2009-12-29 02:45:01     2009R1 & bf561 rev0.3

Document created by Aaronwu Employee on Aug 20, 2013
Version 1Show Document
  • View in full screen mode

2009-12-29 02:45:01     2009R1 & bf561 rev0.3

zhongztc zhongztc (CHINA)

Message: 84021   

 

2009R1,,

 

       i configured like this:                                                 x x

  x x                             CPU (BF561)  --->                                                                           x x

  x x                         [*] Symmetric multi-processing support                                                          x x

  x x                             Silicon Rev (0.3)  --->                                                                     x x

  x x                             System type (BF561-EZKIT)  --->

 

 

 

  CC      libbb/xrealloc_vector.o

  CC      libbb/xregcomp.o

  AR      libbb/lib.a

  LINK    busybox_unstripped

Trying libraries: crypt m

Library crypt is not needed, excluding it

Library m is not needed, excluding it

Final link with: <none>

make[4]:正在离开目录 `/media/??/2009/uClinux-dist/user/busybox/build-busybox-1.13.4'

make[3]:正在离开目录 `/media/??/2009/uClinux-dist/user/busybox'

make[2]:正在离开目录 `/media/??/2009/uClinux-dist/user'

make[1]: *** [all] 错误 2

make[1]:正在离开目录 `/media/??/2009/uClinux-dist/user'

make: *** [subdirs] 错误 1

QuoteReplyEditDelete

 

 

2009-12-29 02:51:06     Re: 2009R1 & bf561 rev0.3

Mike Frysinger (UNITED STATES)

Message: 84022   

 

you didnt post the full build log.  you didnt include the actual error message in the output.

 

run `make single` to force the build to abort on an error.

QuoteReplyEditDelete

 

 

2009-12-29 03:07:08     Re: 2009R1 & bf561 rev0.3

Yi Li (CHINA)

Message: 84024   

 

SMP is not supposed to work on bf561 rev0.3 due to hardware anomalies. You may try bf561 rev0.5.

 

-Yi

QuoteReplyEditDelete

 

 

2009-12-29 03:12:04     Re: 2009R1 & bf561 rev0.3

Mike Frysinger (UNITED STATES)

Message: 84025   

 

what anomalies ?  there are no checks in the kernel for 0.3+SMP that cause a failure.

QuoteReplyEditDelete

 

 

2009-12-29 04:52:11     Re: 2009R1 & bf561 rev0.3

Graf Yang (CHINA)

Message: 84050   

 

Anomally 248 said,

 

BF561 0.3: When one core performs a testset operation to L2 memory it locks out the other core if it is accessing L2 memory.

 

Workaround:

 

Any TESTSET operation to internal L2 memory should be followed immediately by a (dummy) write to L2 memory.

 

     CLI R0;

  NOP; NOP;

  TESTSET(P0);

  W[P1] = R0;   // P1 Points to L2 Memory

  STI R0;

 

 

 

But the dummy write which following testest will cause infinity stall.

 

 

 

There is a task,

 

  blackfin.uclinux.org/gf/tracker/5092

QuoteReplyEditDelete

 

 

2009-12-29 06:31:24     Re: 2009R1 & bf561 rev0.3

zhongztc zhongztc (CHINA)

Message: 84052   

 

but after the cancllelationof supporting smp,

 

only bf561 rev0.3

 

it gave me  the same erro.

QuoteReplyEditDelete

 

 

2009-12-29 11:29:05     Re: 2009R1 & bf561 rev0.3

Mike Frysinger (UNITED STATES)

Message: 84054   

 

you need to post the log like you were already told.  if you dont, then we cant help you.

QuoteReplyEditDelete

 

 

2009-12-29 11:33:15     Re: 2009R1 & bf561 rev0.3

Mike Frysinger (UNITED STATES)

Message: 84055   

 

if the workaround documented in the anomaly text isnt working, then you should raise the issue with the ADI apps team

Attachments

    Outcomes