlackTreeLib.c:584:
>> undefined reference to `__stack_chk_guard'
>>
>> --
>> Rebecca Cran
>>
>>
>> On 8/19/22 18:46, Pedro Falcato wrote:
>>
>> Hi Rebecca,
>>
>> What EDK2 toolchain are you using? And how is your toolchain configured
>> (or where did you get it from?)? It seems that it's trying to use the stack
>> protector automatically...
>>
>> Thanks,
>> Pedro
>>
>>
>> On Sat, 20 Aug 2022, 00:40 Rebecca Cran, <rebecca@bsdio.com> wrote:
>>
>>> ./Features/Ext4Pkg/Ext4Pkg.dsc is also failing - with errors about
>>> __stack_chk_guard and __stack_chk_fail.
>>>
>>>
>>> And I get an error from Andy Hayes' email: "Your message couldn't be
>>> delivered to the recipient because you don't have permission to send to it."
>>>
>>>
>>> --
>>> Rebecca Cran
>>>
>>>
>>> On 8/19/22 17:35, Rebecca Cran wrote:
>>>
>>> I have an armplatbld.sh script that goes through and tries to build as
>>> many of the Arm (AARCH64 and ARM) platforms in edk2-platforms as possible.
>>>
>>> I'm think this used to work for these, but I'm getting some errors now.
>>>
>>> I'm using edk2-platforms 46686eeb7e78efe603badd86f13777d9fb070fb8 and
>>> edk2 e2ac68a23b4954d5c0399913a1df3dd9fd90315d.
>>>
>>>
>>> Drivers/ASIX/Asix.dsc (fails with undefined references to __stack_chk_guard
>>> and __stack_chk_fail)
>>>
>>> Drivers/DisplayLink/DisplayLinkPkg/DisplayLinkPkg.dsc (fails with
>>> undefined references to __stack_chk_guard and __stack_chk_fail)
>>> Platform/Socionext/DeveloperBox/DeveloperBoxMm.dsc (fails with bad
>>> definition for symbol '_GLOBAL_OFFSET_TABLE_'@0x72d8 or unsupported
>>> symbol type.  For example, absolute and undefined symbols are not
>>> supported.)
>>>
>>> Drivers/StandaloneMmCpu/StandaloneMmCpu (fails with undefined
>>> references to __stack_chk_guard and __stack_chk_fail)
>>>
>>> Platform/StandaloneMm/PlatformStandaloneMmPkg/PlatformStandaloneMmRpmb.dsc
>>> (fails with bad definition for symbol '_GLOBAL_OFFSET_TABLE_'@0x72d8 or
>>> unsupported symbol type.  For example, absolute and undefined symbols are
>>> not supported.)
>>>
>>>
>>> --
>>> Rebecca Cran
>>> 
>>>
>>>
>
> --
> Pedro Falcato
>
>

-- 
Pedro Falcato

[-- Attachment #2: Type: text/html, Size: 19719 bytes --]