Date   

Re: [edk2-devel] [edk2-discuss] WriteSections64()

Ard Biesheuvel <ard.biesheuvel@...>
 

On Sat, 15 Jun 2019 at 02:17, Russell Peterson <russell@mellanox.com> wrote:

Here is a sample of the gcc command line:

"/x/sdk/sysroots/x86_64-pokysdk-linux/usr/bin/aarch64-poky-linux/aarch64-poky-linux-gcc"
-I/x/edk2/edk2/MlxPlatformPkg/Include -march=armv8-a+crc -g -fshort-wchar
-fno-builtin -fno-strict-aliasing -Wall -Werror -Wno-array-bounds -ffunction-sections -fdata-sections -incl\
ude AutoGen.h -fno-common -DSTRING_ARRAY_NAME=BaseCryptLibStrings -g -Os -fshort-wchar -fno-builtin -fno-stri\
ct-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-pic -fno-pie -ffixed-x18 -flto -Wno-unused-but-set-variable -Wno-unused-const-variable -mcmodel=tiny -n\
ostdinc -nostdlib -DUEFI_C_SOURCE -DMDEPKG_NDEBUG -c -o /x/RELEASE_GCC5/AARCH64/CryptoPkg/Library/
BaseCryptLib/BaseCryptLib/OUTPUT/Pk/CryptX509.obj <-I include files left out> /x/edk2/CryptoPkg/Library/BaseCryptLib/Pk/CryptX509.c

echo --strip-unneeded -R .eh_frame /x/edk2/Build/RELEASE_GCC5/AARC\
H64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll
--strip-unneeded -R .eh_frame /x/edk2/Build/RELEASE_GCC5/AARCH64/M\
deModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll
echo /x/edk2/Build/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Ne\
twork/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll
Could you dump the section headers of such a .dll using readelf -S please?


Re: [edk2-devel] [edk2-discuss] WriteSections64()

Ard Biesheuvel <ard.biesheuvel@...>
 

On Fri, 14 Jun 2019 at 21:17, Laszlo Ersek <lersek@redhat.com> wrote:

+Ard, +devel

On 06/13/19 17:29, Russell Peterson wrote:
Hello,

We are in the process of upgrading our gcc toolchain to gcc 8.3 (from 7.3). In the process of doing that we have hit an error building edk2:

make[2]: *** [/home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.efi] Error 1
WriteSections64(): /home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll AARCH64 relative relocations require identical ELF and PE/COFF section offsets

There are multiple errors similar to this issued by GenFw. Our edk2 version is fairly recent (UDK2018 upstream branch as of commit 10318063b) but all this did build fine with gcc 7.3. Any clues?
Your Conf/tools_def.txt is probably stale, and does not contain the
-fno-pie option we added a while ago.


Re: [edk2-devel] [edk2-discuss] WriteSections64()

Russell Peterson
 

Here is a sample of the gcc command line:

"/x/sdk/sysroots/x86_64-pokysdk-linux/usr/bin/aarch64-poky-linux/aarch64-poky-linux-gcc"
-I/x/edk2/edk2/MlxPlatformPkg/Include -march=armv8-a+crc -g -fshort-wchar
-fno-builtin -fno-strict-aliasing -Wall -Werror -Wno-array-bounds -ffunction-sections -fdata-sections -incl\
ude AutoGen.h -fno-common -DSTRING_ARRAY_NAME=BaseCryptLibStrings -g -Os -fshort-wchar -fno-builtin -fno-stri\
ct-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-pic -fno-pie -ffixed-x18 -flto -Wno-unused-but-set-variable -Wno-unused-const-variable -mcmodel=tiny -n\
ostdinc -nostdlib -DUEFI_C_SOURCE -DMDEPKG_NDEBUG -c -o /x/RELEASE_GCC5/AARCH64/CryptoPkg/Library/
BaseCryptLib/BaseCryptLib/OUTPUT/Pk/CryptX509.obj <-I include files left out> /x/edk2/CryptoPkg/Library/BaseCryptLib/Pk/CryptX509.c

echo --strip-unneeded -R .eh_frame /x/edk2/Build/RELEASE_GCC5/AARC\
H64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll
--strip-unneeded -R .eh_frame /x/edk2/Build/RELEASE_GCC5/AARCH64/M\
deModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll
echo /x/edk2/Build/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Ne\
twork/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll

-----Original Message-----
From: discuss@edk2.groups.io <discuss@edk2.groups.io> On Behalf Of Russell Peterson via Groups.Io
Sent: Friday, June 14, 2019 5:19 PM
To: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: edk2-devel-groups-io <devel@edk2.groups.io>; Laszlo Ersek <lersek@redhat.com>; discuss@edk2.groups.io
Subject: Re: [edk2-devel] [edk2-discuss] WriteSections64()

Hello, Ard.

The tools_def.txt was updated and both the -fno-pie and -fno-pic are specified when executing gcc. I don’t have access to the code at the moment but I will post the exact gcc command line as soon as I can.

Regards,

Russell


On Jun 14, 2019, at 3:43 PM, Ard Biesheuvel <ard.biesheuvel@linaro.org<mailto:ard.biesheuvel@linaro.org>> wrote:

On Fri, 14 Jun 2019 at 21:17, Laszlo Ersek <lersek@redhat.com<mailto:lersek@redhat.com>> wrote:

+Ard, +devel

On 06/13/19 17:29, Russell Peterson wrote:
Hello,

We are in the process of upgrading our gcc toolchain to gcc 8.3 (from 7.3). In the process of doing that we have hit an error building edk2:

make[2]: *** [/home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.efi] Error 1
WriteSections64(): /home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll AARCH64 relative relocations require identical ELF and PE/COFF section offsets

There are multiple errors similar to this issued by GenFw. Our edk2 version is fairly recent (UDK2018 upstream branch as of commit 10318063b) but all this did build fine with gcc 7.3. Any clues?


Your Conf/tools_def.txt is probably stale, and does not contain the -fno-pie option we added a while ago.


Re: [edk2-devel] [edk2-discuss] WriteSections64()

Russell Peterson
 

Hello, Ard.

The tools_def.txt was updated and both the -fno-pie and -fno-pic are specified when executing gcc. I don’t have access to the code at the moment but I will post the exact gcc command line as soon as I can.

Regards,

Russell

On Jun 14, 2019, at 3:43 PM, Ard Biesheuvel <ard.biesheuvel@linaro.org<mailto:ard.biesheuvel@linaro.org>> wrote:

On Fri, 14 Jun 2019 at 21:17, Laszlo Ersek <lersek@redhat.com<mailto:lersek@redhat.com>> wrote:

+Ard, +devel

On 06/13/19 17:29, Russell Peterson wrote:
Hello,

We are in the process of upgrading our gcc toolchain to gcc 8.3 (from 7.3). In the process of doing that we have hit an error building edk2:

make[2]: *** [/home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.efi] Error 1
WriteSections64(): /home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll AARCH64 relative relocations require identical ELF and PE/COFF section offsets

There are multiple errors similar to this issued by GenFw. Our edk2 version is fairly recent (UDK2018 upstream branch as of commit 10318063b) but all this did build fine with gcc 7.3. Any clues?


Your Conf/tools_def.txt is probably stale, and does not contain the
-fno-pie option we added a while ago.


Re: WriteSections64()

Laszlo Ersek
 

+Ard, +devel

On 06/13/19 17:29, Russell Peterson wrote:
Hello,

We are in the process of upgrading our gcc toolchain to gcc 8.3 (from 7.3). In the process of doing that we have hit an error building edk2:

make[2]: *** [/home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.efi] Error 1
WriteSections64(): /home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll AARCH64 relative relocations require identical ELF and PE/COFF section offsets

There are multiple errors similar to this issued by GenFw. Our edk2 version is fairly recent (UDK2018 upstream branch as of commit 10318063b) but all this did build fine with gcc 7.3. Any clues?

Regards,

Russell





WriteSections64()

Russell Peterson
 

Hello,

We are in the process of upgrading our gcc toolchain to gcc 8.3 (from 7.3). In the process of doing that we have hit an error building edk2:

make[2]: *** [/home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.efi] Error 1
WriteSections64(): /home/x/RELEASE_GCC5/AARCH64/MdeModulePkg/Universal/Network/DpcDxe/DpcDxe/DEBUG/DpcDxe.dll AARCH64 relative relocations require identical ELF and PE/COFF section offsets

There are multiple errors similar to this issued by GenFw. Our edk2 version is fairly recent (UDK2018 upstream branch as of commit 10318063b) but all this did build fine with gcc 7.3. Any clues?

Regards,

Russell


Re: Looking for help to enable eDP1 port on Minnowboard derivative board

Stephano Cetola
 

A couple of questions:

1. Have you updated the Video BIOS Table (VBT) adding the LVDS details
(dot clock, EDID support, etc)?
2. If so, have you followed the DDI1_XXX lines (at the bridge and LVDS
display), and are the signals present?

Cheers,
Stephano

On Mon, Jun 3, 2019 at 6:03 PM <rksaripalli@hotmail.com> wrote:

To all,
I have a board designed on the e3800 SoC. We are reusing the Minnowboard UEFI FW (with some minor changes for the memory controller) and are able to boot this board with the UEFI FW that we are building with these patches.

The board has eDP1 enabled. On this port, we have an integrated LVDS panel using a eDP to LVDS bridge.
a)So essentially, we have DDI1_XXX signals connected to the eDP to LVDS bridge. The bridge translates from eDP to LVDS.
b) The DDI1_DDCDATA and DDI1_DDCCLK are connected directly to i2c on the LVDS panel (for EDID/EEPROM access)
c) On the Minnowboard the eDP1 graphics output port is removed (In the Minnowboard schematics refer to the CPU-DDI/VGA sheet and on the right you will see BayTrail-M remove eDP port)
d) On the DDI0 signals, we have HDMI output.

But with the Minnowboard FW, we are only able to see the display output on HDMI connected through eDP0.
We do not see see any output on eDP1.

Thinking this could be a backlight issue, I even wrote some code to program the DDI_BKLTEN / DDI_BKLTVDD signals (that are tied to GPIO signals in the NC) and that does not work either.

What are the steps/tools I need to do to enable LVDS panel on eDP1?.

Ideally, we would like to have the output be sent to both HDMI and LVDS panel.
I am familiar with UEFI code and I can make any changes needed in the UEFI FW code.
The HDMI is purely for developer support at this point.

Thanks,
Ram



Looking for help to enable eDP1 port on Minnowboard derivative board

rksaripalli@...
 

To all,
I have a board designed on the e3800 SoC. We are reusing the Minnowboard UEFI FW (with some minor changes for the memory controller) and are able to boot this board with the UEFI FW that we are building with these patches.

The board has eDP1 enabled. On this port, we have an integrated LVDS panel using a eDP to LVDS bridge.
a)So essentially, we have DDI1_XXX signals connected to the eDP to LVDS bridge. The bridge translates from eDP to LVDS.
b) The DDI1_DDCDATA and DDI1_DDCCLK are connected directly to i2c on the LVDS panel (for EDID/EEPROM access)
c) On the Minnowboard the eDP1 graphics output port is removed (In the Minnowboard schematics refer to the CPU-DDI/VGA sheet and on the right you will see BayTrail-M remove eDP port)
d) On the DDI0 signals, we have HDMI output.

But with the Minnowboard FW, we are only able to see the display output on HDMI connected through eDP0.
We do not see see any output on eDP1.

Thinking this could be a backlight issue, I even wrote some code to program the DDI_BKLTEN / DDI_BKLTVDD signals (that are tied to GPIO signals in the NC) and that does not work either.

What are the steps/tools I need to do to enable LVDS panel on eDP1?.

Ideally, we would like to have the output be sent to both HDMI and LVDS panel.
I am familiar with UEFI code and I can make any changes needed in the UEFI FW code.
The HDMI is purely for developer support at this point.

Thanks,
Ram

881 - 888 of 888