#1919 Lots of gbz80 failures using --max-allocs-per-node 5

closed-fixed
z80 port (188)
6
2012-01-27
2012-01-26
No

When adding --max-allocs-per-node 5 in spec.mk there are lots of regression test failures on gbz80:
bug1115321
funptrs_type_long
gcc-torture-execute-divmod-1
gcc-torture-execute-postmod-1
libmullong_type_asm
libmullong_type_c
onebyte_attrL_volatile_attrR_none
onebyte_attrL_volatile_attrR_volatile
swap
uminus_resulttype_long_lefttype_signed_char_storage_none_attr_volatile
uminus_resulttype_long_lefttype_signed_char_storage_static_attr_volatile
uminus_resulttype_short_lefttype_signed_char_storage_none_attr_volatile

While --max-allocs-per-node 5 can be considered a very exotic setting, all these are bugs that might reappear using more common settings in larger functions with more register pressure.

Philipp

Discussion

  • Philipp Klaus Krause

    • summary: Lots of gbz890 failures using --max-allocs-per-node 5 --> Lots of gbz80 failures using --max-allocs-per-node 5
     
  • Philipp Klaus Krause

    • assigned_to: nobody --> spth
     
  • Philipp Klaus Krause

    As of revision #7261, gbz80 regression when using --max-allocs-per-node 5 are down to 67 from 88 when this bug was filed, and down to 0 from 11 with default options.

    Philipp

     
  • Philipp Klaus Krause

    • status: open --> closed-fixed
     
  • Philipp Klaus Krause

    Fixed in revision #7262.

    Philipp

     

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks