>It might be the application who runs, calls kernel for output, drivers used >for output uses floating point. >No application running => no floating point exceptions... Yes even I do agree with this point.....and need ur help how to check this coz the main reason is that the error is not consistently coming from one particular point but the location keeps on changing...... if we try to debug and remove from one point it start coming from some other point in the application...... >Some driver, most likely your if you have made any yourself is >using floating point. >serial, console, ethernet drivers? No I am not running any such drivers which are new and can be suspected all the drivers used are fully tested and working well with all other applications... >pc = 3184 (in hex) >Now check your System.map >The routine with the closest lower address is your main suspect, >the use of inlines can blur this... >(the task pointer is not that useful, they should have converted it >to process id) Even this pc value is very less. I think it should be in the range of c0000000 - c0233e38(range I got from system.map file).... if we think pc is returning the four LSB's even then it is not giving any such clue as there are only ISR's defined in arc/ppc/kernel/head.S which are lying close to this address Regards, Malik
Roger Larsson <roger.larsson at norran.net> wrote: On onsdag 24 maj 2006 10.14, you wrote: > Hi Roger... > Thanks for your response.....but in my case the application is causing > this error.... It might be the application who runs, calls kernel for output, drivers used for output uses floating point. No application running => no floating point exceptions... > The result was exactly what is expected.....so i think this is not > related to floating point but might be some other issue...... Some driver, most likely your if you have made any yourself is using floating point. serial, console, ethernet drivers? > > Are there any other scenarios which can lead to this message???? Not unless the exception is setup wrong - I would not bet on that. Lets trace it: > > floating point used in kernel (task=c0398410, pc=3184) This message is generated in ./arch/ppc/kernel/head.S pc = 3184 (in hex) Now check your System.map The routine with the closest lower address is your main suspect, the use of inlines can blur this... (the task pointer is not that useful, they should have converted it to process id) /RogerL --------------------------------- Yahoo! India Answers Share what your know-how and wisdom Send free SMS to your Friends on Mobile from your Yahoo! Messenger Download now -------------- next part -------------- An HTML attachment was scrubbed... URL: http://ozlabs.org/pipermail/linuxppc-embedded/attachments/20060524/0243fff3/attachment.htm