[U-Boot-Users] _udivsi3 and _umodsi3 from the compiler libraries instead of libarm.a
I had a problem linking u-boot for an IXP425 (ARM) board. The problem was that _udivsi3 and _umodsi3 were being satisfied from the compiler libraries instead of libarm.a.
One fix that works is to change this line in the top level Makefile:
--start-group $(LIBS) $(PLATFORM_LIBS) --end-group \
To this:
--start-group $(LIBS) --end-group $(PLATFORM_LIBS) \
I'm concerned that this might break other stuff. Is there a better way to fix this?
Robert
In message 42112C01.9000003@applieddata.net you wrote:
I had a problem linking u-boot for an IXP425 (ARM) board. The problem was that _udivsi3 and _umodsi3 were being satisfied from the compiler libraries instead of libarm.a.
One fix that works is to change this line in the top level Makefile:
--start-group $(LIBS) $(PLATFORM_LIBS) --end-group \
To this:
--start-group $(LIBS) --end-group $(PLATFORM_LIBS) \
Yes, this is what we changed in recent versions of U-Boot.
I'm concerned that this might break other stuff. Is there a better way to fix this?
Can you please explain your concerns? Why should this break "other stuff"? Why should there be a "better way", i. e. what is it that you dislike with this solution?
Best regards,
Wolfgang Denk
Wolfgang Denk wrote:
In message 42112C01.9000003@applieddata.net you wrote:
I had a problem linking u-boot for an IXP425 (ARM) board. The problem was that _udivsi3 and _umodsi3 were being satisfied from the compiler libraries instead of libarm.a.
One fix that works is to change this line in the top level Makefile:
--start-group $(LIBS) $(PLATFORM_LIBS) --end-group \
To this:
--start-group $(LIBS) --end-group $(PLATFORM_LIBS) \
Yes, this is what we changed in recent versions of U-Boot.
I'm concerned that this might break other stuff. Is there a better way to fix this?
Can you please explain your concerns? Why should this break "other stuff"? Why should there be a "better way", i. e. what is it that you dislike with this solution?
Best regards,
Wolfgang Denk
I was only concerned only because I wasn't sure that some system didn't depend on the compiler libraries being searched twice.
Sorry, I'm reinventing the wheel (I just downloaded this 2 weeks ago).
In message 42113ACB.70208@applieddata.net you wrote:
I was only concerned only because I wasn't sure that some system didn't depend on the compiler libraries being searched twice.
They should not; I've never seen this in any default compiler configuration.
Sorry, I'm reinventing the wheel (I just downloaded this 2 weeks ago).
Close, but not close enough ;-)
Best regards,
Wolfgang Denk
participants (2)
-
Robert Whaley -
Wolfgang Denk