Aarch64 u-boot with gcc .5x
Hello,
I have an arm64 odroid c2 board. If I build u-boot with the current OE 5.3 gcc it hangs at kernel load time. If I use 4.9 gcc it works fine. A 5.3 built kernel boots fine with a 4.9 built u-boot.
The vendor supplied u-boot is hacked up and am leaning it that direction as the culprit so I am looking for data points.
Is anyone using a current u-boot with 5.3 gcc on an arm64 machine? if so, any special tweaks to get that combo to work?
kind regards,
Armin
Hello Armin,
Quoted 9 lines · On Mon, Mar 28, 2016 at 12:51 PM, akuster808 <[email protected]> wrote:
I have an arm64 odroid c2 board. If I build u-boot with the current OE 5.3 gcc it hangs at kernel load time. If I use 4.9 gcc it works fine. A 5.3 built kernel boots fine with a 4.9 built u-boot.
The vendor supplied u-boot is hacked up and am leaning it that direction as the culprit so I am looking for data points.
Is anyone using a current u-boot with 5.3 gcc on an arm64 machine? if so, any special tweaks to get that combo to work?
I am using one ARM 64 SoC with GCC 5.2 at this moment and it works fine. What U-Boot version is the fork based on?
Signature · 3 lines
Otavio Salvador O.S. Systems http://www.ossystems.com.br http://code.ossystems.com.br Mobile: +55 (53) 9981-7854 Mobile: +1 (347) 903-9750
I hit the same thing and ended up using their prebuilt bootloader to get it up for testing. There is a big C++ ABI change with gcc 5.x (I've been using 5.3.0 for a while now) and the gold linker is much less tolerant of underlinking than bfd is. That said, their u-boot is funky but not actually that old, and they also have a blob 1st-stage that is only provided as a binary.
Their wiki:
http://odroid.com/dokuwiki/doku.php?id=en:c2_building_u-boot
says to checkout branch odroidc2-v2015.01 but I didn't dig into the commit history to see when they last pulled from upstream. The error I get with 5.3.0 looks like their u-boot/include/linux stuff is just way too old for gcc5 (since they're still using a 3.10 kernel fork probably). Their build also blows up with gcc-4.8 - can you say brittle-hard-dep on gcc 4.9?
Steve
Quoted 25 lines · Otavio Salvador
Hello Armin,
Quoted 9 lines · On Mon, Mar 28, 2016 at 12:51 PM, akuster808 <[email protected]> wrote:
I have an arm64 odroid c2 board. If I build u-boot with the current OE 5.3 gcc it hangs at kernel load time. If I use 4.9 gcc it works fine. A 5.3 built kernel boots fine with a 4.9 built u-boot.
The vendor supplied u-boot is hacked up and am leaning it that direction as the culprit so I am looking for data points.
Is anyone using a current u-boot with 5.3 gcc on an arm64 machine? if so, any special tweaks to get that combo to work?
I am using one ARM 64 SoC with GCC 5.2 at this moment and it works fine. What U-Boot version is the fork based on?
--
Otavio SalvadorO.S. Systems
http://www.ossystems.com.br http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854 Mobile: +1 (347) 903-9750
--
_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.openembedded.org/mailman/listinfo/openembedded-core
Quoted 7 lines · Otavio Salvador
Hello Armin,
Quoted 4 lines · On Mon, Mar 28, 2016 at 12:51 PM, akuster808 <[email protected]> wrote:
I have an arm64 odroid c2 board. If I build u-boot with the current OE 5.3 gcc it hangs at kernel load time. If I use 4.9 gcc it works fine. A 5.3 built kernel boots fine with a 4.9 built u-boot.
Quoted 8 lines
Quoted 5 lines
The vendor supplied u-boot is hacked up and am leaning it that direction as the culprit so I am looking for data points.
Is anyone using a current u-boot with 5.3 gcc on an arm64 machine? if so, any special tweaks to get that combo to work?
I am using one ARM 64 SoC with GCC 5.2 at this moment and it works fine. What U-Boot version is the fork based on?
2015.01
- armin
Quoted 1 line
Quoted 9 lines · On Mon, Mar 28, 2016 at 5:52 PM, akuster808 <[email protected]> wrote:
Quoted 8 lines · Otavio Salvador
Quoted 7 lines · On Mon, Mar 28, 2016 at 12:51 PM, akuster808 <[email protected]> wrote:
I have an arm64 odroid c2 board. If I build u-boot with the current OE 5.3 gcc it hangs at kernel load time. If I use 4.9 gcc it works fine. A 5.3 built kernel boots fine with a 4.9 built u-boot.
The vendor supplied u-boot is hacked up and am leaning it that direction as the culprit so I am looking for data points.
Quoted 7 lines
Quoted 5 lines
Quoted 2 lines
Is anyone using a current u-boot with 5.3 gcc on an arm64 machine? if so, any special tweaks to get that combo to work?
I am using one ARM 64 SoC with GCC 5.2 at this moment and it works fine. What U-Boot version is the fork based on?
2015.01
Please try to use GCC 5.2 and see if it makes any difference. The BSP I am referring to is currently using it and it is working fine.
The forked version is the same. Of course, it may have changes which hides or solve the issue but it is a hint anyway.
Signature · 3 lines
Otavio Salvador O.S. Systems http://www.ossystems.com.br http://code.ossystems.com.br Mobile: +55 (53) 9981-7854 Mobile: +1 (347) 903-9750
Quoted 16 lines · Otavio Salvador
Quoted 15 lines · On Mon, Mar 28, 2016 at 5:52 PM, akuster808 <[email protected]> wrote:
Quoted 13 lines · Otavio Salvador
Quoted 9 lines · On Mon, Mar 28, 2016 at 12:51 PM, akuster808 <[email protected]> wrote:
I have an arm64 odroid c2 board. If I build u-boot with the current OE 5.3 gcc it hangs at kernel load time. If I use 4.9 gcc it works fine. A 5.3 built kernel boots fine with a 4.9 built u-boot.
The vendor supplied u-boot is hacked up and am leaning it that direction as the culprit so I am looking for data points.
Is anyone using a current u-boot with 5.3 gcc on an arm64 machine? if so, any special tweaks to get that combo to work?
I am using one ARM 64 SoC with GCC 5.2 at this moment and it works fine. What U-Boot version is the fork based on?
Quoted 7 lines
Quoted 1 line
2015.01
Please try to use GCC 5.2 and see if it makes any difference. The BSP I am referring to is currently using it and it is working fine.
The forked version is the same. Of course, it may have changes which hides or solve the issue but it is a hint anyway.
ok. will try this next.
kind regards,
Armin