[patch] Implement reglocs for s390/s390x

Roland McGrath roland at hack.frob.com
Wed Oct 10 18:47:13 UTC 2012

> The most native display of "upper halves" as items is:
>   LINUX                 64  S390_HIGH_GPRS
>     high_r0: 1023, high_r1: 1023, high_r2: 0, high_r3: 0, high_r4: 0, high_r5: 1023
>     high_r6: 0, high_r7: 0, high_r8: 0, high_r9: 0, high_r10: 0, high_r11: 0
>     high_r12: 1023, high_r13: 0, high_r14: 0, high_r15: 0
> But I do not find it much readable.  So I chose 'x' format there:

The display format for each register is chosen by the backend based on the
usual use of those registers.  Registers that are usually used for
addresses get hex, generic integer registers get decimal.

But since these are not register values but high halves of register values,
using hex for them all makes sense to me.

>   LINUX                 64  S390_HIGH_GPRS
>     high_r0: 0x000003ff, high_r1: 0x000003ff, high_r2: 0x00000000, high_r3: 0x00000000, high_r4: 0x00000000, high_r5: 0x000003ff
>     high_r6: 0x00000000, high_r7: 0x00000000, high_r8: 0x00000000, high_r9: 0x00000000, high_r10: 0x00000000, high_r11: 0x00000000
>     high_r12: 0x000003ff, high_r13: 0x00000000, high_r14: 0x00000000, high_r15: 0x00000000
> But then it is too wide so I made an exception for the "register" group width:

The special case on group name is bletcherous.  I don't recall at all why I
set ITEM_WRAP_COLUMN so high.  At the moment I can't think of any good
rationale for ever wrapping past column 79, and 75 seems like a good choice.
Does anything bad happen if you just change ITEM_WRAP_COLUMN to 75?


More information about the elfutils-devel mailing list