drivers/gpu/drm/radeon/radeon_test.c:45: undefined reference to `__udivdi3'
in [Kernel]
|
From: Dave Hylands on 19 Dec 2009 04:20 Hi Jim, On Fri, Dec 18, 2009 at 12:54 PM, Mr. James W. Laferriere <babydr(a)baby-dragons.com> wrote: > � � � �Hello Benjamin , �I am not having a good time compiling 2.6.32.1 . > First RCU 'inline' difficulties then fusion-mpt 'inline' difficulty , �Now > this . > � � � �This one is not a simple as removing 'inline' from a statement . > � � � �Would you please look into this ? �Anything else you need please just > ask . > � � � �Please cc: me as I am not subscribed to linux-fbdev . > > � � � � � � � �Tia , �JimL > ....snip... > drivers/built-in.o(.text+0x11345a): In function `radeon_test_moves': > drivers/gpu/drm/radeon/radeon_test.c:45: undefined reference to `__udivdi3' > make: *** [.tmp_vmlinux1] Error 1 __udivdi3 is a gcc helper function used to do 64-bit division. This form of 64-bit division is not supported from within the kernel (which is why you get the undefined symbol). There is <asm/div64.h> which contains a macro called do_div which can perform a 64-bit dvision by a 32-bit divisor. -- Dave Hylands Shuswap, BC, Canada http://www.DaveHylands.com/ -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Dave Hylands on 19 Dec 2009 21:10 Hi Jim, > � � � �Thank you , �For replying . �Now I being a kernel level nothing , > �How does this help me get past the 'ld' time undefined reference ? �So we > can have a kernel build complete (hopefully) successfully . > > � � � �I've another ? as well how is it that the drm/radeon_test.c is using > (well at least referencing) a non-kernel symbol ? �That alone is rather > interesting . > > � � � �But all I had to do is Disable DRM in the kernel . �Since this system > is a remote device without a real need for DRI/DRM support that should cure > my problem . �But someone else is going to need that support & I am sure > they'll hit that same bug . I'm not familiar with the source code in question, but presumably, someone will need to change the source to remove the 64-bit divide and do it using the do_div macro or something equivalent. -- Dave Hylands Shuswap, BC, Canada http://www.DaveHylands.com/ -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Dave Airlie on 20 Dec 2009 01:20 On Sat, Dec 19, 2009 at 6:54 AM, Mr. James W. Laferriere <babydr(a)baby-dragons.com> wrote: > Hello Benjamin , I am not having a good time compiling 2.6.32.1 . > First RCU 'inline' difficulties then fusion-mpt 'inline' difficulty , Now > this . > This one is not a simple as removing 'inline' from a statement . > Would you please look into this ? Anything else you need please just > ask . > Please cc: me as I am not subscribed to linux-fbdev . > > Tia , JimL > Can you try the attached patch? Dave.
|
Pages: 1 Prev: Compat-wireless releases for 2.6.33-rc1 and 2.6.32.2 Next: Documentation: update mmiotrace.txt |