Timothy – I don’t personally have any programs
that have lots of pointers in memory so that the
effect of defeating the cache can be demonstrated.
But while ever people are building ELF32 binaries
for reasons that they think make sense, I would
like to see “truth in advertising” and for gcc to
use the “-m32” option to demonstrate that the
code itself is 32-bit clean, and for these modules
to run as AM64 so that a z/Linux system runs in
pure 64-bit mode.

Note that in the future it may even be possible
to neuter SAM24/31 in the hardware to ensure
that the software is all running cleanly. Or maybe
the hardware will allow AM31 to be obsoleted
and replaced by AM32.

Besides memory caching there are other reasons
to produce ELF32 modules, such as potentially
being able to run those modules on different
32-bit-only hardware. I don’t think utilities like
“sed” should be required to be 64-bit when they
don’t go anywhere near exceeding the 4 GiB
barrier.

Joe – yes, I read the document you sent me, but
it only detailed ELF64.

BFN. Paul.

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
----------------------------------------------------------------------
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Reply via email to