Dear Stefan,
Stefan Roese wrote:
On Monday 22 June 2009 20:10:23 Dirk Behme wrote:
This is because it's a copy from the Linux code. Not sure if we should change the coding style here. I really prefer to be in sync with the Linux version. This makes future updates easier.
You have to discuss this with Wolfgang, but he wasn't happy about it. Enjoy the time while he is on vacation ;)
Yes, will do. ;)
- The major one: By the new 64bit variables, depending on tool chain,
there are now calls to libgcc introduced. Depending on tool chain, this might work, or fail. As it does with one of my tool chains, which worked totally fine until here. It was my understanding that U-Boot shall not rely on libgcc, i.e. proper tool chain libraries.
Having CONFIG_SYS_64BIT_VSPRINTF not defined, the linker needs "_lshrdi3", which comes from nand_base.o and nand_bbt.o. E.g. from "len = mtd->size >> (this->bbt_erase_shift + 2);" from nand_bbt.c.
Having CONFIG_SYS_64BIT_VSPRINTF enabled, the linker requests _lshrdi3, _udivdi3, _umoddi3 and _clz from libgcc.
Looking into U-Boot's lib_arm/ directory, we have already some low level math functions there to avoid libgcc. E.g. nand_bbt.c has references to _ashrdi3, too, but this is resolved by lib_arm's version what is fine.
I'd like that NAND code is modified that only math is used which U-Boot provides and no libgcc is needed.
this is a already known problem by Stefan and I and solved
two patchs was send to the ML
Thanks for the hint! Do you kindly have the subjects or better links to mailing list archive?
http://git.denx.de/?p=u-boot/u-boot-arm.git;a=commit;h=cf96e690cb9840b4a5 dc7e750be863228550b5ce
and
http://git.denx.de/?p=u-boot/u-boot-arm.git;a=commit;h=c5ae538c9bd359c137 d306e431b2589ba91fbc08
both available in the "testing" branch of u-boot-arm.
Yes, I found them. I was confused because I understood Jean-Christophe that they were posted to the list already in the past, and not while we were discussing about it ;)
Jean-Christophe did post those patches a few days ago.
From you comments on the list I see that you found those patches here too. :)
Do these patches fix the building problem for you?
"testing" branch of u-boot-arm compiled fine for me, but I'm not sure if the NAND changes are in "testing" branch of u-boot-arm, too.
I'm pretty sure they are, since Jean-Christophe used these patches exactly to fix this NAND related problem on ARM. But it should be easy to check if the patches are in here. git is your friend... ;)
Anyway, I would vote for moving these two commits as soon as possible into mainline to fix mainline, too. But there was some discussion about how to deal with Makefile changes to the patches Jean-Christophe sent to the list.
Maybe we can find a solution for the Makefile issues of these patches discussed on this list and then move the fixes asap into mainline?
Yes, I would really like to see those changes in mainline soon as well. I have to admit that I'm not a real Makefile "expert", so I don't have a real expertise to get involved in this discussion. But I'll try to take another look at this thread tomorrow.
Sorry, I know, the first week after vacation is always quite busy ;)
But any news on this?
Many thanks
Dirk