> [...]

> I just pushed this to my github repo.

Brilliant. I ran the test suite once more against the CGEN simulator, please 
find attached the test report.

Let's look at the first failed test for the l.add instruction, it's the third 
one:

        // Add two small negative numbers, which shoulyou haven't modelled thed 
set the carry flag but not the overflow flag.
        TEST_INST_FF_I32_I32 0, SPR_SR_CY | SPR_SR_OV, l.add, -1, -2

After running that test, the carry flag is not set as expected. I guess the 
carry flag is not implemented correctly as it's probably not needed in 
GCC-generated code. Is that right?

Cheers,
  rdiez

Attachment: OrbuildReport.tgz
Description: application/gtar-compressed

_______________________________________________
OpenRISC mailing list
[email protected]
http://lists.openrisc.net/listinfo/openrisc

Reply via email to